Ecology and Conservation of Estuarine Ecosystems: Lake St Lucia As a Global Model 9781107347878, 9781107019751

Citation preview

Ecology and Conservation of Estuarine Ecosystems Lake St Lucia as a Global Model

St Lucia is the world’s oldest protected estuary, and Africa’s largest estuarine system. It is also the centrepiece of South Africa’s first UNESCO World Heritage Site, the iSimangaliso Wetland Park, and has been a Ramsar Wetland of International Importance since 1986. Knowledge of its biodiversity, geological origins, hydrology, hydrodynamics and the long history of management is unique in the world. However, the impact of global change has culminated in unprecedented challenges for the conservation and management of the St Lucia system, leading to the recent initiation of a project in support of its rehabilitation and long-term sustainability. This timely volume provides a unique source of information on the functioning and management of the estuary for researchers, students and environmental managers. The insights and experiences described build on over 60 years of study and management at the site and will serve as a valuable model for similar estuaries around the world. Renzo Perissinotto is Professor of Marine Biology and Academic Leader in the School of Life Sciences, University of KwaZulu-Natal, Durban, South Africa. He has published 110 articles in the primary peer-reviewed literature and is the senior author of a book on South African temporarily open/closed estuaries published in 2010. Derek D. Stretch is Professor of Hydraulics and Environmental Fluid Mechanics, the eThekwini-sponsored Chair in Civil Engineering and the Director of the Centre for Research in Environmental, Coastal & Hydrological Engineering in the School of Engineering, University of KwaZulu-Natal, Durban, South Africa. Ricky H. Taylor was until recently the Park Ecologist of Ezemvelo KZN Wildlife for the iSimangaliso Wetland Park. He has 37 years of experience at the interface between science and management and has completed a Ph.D. on the St Lucia system through the Norwegian University of Life Sciences, A˚s, Norway. He is an Honorary Research Fellow in the School of Life Sciences at the University of KwaZulu-Natal, Durban, South Africa.

Ecology and Conservation of Estuarine Ecosystems Lake St Lucia as a Global Model EDITED BY

RENZO PERISSINOTTO University of KwaZulu-Natal, South Africa

DEREK D. STRETCH University of KwaZulu-Natal, South Africa

RICKY H. TAYLOR Ezemvelo KZN Wildlife and University of KwaZulu-Natal, South Africa

CAMBRIDGE UNIVERSITY PRESS

Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, Sa˜o Paulo, Delhi, Mexico City Cambridge University Press The Edinburgh Building, Cambridge CB2 8RU, UK Published in the United States of America by Cambridge University Press, New York www.cambridge.org Information on this title: www.cambridge.org/9781107019751 © Cambridge University Press 2013 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2013 Printed and Bound in the United Kingdom by the MPG Books Group A catalogue record for this publication is available from the British Library Library of Congress Cataloging in Publication data Ecology and conservation of estuarine ecosystems : Lake St. Lucia as a model / edited by Renzo Perissinotto, University of KwaZulu-Natal, South Africa; Derek D. Stretch, University of KwaZulu-Natal, South Africa; Ricky H. Taylor, Ezemvelo KZN Wildlife, South Africa. pages cm Summary: “This book provides a unique source of information on the functioning and management of the estuary for researchers, students and environmental managers”– Provided by publisher. Includes bibliographical references and index. ISBN 978-1-107-01975-1 1. Natural history–South Africa–Saint Lucia, Lake. 2. Estuarine ecology–South Africa–Saint Lucia, Lake. I. Perissinotto, R., editor of compilation. II. Stretch, Derek D., editor of compilation. III. Taylor, R. H. (Richard Hilton), 1949- editor of compilation. QH195.S6E365 2013 577.70 86096843–dc23 2012031742 ISBN 978-1-107-01975-1 Hardback Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

Contents List of contributors page [xiii] Foreword [xvii] Alan K. Whitfield Preface [xxi] Acknowledgements [xxv]

1 South Africa’s first World Heritage Site

[1]

Roger N. Porter 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 1.10

Introduction [1] Early efforts to protect Lake St Lucia: 1895–1947 [2] Lake St Lucia under divided management: 1947–2001 [6] The controversial ‘St Lucia’ dune mining proposal [10] Decision of the South African Cabinet [12] Towards World Heritage Site listing [13] Outstanding universal values [14] Strengthening conservation measures [16] Future management imperatives [18] Conclusions [18]

2 Management history

[21]

Ricky H. Taylor 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10

Introduction [21] 1850s to 1950s: drainage of the Mfolozi floodplain [21] 1951–1960: closure of the combined Mfolozi–St Lucia Mouth [28] 1960–1968: stabilization of the St Lucia Mouth [30] 1968–1972: severe drought and the canal in the Mkhuze Swamps [33] 1973–1984: hydrological model and the Mfolozi Link Canal [35] 1984–2000: Cyclone Domoina and its aftermath [38] 2000 to present: the current drought [40] Future management [42] Assessment and lessons [44]

vi

Contents

3 Geological history

[47]

Greg A. Botha, Sylvi Haldorsen and Naomi Porat 3.1 3.2 3.3 3.4 3.5

Introduction [47] Synthesis of the geological evolution of the St Lucia basin [48] The late Pleistocene and Holocene evolution of the St Lucia basin [51] Evaluation of St Lucia sedimentation [57] Conclusions [61]

4 The marine environment

[63]

Allan D. Connell and Sean N. Porter 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9

Introduction [63] The Agulhas Current [63] Inshore waters and currents [65] The Central Shelf [67] Marine influences on St Lucia Mouth condition [67] Coastal productivity, nutrients, zooplankton and fish [68] Adult fish migrations offshore in relation to St Lucia Estuary [70] Offshore canyons and their potential impact on St Lucia Estuary [72] Coastal biogeography [73]

5 Catchment hydrology [77] Derek D. Stretch and Andrew Z. Maro 5.1 5.2 5.3 5.4 5.5 5.6

Introduction [77] Hydrological context [78] Annual average runoff and sediment yields [83] Modelling the impact of land-use changes in the Mfolozi catchment Management implications [92] Conclusions [92]

6 The wetlands

[83]

[95]

William N. Ellery, Suzanne E. Grenfell, Michael C. Grenfell, Marc S. Humphries and Kirsten B. Barnes 6.1 Introduction [95] 6.2 Connectivity of coastal plain wetlands and Lake St Lucia [95] 6.3 Diversity and hydrogeomorphology of freshwater wetlands on the coastal plain [96] 6.4 Floodplains of the coastal plain and their relationship with Lake St Lucia [97]

Contents

6.5 Valley-bottom wetlands: blocked valley lakes and wetlands [104] 6.6 Depression wetlands on the coastal plain [109] 6.7 Connectivity revisited: a framework for examining artificial impacts to wetlands [109] 6.8 Future landscape development [111] 6.9 Conclusion [111]

7 Estuary and lake hydrodynamics

[113]

Derek D. Stretch, Clinton P. Chrystal, Robynne A. Chrystal, Christopher M. Maine and Justin J. Pringle 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8

Introduction [113] The water and salt budgets in different management scenarios Wind-driven flows, waves and mixing [126] Tidal inlet hydrodynamics and morphodynamics [130] Sediment dynamics in the St Lucia/Mfolozi complex [135] Modelling the outcomes of management interventions [140] Links between physical and biological functioning [143] Summary and conclusions [148]

8 Groundwater hydrology

[116]

[151]

Bruce E. Kelbe, Ricky H. Taylor and Sylvi Haldorsen 8.1 8.2 8.3 8.4 8.5 8.6 8.7 8.8 8.9 8.10 8.11

Introduction [151] Hydrogeological setting [151] Groundwater distribution and utilization [153] Groundwater characteristics [155] Groundwater recharge [158] Groundwater discharge [160] Groundwater seepage [162] Evaporation and land use [164] The dependence of Lake St Lucia on groundwater [164] Mpate River catchment [166] Lessons learnt about groundwater which can be applied to other systems [167]

9 Physico-chemical environment

[169]

Renzo Perissinotto, Nicola K. Carrasco and Ricky H. Taylor 9.1 9.2 9.3 9.4

Introduction [169] Water depths and system partitioning [170] Salinity [172] Temperature [175]

vii

viii

Contents

9.5 9.6 9.7 9.8 9.9 9.10 9.11

Turbidity [176] Irradiance [177] pH [177] Dissolved oxygen [178] Nutrient concentration [179] Sediment composition and size structure [182] Temporal variability and states of the St Lucia system

10 Microalgae

[183]

[187]

Renzo Perissinotto, Guy C. Bate and David G. Muir 10.1 10.2 10.3 10.4 10.5

Introduction [187] Diversity and community structure Biomass [200] Productivity [203] Exceptional blooms [205]

[188]

11 Macrophytes [209] Janine B. Adams, Sibulele Nondoda and Ricky H. Taylor 11.1 11.2 11.3 11.4 11.5 11.6 11.7

Introduction [209] Distribution of macrophytes [209] Abiotic factors and macrophyte response [213] Recent studies on response to drought and mouth closure Summary of changes in macrophyte habitats [220] Conservation and management [223] Future research [224]

12 Benthic invertebrates

[227]

Deena Pillay, Sarah J. Bownes and Holly A. Nel 12.1 12.2 12.3 12.4 12.5

Introduction [227] Benthic macrofauna [227] Meiofauna [240] Ecological importance of benthic fauna [242] Conclusion [244]

13 Zooplankton

[247]

Nicola K. Carrasco, Renzo Perissinotto and Hendrik L. Jerling 13.1 Introduction [247] 13.2 Diversity and community structure

[248]

[217]

Contents

13.3 13.4 13.5 13.6 13.7

Abundance and biomass [255] Mfolozi–St Lucia link [256] Opportunistic species [258] Trophic interactions [261] System threats [263]

14 Penaeid prawns

[269]

Anthony T. Forbes and Nicolette T. Forbes 14.1 14.2 14.3 14.4 14.5 14.6 14.7 14.8

Introduction [269] Penaeid diversity and distribution [270] Prawns in South Africa: species, distribution, life cycles and biology [271] Immigration and emigration: movement, dispersal and origin of South African prawns [274] Population dynamics [277] The estuarine bait fisheries [278] Ecological significance of prawns in the St Lucia system [286] Status and threats to the major nursery grounds and the offshore habitat: the future of South African prawn stocks [288]

15 Fish and fisheries

[291]

Digby P. Cyrus 15.1 15.2 15.3 15.4 15.5 15.6 15.7 15.8 15.9 15.10

16 Birds

Introduction [291] Fish fauna of St Lucia [291] Species distribution related to physical factors [300] Reproduction [305] Nursery function of St Lucia for fish [305] Predation on the fish fauna [306] Effect of episodic events on the fish fauna [306] Extensive closure and effect on the meta-system [310] Fisheries [311] Importance of St Lucia for fish [314]

[317]

Jane Turpie, Ricky H. Taylor, Meyrick B. Bowker and Caroline Fox 16.1 16.2 16.3 16.4 16.5

Introduction [317] Bird abundance and diversity [318] Avian ecology and seasonal patterns of use [319] Spatial distribution of birds [323] Interannual variation in abundance and community composition

[323]

ix

x

Contents

16.6 The role of the estuary in a regional context [327] 16.7 Health status of the avifaunal community [328] 16.8 Conserving St Lucia birdlife [329]

17 Crocodiles

[333]

Xander Combrink, Jonathan K. Warner and Colleen T. Downs 17.1 17.2 17.3 17.4 17.5 17.6 17.7 17.8

Introduction [333] The St Lucia estuarine system’s physical environment Population size and distribution [334] Nesting [340] Diet and feeding behaviour [342] Movements [343] Threats [349] Conclusions [350]

18 Hippopotamuses

[334]

[355]

Ricky H. Taylor 18.1 18.2 18.3 18.4 18.5 18.6 18.7 18.8

Introduction [355] Hippo distribution patterns [355] Distribution maps [357] St Lucia hippo population size and trends [359] Hippo impacts on the environment [361] Management of the St Lucia hippo population [364] Research [365] Conservation of hippos in Africa [365]

19 Alien and invasive species

[367]

Nelson A. F. Miranda and Janine B. Adams 19.1 19.2 19.3 19.4

Introduction [367] Biodiversity and invasibility [368] Perceived threats [369] Management challenges [379]

20 Food webs and ecosystem functioning

[381]

Ursula M. Scharler and C. Fiona MacKay 20.1 Introduction [381] 20.2 Food webs [383] 20.3 Ecosystem states in the St Lucia estuarine system

[395]

Contents

21 Climate change impacts

[397]

Andrew A. Mather, Derek D. Stretch and Andrew Z. Maro 21.1 21.2 21.3 21.4 21.5 21.6

Introduction [397] Ocean-based processes [397] Terrestrial-based processes [402] Processes at the land–ocean interface [404] Synthesis of key climate change effects [409] Summary and conclusions [412]

References [414] Appendix: Web page database 1900–2010 Nuette Gordon Taxonomic index Index [475]

[465]

[463]

xi

Contributors

View from Fani’s Island Camp, with the coastal dunes in the background. The narrow channel is the only connection between the North and South lakes of St Lucia. (Photo: Ricky H. Taylor, July 2011.)

Janine B. Adams Department of Botany, Nelson Mandela Metropolitan University, Port Elizabeth, South Africa Kirsten B. Barnes School of Chemistry and Physics, University of KwaZulu-Natal, Durban, South Africa Guy C. Bate Department of Botany, Nelson Mandela Metropolitan University, Port Elizabeth, South Africa Greg A. Botha Council for Geoscience, Pietermaritzburg, South Africa Meyrick B. Bowker School of Life Sciences, University of KwaZulu-Natal, Pietermaritzburg, South Africa

xiv

List of contributors

Sarah J. Bownes School of Life Sciences, University of KwaZulu-Natal, Durban, South Africa Nicola K. Carrasco School of Life Sciences, University of KwaZulu-Natal, Durban, South Africa Clinton P. Chrystal Centre for Research in Environmental, Coastal & Hydrological Engineering, School of Engineering, University of KwaZulu-Natal, Durban, South Africa Robynne A. Chrystal Centre for Research in Environmental, Coastal & Hydrological Engineering, School of Engineering, University of KwaZulu-Natal, Durban, South Africa Xander Combrink Ezemvelo KwaZulu-Natal Wildlife, St Lucia Estuary & School of Life Sciences, University of KwaZulu-Natal, Pietermaritzburg, South Africa Allan D. Connell South African Institute for Aquatic Biodiversity, Grahamstown, South Africa Digby P. Cyrus Coastal Research Unit of Zululand, Department of Zoology, University of Zululand, KwaDlangezwa, South Africa Colleen T. Downs School of Life Sciences, University of KwaZulu-Natal, Pietermaritzburg, South Africa William N. Ellery Department of Environmental Sciences, Rhodes University, Grahamstown, South Africa Anthony T. Forbes School of Life Sciences, University of KwaZulu-Natal, Durban & Marine and Estuarine Research, Durban, South Africa Nicolette T. Forbes Marine and Estuarine Research, Durban, South Africa Caroline Fox Ezemvelo KwaZulu-Natal Wildlife, St Lucia Estuary, South Africa Nuette Gordon Department of Botany, Nelson Mandela Metropolitan University, Port Elizabeth, South Africa Michael C. Grenfell Geography, College of Life and Environmental Sciences, University of Exeter, Exeter, UK Suzanne E. Grenfell SEACAMS, Department of Geography, Swansea University, Swansea, UK

List of contributors

Sylvi Haldorsen Department of Plant and Environmental Sciences, Norwegian University of Life Sciences, A˚s, Norway Marc S. Humphries School of Chemistry, University of the Witwatersrand, Johannesburg, South Africa Hendrik L. Jerling Department of Zoology, University of Zululand, KwaDlangezwa, South Africa Bruce E. Kelbe Department of Hydrology, University of Zululand, KwaDlangezwa, South Africa C. Fiona MacKay Oceanographic Research Institute, Durban, South Africa Christopher M. Maine Centre for Research in Environmental, Coastal & Hydrological Engineering, School of Engineering, University of KwaZulu-Natal, Durban, South Africa Andrew Z. Maro Centre for Research in Environmental, Coastal & Hydrological Engineering, School of Engineering, University of KwaZulu-Natal, Durban, South Africa Andrew A. Mather Project Executive: Coastal Policy, eThekwini Municipality, Durban, South Africa Nelson A. F. Miranda School of Life Sciences, University of KwaZulu-Natal, Durban, South Africa David G. Muir School of Life Sciences, University of KwaZulu-Natal, Durban, South Africa Holly A. Nel School of Life Sciences, University of KwaZulu-Natal, Durban, South Africa Sibulele Nondoda Department of Botany, Nelson Mandela Metropolitan University, Port Elizabeth, South Africa Renzo Perissinotto School of Life Sciences, University of KwaZulu-Natal, Durban, South Africa Deena Pillay Marine Research Institute, Zoology Department, University of Cape Town, Cape Town, South Africa Naomi Porat Geological Survey of Israel, Jerusalem, Israel Roger N. Porter Conservation Planner & Ecologist, Ezemvelo KZN Wildlife, Pietermaritzburg, South Africa

xv

xvi

List of contributors

Sean N. Porter Anchor Environmental Consultants, Cape Town, South Africa Justin J. Pringle Centre for Research in Environmental, Coastal & Hydrological Engineering, School of Engineering, University of KwaZulu-Natal, Durban, South Africa. Ursula M. Scharler School of Life Sciences, University of KwaZulu-Natal, Durban, South Africa Derek D. Stretch Centre for Research in Environmental, Coastal & Hydrological Engineering, School of Engineering, University of KwaZulu-Natal, Durban, South Africa Ricky H. Taylor Ezemvelo KwaZulu-Natal Wildlife, St Lucia Estuary & School of Life Sciences, University of KwaZulu-Natal, Durban, South Africa Jane Turpie South African Institute for Aquatic Biodiversity, Grahamstown & Anchor Environmental Consultants, Cape Town, South Africa Jonathan K. Warner School of Life Sciences, University of KwaZulu-Natal, Pietermaritzburg, South Africa Alan K. Whitfield South African Institute for Aquatic Biodiversity, Grahamstown, South Africa

Foreword

View from the Ndlozi Peninsula across North Lake to the coastal dunes in the distance. (Photo: Ricky H. Taylor, July 2011.)

Estuaries are one of the most fascinating and constantly changing ecosystem types on our planet. The size, shape and position of these systems on any coastline is driven primarily by changing sea level and fluctuating climatic conditions, especially river runoff and sedimentation rates. St Lucia is also subject to these primary drivers and it is hard to believe that only 150 000 years ago the estuarine lake system did not even exist! These rapid changes, on a geological timescale, have led to geomorphologists labelling estuaries as ephemeral features of the coast, a reality that biologists seldom take into account when studying the ecology of estuaries. The present-day St Lucia system actually only started taking shape within the last 18 000 years when sea level began to increase once more at the end of the last glacial period. Stabilization at the new increased sea level approximately 6000 years ago gave rise to a recognizable St Lucia system but with two major openings to the sea, one in the north and one in the south. It is only within the last 2000 years that the system was reduced to a single mouth in its present southerly position. Major sedimentation followed this event, particularly during the past few centuries, with

xviii

Foreword

the infilling of the Mkhuze and Mfolozi basins creating extensive swamps in the north and south of the system respectively. Indeed, St Lucia will continue on its trajectory to evolve into a large floodplain – unless global warming causes a dramatic rise in the current sea level. Even then St Lucia will continue to shrink in size due to on-going natural sedimentary processes. Apart from the longer-term evolutionary trends described above, estuaries are also well known as highly dynamic and fluctuating environments on a much shorter timescale. In my opinion there are few natural ecosystems that are subject to the variety and magnitude of external abiotic drivers that estuarine organisms are forced to become adapted to, and the St Lucia system and its biota probably rank very high in this regard. Prolonged droughts, when salinity within its estuarine waters often exceeds twice that of seawater, interspersed by tropical cyclonic deluges which can transform the same waters into an almost freshwater state within a matter of weeks, are experienced by Lake St Lucia every few decades. But it is these large swings of the environmental pendulum that make St Lucia such a fascinating system to study, especially for ecologists. On the basis of the above features we can say with some certainty that Lake St Lucia is an outdoor laboratory that is probably unequalled anywhere else in the world. Not only does it undertake ecosystem shifts that cannot be repeated or simulated within any conventional scientific laboratory, it also pushes plant and animal species to the very brink of their tolerance levels and then back again. This frequent testing of resilience also occurs at the ecosystem level, with species successions and food webs changing dramatically within a matter of months, depending on freshwater inputs. Consequently it is not surprising that St Lucia has captured the dedicated attention of dozens of natural scientists from different disciplines for more than half a century. In attempting to understand the structure and functioning of a complex system such as St Lucia, numerous student theses and hundreds of scientific publications and reports have been produced, each of which has added a piece or pieces to the overall St Lucia jigsaw puzzle. Perhaps what is most surprising of all is that it has taken so long for the above wealth of information to be synthesized and published as a book. In this respect Renzo Perissinotto, Derek Stretch and Ricky Taylor, as well as Cambridge University Press, are to be commended for facilitating the production of an authoritative volume that comes at a very opportune time in the history of St Lucia. Major management changes have recently been proposed and implemented that will alter the evolutionary course of this ecosystem forever. The relinkage of the Mfolozi River to the St Lucia system has long been regarded as an important action for countering the artificially high salinities over the past half century and the almost complete evaporation of the lake earlier this decade. Indeed, it could be argued that unless the permanent relinkage of the Mfolozi takes place soon, the biological and ecological richness of Lake St Lucia may well be permanently impaired or even destroyed. The contents of this book range through all the major biological and earth science disciplines but much of the focus is on St Lucia as an estuary (mainly because most

Foreword

of the chapters are written by estuarine scientists!). Perhaps one of the areas that needs to be studied and highlighted in a future volume is the diverse wetland types located within the greater St Lucia ecosystem. At present there is a diverse array of wetlands on the St Lucia floodplain, ranging from freshwater swamp forests to mangroves, each with its own set of biota. The linkage of Lake Bhangazi South to South Lake through the Nkazana Stream is but one example of floodplain connectivity that is not fully understood from an ecological perspective and is in need of further research. When one realizes that the greater St Lucia system has more than 400 km of wetland interface, the importance of these diverse littoral contact zones becomes even more meaningful. Not only are these estuarine and freshwater wetlands significant habitats to an array of birds, fishes, invertebrates and aquatic plants, they also provide refuge for the single largest population of hippos and crocodiles in South Africa. How many people are aware that there are fewer hippos than white rhinos in the country today – a fact that suggests the original declaration of St Lucia as a ‘game reserve’, more than a century ago, was far-sighted indeed! In conclusion there is little doubt that, as a consequence of this informative new book becoming available to a global audience, there will be an influx of geologists, hydrologists, botanists, zoologists, ichthyologists, ornithologists and other scientists from around the world, eager to study this jewel in the estuarine crown. That interest is likely to grow over time – particularly as St Lucia, the oldest formally protected estuary in the world, is also one of the largest and most productive estuarine water bodies in Africa and is both a Ramsar Wetland of International Importance and a UNESCO World Heritage Site. Indeed, there are great expectations that St Lucia, as part of the recently created iSimangaliso Wetland Park, will live up to this new name (which means ‘Place of Wonder’) and that this book will provide a rich source of information for future work on this fascinating ecosystem. Professor Alan K. Whitfield South African Institute for Aquatic Biodiversity (SAIAB), Grahamstown, South Africa

xix

Preface

Aerial photo of the St Lucia Estuary, with the town, the sea and the mouth in the foreground, the Narrows and South Lake in the distance. (Photo: Ricky H. Taylor, April 2002.)

The St Lucia estuarine lake (sensu Whitfield, 1992, 2000) is the largest estuarine system in Africa (Begg, 1978). It was the first estuary in the world to receive protection, in 1895, and was declared a Ramsar Wetland of International Importance in 1986. In December 1999 it became a core part of South Africa’s first UNESCO World Heritage Site, the Greater St Lucia (renamed iSimangaliso in November 2007) Wetland Park. In Zulu the new name means ‘Place of Wonder’ or ‘A Miracle’, because of its stunning beauty and spectacular views. The purpose of this book is to provide a compendium of accumulated knowledge by scientists and managers that have worked on the system over the last century. The St Lucia system has been the subject of intensive research since the 1950s and through the following decades has attracted research funding and involvement from a wide variety of universities and national research institutions. Almost 200 peer-reviewed scientific contributions and reports had been published on the system by the year 2000 and scores more have been added since. Its main research and management issues have applications to similar systems worldwide. Global change

xxii

Preface

impacts on the system have recently escalated in the form of catchment degradation, land-use changes (e.g. sugar, pineapple and timber), invasion by alien species, increase in climatic extremes and habitat destruction. Currently, the system is at a pivotal stage in terms of the sustainability of its estuarine ecological function. The lessons learnt and the experience acquired in managing the system in a context of strong human-induced changes are particularly relevant on a global scale and should, therefore, be useful to policy makers and managers of similar systems worldwide. It is also intended as a definitive reference for future researchers working on the system and a basis for future multidisciplinary and integrative scientific studies. Students in the areas of marine, estuarine and inland water sciences, as well as environmental managers, resource planners, tourism operators and eco-tourists visiting the park should also regard it as an essential source of information and guidance. This book, however, does not address the complex socioeconomic and political issues associated with management decision making. The book starts with a description of the process of achieving World Heritage status through various stages of conservation protection, starting in 1895. This is followed by a chapter describing the history of management interventions that have been implemented since 1948. This is followed by chapters discussing the St Lucia estuarine lake in the context of its broader meta-oceanic system, its geological evolution, the catchment and groundwater hydrology, and the role of associated wetlands. The hydrodynamic functioning of the system, including the water and salt balances as well as sediments and mouth dynamics, are then reviewed to describe the extreme variability in the physico-chemical conditions that characterize the estuarine system. This is followed by chapters that describe the response of the various biotic components to the environmental drivers of the system. These include the primary producers, their benthic and planktonic consumers and the top predators, including fish, birds and crocodiles. The unique role of hippopotamuses in the system is highlighted in a separate chapter, since their large population is unusual for an estuarine system. A description of the food-web structure is provided in a separate chapter. Key threats, including alien invasive species and climate changes that are globally relevant, are discussed in the final chapters of the book. The estuarine system is essentially made up of three large lake basins in its upper reaches: False Bay, North Lake and South Lake (Figure P.1). These communicate with the Mouth through a narrow channel known as the Narrows. The estuarine lake shoreline on the ocean side is normally referred to as Eastern Shores, while its inland margins are known as Western Shores. The system includes the coastal floodplains of the Mfolozi and Mkhuze rivers (Figure P.1). A wide variety of terminology has been used historically to define the system and its basin components. While most early scientific surveys (e.g. Day et al., 1954; Millard & Broekhuysen, 1970; Begg, 1978) used the general term ‘estuarine system’ or ‘estuary’ to define the whole complex, the local population, conservation authorities and more popular articles have regarded ‘estuary’ as only the lower reaches of the system, essentially from the Bridge to the Mouth. In the official

Preface

FIGURE P.1 Map of Lake St Lucia with details of the iSimangaliso Wetland Park and the geographic position within South Africa.

xxiii

xxiv

Preface

Government Gazetteer, the town itself is reported as ‘St Lucia Estuary’. To complicate matters further, even in the literature the name ‘Lake St Lucia’ has often been used to include only the three lake basins, while the word ‘Estuary’ has popularly been reserved for the area of the Narrows, from the southern end of South Lake to the Mouth (Figure P.1). Because of this rich historical and geographic terminology and, above all, in order to avoid potential confusion among the readers less familiar with this system, we have selected the following terms for the purpose of this book. These will be used as consistently as possible across all chapters. St Lucia estuarine system: includes the estuary, the Narrows, South Lake, North Lake (including Selley’s Lakes), False Bay, Mkhuze Swamps, Mfolozi Swamps, Eastern Shores and Western Shores. The St Lucia system: includes all the above specific areas and the inflowing catchment rivers (Mfolozi, Msunduzi, Mkhuze, Mzinene, Hluhluwe, Mpate, Nkazana Stream, etc.). Lake St Lucia: an estuarine lake and comprises three main lake basins – South Lake, North Lake and False Bay (the Narrows, estuary and the two swamps are not lakes, but they are part of the St Lucia estuarine system – see above). St Lucia catchments or St Lucia catchment areas: used only when the full river catchment area is included. St Lucia Estuary: used to indicate the town of St Lucia, as well as the lower part of the Narrows, from the Bridge to the Mouth. Mouth: the dynamic area where there is tidal sediment movement. iSimangaliso Wetland Park: the broader protected region that includes not only the St Lucia system, but also other important Ramsar wetlands, such as Lake Sibaya and Kosi Bay, among others. Throughout the book, the Practical Salinity Scale (PSS) is used and, therefore, salinity is expressed as a dimensionless unit, as recommended by the Intergovernmental Oceanographic Commission of UNESCO (McLusky and Elliott, 2004; UNESCO-IOC, 2009).

Acknowledgements We would like to express our most sincere gratitude to the team at Cambridge University Press for their courteous and effective coordination and handling of all aspects of this project. In particular, we are very grateful to Dominic Lewis, Commissioning Editor – Life Sciences, and Zewdi Tsegai, Publishing Assistant, for their invaluable assistance and guidance with the editorial process. A comprehensive and demanding peer-review process was undertaken of all contributions submitted for this book. This was brilliantly executed by Edward Bailey, Publishing Assistant – Life Sciences at Cambridge, who managed to collect reports from 33 reviewers with international experience and expertise. A list of all the reviewers, in alphabetic order, is reported here below. Many thanks to Edward and the reviewing team for providing much constructive and useful input towards the improvement of the original drafts. Janine B. Adams, Nelson Mandela Metropolitan University, Port Elizabeth, South Africa Dan Baird, University of Stellenbosch, Stellenbosch, South Africa Guy C. Bate, Nelson Mandela Metropolitan University, Port Elizabeth, South Africa Steve J. M. Blaber, CSIRO Marine and Atmospheric Research, Cleveland, Australia Sven Bourquin, Okavango Crocodile Research, Dibatana, Botswana Chong Ving Ching, Institute of Biological Sciences, University of Malaya, Malaysia Allan Connell, South African Institute for Aquatic Biodiversity, Grahamstown, South Africa Andrew J. G. Cooper, University of Ulster, Northern Ireland, UK Sabine Dittmann, Flinders University, School of Biological Sciences, Adelaide, Australia William N. Ellery, Rhodes University, Grahamstown, South Africa William P. Froneman, Rhodes University, Grahamstown, South Africa

xxvi

Acknowledgements

Charles Griffiths, University of Cape Town, Cape Town, South Africa Sylvi Haldorsen, Norwegian University of Life Sciences, A˚s, Norway Piet Huizinga, ex Council for Scientific and Industrial Research, Stellenbosch, South Africa Herman Hummel, Center for Estuarine & Marine Ecology, NIOO, Yerseke, The Netherlands David Johnson, Hayfields, Pietermaritzburg, South Africa John L. Largier, University of California – Davis, USA Spike McCarthy, University of the Witwatersrand, Johannesburg, South Africa Donal McCracken, University of KwaZulu-Natal, Durban, South Africa Tom Minello, NOAA, Fishery Ecology, SEFSC Galveston Laboratory, Galveston, Texas, USA John Ndiritu, University of the Witwatersrand, Johannesburg, South Africa Christian Nozais, Universite´ du Que´bec a` Rimouski, Que´bec, Canada Dan Parker, Rhodes University, Grahamstown, South Africa Jean-Pierre Pointier, Universite´ de Perpignan, Perpignan, France Simon Pooley, St Antony’s College, University of Oxford, UK Mike Roberts, Department of Environmental Affairs, Cape Town, South Africa Peter Ryan, University of Cape Town, Cape Town, South Africa Ekhart Schumann, Nelson Mandela Metropolitan University, Port Elizabeth, South Africa Ron Uken, University of KwaZulu-Natal, Durban, South Africa Henk Jan Verhagen, Delft University of Technology, Delft, The Netherlands Alan Whitfield, South African Institute for Aquatic Biodiversity, Grahamstown, South Africa Tris Wooldridge, Nelson Mandela Metropolitan University, Port Elizabeth, South Africa

Acknowledgements

Finally, we are very grateful to the University of KwaZulu-Natal Research Office (Durban) for providing a generous contribution towards the publication costs of the colour illustrations contained in the book. We are particularly grateful to Nicola Carrasco and Nelson Miranda for their efforts with organizing the References and Index sections, respectively.

xxvii

Chapter contents 1.1 Introduction 1.2 Early efforts to protect Lake St Lucia: 1895–1947 1.3 Lake St Lucia under divided management: 1947–2001 1.4 The controversial ‘St Lucia’ dune mining proposal 1.5 Decision of the South African Cabinet 1.6 Towards World Heritage Site listing 1.7 Outstanding universal values 1.8 Strengthening conservation measures 1.9 Future management imperatives 1.10 Conclusions

White rhinos on the desiccated lake bed in southern Catalina Bay at the peak of the recent drought. (Photo: Ricky H. Taylor, October 2010.)

1

South Africa’s first World Heritage Site Roger N. Porter

1.1 Introduction The St Lucia system, Africa’s largest estuarine system, lies on the south-eastern coast of KwaZulu-Natal province of South Africa (Begg, 1978). Today this estuarine system forms the major part of the iSimangaliso Wetland Park which gained global recognition when this special place was listed as a natural World Heritage Site by UNESCO in December 1999. This account describes the various events and socio-political circumstances related to the conservation history, the threats, and development of St Lucia as a protected area from the earliest times. It is an account spanning more than 100 years about dedicated people, determined to conserve this important natural wetland system and who opposed the political, social and commercial forces arrayed against them. The area around Lake St Lucia was settled by agriculturists in the Early (AD 250–1000) and Late Iron Age (AD 1000–1840) periods (Maggs, 1984). Shell middens on the coast testify to extensive use of colonies of Perna perna (brown mussel) as a food source (Hall and Vogel, 1980; Maggs et al., 1991). These people may have occupied sites on the coast as early as 1600 years ago and cut fields for crops in the coastal forest. They used fire, burning grassland areas to provide grazing for cattle. The earliest accounts of the St Lucia estuarine system were provided by Portuguese navigators and shipwreck survivors. The name ‘St Lucia’ was originally given to the Thukela River to the south in 1514 by Portuguese sailors and inadvertently

transferred to the lake and estuary in 1575 (Mountain, 1990). Nguni-speaking people that practised shifting cultivation and pastoralism occupied the area in the early nineteenth century having migrated from the Delagoa Bay region in the north (Bryant, 1929; Dominy, 1992). Rivalry between emerging Ndwandwe and Mthethwa chiefdoms and the growth of militarism in the subordinate Sokhulu and Zulu chiefdoms led in 1819 to expansion of amaZulu dominance under the leadership of King Shaka (Wright and Hamilton, 1989). Due to the prevalence of malaria and the cattle disease trypanosomiasis (nagana) extensive areas around the St Lucia estuarine system were uninhabited (Bruton et al., 1980). However, small scattered settlements were present. For example, the French naturalist and hunter Adulphe Delegorgue gave the first written account of the area after his visits between 1838 and 1844 when he encountered the Sokhulu people under Chief Nqoboka Kalanga near the St Lucia Estuary. These people were specialists in iron smelting and collected bog iron from the wetlands. Trees were felled to produce charcoal for their smelters (Hall and Vogel, 1980). King Mpande ceded the St Lucia estuarine system to the British government in 1843 but they did not exercise their rights to the area until after the Anglo-Zulu War of 1879, after which the area was annexed in 1884. In terms of the post-war settlement the southern area of the St Lucia estuarine system

Ecology and Conservation of Estuarine Ecosystems: Lake St Lucia as a Global Model, eds. Renzo Perissinotto, Derek D. Stretch and Ricky H. Taylor. Published by Cambridge University Press. © Cambridge University Press 2013.

2

Roger N. Porter

was allocated to Chief Somkhele of the Mphukunyoni clan and a supporter of King Cetshwayo. Zibhebhu kaMapiyha, an opponent of the king, attacked Somkhele who fled into the St Lucia swamps. Annexation by the British Crown was in response to the Transvaal Boer Republic’s endeavour to establish a port at St Lucia Estuary by expanding their territory and gaining access to the Indian Ocean, and their concern about feared German expansionism in southern Africa. The Royal Navy ship HMS Goshawk sailed to the St Lucia Estuary and the Union Jack was hoisted on Sugar Loaf Hill. Following the Zulu civil war (1883–1885) and the death of King Cetshwayo, Zululand became a British colony on 14 May 1887 with an administration based at Eshowe. Ten years later the Colony of Zululand was incorporated as part of the Colony of Natal on 29 December 1897. The earliest accounts of the Zululand wilderness, written before the outside world interfered, describe the diversity of habitats and the rich complement of plant and animal species. Such accounts were written by a few hunters and naturalists who ventured into this region in the nineteenth century and have been explicitly detailed by Donal McCracken in his book Saving the Zululand Wilderness: An Early Struggle For Nature Conservation (2008). He quotes the poem by Barter written in1897, ‘Of Zululand the paradise/a land of forests, streams, and plains,/Of verdant meads, and gentle rains’. Robert Plant describing elephant in the St Lucia region in 1852 wrote ‘Elephant seen in great plenty all over the district, as we frequently saw herds of them.’ McCracken states ‘large numbers of hippopotamus were

concentrated … in the greater St Lucia area’ in the 1830s, and that there were concentrations of black rhinoceros ‘around St Lucia’. The naturalist George French Angas discovered the nyala antelope just north of St Lucia in 1847 and in his book he illustrated a group comprising adult male and female nyala and calf. Such accounts – for example, John Dunn boasted of shooting 203 hippos in a season, 23 in one morning ‘before 10 o’clock’ (McCracken, 2008) – and those later by Drummond (1875) and Baldwin (1894) who hunted in the area in 1851 and 1853, attracted hunters to Zululand and ultimately resulted in the overhunting and slaughter of the game, with elephant, buffalo, rhinoceros and hippopotamus being the main victims. Ivory and hides were Natal’s principal export commodities to Britain and Europe until 1862. Wildlife was seen as an exploitable and unlimited resource and by the 1880s hunting had severely and to a large extent depleted the ungulate and carnivore fauna, with commercially valuable species such as elephant, hippopotamus and buffalo populations being radically reduced. Awareness and recognition of wild animal exploitation in Zululand grew and received the support of Sir Michael Clarke, Resident Commissioner of Zululand, who commented that ‘we may be within measurable distance of the total extermination of game in Zululand’ (Ellis, 1975). The British colonial administration of Zululand acted and promulgated legislation to regulate hunting in the region during the 1890s. In 1906, the Natal and Zululand game laws were consolidated, and in 1912, the laws regarding coastal fishing were extended to include the Zululand coast.

1.2 Early efforts to protect Lake St Lucia: 1895–1947 In addition to the laws on hunting proclaimed initially in 1890 and subsequently in 1893, 1897 and 1906 (McCracken, 2008), there was a realization that areas where no hunting could take place were needed

if the wildlife was to recover. The Zululand colonial government resolved to set aside five such ‘reserves’ in 1895 which included the ‘St Lucia Reserve’ where hunting was prohibited and periodic patrols were

South Africa’s first World Heritage Site

hunted (McCracken, 2008). These protected areas, together with the Pongola Game Reserve, are therefore the oldest extant game reserves on the African continent and amongst the oldest in the world (Ellis, 1975). However, in 1898 the South Africa General Mission established the Mount Tabor Mission Station on the Eastern Shores headed by a Norwegian, the Rev. L. O. Freyling (R. H. Taylor, pers. comm.). A chapel, school, health facilities and accommodation for staff were constructed. The mission continued to function until the mid 1950s when the Department of Forestry took over the Eastern Shores, demolishing structures and replacing them with new accommodation for their staff (Dominy, 1992). This site is now known as Mission Rocks. During the ensuing decades additional areas were identified and proclaimed, thus enlarging the area protected. These were:

FIGURE 1.1 The first diagram of the St Lucia Reserve was published as part of Zululand Government Notice 16 April 1897. Note that the Eastern Shores were part of the protected area.

undertaken to enforce the hunting laws (Zululand Government Notice 12 of April 1895). Later in terms of a new law that provided for the establishment of ‘Game Reserves’, the status of four of the original reserves, including the St Lucia Reserve, was reaffirmed by Zululand Government Notice 16 of April 1897 (Figure 1.1). However, hunting was allowed subject to restrictions under the law and regulations such as a limit on the number of animals that could be shot, closed seasons, and accordance with game schedules. The schedules had four categories: game that had no protection, game that was protected during a closed season, royal game (species such as hippopotamus that were subject to a higher hunting licence fee), and royal game special category (e.g. elephant) that could not be

• St Lucia Game Reserve (36826 ha) Provincial Notice 74 of 1916, 59 of 1917, 108 of 1935, 140 of 1939, and 35 of 1939. • Hlabisa Game Reserve, proclaimed in 1905, that stretched from False Bay up to and including the Hluhluwe and Imfolozi game reserves, and south to Cape St Lucia. • Mkuzi Game Reserve (37 985 ha) – Provincial Notice 23 of 1912, 28 of 1912, 74 of 1916, 57 of 1917, 266 of 1918, 140 of 1939, and 131 of 1941. • St Lucia Bird Sanctuary – the entire lake and its eastern shores – Provincial Gazette proclamation 7 of 1927. • St Lucia Park (12 545 ha) – proclamation 35 of 1939, 11 of 1944, and 36 of 1957. • False Bay Park (2247 ha) – proclamation 9 of 1944 and 111 of 1952. The proclamation of the St Lucia Bird Sanctuary in 1927 is particularly noteworthy as, for the first time, there was legal recognition to preserve an entire lake ecosystem; this at a very early stage in South Africa’s conservation history and development.

3

4

Roger N. Porter

The first game conservator in Zululand, Mr W. E. Pettie, was appointed in 1896 and was responsible for the protection of the four Game Reserves, including St Lucia Reserve No.1, and for implementing the game laws (McCracken, 2008). He was assisted initially by the Zululand Police and game constables Mr D. D. Tweedie and Mr S. Silverton. The Lands Delimitation Commission of 1902–03 demarked areas in Zululand that became available for settlement by white people, mainly cattle farmers. Many of these areas were near the game reserves and had been avoided by the Zulu people because of the prevalence of tsetse fly that carried nagana disease. In 1910 the four British colonies joined together to form the Union of South Africa and, under the Union constitution, the control of game and fish was delegated as a competence to each of the four provinces. The Natal Provincial Administration appointed Mr F. Vaughan-Kirby as Game Conservator for Zululand in 1911 and was stationed far away from the game reserves at Nongoma where he had access to a telephone and was able to visit the game reserves by motorbike. He was assisted by a small staff of Zulu men and served until his retirement in 1928. He was succeeded by Mr R. Symons who resigned two years later. In 1929 Captain H. B. Potter was appointed to the post and was stationed at Hluhluwe Game Reserve from where he travelled to all the Zululand game reserves to engage with his field ranger staff until his retirement in 1950. During this time the incidence of nagana in cattle increased and led the cattle ranchers to carry out the wholesale destruction of game outside the game reserves from 1910. They pressed for greater control of game populations, even calling for their extermination, especially following the establishment of the Ntambanana settlement near Umfolozi Game Reserve in 1919. In a study of the disease conducted in the 1890s in the Pongola– Mkhuze game reserve areas, Surgeon Major David Bruce established that nagana was transmitted by the

tsetse fly. Special shooting areas were established in 1916, many bordering the game reserves (McCracken, 2008). The Veterinary Department was brought in to undertake a research programme to determine methods to eradicate the tsetse fly (Glossina spp.) and thereby eliminate the nagana disease from Zululand in 1920. Except for Hluhluwe Game Reserve, all other reserves in Zululand were under the control of the Veterinary Department who implemented a policy to eradicate the larger ungulates, the host of the trypanosome, except for rhinoceros. Extensive game eradication programmes which included periodic game drives intensified from about 1915 until 1950 (McCracken, 2008). Umfolozi Game Reserve was de-proclaimed in 1920 to allow for this eradication of game animals. Later in 1930 the game reserve was re-proclaimed but continued under the control of the Veterinary Department. After 33 years as the St Lucia Game Reserve, the area was de-proclaimed in Proclamation 20 of 1928 to allow for the Veterinary Department to take over and implement their programme to eradicate the tsetse fly. Woody vegetation was also removed from bush-cleared zones in an attempt to confine the fly to the game reserves. From the mid 1940s aerial spraying of the insecticides DDT and BHC particularly over the game reserve areas and in the river catchments of the lake led to the eradication of the disease in Zululand (Minnaar, 1989). These circumstances soon brought about a conflict between the Natal Provincial Administration responsible for the preservation of fish and game in the province, and the departments of Agriculture, Lands and Veterinary Research which were responsible for the study and the elimination of the source of the disease (Ellis, 1975). Also concern about the on-going slaughter of the game animals and the fate of the former Zululand game reserves grew amongst the Natal public and legendary personalities such as Dr G. Campbell of the newly formed Natal branch of the Wildlife Society, Deneys Reitz, J. Stevenson-Hamilton, Captain H. B. Potter, Dr E. L. Gill, and government Minister P. G. W Grobler (McCracken, 2008). In response a

South Africa’s first World Heritage Site

‘Game Reserves Commission’ was appointed in 1935 by the Administrator of Natal. Importantly, the commission recommended the establishment of the Zululand Game Reserves and Parks Committee which when established in terms of Natal Provincial Ordinance of 1939 was known as the Zululand Game Reserves and Parks Board under the chairmanship of Mr W. Power. The St Lucia Game Reserve had been re-proclaimed in March 1938 (Natal Provincial Notice 108 of 1938) but with reduced boundaries from those of the original protected area; that is, the Eastern Shores that had previously formed part of the 1927 St Lucia Bird Sanctuary was not included in the re-proclaimed protected area. The new boundary became the high water mark of the estuarine system including the islands, but excluded the surrounding terrestrial areas. It was subsequently enlarged in 1939 when a narrow terrestrial area that incorporated the ‘half-mile wide’ strip of land following the lake shore was added and proclaimed as St Lucia Park. Then in 1944, during World War II, the False Bay Park was added and proclaimed (Ellis, 1975). A large expanded area of the Mfolozi River floodplain, known as the Mfolozi Flats, was opened up for sugar cane plantations in 1927. Little or no thought was given to potential impacts on the St Lucia Estuary or that the river was subject to periodic largescale flooding, inundation and sedimentation. In an attempt to prevent flooding of the farms, the Mfolozi Swamps were canalized in 1936, and the water drained off. Later the river was confined by the construction of large levees. The mouth of the Mfolozi was separated from the St Lucia Mouth in 1952. River-borne sediment therefore was carried further downstream, deposited at the mouth of the Mfolozi River, and carried by the north-flowing sea current closing the estuary mouth. The St Lucia and Mfolozi estuary mouths closed from April 1951 to April 1956 causing fish populations in the lake to decline and leading to an outcry by recreational fishermen. Dredging operations to keep the mouth open commenced in 1957 and continued on a regular basis for the next four to five decades (Hill, 1975a; Orme, 1990).

During World War II a radar station and Royal Air Force base were constructed at Mount Tabor near Mission Rocks. In November 1942, a jetty was built at Catalina Bay, and Catalina flying boats operated out of Lake St Lucia undertaking coastal patrols from February 1943. Between June 1943 and October 1944 these aircraft conducted extensive offshore antisubmarine patrols. Several German and Japanese U-boats were attacked with varying success and Catalina FP226/J suffered damage to the port engine when U 198 returned fire after the aircraft had dropped its depth charges on the night of 19 May 1943. FP226/J flew back to Lake St Lucia on one engine. Flying boat Catalina FP275/E of 259 Squadron crashed into Catalina Bay on 7 June 1943 when an engine failed while landing. A second accident occurred on 25 June 1943 when Catalina FP265/H of 262 Squadron crashed into the lake probably as a result of a stall on take-off. Due to concerns about shallow water levels in the lake causing an accident when aircraft were at full war load on take-off, the squadrons were relocated to Lake Msingazi, Richards Bay, by February 1945 (Spring, 1995). The 1939 Ordinance was the precursor of the Natal Parks Game and Fish Preservation Board established in terms of Natal Ordinance 35 of 1947. This body was known as the Natal Parks Board and assumed control of the game reserves including the St Lucia protected area complex. Up until 1997 it was responsible for the implementation of nature conservation management and development programmes in the province. During these years nature conservation in Natal was served by and had the support of several influential members of the Natal Provincial Administration, in particular Mr A. E. Charter (Provincial Secretary appointed in 1928) after whom Charter’s Creek was named, Mr W. M. Power and Mr D. E. Mitchell (Administrator of Natal from 1944 to 1948 and after whom Mitchell Island was named).

5

6

Roger N. Porter

1.3 Lake St Lucia under divided management: 1947–2001 From the time of the formation of the Natal Parks, Game and Fish Preservation Board (NPGFPB) in 1947 and the appointment of Colonel Jack Vincent as the first secretary (later the post was upgraded to that of director) in October 1949, management staff was stationed at the St Lucia estuarine system and tourist facilities were constructed at Charter’s Creek, Fani’s Island and False Bay (Crass, 2011). Captain H. B. Potter, Chief Conservator for Zululand, had become concerned about the natural state of Lake St Lucia as a consequence of the effects of a drought period during 1946 and 1947. The matter was brought to the attention of the NPGFPB which provided ₤1000 in funding, and together with a grant from the Carnegie Foundation, a scientific investigation by eight biologists under the leadership of Professor J. H. Day took place in July 1948. For the first time a critical, although preliminary, scientific analysis of the various biotic and abiotic factors that influence the ecology of Lake St Lucia was officially reported (Day, 1948). He concluded his report by stating ‘We have made one brief visit to St Lucia and our conclusions are necessarily tentative. We have found high salinity in the lakes and silting at the mouths. These factors contribute to the scanty aquatic vegetation, poor bottom fauna and thus the small number of fish’. In spite of such early warnings, infrastructure and agricultural developments and land-use change in the upstream catchments over the next almost half a century were to have a profound environmental effect on the St Lucia estuarine system. However, in an attempt to create a larger protected area, additional conservation areas were set aside and proclaimed under both national and provincial legislation. Subsequently these subcomponent areas have been consolidated to form a single protected area named the Greater St Lucia Wetland Park. The name was later changed after a process of public consultation and given the culturally significant name iSimangaliso Wetland Park. The meaning of

this isiZulu word is ‘If you have seen miracles, you have seen the flat land and lakes of Thonga.’

1.3.1 Development of afforestation In 1956 the Department of Forestry commenced an aggressive commercial afforestation programme in the Eastern Shores, Nyalazi and Dukuduku State Forests. A plantation exceeding 5244 ha mainly of Pinus elliottii was established on the Eastern Shores of the lake (Stubbings and Venter, 1992). With the exception of the Dukuduku State Forest which was gazetted in 1930 (10 125 ha – Government Notice 1479 of 1930), the following State Forest areas were gazetted at this time: • Cape Vidal State Forest (11 313 ha) – Government Notice 1408 of 1956. • Sodwana State Forest (47 127 ha) – Government Notice 1408 of 1956. • Eastern Shores State Forest (12 873 ha) – Government Notice 1408 of 1956, Government Notice 14246 of 1992. • Nyalazi State Forest (1367 ha) – Government Notice 1408 of 1956.

1.3.2 ‘Kriel’ Commission of Enquiry The controversial development of plantation forestry in the river catchments of Lake St Lucia and government’s plan to build a large storage dam on the Hluhluwe River provoked a public outcry in the 1960s. The government responded by appointing a ‘Commission of Enquiry into the alleged threat to animal and plant life in St Lucia Lake’ (Kriel Commission) in 1963. The commission’s report was tabled in parliament in 1966 in which Lake St Lucia was described as a ‘unique environment’ with great potential for conservation and tourism. The report recommended that the protected area be increased in size including the incorporation of the Eastern Shores State Forest, as a matter of ‘extreme importance’; as

South Africa’s first World Heritage Site

well as the phasing out of plantation forestry with no new plantations to be established especially on the Eastern Shores area. Also recommended was that the enlarged protected area should be managed by a ‘single body’ with the necessary experience and executive powers (Kriel et al., 1966). Although the Kriel Commission marked a turning point in the environmental history of the St Lucia estuarine system because it commissioned a number of specialist scientific investigations and recommended a larger protected area, none of the important recommendations were implemented, being simply ignored by the state (Frost, 1990). The protected area was not enlarged, the Eastern Shores was not incorporated, plantation forestry was aggressively implemented from 1968 including the planting up of the Eastern Shores, water quotas were not provided (for example, from the Hluhluwe Dam), and control of the area continued to be divided with national departments holding sway over the management advice and recommendations of the Natal Parks Board. Worst of all, the Commission’s recommendation that the whole area should be given legal protection in a consolidated conservation area under a single management policy was not implemented (Frost, 1990). Communities numbering approximately 5000 people that were living within the State Forest areas, mainly the Eastern and Western Shores and Sodwana area, were evicted, relocated and resettled in areas west of the St Lucia estuarine system.

1.3.3 St Lucia Advisory Council Although the scientific study of the St Lucia system had tentatively begun in 1948, it was the outcome of the Kriel Commission 20 years later (1968) that led to the appointment of the St Lucia Scientific Advisory Council (SCADCO). This was at a time when severe drought in Zululand caused high salinity conditions in the lake as a consequence of the estuary mouth closing combined with high rates of evaporation. Importantly many biological, hydrological and engineering studies were undertaken in ensuing

years building a body of knowledge and understanding of this diverse, outstanding, natural ecosystem. Many recommendations of SCADCO were implemented including scientific studies and dredging operations to keep the estuary mouth open, the excavation of the ‘Mercer Canal’ and Intake Works to allow for fresh water to flow into the Narrows from the Mfolozi River, and the ‘Van Niekerk Canal’ to increase flow of fresh water through the Mkhuze Swamps into the northern compartment of the lake. These engineering solutions were undertaken in an absence of an environment impact assessment. Adverse environmental impacts resulted, especially as a consequence of the huge inflows of seawater, which with evaporation created hypersaline lake conditions, causing a die-off and loss of biota as the drought continued, and extensive erosion and drying out of large parts of the Mkhuze Swamps.

1.3.4 A ‘greater’ protected area The Natal Parks Board was assigned nature conservation and recreation duties under certain conditions, and the responsibilities to manage and control certain demarcated areas of the Eastern Shores State Forest by government in terms of the Forest Act in June 1977. The areas included specified locations such as Cape Vidal, Lake Bhangazi South and areas outside plantations but excluded the military area. Thus a period of joint management of the Eastern Shores between the Department of Forestry and the Natal Parks Board commenced, requiring wildlife management activities to be negotiated well in advance. Then after years of neglect to protect coastal and marine living resources, the government decided to establish several marine protected areas on the country’s coast. In 1979 the St Lucia Marine Reserve (44 280 ha) was proclaimed and the Natal Parks Board appointed as the management authority for the area (Proclamation 35 of 1979 and 48 of 1986). In 1982 the government decided to excise the Ingwavuma district and hand the area to Swaziland

7

8

Roger N. Porter

as compensation for their land that was now part of the inundated area of the Pongolapoort Dam. In so doing the government would create a large game reserve between Mkhuze and Lake St Lucia, incorporate the Eastern Shores, and transfer the whole area to the Natal Parks Board. The matter was hugely controversial and taken to the Appellate Division of the Supreme Court. The court ruled against government as the rights of the inhabitants had been infringed and the KwaZulu Government had not been consulted. Government cut-backs on budget allocations to several departments so as to increase the military budget led the State President to transfer control and management of the following State Forests (in terms of the Forest Act No. 122 of 1984) to the Administrator of Natal in November 1986 (State President’s minute No. 1109): Cape Vidal, Sodwana, and Mhlatuze (including the Mapelane Nature Reserve). The Administrator then assigned these to the Natal Parks Board (Executive Committee Resolution, 1987). By 1991, plantation forestry was no longer commercially viable. The Department of Forestry decided to phase out the plantations on the Eastern Shores and part of the Nyalazi Plantation, particularly trees that were within the Mpate River catchment growing on the hydromorphic soils. The following state forests were transferred to the Administrator of Natal in 1993 to be used for the purpose of nature conservation and to be incorporated as part of the then Greater St Lucia Wetland Park: • Mhlatuze State Forest and Mapelane Nature Reserve (1103 ha) – Government Gazette 402 of 1916 and 9512 of 1984. • Dukuduku State Forest, Nyalazi State Forest and Umfolozi Swamps (10 125 ha) – Government Notice 1479 of 1930. With the establishment of the KwaZulu Bureau of Nature Conservation, protected areas were

proclaimed to the north of the Sodwana State Forest following the proclamation (Government Notice 404 of 1986) of the Maputaland Marine Reserve (39 740 ha), as follows: • Coastal Forest Reserve (21 772 ha) in 1992 • Lake Sibayi Freshwater Reserve (7218 ha) in 1994.

1.3.5 Military activities The South African Defence Force established a small missile testing facility on the Ndlozi Peninsula at Hell’s Gate in 1968 and their armourment companies Kentron and Denel undertook development and testing of missiles including those with destructive charges during the 1970s and 1980s which were dropped on the land and within the lake where these could be recovered. During such test periods the lake had to be closed to tourism and management staff warned not to be operating in the test area. The Defence Force also conducted field training exercises for soldiers who also used live weapons. Army personnel were later withdrawn, the base was subsequently little used after about 2000 and the Ndlozi Peninsula area was incorporated into the consolidated World Heritage Site by proclamation in November 2000. The base was progressively closed down and buildings demolished by about 2008.

1.3.6 Wetlands of International Importance Given their national wetland biodiversity importance and conservation status the Department of Environmental Affairs supported the Natal Parks Board’s submission for two sites, namely the ‘St Lucia System’ and the Turtle Beaches and Coral Reefs of Tongaland. The submission was compiled by Mr R. Porter and Dr H. Grobler and presented at the Ramsar Convention’s Conference of Parties in 1986 as candidate sites to be designated as Wetlands of International Importance. South Africa was the fifth

South Africa’s first World Heritage Site

contracting party to the Ramsar Convention (March 1975) and thus had accepted certain international commitments as well as to conserve wetlands and make wise use of them within the country. These commitments included the promotion of international cooperation in wetland conservation, the fostering of communication about wetland conservation and supporting the work of the Convention. This Convention recognizes the interrelationship of wetland systems with their supporting systems and surrounding areas. The submission for the two sites was accepted and the two sites were designated by the Ramsar Convention as Wetlands of International Importance in October 1986. The St Lucia System (155 000 ha) qualified as a Wetland of International Importance in that it satisfied all three of the Convention’s criteria (and most of the subcriteria) for designation, as follows: • The St Lucia System is an excellent representative example of a natural wetland characteristic of its biogeographic region – Criterion 1(a). • It is a particularly good representative example of a wetland, which plays a substantial hydrological, biological and ecological role in the natural functioning of the south-eastern African coastal system – Criterion 1(c). • The St Lucia System supports an appreciable assemblage and populations of rare, vulnerable, and endangered species of plants and animals – Criterion 2(a). • It is of special value in maintaining the genetic and ecological diversity of the region given its substantial fauna and flora – Criterion 2(b). • It is of considerable value as a habitat of plants and animals at a critical stage of their biological cycle – Criterion 2(c). • The St Lucia System regularly supports 20 000 waterfowl – Criterion 3(a). • It regularly supports substantial numbers of individuals from several groups of waterfowl, indicative of wetland values, productivity and diversity – Criterion 3(b).

• It regularly supports 1% of the individuals in a population of one (or more) species of waterfowl – Criterion 3(c). At the 1990 Montreux conference it was agreed that a Register of Ramsar sites would be compiled (named the Montreux Record) where a ‘change in ecological character of wetlands sites is taking place, may take place, or already had taken place’ so as to give priority to the international monitoring of these sites. Following South Africa’s report at this conference, contracting parties expressed grave concern at the potential impacts of mining on the St Lucia System site and resolved to add it to the Montreux Record.

1.3.7 Divided control by government continued Decisions by the South African government during this time were contradictory and led to management of the area being divided between state and provincial bodies. The government’s Department of Mineral and Energy Affairs had granted a mining lease in the 1970s. In 1986 the Department of Environmental Affairs supported the designation of the area as a Wetland of International Importance while the threat of mining still remained and the government’s Department of Forestry was aggressively expanding plantation forestry while the lake system was suffering significant freshwater shortages. The Department of Defence tested missiles with live war heads and allowed for the area to be used for nature-based tourism under the management of the Natal Parks Board. Although Richards Bay Minerals indicated (1989) that they would exercise their right to mine the Eastern Shores, the government announced their intention on 9 February 1990 to assist the Natal Parks Board to establish a major protected area to be called the Greater St Lucia Wetland Park. In August 1992 management of the Eastern Shores State Forest was transferred from the Department of Forestry to the Natal Provincial Administration.

9

10

Roger N. Porter

1.4 The controversial ‘St Lucia’ dune mining proposal 1.4.1 The environmental impact assessment process In 1972 and 1976, dune sand-mining prospecting leases in the 3460 ha coastal dune area of the Eastern Shores State Forest were granted to the mining company Richards Bay Minerals. Results of mineral exploration showed substantial deposits of titanium ore (ilmenite) and the company applied to the Department of Mineral and Energy Affairs in 1989 for a mining authorization. Considerable public opposition resulted given the numerous environmental concerns that were raised; a public petition received 222 667 signatures. The government announced that an environmental impact assessment (EIA) would be undertaken and Coastal and Environmental Services were contracted by Richards Bay Minerals to undertake the work. In March 1990 Dr J. P. Raimondo of the University of Cape Town was appointed as specialist consultant for the coordination and management of the EIA process. A Coordinating Committee chaired by the Department of Environmental Affairs was formed and was responsible for reporting back to the South African Cabinet. Specialists were appointed to report on 23 identified matters of concern. These were: landform, geomorphology and geology, hydrology, water supply, sedimentology, climate, soils and land-use potential, vegetation, rehabilitation, animals, wetlands, archaeology, history, visual impacts, landscape evaluation, noise, tourism, economic impacts, and the Ramsar Convention. The Specialist Reports was sent to 14 lead interested and affected parties (I≈ total number of registered was 120) in late 1991 (Volume 1 of the CSIR). On 1 December 1991 the Council for Scientific and Industrial Research (CSIR, 1993a, 1993b, 1993c) replaced Coastal and Environmental Service as the lead consultants responsible for compilation of Key

Issue Report (Volume 2 of the CSIR) and the final EIA (Volume 3 of the CSIR). Dr A. v. B. Weaver was project manager assisted by Dr B. W. van Wilgen and Dr K. J. Kruger as principal consultants. Key issues identified were: functioning of terrestrial ecosystems, wetlands, estuarine and marine systems, biotic diversity, effects on tourism, economy, community life and social services, plans, policies and laws, transport of heavy metal concentrate, economic evaluation of mining and ecotourism alternatives.

1.4.2 The nature conservation and tourism alternative land use to mining In a bold move the Natal Parks Board announced in 1991 its concept development plan for the Greater St Lucia Wetland Park that identified the Eastern Shores as a strict nature reserve without forestry, which was to be phased out, and which also precluded resettlement (Bainbridge et al., 1993). The document was therefore a formal alternative government policy to mining at St Lucia. This allowed for the EIA process to compare the environmental implications between two alternative land uses, namely nature conservation and tourism, or mining with nature conservation and tourism. The nature conservation and tourism option called for the removal of forestry operations and the rehabilitation to natural vegetation after clear felling of the pine plantations, the reintroduction of game, and the development of tourism facilities. The mining option also called for the removal of forestry operations and the rehabilitation of the natural vegetation of the area, but allowing mining to proceed in conjunction with the nature conservation and tourism activities where feasible. After mining, the area would be rehabilitated to natural vegetation and would revert to conservation and tourism.

South Africa’s first World Heritage Site

Given South Africa’s obligations in terms of the Ramsar Convention, the Department of Environmental Affairs in a formal statement informed the Bureau and the Conference of Contracting Parties held at Montreux in 1990 that there was a potential for the ecological character of the St Lucia System Ramsar Site to change as a consequence of possible future titanium mining on the Eastern Shores. The Conference called upon the South African government to (a) prohibit any mining activity which would damage the ecological character of the St Lucia System site, and (b) ensure that the St Lucia System be retained as a protected site because of its national and international importance. An international Monitoring Mission undertook a site visit during April and May 1992. Their report was received by the Minister of Environmental Affairs in February 1993 in which it was recommended that the South African government should consider that the mining be refused on principle, that a broad plan for the wise use of the whole of the Maputaland coastal plain be developed, and should the government decide to give further consideration to the possibility of mining, special consideration be given to: alternative sources of ore, undertaking an environmental cost/benefit analysis, consideration of aspects where critical impacts on the St Lucia System are identified, and taking note of the difficulties of restoration, amongst other concerns (CSIR, 1993d).

assessment, and additional factual information. Written comments were also submitted by 135 organizations and individuals, and general comments mostly opposing the mining alternative were made by a further 248 people in letters sent to the consultants. Considerable public sentiment and opposition to mining at St Lucia grew substantially with the formation of action and activist groups. These included the Natal Branch of the Wildlife Society led by Dr E. A. (Nolly) Zaloumis, which published many articles and letters, and the Zululand Environmental Alliance (ZEAL) with a membership of several thousand persons, which continuously wrote letters to the media, contacted decision makers and held public meetings. These organizations joined forces to form the ‘Campaign for St Lucia’. It was a loose association of people and organizations bound only by a Mission Statement. They held press conferences, wrote letters to politicians, met with the chairman of the parent mining company, instigated antimining petitions including a postcard petition where the public voted to prevent the mining, and published information brochures on St Lucia (Senogles, 2012). By July 1993 communication by the media resulted in some 1155 reports which appeared in 148 publications. The level of public participation and concern, both local and international, was therefore unprecedented in EIA processes in South Africa.

1.4.3 Involvement of interested and affected parties

1.4.4 Review Panel hearings

The EIA was undertaken by the CSIR and the report released for public comment in March 1993. Ten lead I&AP provided written comment with substantive comment being submitted by the Natal Parks Board, KwaZulu Bureau of Natural Resources, Richards Bay Minerals, Wildlife Society of Southern Africa and the Zululand Environmental Alliance. Such comment focused on matters of process, omissions, inaccuracies, risks, consistency, criteria, impact

Following the EIA report review process, public hearings were held by an appointed independent, five member Review Panel under the chairmanship of former Justice R. Leon. The Review Panel was required to evaluate the EIA process, review the Environmental Impact Report, ensure full public disclosure, conduct public hearings, and give judgement on the relative significance of the issues contained in the report. It was to present its report to the Cabinet and recommend on (a) whether or not

11

12

Roger N. Porter

it found any ‘unacceptable damage’ caused by either of the land-use options, (b) the preferred land-use option, and (c) the conditions under which the preferred land-use option should proceed. Both the Chairman Mr P. Goss and Chief Executive Officer Dr G. Hughes of the Natal Parks Board presented concluding arguments to the Review Panel. Five principal arguments were given in opposition to the mining alternative, namely: St Lucia is a special place; nature conservation is not compatible with mining; there is no substitute for the Eastern Shores; impacts from mining would be significant and major on the natural environment; and there were many risks and uncertainties associated with mining. The panel concluded that no mining should be allowed in the Greater St Lucia area (Leon et al., 1993). In the panel’s judgement, mining the Eastern Shores would cause unacceptable damage to a place which is special because of its rich history, ecological and biological diversity and the significance it has

for many visitors. Referring to this unique combination, the panel stated that there was no substitute for St Lucia and recommended that the necessary steps be taken for the area to acquire World Heritage Site status. Mindful of the plight of the original inhabitants and the treatment they had received from Apartheid government departments and officials, the panel recommended that the area be managed by a new Board, established by Act of Parliament, and that in the interim a new name (i.e. National Heritage Park) be used replacing that of the Greater St Lucia Wetland Park. These latter recommendations were to ensure that direct benefits would accrue to both local communities and the nation. The government considered the panel’s report and decided, in view of the upcoming elections, that the final decision would be left for the new government to take. In April 1994 the Nationalist Party lost the elections to the African National Congress, issuing in a new government of national unity.

1.5 Decision of the South African Cabinet The Minister of Agriculture appointed the Land and Agriculture Policy Center (LAPC) to advise on the findings contained in the Review Panel report. Consultants were appointed to submit reports on several specific matters that had developed in the two to three years subsequent to the panel’s recommendations. A synthesis report was then compiled for submission to Minister D. Hanekom. The Natal Parks Board then submitted a report on the present level of development and possible future development of St Lucia to the Department of Environmental Affairs for Minister D. de Villiers’s attention in preparation for his select Cabinet meeting with Ministers D. Hanekom, K. Asmal (Water Affairs and Forestry), B. Ngubane (Arts and Culture), and P. Botha (Mineral and Energy Affairs). It was again recommended that there be no mining nor

resettlement in the area, the findings of the LAPC be endorsed, and that it was the submission of the Natal Parks Board that nature conservation and tourism should be developed so as to achieve its potential to be a lead economic sector in the subregion (Porter, 1996). On 6 March 1996 and after years of controversy over the prospect of dune mining at St Lucia, the Cabinet announced its considered decision to disallow mining on the Eastern Shores. It was in favour of an integrated development and land-use planning strategy for the St Lucia region that would enable various sectors to work collectively towards the common goal of eradicating the region’s poverty and promoting sustainable development. The Cabinet considered that the tourism potential of the region could now be fully exploited and it also

South Africa’s first World Heritage Site

decided that the submission for St Lucia to be listed as a World Heritage Site must go ahead

urgently (Government of South Africa: Joint Statement, 1996).

1.6 Towards World Heritage Site listing During the period while the EIA was being undertaken, the Natal Parks Board was of the view that the Greater St Lucia Park was of international importance in addition to its Ramsar status and, as such, a potential natural World Heritage Site (CSIR, 1993d). The matter was also raised by a number of people and organizations in their opposition to the area being mined. The response by the CSIR consultants was ‘that the area with or without mining would presumably qualify as a World Heritage Site’ (CSIR, 1993c). The Review Panel in its report stated ‘that it is important that the necessary steps be taken as soon as possible to have the area declared a World Heritage Site’. The initial nomination dossier for the Greater St Lucia Wetland Park to be listed as a natural World Heritage Site was prepared by the Natal Parks Board with contributions from the KwaZulu Department of Nature Conservation and submitted under the signature of Dr F.T. Mdalose, Premier of KwaZulu Natal, to the Department of Environmental Affairs in 1994 (Porter et al., 1994). However, the submission was made only a few months after South Africa’s democratic elections and change of government. Only after South Africa was readmitted into the United Nations could the country ratify several international conventions. The UNESCO World Heritage Convention was ratified in July 1997 and the government established policy, legal and institutional systems for the implementation of the convention and the management of future possible World Heritage Sites following a visit by Dr B. von Droste, Director of the Paris-based World Heritage Centre. The Minister of Environmental Affairs was appointed the cabinet minister responsible for the implementation of the convention in South Africa. The Department of Environmental Affairs together with provincial structures and through the South

African World Heritage Convention Committee advises the Minister on all matters associated with the implementation of the convention both nationally and internationally. This is done within the framework of the World Heritage Convention Act passed by the parliament in 1999 (Porter et al., 2003). A period for public review and comment on the nomination dossier for World Heritage Site listing was provided. Many comments were received from individuals and organizations with about 70% in favour of the proposal. A Response Report was compiled and submitted to the Department of Environmental Affairs (Porter, 1995). The 1994 nomination dossier for St Lucia was withheld by the Department of Environmental Affairs until the first meeting of the South African World Heritage Committee in 1997 when it was agreed that several nomination dossiers including those for Robben Island and the Cradle of Humankind candidate sites be finalized and jointly included in South Africa’s submission. In the meantime the format for a World Heritage Site nomination had changed substantially, requiring a rewrite of the dossier for the Greater St Lucia Wetland Park (Porter et al., 1999). The three nomination dossiers were submitted by the Department of Environmental Affairs and South Africa’s Ambassador to France, to the World Heritage Centre, Paris, by July 1998. Following South Africa’s democratic elections in 1994 and the scrapping of Apartheid laws and homelands, a process was undertaken that led to the amalgamation of the Natal Parks Board and the KwaZulu Department of Nature Conservation in December 1997. A new conservation authority, the KwaZulu Natal Nature Conservation Service was born. This organization is generally referred to as Ezemvelo KZN Wildlife (EKZNW).

13

14

Roger N. Porter

The IUCN appointed its Senior Advisor, Dr J. Thorsell, to undertake the Evaluation Mission and report to the Director of the World Heritage Centre, which he did at the meeting of the World Heritage

Convention Committee held in Marrakech in December 1999. The Greater St Lucia Wetland Park became South Africa’s first World Heritage Site to be accepted and listed by UNESCO.

1.7 Outstanding universal values The World Heritage Convention’s general statement of the outstanding universal values of the iSimangaliso Wetland Park is as follows: The iSimangaliso Wetland Park is one of the outstanding natural wetland and coastal sites of Africa. Covering an area of 239 566 ha, it includes a wide range of pristine marine, coastal, wetland, estuarine, and terrestrial environments which are scenically beautiful and basically unmodified by people. These include coral reefs, long sandy beaches, coastal dunes, lake systems, swamps, and extensive reed and papyrus wetlands, providing critical habitat for a wide range of species from Africa’s seas, wetlands and savannahs. The interaction of these environments with major floods and coastal storms in the Park’s transitional location has resulted in continuing speciation and exceptional species diversity. Its vivid natural spectacles include nesting turtles and large aggregations of flamingos and other waterfowl. The Wetland Park satisfied three of the World Heritage Site criteria and also met all the conditions of integrity. The conditions of integrity were: protection in South African law was adequate, the core area was proclaimed, conservation management plans had been adopted and were being implemented, there was political support at all levels of government, local communities and NGOs were also in support, management was to be undertaken by established and professional authorities adequately capacitated and resourced, and research and monitoring systems were in place. Presently about 100 000 people from 48 tribal groups live in villages surrounding the Park and community conservation programmes are in place to minimize conflicts and

maximize benefits. A neighbour-relations policy fosters good relations with communities that live near the Park to ensure that they derive direct benefits from the World Heritage Site such as free access, business and employment. The Park was listed in terms of the following criteria: • Superlative natural phenomena and scenic beauty (Criterion 7). The Review Panel in its summing up of the St Lucia EIA stated that the Wetland Park is a very special asset for the nation and that there was no substitute (Leon et al., 1993). The Wetland Park has exceptional aesthetic qualities in its striking natural land and seascapes with phenomenal scenic marine, coastal and wetland vistas. These range from clear Indian Ocean waters, brightly coloured coral reefs and fishes, to wide uninhabited sandy beaches and rocky shores, from a high forested dune cordon to a mosaic of differently coloured and textured wetlands, grasslands, forests, lakes and savannah types. Several natural phenomena are judged to be outstanding, such as the shifting salinity states from low saline to hypersaline within Lake St Lucia, linked to wet and dry climatic cycles, with the lake’s biota responding accordingly. There is an amazing spectacle of large numbers of nesting turtles on the beaches, along with migrating whales, dolphins and whale sharks, as well as large breeding colonies of pelicans, nesting sites of crocodiles, herds of elephants and other ungulate species, and pods of hippopotamuses. • On-going ecological and physical processes (Criterion 9). The Park’s geographic location at the

South Africa’s first World Heritage Site

transition between tropical and subtropical regions of south-eastern coastal Africa has resulted in exceptional diversity in plant and animal species. The combination of the fluvial, marine and aeolian processes that began in the early Pleistocene, and continue to the present day, produce an extensive coastal plain. Erosion, sedimentary and scouring processes shape and maintain a variety of coastal (dunes, sandbars and beaches), wetland (floodplains, swamps, pans, lakes and estuaries) and terrestrial landforms and niche habitats. The interplay of these processes is further affected by major floods, droughts and tropical cyclonic storm events. The absence of large rivers along a 220 km length of coast north of the Mfolozi River mouth, together with marine processes, create different ecological determinants such as the trapping of sediment loads by deep submarine canyons on the continental shelf, borne by the north flowing Agulhas Current, which results in clear waters and favourable environmental conditions for an expression of a wide diversity of coral reef organisms. Linked to a climatic cycle of wet and dry periods, salinity states within Lake St Lucia respond accordingly, ranging from freshwater to hypersaline conditions. With this change in the aquatic conditions, there is a corresponding shift in the biota of the system. Under freshwater to low salinity states, submerged macrophytes increase, attracting large numbers of ducks and other waterfowl. During medium salinity states, populations of benthic organisms increase and fish, fish-eating birds and crocodile populations expand. During high salinity states both phytoplankton and zooplankton increase, attracting large concentrations of feeding flamingos (Taylor, 1993). It is within this environmental heterogeneity of ecological determinants and environmental conditions that on-going evolutionary processes result in speciation events within this region, usually referred to as the Maputaland Centre of Endemism. The Wetland Park was therefore recognized as

being of sufficient size for the continued functioning in perpetuity of all these subcontinental to local on-going processes in association with the diversity of essential and key biotic elements. • Biodiversity and threatened species (Criterion 10). The Wetland Park supports viable populations of various species which are of international and national importance, including feeding and breeding areas for endangered and endemic species. Population sizes, particularly of resident breeding species, are believed to be sufficiently large and genetically heterogeneous to ensure their integrity. In addition to iconic African species such as elephant, rhinoceros, leopard, wild dog, as well as marine turtles, crocodiles and hippopotamuses, the Wetland Park is habitat to one of the world’s most remarkable marine species. This is Latimeria chalumnae, the coelacanth, known originally from fossils and thought to be extinct for some 70 million years until 1938, when a dead specimen was discovered on a trawler at East London. In October 2000 this ancient fish was found inhabiting a cave in Jesser Canyon at a depth of 104 meters. Subsequently an expedition found 24 coelacanths in three canyons over a 45 km length of coast in the Wetland Park (Sandwith and Porter, 2001; Zaloumis et al., 2005). The rich flora has 734 genera and more than 2180 species of plants of which there are 46 endemic species recorded. Marine species richness is also remarkable with 129 coral species, 812 mollusc species and 991 marine fish species. A rich terrestrial fauna is demonstrated by the occurrence of 50 amphibian species, 109 species of reptiles (including five marine turtle species) including endemic and threatened species, 526 bird species with many migratory and threatened species, 97 terrestrial and 32 marine mammal species. The total of threatened or endangered species (IUCN Red Data Species) recorded for the Park was 147 species.

15

16

Roger N. Porter

1.8 Strengthening conservation measures Under the World Heritage Convention Act of 1999, Regulations were published in the Government Gazette No. 21778 of November 2000 that consolidated the 16 individual protected areas into a single proclamation, forming the approximately 325 000 ha Greater St Lucia Wetland Park that extends from the Mozambique border southwards for 230 km and incorporating the uMkhuze Game Reserve. Additional areas were also incorporated; these included parts of the Dukuduku and Nyalazi state forests and the military area of the Ndlozi Peninsula, and the marine area was extended southwards to Cape St Lucia (Figure 1.2). The Greater St Lucia Wetland Park Authority was also established as the mandated institution responsible for the management of the World Heritage Site. This authority is governed by a board comprising representatives from both government and civil society including traditional leaders. The Regulations also appointed EKZNW as the body responsible for the day-to-day conservation management of the area. To give effect to this arrangement a Management Agreement was signed by the Chief Executive Officers of EKZNW and the Wetlands Authority in May 2002. The Park Authority has faced major management challenges since it was formed. Possibly the first of these concerned several land claims against those areas comprising the western, eastern and northern shores of Lake St Lucia, being about 60% of the area of the Park. The people who were forcibly removed by the Apartheid government in the past from the state forest areas to allow for afforestation, missile testing and military training, claimed their rights after the repeal of the Group Areas Act and when the Restitution of Land Rights Act had been passed into law. Protracted negotiations followed before the claims were eventually settled. A final agreement ensured that the claimed areas would remain

unsettled by people, would continue to be part of the protected area thus not affecting its ecological integrity, the claimants would be compensated, and would gain specific access rights to sacred sites. The phased removal of some 12 000 ha of alien tree plantations, mainly Pinus spp., on both the Eastern and Western Shores of the lake also required protracted negotiation with the state-owned timber company. In order to rehabilitate these areas, they were clear felled, the timber removed, and then followed up by treatments to eliminate volunteer seedlings and ensure recovery of the natural vegetation. Such recovery was startlingly dramatic and fast, with the reappearance of freshwater wetland systems that had not functioned for many decades creating extensive rehabilitated wildlife habitats adjacent to the lake. Environmental legislation requires an Integrated Management Plan (IMP) to be in place and formally adopted for all protected areas. A management plan was compiled that addresses matters such as improved fencing, zoning, rehabilitation of roads and tracks, reintroduction of species that had been lost to the system, setting of carrying capacity limits for certain animal species, and establishing a more comprehensive scientifically based monitoring system. A programme that provides for the on-going expansion of the Wetland Park through the incorporation of private and communal land has been developed and implemented. This has resulted in the inclusion of about 20 000 ha of privately owned land on the Western Shores of the lake to be included as well as areas previously used by the South African Defence Force and the state-owned forestry company. A Social, Environmental and Economic Development (SEED) unit forms part of the Wetland

South Africa’s first World Heritage Site

Coastal Forest

Maputaland Marine

Lake Sibaya

Sodwana Bay

Mkhuze

Sodwana

Lower Mkhuze

St Lucia Marine

False Bay

St Lucia Game Reserve

Cape Vidal

Additional areas included with proclamation in 2000

Nyalazi

Eastern Shores

N

Dukuduku

St Lucia Park Mapelane

0 2.55

10

15

20

25

Kilometers

FIGURE 1.2 The iSimangaliso Wetland Park World Heritage Site showing the various components protected and other areas that were consolidated by Government Gazette Notice of November 2000 into one area.

17

18

Roger N. Porter

Authority team that focuses its activities on alleviating widespread poverty in areas adjacent to the Park. Nature-based tourism and infrastructural projects such as road, fences, beach facilities, jetties and water supply have been undertaken by employing and training local people and small local

contractors from nearby villages. This approach has achieved the balance between biodiversity protection and ecosystem rehabilitation on the one hand, with a commitment to social equity and economic development on the other (Zaloumis et al., 2005).

1.9 Future management imperatives Given that the St Lucia estuarine system lies at the bottom end of several major river systems as well as on a coast where there are major shipping lanes, Park management needs to be continually vigilant to identify potential external and internal threats that could adversely impact on the outstanding universal values of this World Heritage Site. Possible future exploitation, such as water abstraction from rivers and groundwater and potential pollution of inflowing water, is a constant and on-going threat. Mfolozi Swamps and the Mfolozi River water without significant sediment loads are needed to restore the natural functioning of the estuarine system. Negotiations to relocate shipping lanes further eastward, as far as possible from the coast, need to begin so as to avoid future oil spills or ships running aground in the Park as have happened many times over the years. Also important is updating the Integrated Management Plan, securing the protected area and its complement of species, disease prevention and monitoring, and the eradication of terrestrial and aquatic alien invasive species. Such vigilance must also be applied within the ‘zone of influence’ or buffer zone to ensure that

inappropriate activities and or land-use developments do not occur that would adversely affect the integrity of the Wetland Park. Conservation programmes that encourage communal and private landowners to establish and maintain protected areas, of one kind or another, adjacent to the Wetland Park need to be championed and encouraged with incentives. Authorities in Mozambique need to be encouraged and advised in the establishment of a coastal protected area from the international border northwards to Maputo Special Reserve and Inhaca Island. The Wetland Park remains underdeveloped particularly with regard to tourism infrastructure and facilities, but any future development must be within the constraint of limited tourist density criteria. Access to the World Heritage Site and its natural resources is a fundamental imperative of its listing. Building and growing support of the Park by visitors, communities and politicians is crucial to its sustainability. Community outreach programmes need to be on-going and regularly updated to ensure their relevance, effectiveness and, importantly, the flow of benefits across the boundaries of the Park.

1.10 Conclusions The 104 year long historical path that led from pre-1895 to the listing of St Lucia as a natural World Heritage Site in 1999 reveals a pattern

of neglect and abuse by imperial, colonial and national administrations, largely uninterested in the protection of the St Lucia

South Africa’s first World Heritage Site

ecosystem, but more concerned with political, economic and ideological issues. It also reveals a destructive exploitation of the area’s biodiversity, natural resources and its indigenous inhabitants. In opposition and contrast there has evolved a growing determination and struggle to

ensure its long-term conservation security, survival and international recognition during this time, to preserve this outstanding natural heritage and its globally important biodiversity and sense of place for the people of the world.

19

Chapter contents 2.1 Introduction 2.2 1850s to 1950s: drainage of the Mfolozi floodplain 2.3 1951–1960: closure of the combined Mfolozi–St Lucia Mouth 2.4 1960–1968: stabilization of the St Lucia Mouth 2.5 1968–1972: severe drought and the canal in the Mkhuze Swamps 2.6 1973–1984: hydrological model and the Mfolozi Link Canal 2.7 1984–2000: Cyclone Domoina and its aftermath 2.8 2000 to present: the current drought 2.9 Future management 2.10 Assessment and lessons

Breaching the Mfolozi Mouth. (Photo: Ricky H. Taylor, 22 July 2010.)

2

Management history Ricky H. Taylor

2.1 Introduction This chapter describes the main management actions that have been implemented at St Lucia to retain it as an estuarine system. The biota of the ecosystem respond to the physical environment, and because of this much of the management of the system has been focused on the management of water and sediments. Much of the management has been initiated as a response to flood and drought events and then tempered by public perceptions on how the system should be managed. As an historical account, it is necessary to identify the driving events, to describe the scientific understanding at the time, to mention the public pressures on the managers and to identify those individual people who have been the main drivers of the management interventions. This review focuses on the management of water and sediments, but there have also been other management actions which are not discussed in this

chapter. These include the management of the fisheries and the controlled harvesting of its other natural resources, the zonation of the estuarine system for public use, the provision of recreational facilities, the maintenance of the integrity of the protected area and the prevention of poaching. Many of the records available are unpublished documents; records of meetings, unpublished reports detailing scientific investigations and the plans and records of work being done. Taylor (2011a) provides a historical perspective of the progression of understanding of the processes that drive the St Lucia–Mfolozi connection. This chapter complements this, by describing the management that was implemented at the mouth of St Lucia and elsewhere in the lake to counter the effects of freshwater starvation and accelerated sedimentation.

2.2 1850s to 1950s: drainage of the Mfolozi floodplain In the mid to late1800s hunters, traders and missionaries from Port Natal (Durban) travelled to St Lucia. At that time much of the economy of the colony was based on the export of wildlife products. To obtain these products, animals such as hippos and elephants were hunted extensively in Zululand (McCracken, 2008). Wild areas were regarded as places to harvest game, fish and forest products. In addition, large

numbers of animals were shot by more affluent members of society for ‘sport’ hunting. By the late 1800s, it was realized that animal numbers were being depleted. In response, the first management intervention to look after the well-being of St Lucia was to proclaim parts of it as a Game Reserve. This was in 1895, making it (together with Hluhluwe Game Reserve) the oldest existing protected area in Africa.

Ecology and Conservation of Estuarine Ecosystems: Lake St Lucia as a Global Model, eds. Renzo Perissinotto, Derek D. Stretch and Ricky H. Taylor. Published by Cambridge University Press. © Cambridge University Press 2013.

22

Ricky H. Taylor

FIGURE 2.1 Detail of the map of the St Lucia Estuary in 1903 or 1904 from Crofts (1905). It shows the combined St Lucia–Mfolozi Mouth (north is to the right).

The next focus on St Lucia was in September 1902, when the St Lucia Bay was assessed for its suitability as a harbour (Methven, 1903). Charlie Crofts, Harbour Engineer from Durban, was commissioned to undertake a detailed survey of the area, resulting in the first high quality map of the St Lucia–Mfolozi Mouth area (Crofts, 1905). The map (Figure 2.1) clearly shows the confluence of the Mfolozi and St Lucia forming what was known as the St Lucia Bay; with a single outlet to sea. The site was found to be unsuitable for harbour development, because the mouth was unlikely to remain deep (Methven, 1903). So, even at that time it was understood that the mouth state was variable. In the first half of the twentieth century, there were some notable events that affected St Lucia. Sugar farming started on the Mfolozi floodplain in 1911 (Dobeyn, 1987). However, soon after this the nascent sugar industry suffered successive setbacks as a result of the huge floods of 1918 and of 1925 (Harrison, 1989). These damaged the fields and the latter flood demolished the newly constructed sugar mill (Anon., 1973). The resilient farmers started again, draining their fields and rebuilding the mill. This time it was situated on the high ground at Riverview (Table 2.1), where it is today. The legacy of this is the respect for the power of flooding ingrained in the sugar farming community. The farmers know that at any time a

large flood can wipe out their entire livelihoods within a space of 24 hours. This respect for the floods influences their approach towards management of the floodplain. In 1932, the combined St Lucia–Mfolozi Mouth closed. Water, from all the catchment areas flowing into St Lucia, backed up behind the beach berm. This, through wind and wave action, had built up to a height of 14 ft (4.3 m) above mean sea level (Harrison, 1989). The backing-up water flooded low-lying areas around St Lucia and the Mfolozi floodplain. Sugar fields were inundated – this time from the backing up of water behind a closed mouth and not a large river flood. To alleviate this, the Mouth was breached by local farmer George Perrier and ‘a gang of about 40 Africans’ (Harrison, 1989). On breaching, the outflow was impressive; the water carrying sediments and trees out to sea for two weeks and diverting shipping far offshore. This record illustrates the process of breaching by overtopping of the beach berm (assisted in this case) and the associated scouring effects. It provides insights as to how important this breaching process was likely to have been in the past. The higher the elevation of the backed-up water, the greater the scouring of accumulated sediment deposits that occurs on breaching. This erosion rejuvenated the St Lucia Bay after each breaching event.

Management history

Table 2.1. Gazetteer of place names with their coordinates Place

Latitude

Longitude

Comment

2nd Narrows

28.3444

32.4082

The second constriction north of the Bridge

Back Channel

28.3922

32.4094

Excavated in the late 1960s

Brodie’s Crossing

28.2458

32.4509

Named after Brodie, the owner of the cattle trading store run by Challis and Brodie at the site of the present-day Makakatana Township

Brodie’s Shallows

28.2583

32.4507

Charter’s Creek

28.1994

32.4162

Demezane Pan

27.7532

32.5125

Durban

29.8761

32.0282

Eastern Shores

28.2107

32.5006

False Bay

27.9916

32.3857

Hluhluwe Dam

28.1238

32.1719

Hluhluwe River

28.0756

32.3393

Honeymoon Bend

28.3871

32.4032

Intake Works

28.4231

32.3673

Lake St Lucia

28.0622

32.4626

Link Canal

28.3945

32.3943

Links the Mfolozi River to St Lucia

Makakatana

28.2369

32.4141

A private township which is an enclave within the park

Maphelane

28.4065

32.4226

Mfolozi Flats

28.4923

32.3158

Mfolozi floodplain

28.4923

32.3158

Mfolozi Mouth

28.3996

32.4235

Mfolozi River

28.4226

32.3858

Mfolozi–St Lucia Link Canal

28.3945

32.3943

The ‘Link Canal’

Mitchell Island

28.2375

32.4510

Named after Douglas Mitchell, Administrator of the Province of Natal and the founder of the Natal Parks Board

Mkhuze Canal

27.7105

32.4824

Excavated in the early 1970s

Mkhuze floodplain

27.7160

32.5039

Named after A. E. Charter, Provincial Secretary

Formerly known as Port Natal, named in 1835 in honour of Sir Benjamin D’Urban (Governor of the Cape Colony)

The control structures on the Mfolozi River at the inlet of the Link Canal

23

24

Ricky H. Taylor

Table 2.1. (cont.) Place

Latitude

Longitude

Comment

Mkhuze Mouth

27.8834

32.4839

Mkhuze River

27.5805

32.2324

Mpate River

28.2952

32.3931

Mpempe Pan

27.6816

32.4659

Msunduzi River

28.5097

32.2880

Mtubatuba

28.4165

32.1874

Mzinene River

27.8675

32.3611

Narrows

28.334

32.4047

North Island

27.9276

32.5049

North Lake

28.0554

32.4714

Nyalazi River

28.1258

32.3663

Port Natal

29.8761

32.31.28

Now known as Durban

Potter’s Channel

28.2526

32.4389

Named after Captain H. B. Potter, the first Conservator of St Lucia

Riverview

28.4442

32.1827

Selley’s Lakes

27.9291

32.5639

South Lake

28.2141

32.4476

St Lucia Bay

28.3946

32.4128

St Lucia Bridge

28.3700

32.4105

St Lucia Estuary

28.3877

32.4140

St Lucia Mouth

28.3830

32.4250

St Lucia Pont

28.3700

32.4105

The Forks

28.2810

32.4091

Tshanetshe Pan

27.6750

32.4391

Van Niekerk’s Canal

27.7105

32.4824

Warner’s Drain

28.4474

32.3199

Western Shores

28.1407

32.4058

Wilson’s Drain

28.5065

32.2832

Named after Chief Mtubatuba

Named after Jeff Selley, an inhabitant of St Lucia in the 1940s

Named by the Portuguese navigators after Santa Luzia

Named after Conservator Nick van Niekerk who promoted the construction of the canal

Management history

St Lucia Mouth

N

Mfolozi Mouth l

Lake Futululu

nk

a an

c

Li

M fo lo

zi

Warner’s Drain

FIGURE 2.2 Map of the Mfolozi floodplain showing the present-day extent of the drainage canals. It also shows the spillway created after the 1984 floods. (Map by I. L. van Heerden, unpubl., after Begg, 1988.)

y rb he et in W dra

w r ro Bo

t pi

n ai dr

Mo

kan

a

Lake Eteza

Mavuya

14 drain

Flood diversion spillway Spillw ay

Natural drainage channels Artificial drainage channels Artificial spillway

FIGURE 2.3 In the 1930s and 1940s St Lucia became a favoured fishing destination as huge catches of fish were to be made. (Date and photographer unknown.)

In response to river floods and back-flooding behind a closed mouth, the farmers cooperatively, as the Umfolozi Cooperative Sugar Planters Ltd. (UCOSP), initiated ambitious schemes to drain the Mfolozi Flats (Figure 2.2). First Wilson’s Drain was excavated along the Msunduzi watercourse (Table 2.1), followed by Warner’s Drain ‘to tame the Mfolozi River’ (Dobeyn, 1987). The latter canal was completed in September 1936 (Director of

Irrigation, 1948) to prevent floods from inundating the sugar fields. One effect of this is that, from this time onwards, sediments which would otherwise have been deposited in the upper reaches of the floodplain are carried via the canals to the lower parts of the floodplain. These are the bedload and suspended sediments that are deposited beyond the river channel whenever the river overtops its banks. Thus, the point of deposition has been shifted eastwards by about 30 km (Taylor, 2011a), to be about 6 km from the combined St Lucia–Mfolozi Mouth. By the 1930s, St Lucia had become well known throughout South Africa as a fisherman’s paradise (Figure 2.3). The anglers had come to expect large catches, but had little understanding of the variable nature of estuaries or how the ecological dynamics responded to the quasi-decadal wet and dry periods experienced in the region. In the late 1930s to the early 1940s, there was a series of drought years. Conditions at St Lucia Mouth and in the lake changed. Sediments from the sea and the Mfolozi accumulated in the mouth. There was a freshening of the estuary due to the natural diversion of the Mfolozi River into St Lucia as the mouth constricted. Large beds of submerged macrophytes formed and there was a decline in fishing. These drought-induced responses alarmed the custodians of the area and the general public.

25

26

Ricky H. Taylor

FIGURE 2.4 Captain H. B. Potter – the first conservator at St Lucia. It was through his actions that the first investigations into the sediments and ecology of St Lucia were made in the late 1930s and 1940s. (Photo provided by Derek Potter.)

The Conservator of the area was Captain H. B. Potter, after whom the Potter’s Channel is named (Figure 2.4). He played an important role in creating an awareness of the area and initiated the first conservation measures. In 1938 he invited the Harbour Engineer, Colonel D. E. Patterson, to visit St Lucia to advise on sediment accumulation in the mouth. Patterson suggested that a new mouth should possibly be made to separate the Mfolozi from St Lucia, to prevent too much fresh water entering St Lucia. This indicated that at this stage Mfolozi water was diverting into the lake. He also identified that there was accelerated sedimentation, which he ascribed to the canalization of the Mfolozi Flats (Von Bonde, 1940). No action was taken at this time to implement these recommendations. From 1940 to the mid 1950s, Zululand was affected by a severe drought. Through Potter, the Provincial Secretary A. E. Charter requested Dr Cecil von Bonde, Director of Sea Fisheries, to visit St Lucia in July1940 to advise on management of the area. Von Bonde’s brief was to provide advice on: • sediment accumulation in the lower reaches of the Mfolozi and St Lucia Estuaries;

FIGURE 2.5 The pont at St Lucia. The influx of marine sediments at the early stages of the drought of the 1940s grounded the pont and a causeway was built across the estuary for access. (Date and photographer unknown.)

• fisheries management in St Lucia; and • the rampant growth of weed in the Brodie’s Crossing area at the northern end of the Narrows (Table 2.1). He spent nine days in the area. It was a time when the mouth was severely constricted and Mfolozi water was being naturally diverted northwards via the Narrows into St Lucia. There was little tidal movement and the water in the estuary was so fresh that Von Bonde reported seeing several kob (Argyrosomus japonicus) ‘that had died from the freshness’. Fine sediment from the Mfolozi was settling in the estuary mouth area – up to the area of the pont (Figure 2.5). At the extreme northern end of the Narrows, in the Brodie’s Crossing area, there was concern that the extensive growth of ‘weed’ was blocking the migrations of fish and prawns (Von Bonde, 1940). Von Bonde made a number of recommendations which included: • the prawn sanctuary that had been established by Captain Potter should be retained; • fish bag and size limits should be set; and • advice on issues relating to the management of hydrology and sedimentology should be sought. In hindsight, the most important of these recommendations relate to the water and sediment.

Management history

FIGURE 2.6 Oblique aerial photograph from Von Bonde’s 1940 report. Looking northwards, the aircraft being over the point where the Mfolozi splits, the photo shows the Western Channel and the main eastern course of the Mfolozi, the beach berm and the single mouth. Clearly visible is the sediment colouring the sea from the mouth southwards. Also note the flood-tidal sandbars in from the mouth. (Photo taken c. 1937, photographer unknown.)

Von Bonde considered a new mouth for the Mfolozi, but thought it should not be excavated at this time. He wrote: ‘The suggestion that a new exit for the Umfolozi River should be cut further southwards than the present mouth through the sandy beach so as to keep the freshwater away from the Estuary was also raised, but I do not consider this a practical solution owing to the expense of cutting a channel about 350 yards long through the sand would be too great and there is no guarantee that the new mouth would not soon become silted up’ (Von Bonde, 1940). Von Bonde also supported the closure of the Western Channel (Figure 2.6), to prevent sediment inputs and to reduce the freshwater inputs into St Lucia and suggested a dredger owned by UCOSP could be used to do the job. At the time of his visit, there were dense beds of what is likely to have been the plant Stuckenia pectinata at Brodie’s Crossing – ‘The presence of these weeds tends to form a virtually impenetrable barrier to the migrations of fish both to and from

the lake.’ He recommended that no action should be taken to control the ‘weed’ as it would just grow back again unless properly cleared (including roots). It seems as if none of these recommendations were implemented at the time – but it shows the way that the thinking was going. The drought of the 1940s proved to be one the longest and one of the most severe on record and, in 1948, Colonel Patterson was again invited to visit the area to investigate the silting of St Lucia. In the intervening ten years there had been ‘a phenomenal silting up of the estuary’ (Patterson, 1948). He expressed his ‘grave concern’ about the consequences of the silting to the well-being of Lake St Lucia and recommended that a separate mouth for the Mfolozi should be excavated near Maphelane (Table 2.1) as soon as possible. He also recommended that suitable vegetation be planted to stabilize the wind-blown beach dunes. In that same year the newly formed Natal Parks, Game and Fish Preservation Board invited Professor John Day of the Zoology Department of the University of Cape Town to conduct an ecological survey of the area. This was to be the first comprehensive ecological survey of St Lucia. His party of eight zoologists spent three weeks in July 1948 describing the topography, hydrology, sediments, botany and zoology of the system. He described a constricted mouth condition with little tidal exchange. He noted that there was not much water in the Mfolozi River and that little fresh water was flowing into St Lucia. At this stage of the drought, hypersaline conditions of 52.6 were measured in the False Bay area (Day, 1948). Day identified that the accumulating sediments were a result of the canalization of the Mfolozi River and he supported the concept that the Mfolozi should be given a new mouth. Considering the need for the fish and prawns to move between the estuary and the sea, he recommended the mouth be kept open.

27

28

Ricky H. Taylor

level and an accumulation of sediments in the lower parts of the estuary. • The constriction before closure, and the subsequent closure, naturally diverts the Mfolozi into St Lucia. • After closure water backs up behind the beach berm – raising the water level in the system to above that of mean sea level. • When the berm is overtopped, breaching occurs and the outflowing water flushes sediments to the sea.

This was an important period for St Lucia. The area was set aside for conservation and developed into a fishing mecca. Sugar farming modified the Mfolozi floodplain, causing sediments to be deposited in the lower parts of the floodplain rather than in its upper reaches where the river enters it. This is the only period for which we have records of how the combined mouth configuration functioned and from which we are able to gain the following insights on its natural functioning: • The closure of the combined mouth occurs in response to drought – due to a lowering of lake

2.3 1951–1960: closure of the combined Mfolozi–St Lucia Mouth In 1951 the combined mouth closed – caused by the accumulation of sediments and the effects of the prolonged drought. ‘By 1951 the waterway had become so restricted that the tidal flow became extremely sluggish … The reduced tidal flow was unable to keep the mouth clear and in May, 1951, it was finally closed by a sand bar’ (Kokot, 1959). The sediment accumulation and low lake levels caused shallowing some distance up the estuary, and the pont could only be operated when a north wind raised the water level. Eventually the pont was shut down and in its place, to allow vehicle access to St Lucia, an earth causeway was placed across the estuary blocking all connection between the lake and the sea. Through the interventions of Douglas Mitchell, a former Administrator of Natal, a cooperation between the Natal Provincial Administration and the Central Government was established and a Technical Advisory Committee was appointed to study the problem and draw up a plan for the reclamation of the estuary (Kokot, 1959). This small committee of experts reported on the problem at a conference at St Lucia on 18 August 1951 (Kokot, 1952). Their recommendations were (Kokot, 1959): • The cause of the silting had to be eliminated. The Mfolozi River was bringing in silt faster than it

•

•

•

•

could be removed. Since the accelerated soil erosion in the catchment could not be stopped, the only other solution was to divert the Mfolozi into the sea via a new mouth. A large tidal basin was to be excavated close to the St Lucia Mouth to maximize tidal exchange, which they hoped would keep the mouth from closing. Six miles (almost 10 km) of the Narrows that had constricted had to be opened in an effort to restore the flow of seawater between the lake and the sea. After opening the new mouth to the Mfolozi, an attempt should be made to stabilize its position by groynes or piles. The mobile dune sands adjacent to the mouth should be vegetated to prevent wind-blown sand from entering the estuary.

This committee raised two questions that are still relevant today: • Would the St Lucia Estuary mouth remain open without the scouring effect of the Mfolozi? • Without the Mfolozi linkage, would the salinity in the lake rise to a level high enough to injure aquatic life in times of drought (Kokot, 1952)? Most of the recommendations of the committee were implemented and set a baseline for management to be

Management history

implemented over the forthcoming decades. Eric Yeld was appointed as the person to undertake the work, and the new mouth was cut through the 40 foot (12 m) high dune near Maphelane. This was hand dug, using coco-pans on rails to move the sand. The approach to the cutting was made using an excavator mounted on pontoons borrowed from UCOSP, while later a floating sand-pump was used (Kokot, 1959). The mouth of the Mfolozi was opened to the sea in 1952. The next step was to focus on the closed St Lucia Estuary. Here, a tidal basin was excavated using three floating sand-pumps assisted by an excavator floating on a pontoon (Kokot, 1959). In 1955 the mouth was breached but closed again eight days later. Dredging was initially up to the causeway that had been put across the estuary to replace the pont. It was realized that the connection to the lake was needed to maintain an open St Lucia Mouth, so the causeway was breached and a single-span bridge erected (followed a few years later by a second span). Attempts were made to stabilize both the Mfolozi and the St Lucia Mouths, using railway lines driven into the sand to prevent northward drift and merging – but this was to no avail. The sand dunes between the two mouths were stabilized by planting casuarina trees to ‘beautify the dreary landscape as well as to fix the sand’ (Kokot, 1959). Some of these trees had to be removed some 20 years later, when it was realized that they hindered the outflow of Mfolozi floodwaters, impeding the drainage of the sugar fields. Most of the rest have been taken out since the 1980s, as a part of the alien plant management programme. A levee was built across the floodplain separating the Mfolozi from St Lucia, to prevent floods from overtopping into the St Lucia Estuary (Brown, 1969). The opening of the causeway on 17 April 1956 finally linked the sea and the lake (Crass, 1982). By 1962 the tidal basin was dredged clear up to Honeymoon Bend (Table 2.1), and in the same year work started on the widening of the Narrows up to the bridge (Brown, 1969). Opening the St Lucia Mouth had taken many years of intensive dredging to clear the

FIGURE 2.7 Aerial photograph from June 1960 showing the progress with the dredging to clear the sediments that had accumulated during the drought period leading up to the closure of the combined St Lucia–Mfolozi Mouth. (Photo: Chief Directorate of National Geo-spatial Information, South Africa.)

sediments that had accumulated in the St Lucia Mouth. This resulted in piles of deposited dredge spoil on the banks of St Lucia, on both sides of the mouth area which filled what had originally been intertidal mudflats (Figure 2.7). In 1984, during the Domoina floods, this dredge spoil was to act as a plug which impeded the free-flow of floodwater to the sea. The mouth closure had been from 20 April 1951 to 17 April 1956. During this period there had been no water exchange between the lake and the sea and water levels in the lake had ranged from 2 ft to +3.5 ft. The maximum height of the water level was due to a flood in 1955. This water was retained in the system, so that when the Mouth was finally opened in 1956 it was with a strong outflowing flush (Crass, 1956, 1982).

During this period, the Mfolozi River was separated from St Lucia and the philosophy that the mouth should stay open at (almost) all times was prevalent. After several years of dredging, the St Lucia Mouth was opened by the newly formed dredge unit.

29

30

Ricky H. Taylor

2.4 1960–1968: stabilization of the St Lucia Mouth The Technical Committee under the guidance of the provincial water engineer, E. A. Middleton, contracted the CSIR in 1961 to investigate the possibility of maintaining a permanent link with the sea. The scientist assigned to undertake the work was H. B. Sauermann. Starting with a void of scientific knowledge, he approached the problem by doing extensive careful field measurements, followed by the construction of a physical model. The field measurements included: • • • • • • •

• • • • • • •

offshore and beach topography aerial photography and surveying of the estuary sea swell and wave patterns wave refraction studies energy distribution of incoming swell float tests to determine the influence of the Agulhas Current tide and tidal propagation (e.g. he found that the lag between sea and bridge is 30 to 100 minutes, and up to the Mpate River 3.5 to 5 hours) wind patterns wind effects on the beach wind effects on water movement in the estuary and in the lake beach material (sediment grain size) measurement of littoral drift along the beach St Lucia lake level lake salinity.

With this information he gained a holistic picture of the dynamics of the St Lucia Mouth area and gave his ‘diagnosis of the problem’. With the disconnection of the Mfolozi from St Lucia, marine sediments enter St Lucia on a flood tide. The flood tide has a duration of 5.4 hours, whereas the ebb tide lasts approximately 7.0 hours. Thus the inflow velocity is greater than the outflow velocity, and this affects the bedload transport – with a greater amount of sediment being transported by the higher velocities. Wave action too is important, as this suspends the sediment in the sea prior to it being transported into the estuary. He also recognized that

when the Mfolozi was still connected the additional flow transported sediment out of the estuary. However, this water carries silt (especially in floods), which on mixing with salt water flocculates and deposition occurs. Sauermann’s data collection was thorough and it formed a solid foundation for management and future studies at St Lucia. He was of the opinion that ‘St Lucia Estuary … can only be saved if a good stable estuary mouth exists to provide for the needs of the flora and fauna in the area’ (Sauermann, 1963). This led him to the second phase to the project, in which he constructed a physical model of the sea and mouth in a large wave tank at the CSIR in Pretoria to test schemes for the stabilization of the mouth (Figure 2.8). The model was constructed to a horizontal scale of 1:300 and a vertical scale of 1:72, in a wave basin measuring 70 × 40 feet (21 × 12 m). Using this physical model he conducted numerous simulations; testing various mouth locations, orientations and configurations – and how structures could be used to stabilize the mouth under a variety of wind, wave and current conditions. His

FIGURE 2.8 The physical model of St Lucia constructed in Pretoria by Sauermann (1963). He used this model to develop his groyne-berm scheme for the St Lucia Mouth and to orientate it so that it would be best aligned to withstand wave action and have the least influx of sediments.

Management history

recommendation was to develop the ‘groyne-berm scheme’ that was later implemented, but never fully completed. The concept was to construct a pair of groynes 350 feet (107 m) apart and inclined northwards by about 45 degrees from a line perpendicular to the coast. All but the central 150 ft (45 m) of the width between those groynes was to be stabilized with mats of stone-filled wire baskets (the berms). The upper surface of the berms was to be approximately at mean sea level, with a stepped crosssection to take variable amounts of water. The central section was to be a scour channel. The rationale was to create a self-scouring mouth orientated in a direction that would be least affected by waves bringing in marine sediments. The shelf of the berm allowed the faster flood tide to use the maximum width of the channel, while the ebb tide, which is over a longer period, would be more concentrated in the central channel and hence would scour out accumulated sediments. The groynes were constructed using bags containing a cement-sand mixture and dolosse that each weighed 1500 pounds (680 kg). The berms were never started (Brown, 1969) (Figure 2.9). In 1971 there were concerns about the effectiveness of the scheme and Omar J. Lillevang, an American coastal engineer, was brought in to advise on the management of the estuary mouth (Lillevang, 1971). He noted that Sauermann’s groyne-berm scheme was an expensive strategy and he proposed ‘a better scheme’. His recommendation was to develop a short channel of restricted width between an enhanced tidal prism and the sea, and then to dredge to remove incoming sediment. This seems to have been the germ of the concept that was to be applied after the 1984 floods. In the 1950s the Government established pine plantations on the Western Shores and in the 1960s were preparing to plant on the Eastern Shores as well. The Hluhluwe Dam was built in 1961 and public sentiment was that these developments would impact on St Lucia. To allay fears, the State President established a team under the chairmanship of J. P. Kriel to conduct a ‘Commission of Inquiry into the alleged threat to animal and plant life in St Lucia Lake’ (Kriel et al., 1966). Its purpose was to review

FIGURE 2.9 Photo of the St Lucia Mouth with the partially constructed groyne-berm scheme. The water flows to the sea in the lower right corner of the photograph. The berm on the north bank is constructed from cement-sand in bags, and nearer the sea dolosse were also used. The berm on the south bank is much smaller, but was being extended at the time of the 1984 flood, which washed all the structure into the sea. In the mid left of the picture is a stockpile of dolosse. (Photo: R. H. Taylor, 1976.)

existing knowledge, obtain public submissions and collect field data before synthesizing what was known about the system and making recommendations. It also commissioned 15 special investigations required to inform the Commission. One of these investigations was the survey of the estuarine ecosystem that was conducted in July 1964 and January 1965 (Millard and Broekhuysen, 1970). This built on the study undertaken by Day et al. (1954). During this period the conditions in the lake were ‘good, with lush vegetation throughout and a thriving population of prawns and fish’. In their report, Millard and Broekhuysen (1970) state that the deleterious effects of mouth closure are ‘only too well known. Prawns and fish cannot escape to the sea to breed, there is no restocking of the plankton and young forms, and the larger fish soon disappear.’ It was this perspective that was accepted at the time and reinforced the philosophy that the mouth should stay open at all times – an opinion that was to influence the engineering approaches throughout the 1970s.

31

32

Ricky H. Taylor

The Commission also produced the first reasonably accurate calculations of water balances and of quantities of salt and sediments. But, without computers the calculations were not able to adequately describe the dynamics of the system or give probabilities for the occurrences of the various states. Of the many recommendations made by the Commission, those relating to hydrology, sediments and mouth management included: • A quantity of freshwater flow from the catchments should be reserved to prevent the salinity levels in the lake from becoming too high. In addition the option of bringing in water from the Pongolapoort Dam (north of the Mkhuze catchment) should be investigated. • Accelerated sedimentation should be avoided – through the application of better soil-conservation measures in the catchment, and by retaining the integrity of the Mkhuze Swamps at the northern extremity of St Lucia. The separation of the Mfolozi from St Lucia should be retained to prevent the influx of Mfolozi sediments. • The mouth ‘must be kept open for the protection of the lake and the aquatic life within it.’ • The Narrows should be deepened and enlarged to facilitate water movements between the lake and the sea. Work continued with the construction of the groyne/ berm scheme at the mouth and in 1967 work was started to dredge the Narrows north of the St Lucia Bridge. The spoil dredged from the 300 ft wide (~100 m) and 6 ft deep (~2 m) channel being deposited alternatively on one side then the other. Neither the effectiveness of the channel to maintain a better flow between the lake and the sea, nor the necessity of improving this flow has been fully assessed. However we do know that this channel and the lateral spoil depositions had a deleterious ecological impact on the banks of the Narrows. Vegetation colonization of the spoil has been slow, and is still eroding in many places where it has yet to reach its natural angle of repose. The dredged soils are saline and the material anaerobic. The margin of the Narrows next to the channel has lost intertidal area, and hence

this affects the growth of fringing vegetation, including mangroves. In the channel the disturbed sediment is slow to be colonized by benthic fauna (Hay, 1985a). The dredging of the Narrows was completed when the channel reached Makakatana Bay, on 29 December 1969 (Crass, 1982). Had it gone further north into South Lake it would have cut through the sandbar that stretches between Makakatana and Mitchell Island. In extreme drought conditions, the lake level drops to the extent that the lake separates into several discrete basins. The Makakatana–Mitchell Island sandbar separates the Narrows from the Charter’s Creek Basin. Because the Narrows has an input of water from the Mpate River, it retains a moderate water level and water tends to flow from it into the lower Charter’s Creek Basin. Had a channel through this sandbar been dredged in the late 1960s, the Narrows would have drained to the level of the Charter’s Creek Basin during the current drought. This would have left the tour boats from St Lucia Town high and dry and would have caused the collapse of tourism to the town. To guide future research and the technical application of the Commission’s recommendations, the Lake St Lucia Scientific Advisory Council and its Technical and Ecological Committees (SCADCO) were established in 1969. It was an advisory body to the Natal Provincial Administration and to the Natal Parks Board which was to provide high quality advice for the next three decades. During this period, careful field measurements were made and a physical model used to design the groyne-berm structure to maintain an open mouth. The groynes were constructed, but the full extent of this engineering scheme was never completed. The Kriel Commission was a coordinated effort to understand the hydrology and ecology of St Lucia and its management – but many of the recommendations were never implemented. In the Narrows, the channel from the Mouth to Makakatana Bay was dredged.

Management history

2.5 1968–1972: severe drought and the canal in the Mkhuze Swamps The severe and prolonged drought of 1968–1972 highlighted the problems related to freshwater deprivation, which were to become a pattern to be repeated with every subsequent drought. In the mid to late 1960s, the public had been exposed to the richness of the St Lucia system. This richness possibly increased in the early stages of the drought, as surrounding water bodies dried up and birds from them congregated in St Lucia. With increasing salinity the beds of the submerged macrophyte Stuckenia pectinata died, releasing large quantities of detritus into the system. They were replaced by beds of the more salinity tolerant Ruppia cirrhosa and Zostera capensis (Chapter 11). The detritus and the slightly higher salinity would have boosted the fish and prawn populations, while the lowering water level would have exposed lake margins that would have attracted large numbers of wading birds. This is the typical progression of a water body, as it goes through a ‘dry-down’ phase in the early stages of a drought. Then conditions changed as the drought progressed. Salt concentrated to hypersaline levels in the northern parts of the lake, fringing reed beds browned and died, and fish and bird populations dwindled. By August 1970 both the St Lucia and the Mfolozi Mouths were closed and a huge quantity of salt had been trapped in the system. The ecosystem responses perturbed the public, and when dehydrated and emaciated crocs had to be airlifted from the Mkhuze Mouth to lower-salinity parts of the lake (Anon., 1970), the public opinion was that the lake was dying. During the drought the mouth could not be kept open all the time. As the rivers stopped flowing and water evaporated, the lake level dropped and the St Lucia Mouth closed. It was dredged open and, as the lake level had fallen to below that of the sea, seawater entered to replace water lost through evaporation. A strong salinity gradient formed – ranging from seawater at the mouth to salinity three times as salty as the sea in the north (Chapters 7 and 9). The shallow St Lucia, with a small volume of water in relation to

its large surface area and without the influx of water from the Mfolozi River, is very sensitive to an increase in evaporation loss relative to freshwater gains. With an open mouth during droughts, water level in the system remains close to that of the sea and salinity rises to be saltier than the sea. This is a lethal combination for the fringing vegetation. Under this condition wave action washes hypersaline water onto the plants and kills them, leaving the shorelines of the lake and the islands unprotected (Chapter 11). This results in erosion. During the drought of the 1970s a significant amount of the island habitat of St Lucia was lost and the entire North Island eroded away. It is estimated that the mass of salt in St Lucia at the peak of the 1970s drought exceeded 20 million tonnes (Taylor, 2006). When a large salt load has accumulated, the probability is that there would be a very slow flushing and hence a delayed recovery of the system, after a saline period. However, this was not to be the case in the 1970s, as the drought was terminated by a flood. The system flushed over a short period and recovery was rapid. As the drought progressed, the engineers Chew, Bowen and Mercer (1969) were commissioned to investigate the feasibility of introducing fresh water into St Lucia from the Mfolozi River. They suggested four possible schemes to transfer Mfolozi water into St Lucia. Two of these would involve the trans-catchment movement of water from different parts of the Mfolozi into the Nyalazi River, which drains into False Bay, and two schemes to transfer water from near the Mfolozi Mouth into different points in the Narrows. One of the schemes suggested was to take Mfolozi water along the ‘Back Channel’ from near the mouth to enter St Lucia near Honeymoon Bend. This Back Channel was opened as a temporary measure to bring Mfolozi water into St Lucia from 1970 to 1973. It was during this period that a minor flood came down the Mfolozi providing a ‘life-giving infusion’ of fresh water to the southern part of St Lucia (Crass, 1982). Chew, Bowen and Mercer (1969) were also commissioned to investigate the feasibility of

33

34

Ricky H. Taylor

obtaining fresh water from the Mkhuze River. In a flight over the river they reported that there were about 30 timber blockages in the river channel. These obstructions had resulted from the felling of trees while clearing land for agriculture. Chew, Bowen and Mercer (1969) remarked on the ‘consumptive use’ of water by the Mkhuze Swamps and its pans. To minimize the evaporation of water from the swamp, they recommended that the river channel in the swamp be restored by the removal of these blockages and consideration should be given to the excavation of a canal (possibly concrete lined) along the side of the Mkhuze floodplain. An exercise was started to clear the blockages in the river channel. Initially the managers used manual labour and chainsaws. Some of the blockages were logjams. These were blasted with dynamite, and an excavator was brought in to clear the logs. The Conservator of St Lucia, Nick van Niekerk, was convinced that the swamp was capturing Mkhuze River water that would be more effective if it was in St Lucia, and championed the cause to excavate a channel through the swamps. SCADCO debated the issue and, after revisiting the report of Chew, Bowen and Mercer (1969), the decision was taken to open a channel that would allow minor flows an unobstructed path into St Lucia (Crass, 1982). This canal, excavated by the Reclamation Unit in 1971, was13.5 km long and extended from near Mpempe Pan to 1 km south of Demezane Pan (Alexander, 1976). However, as soon as the drought had broken and there were significant flows in the channel, it started eroding, draining the Mpempe Pan in the process. Alexander (1973) warned of the consequences of the impacts of the canal on the swamps and on three separate occasions attempts were made to close the outlet at Mpempe. In 1974 an earth wall was constructed, but was breached in 1975; in 1979 it was rebuilt, but breached during the Domoina floods of January/February 1984. The final attempt was made in January 1985, but it was breached a month later (Ellery et al., 2003). The impact of the canal was aggravated by a small

channel dug by a farmer to link the Mkhuze River to the Tshanetshe Pan. This channel enlarged rapidly, diverting the water through Tshanetshe Pan. From there it flowed into the dry Mpempe Pan and then down the eroded watercourse of what is now known as the Van Niekerk’s Canal. Ellery et al. (2003) considered that the Tshanetshe canal had been the more damaging of the interventions, so in 2004 an armoured weir was placed across the breach at its outlet. The Van Niekerk’s Canal, excavated as an urgent reaction to the peak drought conditions of 1973, proved to be a very costly mistake that more than three decades after the end of the drought continues to require rehabilitation, and has caused the permanent loss of Mpempe Pan and the draining of a portion of the swamp. In September 1971 (at the peak of the drought) the Reclamation Unit had three dredgers (Ilanda, Inkwazi and Ivubu) and four excavators, of which two were in St Lucia and two in the Mkhuze Swamps (Blok, 1971). On the staff was Tom Blok, a permanent site engineer, the manager of the unit, Gus Brits, a surveyor, dredge operators and staff to maintain and support the equipment. Had the unit and its infrastructure expanded to create a situation where the equipment, infrastructure and staff now had to be kept busy? At the termination of the drought Bob Crass, Chief Professional Officer of the Natal Parks Board, made a case for the mouth to be maintained open permanently. Understanding the larger scenario, he recognized that whenever there is a drought salinity in the north will rise to unmanageable levels which will ‘write off’ North Lake temporarily, but conditions will be satisfactory up the Narrows and in South Lake (Crass, 1970). At this stage there was no effort to consider ways to reconnect the Mfolozi River to St Lucia to counter this.

The drought was a ‘wake-up’ call to remind the managers about the impacts of a severe drought. The dredge equipment was on hand and actions were taken, such as excavation of the canal in the Mkhuze Swamps, that were ill considered.

Management history

2.6 1973–1984: hydrological model and the Mfolozi Link Canal The use of computer modelling to simulate the hydrological functioning of the St Lucia system is an important tool to guide management. The first efforts at computer modelling of the system were by Dr W. James of the University of Natal. In 1967 he and C. Horne developed a model to simulate tidal propagation in the St Lucia Narrows (Horne, 1967). The object of this was to estimate the net inflow of salt water into the lake system during each tidal cycle. The model: • was to be used to show the ‘success of the dredging operations’; • ‘can be used to predict the effect of any channel widening or deepening’; and, • ‘if so desired, could also be extended to include the whole of the St Lucia Lake’ (James and Horne, 1969). Since then several computer models have been developed and have proved to be valuable tools for gaining an understanding of the functioning of the system. In 1970 Professor Des Midgley, the Director of the Hydrological Research Unit at the University of the Witwatersrand, recommended to SCADCO that a mathematical model to simulate the water and salt balance within the estuary and the lake be built (Hutchison and Pitman, 1973). This model, shown schematically in Figure 2.10, was developed by Ian Hutchison (Hutchison, 1976; Hutchison and Pitman, 1977; Hutchison and Midgley, 1978). The main features of the model are: • Gains of fresh water are from river inflows, direct rainfall and local seepage from groundwater. In addition salt water may enter from the sea when the mouth is open and the lake water level is low enough for seawater to enter. • Water losses are due to evaporation and outflow to the sea when the mouth is open and there is a head of water in the lake. • Given the morphology of the lake, a volume:level curve was established, so the relationship between

water inflows, water volume and water outflows could be simulated. The lake itself was divided into several interacting compartments. • The salinity was derived by calculating the mass of salt that enters when carried in by inflowing seawater (at 35 grams of salt for each litre of seawater), and by calculating the mass lost when water, of a known salinity, flows seaward. At St Lucia, Hutchison set up an extensive monitoring programme to gain data to calibrate the model. The model used rainfall and river inputs dating back to 1918. This enabled him to show the probabilities of the occurrence of specific conditions – such as hypersaline events. Once the model had been set up, calibrated and verified, it was used to simulate the impacts of a number of different management options: ranging from bringing in Pongola River water; establishing a new mouth to the sea in the north at Selley’s Lakes (Table 2.1); reduction of the surface area of the lake by dredging new islands; and the importation of Mfolozi water. The guidelines for a desired state were not well described by the biologists, but periods of hypersalinity were to be minimized and the mouth was to stay open most of the time. Of the various options tested using the model, it was shown that the most efficient way to bring fresh water into St Lucia was via a canal from the Mfolozi River. In 1975 the decision was taken to proceed with the construction of the Mfolozi Link Canal. This proposal was further developed by Weiss et al. (1975), and an ‘Intake Works’ was designed by Basil Lund (Midgley, 1982). The Intake Works, with six sluice gates, was built 9 km from the sea where the bank of the Mfolozi first had to be protected with sheet piling capped with concrete. The Intake Works had three automatic gates designed to limit flows to 30 m3 s 1 and to shut off completely during floods. From here a 12 km canal was excavated to enter St Lucia 8 km from its mouth, with the intention to lengthen it at a later stage to enter St Lucia on the

35

Ricky H. Taylor

evaporation rainfall

river inflow evaporation

model parometers irrigation

channel type outlet

runoff

SWAMP one cell model

afforrestation rainfall

FIGURE 2.10 The schematic structure of Hutchison’s model. Coupled to this were models of the catchments and of the Mfolozi (Hutchison, 1976).

outflow form swamp model

CATCHMENT semi-empirical (monthly time step) river inflow

LAKE two-dimensionsol cell model (monthly time step)

DISCHARGE LEVEL

ESTUARY one-dimensional tidal model (1/2 hourly time step)

INTEGRATION YIELDS

TIME

TOTAL INPUT

LAKE LEVEL

LAKE BOUNDARY CONDITION

OUT OF LAKE INTO LAKE AVERAGE ESTUARY DISCHARGE

SEA LEVEL

36

estuary mouth

TIME

north bank of the Makakatana Peninsula (Table 2.1). The canal was excavated to be 20 m wide and 2 m deep. In the canal, immediately downstream of the Intake Works, there was to be a sediment-settling basin designed to settle out 50% of the sediments by delaying water flowing at 30 m3 s 1 by 3 to 4 hours. Nevertheless, it was still accepted that large quantities of silt would go through to the lake. As large quantities of sediment would accumulate in the settling basin, this would have to be dredged periodically (Midgley, 1982). The canal was designed to carry Mfolozi water for 65% of the time (Lund, 1976) and this flow would

carry an estimated 175 000 m3 per annum of sediment. The settling basin would trap an estimated 100 000 m3 per annum of the heavier component of this, which would then be removed by dredgers – a job it is estimated that would take 1000 hours a year using the small dredger. Where this enormous volume of material was to be deposited had not been adequately addressed when the canal was designed (Taylor and Collings, 1988). The settling basin was never completed as boulder-beds were unexpectedly found in the area to be excavated, which delayed the excavation of the basin. The canal and Intake Works were commissioned in 1983 and trials were

Management history

FIGURE 2.11 Construction of the Intake Works to allow Mfolozi River water to enter the Link Canal. The construction had automatic sluice gates which would close off automatically to prevent floodwaters from entering the canal. (Photo: R. H. Taylor, 1983.)

conducted (Figures 2.11 and 2.12). Shortly thereafter the Cyclone Domoina flood damaged the construction and it was never used. The concept had been that the Link Canal (Table 2.1) would transfer Mfolozi water into St Lucia, to reduce the extent and duration of high salinity conditions during drought periods. But, with hindsight, the impact of hypersalinity had possibly been overstated and that of sediment understated. Duncan Hay, who had been studying the effects on the benthic fauna of the Narrows of the fresh water that would be brought in by the canal, expressed his concerns about silt from the Mfolozi River in a report to SCADCO (Hay, 1985b). He considered that an unacceptable amount of Mfolozi silt would be brought into St Lucia by the quantity of water that needs to be introduced to reduce hypersalinity in St Lucia. Had there not been such a strong directive to keep the mouth open, and had it had been accepted that the estuary mouth could be allowed to remain closed for a period after each closure, it is likely that the specifications for the operation of the canal would have been quite different. Less fresh water would have been required to achieve the same effect in a situation where there is no tidal loss of fresh

FIGURE 2.12 Oblique aerial photo showing the partially constructed Intake Works. In the foreground is the Mfolozi River and in the background the Link Canal which was still being excavated. (Photo: R. H. Taylor, 1983.)

water, and if the lake level were allowed to drop to some extent, reducing the evaporative area of the system. This had been a period when detailed data had been collected to be used to develop a salt and water balance computer model. This model was used to simulate salinity and water levels in St Lucia given a number of different possible management scenarios. The scenario deemed to be the best was the construction of the Mfolozi Link Canal, which was then constructed to bring Mfolozi water into St Lucia.

37

38

Ricky H. Taylor

2.7 1984–2000: Cyclone Domoina and its aftermath Cyclone Domoina passed over the catchments of St Lucia in January/February 1984. Associated with it was a large amount of rain and its path was from the upper catchment towards St Lucia (van Heerden, 1984). The resulting flood peak, estimated to have been 16 700 m3 s 1, was the largest ever recorded in the Mfolozi River (Lund, 1984). This flood filled the entire basin of the Mfolozi floodplain and deposited large quantities of sand in the upper reaches of the floodplain near Riverview (van Heerden, 1984). The water flowing from the Mfolozi floodplain split when it reached St Lucia – part of it flowing out to sea and part northwards into the lake (Figure 2.13). A boat from St Lucia was washed northwards to be deposited on a mangrove tree at The Forks (Table 2.1), 16 km from the mouth. Later the flow direction in the Narrows reversed as large volumes of water were flowing from the Mkhuze and all the water then went seawards. The water flowing through the mouth was erosive, cutting a part of the channel downstream of Honeymoon Bend to a depth of 18 m (Taylor, 2011a). Dredge spoil piles that had accumulated on the south bank of the St Lucia Estuary since dredging operations first started in the 1950s acted as an impediment to the smooth flow of the floodwater. It is likely that this formed a bottleneck focusing the water flowing through the Mfolozi Mouth against the base of the 150 ft (50 m) Sugarloaf Dune at Maphelane. It eroded this dune away. The outflowing water washed away the hard structures at the St Lucia Mouth, washed the dredge out to sea, where it sank, and severely damaged the almost-completed Link Canal

and its Intake Works. Two weeks later, the coast was hit by Cyclone Imboa which caused exceptionally high seas that coincided with high tides. This eroded the beach, undercutting the base of the coastal dunes, and enlarged the already damaged mouths of the St Lucia and Mfolozi Estuaries. The cyclones had damaged much of the sugarfarming infrastructure on the Mfolozi floodplain and had wiped the slate clean for the management of the St Lucia Mouth. Immediately after the cyclones, Dr Ivor van Heerden, a coastal geologist, and Dr Harry Swart, a coastal engineer, were contracted to advise on future management. They conducted detailed field surveys (van Heerden and Swart, 1986) and advised on new ways to manage the St Lucia Mouth. Their approach was that we should work with the natural coastal processes. They identified that ‘the key to successful management of the Mfolozi and St Lucia Estuaries is the correct management of the Mfolozi Flats’ and recommended against any further canalization or diversion works in the lower Mfolozi area. On their advice the two mouths were still to remain separate and no hard structures should be built to stabilize either mouth (van Heerden and Swart, 1986). Natural spit/bar configurations should be allowed to develop in the St Lucia Mouth and dredging was to be done only when needed. To implement this, a new cutter-suction dredger was purchased and the focus of the dredging was to excavate a sediment trap in the distal margin of the flood tidal delta to remove marine sediments that FIGURE 2.13 The Domoina flood covered the floodplain area on either side of the Narrows. This photo shows the flooded approach road to St Lucia looking eastward – as a sheet of water of over 1 km. (Photo: R. H. Taylor, February 1984.)

Management history

were settling there. These sediments were pumped northwards to be deposited in the surf zone several hundred metres north of the mouth from where they would be carried northwards by the longshore drift. The CSIR was commissioned to monitor the dredging in 1987–1988, to determine the relationship between the amount of dredging done and the size of the estuary mouth (Badenhorst, 1988). This study also showed that marine sediments had entered the first 800 m from the mouth and beyond that there were estuary-derived sediments. Complementing this was the study by Ian Wright, who worked out the sediment dynamics to obtain a sediment budget for the various sedimentary processes (Wright, 1995). With the Mfolozi separated from St Lucia, dredging was a necessary management tool and without it the St Lucia Mouth would ultimately close during low-rainfall years (Wright and Mason, 1993). The Mfolozi-St Lucia Link Canal had been severely damaged by the Domoina floods, which had eroded parts of the canal and had deposited sediments in other parts. It was closed after the 1984 flood, but was breached again by the September 1987 floods. From September 1987 until January 1988, the Mfolozi River flowed uninterrupted into the lake via the Link Canal. The state and need for the Link Canal was assessed and the decision taken to ‘put it in cotton wool’ (Taylor and

Collings, 1988). This required the construction of several cross-berms to prevent flows in the canal and to hopefully raise the water table to prevent drying of the swamp on either side. The main purpose was to prevent future Mfolozi floods from using the canal as a preferential flow pathway. On the Mfolozi floodplain, some of the farmland had been covered by more than a metre of sand by the Domoina flooding (Figure 2.14). This land was decommissioned and purchased by the state. The floodwaters had burst out of the Mfolozi canal and expanded to fill the floodplain. Much of the water passed down the Msunduzi channel and through Wilson’s Drain to rejoin the Mfolozi River at the confluence of the two rivers at Maphelane. Once floodwaters subsided, this Msunduzi channel became the dominant pathway for the Mfolozi River. The Mfolozi had to be relocated into the Mfolozi canals. A large spillway, about a kilometre in length, was constructed near Riverview to split future floodwaters, so that all water higher than about 800 m3 s 1 would be shunted down the Msunduzi (Taylor, 2011b). In the aftermath of the 1984 and 2007 floods, the conservation value of St Lucia, as a nature conservation area, was reaffirmed by its designation a Wetland of International Importance under the Ramsar Convention (Chapter 1). Part of the

Thick sand deposits Mtubatuba road bridge

Thick mud deposits

St Lucia Mouth

Mfolozi Mouth

High ground Riverview

High ground

lozi Mfo

Flood diverted at this point

High ground

nd Msu

High ground

uzi

High ground Flood channel of the Mfolozi River

FIGURE 2.14 The sediment deposits on the Mfolozi floodplain after the Cyclone Domoina floods. The floodwaters spread to fill the whole basin, but much of the flow was directed down the Msunduzi River channel. The main deposits of sediment follow this pathway (van Heerden, 1984).

39

40

Ricky H. Taylor

Convention is a national commitment to conserve the wetlands of St Lucia, including the St Lucia estuarine system. The Natal Parks Board also prepared the submission for the park to be declared a World Heritage Site – which was attained in 1999, placing St Lucia in that elite category of global natural sites that are considered to be the ‘Best of the Best’. With this status comes the obligation for South Africa to ensure the site is adequately conserved in perpetuity for all the people of the world (Chapter 1).

The legacy of the 1984 Domoina flood was that it removed the hard structures at the mouth. There had been concerns that the design was flawed, but no decision had been taken to alter the strategy of having a stabilized mouth. The Domoina flood allowed a reappraisal of the management strategies and a change in philosophy, where we were to work with the natural processes. As this was underpinned by scientific knowledge of these processes, further research was needed to better understand the system.

2.8 2000 to present: the current drought 2001 was a particularly wet year, terminating a series of above-average rainfall years. After this, rainfall decreased drastically, rivers stopped flowing and the lake level dropped. This was the onset of the current (in 2012) drought. When the mouth closed in 2002, the management decision was taken to leave the mouth closed for the duration of the drought – which was expected to last for three to five years, as had all previous droughts since the Mfolozi had been separated from St Lucia. This was a deviation from the management philosophy of keeping the mouth open that had been in place since the 1950s. It soon was evident that as long as the Mfolozi is not connected to St Lucia, then drought impacts on St Lucia are severe. The management choice is to either open the mouth artificially, causing the mass of salt in the system to build up as water is lost to evaporation and a large influx of marine sediments into the mouth area; or to keep the mouth closed so that the salt loading does not increase as there is no addition of seawater, but then water levels do fluctuate wildly, and parts of the lake dry up (Taylor, 2006). An additional consideration is that when the mouth is closed, all movements of biota between the estuary and the sea are prevented, effectively stopping the nursery function of the estuary (Whitfield and Taylor, 2009). The initial responses to the drought were a lowered lake level, the expansion of fringing vegetation onto the exposed beaches and promotion of groundwater-

dependent habitats – especially along the eastern shoreline, where there was a lot of seepage. In the lake, beds of Stuckenia pectinata flourished and were replaced by Ruppia cirrhosa as water evaporated and there was a rise in salinity (Chapter 11). The exposure of large portions of the lake bed caused the mass mortality of bivalves. With more water loss there was a die-off of many of the fish and a disappearance of most of the fish that have to breed at sea and were now unable to recruit (Chapter 15). Very large populations of tilapia, Oreochromis mossambicus, developed, as these fish are tolerant of an exceptionally wide range of salinity conditions. These fish have supported large populations of great white pelicans which, with the abundant food supply, have bred well (Chapter 16). As salinity concentrated still further, a very simple halotolerant food chain established itself in the northern lakes (Carrasco and Perissinotto, 2012; Chapter 13). Water levels dropped further and in 2006 up to 80% of the lake bed was exposed and dry. Prior to this, there was an accumulation of salt in the northern reaches of the system; salinity concentrations of well over 200 were measured before salt precipitated and then dried out. Winds shifted fine sediments and large quantities of dust (and salt) were blown out of the system (Bate and Taylor, 2008). On several occasions the Mfolozi Mouth closed. Whenever this occurred it reached elevations high enough to allow it to flow through the Back Channel

Management history

into St Lucia. The Mfolozi flowing into the flooded Mfolozi/Msunduzi area loosens all its sediments in the static water, and the water that enters St Lucia is then sediment free (Kelbe and Taylor, 2011). The more this water is able to back up on the floodplain the more effective is the sediment trapping. As the water level rises, greater volumes of water can enter St Lucia. However, the backing-up floods some of the lower farms and affects the drainage of others resulting in expensive damage to the sugar cane. This situation is a short-term solution that allows some water to transfer from the Mfolozi. An important workshop was held in St Lucia in 2004. The National Water Act (Act 36 of 1998) specifies that a quota of water, the Ecological Reserve, shall be set aside for the ecological maintenance of water systems. The workshop was set up to establish the Ecological Reserve at a ‘rapid level’, which is based on published or readily available data and information (Van Niekerk, 2004). For this study, simulated runoff data were provided to the study team by the Department of Water Affairs. The ‘Present State’ mean annual runoff (MAR) for the combined Mkhuze, Mzinene, Nyalazi, Hluhluwe and Mpate rivers was estimated to be 362.26 × 106 m3 yr 1, which is 86% of the MAR under the ‘Reference Condition’ (i.e. 417.89 × 106 m3 yr 1). The ‘Reference Condition’ is a simulation of the flows that would have occurred prior to development of the catchment areas. Using these data, the ‘Present Ecological Status’ is determined using the ‘Estuarine Health Index’, which consists of a ‘Habitat Health’ score and a ‘Biological Health’ score. The scores provide a percentage deviation from the ‘Reference Condition’. For the St Lucia estuarine system, a ‘Category D’ was assigned whereas the recommended ‘Ecological Category’ should be a ‘Category A’. This low category was due to the ecological deterioration caused by developments such as the drainage and canalization of the Mfolozi Swamps, the construction of weirs on the Nyalazi, Hluhluwe and Mpate rivers (Table 2.1), the overfishing due mainly to the illegal netting of mullet and an overall reduction in bird habitat. The aim for such a highly rated estuary is to manage it to

improve its condition – and the managers should strive for a ‘Category A’. This would be attainable only if the Mfolozi River is once again connected to St Lucia (Whitfield et al., 2012). This current drought has led to new insights into the management of St Lucia. There is the realization that we know little about the state and functioning of St Lucia, as it was before the Mfolozi was separated from it. Virtually all the research that has been done has been since then. There has been the realization that strong outflows can occur – either during big floods or when the mouth breaches after a considerable amount of water has backed up behind the beach berm. This is the main process that scours the basin of the combined Mfolozi–St Lucia Mouth area. There is also the understanding that flow of the Mfolozi River through a combined mouth is important in reducing the influx of marine sediments (Chapter 7). This sort of conceptual understanding has been developed by Taylor (2006) and by Lawrie and Stretch (2011a, 2011b). With the acceptance that it is necessary to reconnect the Mfolozi to the St Lucia Estuary, a review of studies relating to the Mfolozi Estuary and the associated floodplain was held in 2010 (Bate et al., 2011). This is a useful compilation of the current state of knowledge. Currently, a Global Environment Facility (GEF) programme, funded by the World Bank, is being implemented to consider various options to bring Mfolozi water into St Lucia. At the time of writing (early 2012) the drought is still severe, St Lucia contains only a small volume of water and the mouth is still closed. The decision has been taken to allow the Mfolozi to link naturally to St Lucia, should this occur.

From the responses of the system to the drought that has persisted since 2002, the understanding of how the system functions has increased to a large degree. The Ecological Reserve Determination workshop in 2004 identified very clearly that for the system to function well the link between the Mfolozi and St Lucia must be reconnected (Whitfield et al., 2012).

41

42

Ricky H. Taylor

2.9 Future management There are many threats to the integrity of the St Lucia system, which need to be addressed in the near future. The priority must be to prevent further deterioration of the system. One area to focus on is the Mkhuze Swamps. Parts of the Mkhuze Swamps were damaged by the canal excavated in the 1970s. Only the eastern half of the swamp falls within the iSimangaliso Wetland Park and currently there is the threat that the remaining swampland, that to the west of the river course, will be permanently damaged by the unauthorized planting of sugar and the uncontrolled establishment of informal agriculture. The St Lucia estuarine system is possibly much more sensitive to the loss of the Mkhuze Swamps than to the loss of the Mfolozi Swamps, as the connectivity with the former is more direct. Loss of the sedimenttrapping function of this swamp is one of the largest threats facing St Lucia.

2.9.1 Management of sediments The recommendation is to allow the Mfolozi to once again link with St Lucia, so that its fresh water enters St Lucia during dry periods. This is detailed in Chapter 7. But, with the linking there will be issues relating to sediment depositions that will need to be addressed. Since the late 1930s, there have been concerns about the influx of sediments into St Lucia. One concern is that that there has been an increase in sediment yield from the catchment area, due to a deterioration of catchment condition. Another is that the canalization of the Mfolozi floodplain has reduced its ability to trap sediments. There has been sediment accumulation in the mouth area of St Lucia, especially during drought periods, which has been attributed to these land-use changes. However, one of the problems is that the distinction has seldom been made between sediments brought in that are of marine origin (i.e. sand) and those from the Mfolozi

River (which are either silt carried in suspension or bedload sand which is often carried by the higher flows). A better understanding of sediments is still one of the important knowledge gaps for the system. The trapping of sediments in the Mfolozi floodplain basin occurs when there is river avulsion. This occurs in the larger floods, such as Domoina, but the minor floods are nowadays contained within the canalized river channel. Thus the point of deposition of these sediments has been transferred, by the river passing through some 30 km of canal, from near Riverview to a point close to the mouth, where the canal ends. Are these sediments now being accommodated in the lower parts of the Mfolozi–Msunduzi swamp by the slumping of the deep sediment deposits, are they being washed out to sea (when the mouth is open), or do they accumulate in the lower reaches of the St Lucia Estuary? This is something that needs further investigation, so decisions can be taken as to whether management action should focus on the restoration of the sediment-trapping abilities of the Mfolozi swampland or not. Marine sediments are brought into the St Lucia Estuary with each flood tide. When there was a combined Mfolozi–St Lucia Mouth, strong river flows would have removed much of this sediment with the ebb tide. This is a strong benefit of relinking the Mfolozi to St Lucia – especially as we know that nowadays about 20% less water enters St Lucia from its catchments (and hence flows out of its mouth). In the drought that lasted from the late 1930s until the mouth closure in 1952, there was a net loss of water from St Lucia and the Mfolozi water diverted naturally to replace it. The scouring of the mouth by fluvial outflow diminished and more marine sediments entered. With our current knowledge of the situation, it was likely that most of the sediment

Management history

accumulation reported during this period was marine and very much less was Mfolozi sediment. Although there may have been relatively much less Mfolozi sediment, the fine component (i.e. the clay and silt) is much more evident to the public. Hence the perception is that the Mfolozi sediment was much more important than the marine sediments. This fine fraction also has a much greater impact on the biota than the sand. Conceptually, the sediments of the mouth area were periodically flushed from the system during a mega-flood, or even more so during a breaching event, once water levels in the system had risen to be above mean sea level (as occurred in 1932). We are entering a management phase where we will allow the Mfolozi to once again link to St Lucia. These sediment issues will need to be addressed in much greater detail than they have been in the past. With overall less fresh water, there will be less flushing of sediments during normal tidal exchange – and so dredging will still be needed to remove marine sediments. The situation where the Mfolozi diverts naturally into St Lucia will occur more frequently than in the past and, as it is unlikely that the public will tolerate a large rise in the water level behind the beach berm, so artificial breaching at a specified level is inevitable. The loss of sediments on a breaching event will be less. Possible mitigating measures would include judicious breaching of the mouth, so that it remains closed for shorter periods – and thereby having less sediment accumulation, or the removal of sediments by dredging.

2.9.2 Management of water quality To date there have been few problems relating to the addition of nutrients and toxins in the St Lucia water. However, considering international trends, there is a strong likelihood that water quality in St Lucia will deteriorate in the future. This is something where there has been little research or monitoring. There

are a number of activities occurring in the catchment areas that could impact on the water quality. As agriculture increases and becomes more intensified, so there is more use of fertilizers, ripening hormones, herbicides, insecticides and other agrichemicals – this has been the case in the sugar industry. With increasing numbers of people living in the catchment, there is sewerage pollution causing nutrient enrichment of water, and with industrialization and urbanization there is a suite of additional chemicals that could pose problems in the future. The leachate from the tailings dumps of the coal mines in the upper reaches of the catchments is lowering the pH of the water (Mackay, 1993). The indications are that reduction in water quality is likely to become a large problem in the future. This is likely to manifest itself as eutrophication, which will promote algal growth – especially at times of reduced river inflows. Algal blooms and scums in rivers entering the park provide informal evidence of the early stages of this. It could lead to diseases such as botulism and microcystis poisoning, diseases in fish and possibly the pansteatitus that has been evident in the crocodiles in rivers of the Kruger National Park (Ashton, 2010). There has been a scare when ducks and other waterbirds died of unknown causes at the St Lucia sewage works in February 2008. Unfortunately, the less the flushing of the system, due to reduced river inflows, the more susceptible it is to the accumulation of pollutants – both in the water column and in the sediments. Pollutants are trapped to some degree in the swamps, so the filtering effect of both the Mfolozi and Mkhuze swamps is important.

2.9.3 Management of the catchment areas Most of the problems affecting St Lucia are a result of poor catchment management. The number of people in the catchment area has expanded

43

44

Ricky H. Taylor

by an order of magnitude in the past century. Wise water management is needed to

reduce the 20% drop in the MAR and declining water quality.

2.10 Assessment and lessons St Lucia is a system that is at the stage in its geological evolution where it easily flips from the state of being an estuary, where connectivity with the sea is a major feature, to a coastal lagoon which may be disconnected from the sea for long periods. The system has gained sediment and shallowed to the point where it is extremely sensitive in its responses to changes in rainfall and runoff – leading to a highly variable ecology. Management interventions should aim to reinstate processes that have been damaged or to prevent the acceleration of human-induced processes, and not control the natural variability which is a feature of the St Lucia estuarine system. This has not always been recognized and several of the management schemes implemented at St Lucia have failed. With hindsight, it is easy to be critical of past initiatives. However, it should be remembered that the interventions were undertaken by people who were passionate about the welfare of the area and who considered that what they were doing was the best course of action at the time to conserve this wonderful system. In the past there has often been a single individual who has had an excessively large influence on the management of the system. This can be to its benefit, if the person is on the right track, but can be detrimental if not. Management of the system needs the checks and balances of having an advisory team of experts who are familiar with the system. That team would ensure continuity of knowledge, counter the overwhelming influence of a single strong-willed person and bring the best available technical and scientific knowledge to support the implementation of management actions. In the 1940s and 1950s, the style of management was that of ‘command and control’. This has evolved

to a style which takes the softer approach of working with the natural processes. In South Africa, a mandatory Environmental Impact Assessment (EIA) process was only introduced in 1993. This means that management actions now need to go through a formal assessment prior to implementation and there is less opportunity for an individual to dominate the decision-making process. St Lucia, as is the case with all estuaries, is at the bottom end of a catchment and hence is affected by upstream modifications. This is taken into account at St Lucia where, within the World Heritage Site Act, a ‘zone of influence’ is recognized. This is a good concept, as it allows for the creation of upstream ‘buffer’ zones where various development activities which may impact on the World Heritage Site need to be scrutinized as part of the EIA process. At St Lucia, the management interventions have been guided by the best available knowledge of the time and, as this knowledge has improved, so the management implemented has changed. An adaptive management strategy has been applied, which is informed by continually improving scientific knowledge of how the system functions and of lessons gleaned from similar systems elsewhere in the world. The key imperatives for adaptive management are: • To have a clearly stated conceptual model, which describes the dynamics of the system and allows for the prediction of how it is likely to respond to management interventions. As the knowledge increases, this conceptual model is refined and when possible the understanding is incorporated into computer models.

Management history

• Then the management is implemented – and good records of the implementation are maintained. • The responses of the system are monitored, and compared with what was predicted prior to implementation. Where the actual response is different from the expected response, the original conceptual understanding must be interrogated and modified to incorporate any new understanding.

• Future management is then adapted to incorporate the new understanding. It is through this process of ‘learning by doing’ that management is continually improved. Over the years, St Lucia has been an excellent example of adaptive management in action where there has been learning, based on the outcomes of previously implemented management strategies.

45

Chapter contents 3.1 Introduction 3.2 Synthesis of the geological evolution of the St Lucia basin 3.3 The late Pleistocene and Holocene evolution of the St Lucia basin 3.4 Evaluation of St Lucia sedimentation 3.5 Conclusions

Exposed ammonite at Hell’s Gate. (Photo: Ricky H. Taylor, December 2001.)

3

Geological history Greg A. Botha, Sylvi Haldorsen and Naomi Porat

3.1 Introduction The evolution of the St Lucia estuarine lake basin and the other coastal lakes in Maputaland bears testimony to the cyclical environmental changes operating across a range of temporal scales during the past 2.6 million years (Ma) of the Quaternary Period. The lake environment is constantly changing in response to fluctuating discharge and sediment flux from the five main catchments, evaporation from the shallow lake, groundwater level change, and the influence of the littoral marine environment on the estuary. The impact of these factors during the Quaternary Period can be interpreted from the geological record in the surrounding dunefields and sediments below the lake bed. Sedimentary deposits show that the St Lucia lake has been subjected to frequent marine transgressions and regressions and that the present shoreline morphology was only achieved within the past ~1000 years. The system has evolved in response to cyclical fluctuations of marine base level and the shifting confluence of the fluvial systems with the Indian Ocean. This has led to the St Lucia lake basin alternating between an open marine embayment or shallow continental shelf environment during sea-level highstands and a subaerially exposed shallow valley lowland incised by river channels during sea-level lowstands (Orme, 1973; Wright et al., 2000; Green and Uken, 2005; Botha and Porat, 2007; Porat and Botha, 2008).

The St Lucia lake basin was elevated above sea level as a terrestrial ecosystem for longer periods during the Pleistocene and Holocene than it was inundated to levels similar to its present lacustrine morphology. The resilience of this environment and its biota to short-term changes such as periodic salinity increases has been influenced by the evolution of the system over geological temporal scales. The system has, in addition, been subjected to anthropogenic stresses over the past few centuries imposed by catchment degradation, rapidly increasing fluvial sediment load, canalization of the lower river channels, dredging of the narrow marine interface and rising sea level. In this chapter we present the evidence for geological and ecological processes that have shaped the St Lucia estuarine lake basin, emphasizing the development during the last glacial cycle, particularly since the Last Glacial Maximum (LGM), about ~20 000 calibrated radiocarbon years before present (cal. yr BP). The depositional environments and sedimentology of the main geological units are described to serve as the basis for interpreting the likely responses of the catchment, lake and estuary system to the impacts of altered precipitation, catchment discharge and rising sea level caused by global change.

Ecology and Conservation of Estuarine Ecosystems: Lake St Lucia as a Global Model, eds. Renzo Perissinotto, Derek D. Stretch and Ricky H. Taylor. Published by Cambridge University Press. © Cambridge University Press 2013.

48

Greg A. Botha et al.

3.2 Synthesis of the geological evolution of the St Lucia basin The St Lucia Estuary basin occupies the interface between the high energy Indian Ocean coastline and the highly variable terrestrial environment comprising five main river catchments with irregular topography influenced by diverse bedrock types and strongly seasonal atmospheric conditions. The geology of the region represents over 3.2 billion years (Ga) of earth history although the Indian Ocean coastline and St Lucia catchments owe their origin to geomorphic processes that began with the incipient rifting of the supercontinent Gondwana some 180 million years ago followed by the gradual opening of the Indian Ocean after ~140 Ma (Broad et al., 2006). In the catchments of the Mfolozi and Mkhuze rivers and the smaller Hluhluwe, Nyalazi and Mzinene rivers, the oldest rocks are the 3.2 Ga Kaapvaal craton granites and the overlying Pongola Supergroup sedimentary and volcanic rocks (~2.9 Ga). These basement rocks are unconformably overlain by the Permo-Triassic Karoo Supergroup sedimentary succession (~260–210 Ma). The ancient continental crust beneath this region has been part of several continental assembly and rifting processes driven by global plate tectonics. Pangaea split in the early Jurassic to form Gondwana which subsequently rifted, beginning the evolution of the Indian Ocean and emergent south-east African continental margin. Gondwana break-up was initiated by the injection of Karoo magmas around 182 Ma and final rifting occurred about 40 million years later, associated with sea-floor spreading linked to the opening of the south Atlantic and development of Cretaceous volcanism east of the Lebombo Mountains (Watkeys, 2000, 2006). The Lebombo Group volcanic rocks (~184 to 133 Ma) (Duncan et al., 1997; Watkeys, 2006) separate the Permo-Triassic sedimentary succession from the overlying Mesozoic and Cenozoic marine deposits that underlie the coastal plain. The basal Letaba

Formation basalt underlies plains west of the Lebombo Mountains which are formed by the ~2000 m thick Jozini Formation rhyolite and pyroclastic volcanic rocks. The final Bumbeni volcanic event on the eastern Lebombo foothills within Mkhuze Game Reserve can be traced north-eastward under Mesozoic cover, forming a palaeo-ridge that influenced subsequent Cretaceous sedimentation (Figure 3.1). The rhyolites lie at 2015 m depth under Charter’s Creek (borehole ZU-1) where they were downfaulted during the late Jurassic or early Cretaceous about 1000 m in a north–south aligned rift valley graben relative to the same rocks that lie at 970 m beneath Makakatana point to the east (borehole ZH1). This narrow rift valley was probably surrounded by terrain that resembled the Red Sea of today. The linear western shorelines of St Lucia and False Bay are aligned above deeply buried, ancient fault lines raising the question as to whether there has been periodic rejuvenation of these faults. Sedimentation in the Mesozoic Zululand Basin below lake St Lucia and the surrounding wetlands was initiated as a response to uplift of the emergent continental margin which led to sedimentation into subsiding rift valleys and accumulation of the Makatini Formation conglomerate and sandstone (Barremian to Aptian age; ~130–112 Ma) of the Zululand Group (Kennedy and Klinger, 1975; Dingle et al., 1983). Post-rifting erosion stripped 1–3 km of mainly Karoo Supergroup cover rocks off the land surface which resulted in deposition of thick offshore sediments onto the subsiding continental shelf. During the subsequent marine transgression and regression cycles the continental margin was submerged as a shallow continental shelf for much of the Cretaceous Period (late Barremian to late Maastrichtian, ~130–65 Ma). An erosional or non-depositional hiatus spanning the Aptian–Albian boundary marks the unconformity above which the Mzinene Formation (Albian to Cenomanian) shallow

Geological history

FIGURE 3.1 Geological map showing the Jurassic and Cretaceous bedrock underlying the Cenozoic Maputaland Group sediments forming the coastal plain dunefields and barrier dune. (After Geological Survey, 1985.)

marine, glauconitic siltstones and fine-grained sandstones accumulated (Kennedy and Klinger, 1975). The diversity of biota increased dramatically in the Albian and includes infaunal and epifaunal bivalves, gastropods, ammonites, nautiloids and echinoids (Kennedy and Klinger, 1975; Tankard et al., 1982; Shone, 2006). This unit comprises a succession of upward-fining, bioturbated, glauconitic sandstones and siltstones with the base of

each cycle characterized by a layer of Lithophagabored limestone concretions, overlain by a 20–30 cm thick shell lag (Tankard et al., 1982), exposed in the cliffs around parts of False Bay. The major mid-Cretaceous hiatus separates the Mzinene Formation from the overlying late Cretaceous, St Lucia Formation (Coniacian to Maastrichtian) which represents a renewed transgression across the Zululand Basin (Kennedy and

49

50

Greg A. Botha et al.

Klinger, 1975). The cliffs around the Nibela Peninsula and the western shoreline near Fani’s Island expose the low eastward dip (< 3 ) of the St Lucia Formation deposits (Kennedy and Klinger, 1975; McCarthy and Cooper, 1988). The buff and greenish grey, richly fossiliferous glauconitic siltstone and fine-grained sandstone with large calcareous concretions contain plant fossils and drifted logs. The succession comprises numerous upward-fining sedimentary units. These consist of basal concretionary limestone (c. 40 cm, upper surface bored and serpulid encrusted) containing abundant, large mollusc fragments (especially Inoceramidae and ammonites) (see Kennedy and Klinger, 2006, 2008a, 2008b, 2010; Walaszczyk et al., 2009). Upward-fining, glauconitic current-aligned shell lag (c. 15 cm) and glauconitic sand with few shell fragments (c. 5–10 cm) grade upwards into glauconitic siltstone beds 2–3 m thick (Dingle et al., 1983). The development of the drainage systems flowing from the proto-Drakensberg escarpment is an ongoing geomorphological process than can be traced back some 140 million years since the rifting of East and West Gondwana. The catchment history during this period has been dominated by erosional degradation of a succession of land surfaces of which only isolated remnants of the oldest, the African Surface, remain in the St Lucia catchments (Partridge and Maud, 1987, 2000). The development of the lake basin during the late Quaternary Period is preserved in Maputaland Group coastal lake, wetland and dune deposits (Figures 3.1, 3.2). The continental margin was inundated by rising eustatic sea level during part of the Eocene (56–34 Ma), followed by the regression of the Oligocene (34–23 Ma). Deposits associated with this period are not exposed at the surface in the St Lucia catchment. During the Neogene Period (23–2.6 Ma) the midMiocene marine highstand was followed by the regression in response to eustatic sea level lowering superimposed on pulses of epeirogenic uplift during the late Miocene to Pliocene (~12.5–2.6 Ma) (Partridge and Maud, 1987, 2000). The regression left a stepped, marine-planed bedrock surface overlain by a series of strandline deposits. The cliffs forming

parts of the linear western margins of the Nibela–Ndlozi Peninsulas and False Bay expose fossiliferous Cretaceous siltstones. These are capped by the calcareous, Mio-Pliocene Uloa Formation shell coquina, conglomerate and coarse-grained sandstone that were deposited in shallow marine and high energy beach and nearshore environments. The overlying cross-stratified Umkwelane Formation aeolianite forms a pronounced ridge along the western lake shoreline and the high ground west of False Bay and is generally decalcified and weathered to produce clayey reddish-brown soil (Figure 3.1). The evolution of the coastal zone and hinterland of South Africa during the Quaternary Period (last 2.6 Ma) was dominated by denudational processes brought about by repetitive base-level changes stimulated by glacio-eustatic sea level fluctuations due to Pleistocene and Holocene climate variability. There are no coastal deposits exposed that record depositional environments during the ~2.2 million years after the Pliocene regression although there were numerous global climatic cycles during this period. Before ~350 ka, the late middle Pleistocene, Port Durnford Formation fossiliferous estuarine deposits accumulated, linked to a sea level slightly lower than the present. Between 300 and 100 ka a series of coastal dune and wetland systems accreted as the composite Kosi Bay Formation succession. Over much of the coastal plain these weathered dune sands and wetland lignite deposits are buried by the surficial dunes of the KwaMbonambi Formation. The northward moving parabolic dunes were mobile in the Eastern Shores area during the terminal Pleistocene and Holocene from 11.7 ka to 8 ka (Porat and Botha, 2008). During the penultimate and last glacial cycles (last ~200 000 years) several generations of calcareous coastal dunes accreted to form the Isipingo Formation aeolianite, locally overlapping onto the Kosi Bay Formation barrier and coastal plain dunes (Figures 3.1, 3.2). At least four phases of Sibaya Formation parabolic dunes formed stacked accretion ridges on the high coastal barrier dune during the Holocene from ~10 ka to 2 ka.

Geological history

East

100

West

boreholes

100

Sibayi FM Kosi Bay FM 0

Reworked PD ?

Isipingo Formation

0

Kosi Bay FM

Port Durnford FM

Alluvium Cretaceous –80

0

–80

500

1000

1500

2000 Coastal barrier cordon - north of Mission Rocks (BHA5 to BH AA1; after Davies lynn and Partners, 1992)

FIGURE 3.2 Cross-section through the composite coastal barrier dune showing the thin surficial cover of Sibayi Formation parabolic dunes burying several generations of Isipingo Formation aeolianite and weathered Kosi Bay Formation dunes. This succession suggests periodic aeolian sand accretion against the barrier dune over a period of at least 350 000 years. Section east of Catalina Bay, section line A. (After DLP, 1992.)

3.3 The late Pleistocene and Holocene evolution of the St Lucia basin The last glacial cycle, spanning the late Pleistocene to Holocene, c. 125 000 to 11 000 yr BP, provides the best documented examples of a range of geomorphological processes that shaped the landscape during the numerous global climatic cycles of the preceding 2.6 million years. The Last Interglacial sea level highstand at 130–120 ka (Marine Isotope Stage 5e) exceeded the level achieved during preceding Quaternary climatic optima (Hearty et al., 2007). The St Lucia coastal plain morphology and evolution of the lake/estuary system was most profoundly influenced by the sea level response to the stepwise growth of continental ice sheets and environmental desiccation after the Last Interglacial. This period included the marine regression associated with MIS 5–2 leading up to the coldest, driest phase around 20 000 cal. yr BP (Waelbroeck et al., 2002; Siddall et al., 2010) and the climatic amelioration during the terminal Pleistocene and Holocene. Remobilization of aeolian sand cover led to the northward migration of KwaMbonambi Formation parabolic dunes from drained coastal lake basins and river floodplains from ~60 ka to ~7 ka which has been documented using luminescence dating techniques (Porat and Botha, 2008). The dune systems forming the present surface of the Eastern Shores, Ozabeni area and the polyphase accretion

against the coastal barrier dune represent periods of dune activity linked to climate variability over the past ~300 000 years. Sea level fluctuations during the past 125 000 years have been tracked using a radiocarbon and uranium series chronology derived from coastal beachrock or littoral zone deposits (Ramsay, 1996; Wright et al., 2000). The thick unconsolidated dune sand cover of the Maputaland coastal plain is the result of the longterm evolutionary connections between the eroding catchments, the littoral marine environment and complex terrestrial coastal plain dune and wetland environments. In the interior, episodic hillslope denudation and sediment transport into the Mfolozi and Mkhuze river channels over the past ~130 000 years can be gauged from the hillslope deposits and the interbedded palaeosol succession of the Masotcheni Formation (Botha et al., 1990, 1992; Botha, 1996; Botha and Partridge, 2000; Clarke et al., 2003). The cyclical erosional history shows catchment-scale pulses of erosion and high sediment loads carried to the coast by the major rivers in response to global climate changes. Littoral zone sediment is transported northwards by longshore currents whereas offshore sediment is moved southwards over the narrow continental shelf by the Agulhas Current (Flemming, 1981; Birch, 1996).

51

52

Greg A. Botha et al.

Ramsay (1995, 1996) and Ramsay and Cooper (2002) generated a late Pleistocene and Holocene sea level curve for south-east Africa. Variations and short-term excursions from the general sea level trend could reflect the differing relative sea level indicator precision of dated sea level indicators used to derive the curves. The influence of the last glacial cycle on the St Lucia basin is outlined below. It is worth bearing in mind that this glacio-eustatic cycle developed in response to cyclical global climate change is but the last of many cycles resulting from changes in the eccentricity of the earth orbit (100 and 400 ka), the obliquity of the ecliptic (41 ka) and the axial precession of the equinoxes (23 and 19 ka) (Patience and Kroon, 1991). Isipingo Formation dune accretion under the present coastal barrier started around 179 ka (Red Sands) and the aeolianite forming the barrier dune core south of Cape Vidal accumulated at 132 ka during the Last Interglacial. Younger, less cemented aeolianite deposited at 64 ka forms the eroding face of the vegetated barrier dune (Porat and Botha, 2008). Calcareous beach and dune facies sandstone forms First Rocks, Mission Rocks and the Bat’s Cave cliffs. During the Last Interglacial sea level highstand, the coastal barrier dunes were breached and the St Lucia Estuary and tributary valleys were inundated by the Indian Ocean. The linear, western cliffs formed the ocean shoreline with False Bay forming a coastline embayment. There was an open marine connection through the bluffs defining Hell’s Gate where the corals in growth position indicate free marinewater circulation (Hobday, 1976, 1979; Wright et al., 2000). Fossiliferous shell coquina containing molluscs, colonial corals in growth position and solitary corals is preserved from 1m to +3.4 m msl at Lister’s Point and Picnic Point in False Bay (Figure 3.1) and between Hell’s Gate (Hobday, 1976). Further westward in the Munywana Stream valley upstream of the northern end of False Bay an oyster attached to bedrock yielded a uraniumseries age of 95 700 ± 4200 yr BP (Pta-U565) (Ramsay and Cooper, 2002).

3.3.1 Continental shelf aeolianites and canyons After the MIS 5e interglacial, sea level receded rapidly and the lowered base level led to incision of the exposed St Lucia basin. A series of linear aeolianite ridges and beachrock strandlines form at least four submerged palaeocoastlines across the present narrow continental shelf. These record the sea level drawdown and the formation of beaches and coastal dunes now submerged at depths of 40–60 m during short stillstands between 55 000 and 40 000 yr BP (Ramsay, 1995, 1996; Ramsay and Cooper, 2002). Steep-sided submarine canyons that incise the continental shelf-edge wedge begin at depths of 30–40 m within 1–2 km of the shore and have incised the shelf and slope to depths of 650 m abut 8 km offshore (Ramsay, 1996; Green and Uken, 2008; Green et al., 2008). The steep canyons do not exhibit shelf-breaching canyon heads linked to modern fluvial channels, though seismic evidence shows that these canyons were once connected to onshore drainage during the LGM (Green et al., 2009). Intertidal erosional features at depths of 106 m, 124 m and 130 m on the sides of a submarine canyon and conglomerate deposits in caves at 124 m have been associated with the LGM (~21 000 cal. yr BP) sea level lowstand identified by Green and Uken (2005). Despite the close relationship between the lake and sea level it is clear that during the long period when sea level recession exposed the continental shelf, isolated wetlands remained on elevated interfluves. Diatomite accumulated periodically under open water conditions within interdune wetlands near Mbazwana and Mseleni in the Lake Sibaya basin. These wetlands existed from ~53 ka during MIS 3–2 until the wetlands were buried by northward migrating parabolic dunes from 15–10 ka (Miller, 2001; Porat and Botha, 2008). Within palaeotopography defined by the coastal barrier dune and the exposed late Pleistocene deposits beneath the Sibomvini highland, the Mfabeni wetland

Geological history

accumulated organic sediments after ~44 000 yr BP with swampy conditions replaced by dry grassland around 21 000 cal. yr BP, with forest expansion during the Holocene Altithermal (~8000–6000 cal. yr BP) (Finch and Hill, 2008). It is pertinent to note that the youngest pulses of dune mobilization in the Eastern Shores and similar areas to the north ceased around 6 ka due to vegetation growth and rising groundwater that accompanied organic sedimentation within interdune wetlands (Grundling, 1996, 2004; Thamm et al., 1996; Grundling et al., 1998; Porat and Botha, 2008). Thick, well-indurated nodular ferricrete deposits that line the shore along parts of Dead Tree Bay are evidence of fluctuating groundwater seepage that predates the overlying late Pleistocene dune sand cover. Clay-rich, gleyed hydromorphic subsoil with ferruginous nodules and rhizocretions is exposed in the upper Narrows on the southern side of Makakatana Point where this wetland soil is buried by sandy raised beach ridges. Augered transects across the shallow crossing revealed gleyed clay overlain by peat dated at 23 687–24 269 cal. yr BP (Taylor et al., 2004) buried by sand and surface organics with a shallow peat layer yielding an age of 1881 cal. yr BP. The south-eastern margin of Catalina Bay in the Makakatana Point–Brodie’s Crossing area was a stable wetland during a period when sea level was over 100 m lower than at present. Following the coldest, driest climatic conditions during MIS 2 there was a rapid post Last Glacial Maximum climatic amelioration when rapidly rising sea level in response to ameliorating global climate had a profound influence on the low-lying St Lucia basin, the river channels and the marine linkage (Taylor, 2006). The Eastern Shores area, shielded behind the early coastal barrier dune, was an active dunefield for much of this period. Dominant southerly winds drove narrow extended parabolic dunes northwards out of the Mkhuze floodplain from ~19–15 ka. In the Ozabeni area north of St Lucia dunes from the same period were truncated by younger dune migration around 6 ka.

3.3.2 Holocene marine transgression The present St Lucia lake basin has developed from an open lagoon created by the early Holocene (Flandrian) marine transgression (Orme, 1990) that has since been largely transformed into a wetland after 5000 years of sedimentation, segmentation and reed swamp encroachment. At the peak of the mid Holocene there was a marine transgression into the St Lucia proto-lagoon. Orme (1975, 1990) speculated that the lagoon occupied an area of 912 km2 and together with 253 km2 of the inundated lower Mfolozi River wetland covered a total area of 1165 km2 along 112 km of coastline. During the late Holocene the lake area was reduced by about 60% and today it is a shallow wetland about 60 km long and 1–20 km wide, with a 347 km shoreline, covering 350 km2 with an average depth of 0.9 m (Orme, 1990; Taylor et al., 2006a). Evidence of the rising Holocene sea level along the coast is preserved in submerged dune ridges and beachrock, intertidal zone deposits and valley infill deposits on many rivers near the KwaZulu-Natal (KZN) coast (Ramsay, 1995, 1996; Scott and Steenkamp, 1996; Maud, 2000; Grenfell et al., 2010). Calibration of the radiocarbon dates used by Ramsay and Cooper (2002) to compile their sea level curve shows the Holocene transgression reached present sea level around 5464 cal. yr BP and then rose from 4813–3090 cal. yr BP, falling again to the present level around 2294–1611 cal. yr BP. The sandy beach ridges that elevated up to 5 m msl along sections of coastal lakes and estuary shorelines in northern KZN have been described by Orme (1973, 1974), van Heerden (1987), Sydow (1988), Porat et al. (2002) and Botha et al. (2004). The deposits were described as a complex of shorelines, beach deposits and chenier ridges by van Heerden (1987) who regarded them as reflecting environmental changes during the last 5000 years. Hill (1975a) and Sydow (1988) stated that during the climatic optimum 5000 years ago, Lake St Lucia was connected to the sea via an estuarine inlet near Leven Point. Wright (1995) considered Leven Canyon to have been the possible former outlet

53

54

Greg A. Botha et al.

FIGURE 3.3 Fluvial channels incised into the drained proto St Lucia lake bed during the Last Interglacial sea level lowstand. It is postulated that the 40 m deep channel drained into Leven Canyon across the shelf edge. (After Sydow and van Heerden, 1988.)

of the St Lucia/Mkhuze River. What appears to be a buried tidal channel on the inner shelf south of Leven Point (Figure 3.3) was described as signifying a migratory estuary mouth, similar to the Mfolozi– St Lucia estuary mouth before man’s intervention (Sydow, 1988). Foraminiferal evidence from lake-bed sediment cores shows that intermittent marine incursions were common during the sedimentation history of the lake (Phleger, 1976a). The sandy beach ridge sequences in embayments along sections of the Lake St Lucia shoreline were recently described and dated using optically stimulated luminescence (OSL) dating techniques (Porat et al., 2002; Botha et al., 2003, 2004). The beach ridges and swales form truncated points at Makakatana and Sengwana that indicate a gradual shrinking of the water level which mimics the general

sea level curve for south-eastern Africa. Beach ridges at Meme north of Selley’s Lakes and northern False Bay isolate tributary streams and wetlands that were inundated prior to ~3500 ka. A prominent beach ridge extends across much of the confluence of the Mkhuze Swamps and the north-eastern shallows, suggesting that the mid-Holocene marine transgression inundated the lower course of the Mkhuze River (Figure 3.1). Marine fossils beneath Lake Eteza confirm the marine inundation of the lower Mfolozi River valley (Scott and Steenkamp, 1996). The sequence of progressively younger shorelines (Figures 3.1, 3.4, 3.5, 3.6) record the declining water level and shrinking area of St Lucia with stepped regressions after 6.1–5.4 ka, ~4–3.7 ka, 2.4–2.2 ka, 1.3 ka and 0.8–0.6 ka (Porat et al., 2002; Botha et al., 2003, 2004). These high-water indicators show the marine inundation of the St Lucia basin during a period when a lack of other elevated sea level indicators led Ramsay (1995) and Ramsay and Cooper (2002) to suggests that sea level was lower than present for much of the period 1500–500 cal. yr BP (Figure 3.4). The youngest ridge represents stabilization of the lake near to its present mean level around 900 to 600 years ago. Although the sand ridges reach elevations of up to 4.5 m asl, the sandy beach face deposits could be capped by aeolian sand and do not represent accurate lake/sea level indicators. Opposite Makakatana Point the eroded shoreline on the eastern side of Brodie’s Crossing is incised into KwaMbonambi Formation dune deposits dated at 11.7–8.1 ka in the area. Dated sediment cores from Catalina Bay suggests the estuarine lake was at least 3 m deep around 5666 cal. yr BP following the mid-Holocene highstand when the Makakatana beach ridge spit formed. Cooper (1999) described enigmatic, poorly consolidated, fossiliferous Mduku Formation deposits at an elevation of 4 m above present lake level where the Mzinene River valley enters False Bay. Basal conglomerate contains barnacles and serpulid encrusted pebbles, overlain by silty sand and sandy silt rich in extant brackish-marine bivalves and gastropods. The overlying poorly fossiliferous

Geological history

metres msl 4

relative sea-level curve derived mainly from shell/coral associated with intertidal beach rock (after Ramsay, 1995)

3 2

St Lucia beach ridges, back barrier wetland elevations

1

calender years A.D. 2000

1000

1000

2000

–1

3000

4000

5000

6000

7000

14

C dated marine intertidal indicators

–2 –3 –4

Ramsay (1995) presumed sea-level below present due to lack of data

8000

calender years B.C. 9000

FIGURE 3.4 Sea level curve showing the late Pleistocene and Holocene marine transgression (Flandrian) based on radiocarbondated beach rock and shell (after Ramsay, 1995). The OSL-dated beach ridges suggest that marine-influenced water level in St Lucia declined from ~6 ka to reach the present shoreline morphology about 900 years ago.

–5 –6 –7 –8

Legend Tidal range envelope for marine intertidal fixed biological indicators (14C dates) Tidal range envelope for St Lucia beach ridge, back barrier wetland elevation (OSL dates) 2m tidal range indicator

–9 –10 metres msl

clay-rich swamp deposit contains only freshwater snail fossils. Cooper (1999) regards the deposit as representing an earlier extent of Lake St Lucia when sea levels (and hence lake levels) were +5 m above present and aerial extent of ‘Lake Mduku’ was 10% greater. This deposit lies close to post mid-Holocene raised beach ridge deposits and radiocarbon dating of the shells would reveal whether the deposit is Holocene or older. Hill (1975a) indicated that the Flandrian marine transgression blocked the former outlet at Leven Point where fluvial discharge through the 40 m deep channel and estuary mouth was apparently unable to compete with spit accretion by littoral drift (Sydow, 1988) (Figure 3.3). Dating of the youngest phase of dune accretion forming the coastal barrier north of Cape Vidal has shown that this narrow section of the barrier dune was still mobile about 2000 years ago (Porat and Botha, 2008). This youngest of four generations of Sibaya Formation dunes has accreted

against the core of the Last Interglacial island which comprises Kosi Bay and Isipingo Formation dunes covered by three older phases of Holocene Sibaya Formation dunes.

3.3.3 Lake depth change The relevance of the sequence of Holocene beach ridges becomes most significant when viewed in the context of sediment accretion and the decreasing water depth within the basin during this period. A series of seven boreholes drilled in the lake bed records the changes in substrate during the period of accumulation (Kriel et al., 1966; van Heerden, 1975; Hobday, 1976). Shallow boreholes traverse Brodie’s Crossing and a borehole grid on the north shore of Catalina Bay also provides dated sediments that can be used to gauge sedimentation rates during the Holocene. In the north-eastern shallows the basal aeolian sand is buried by shelly estuarine sand

55

56

Greg A. Botha et al.

FIGURE 3.6 Beach ridges dated from 4 to 0.8 ka that isolated Meme Swamp as sea level fell after the mid-Holocene sea level highstand around 6 ka, during which the north-east shallows were isolated from the sea as the coastal barrier dune in the area north of Cape Vidal developed. (Photo G. A. Botha.)

FIGURE 3.5 Proto St Lucia estuarine lake margin during the mid-Holocene sea level highstand (after Orme, 1990) showing the marine link in the vicinity of Leven Point and the inundated Mkhuze and Mfolozi floodplains. The beach ridge complexes track the shrinking lake during the late Holocene until the present shoreline morphology was attained about 900 years ago.

containing lagoonal macrofauna that indicate a tidal marine influence. In all sites the upper deposits become silty and muddy marking the evolution of the system from tidal influence towards fluvial sediment load dominance as the estuarine link became restricted in the late Holocene. Erosion of the early Holocene dune systems forming the eastern shorelines, with possible addition of flood tidal marine sand deposits, actively promoted shallowing and compartmentalization of the lake. Truncated

parabolic dunes and extension of Sengwane Point as a spit indicates the active shoreline erosion and lake segmentation process that accompanied the midHolocene marine incursion. St Lucia achieved a less advanced level of segmentation than the Kosi Lakes (Orme, 1973) possibly due to the bedrock-controlled western shoreline. The water depth that existed in the lake and other coastal wetlands in the area can be inferred from the cored and dated St Lucia lake bed deposits (van Heerden, 1975; Hobday, 1976) and the infill beneath the bed of Lake Eteza in the Mfolozi River wetlands near Mtubatuba (Scott and Steenkamp, 1996; Neumann et al., 2010). The shortterm sedimentation rate trends derived from shallow sediment cores from the St Lucia lake bed (Kriel, 1965; van Heerden, 1975; Vogel and van Urk, 1975) do highlight the dramatic increase in lacustrine sediment deposition during the past two thousand years. These trends can only be regarded as semiquantitative due to measurement errors associated with the radiocarbon dating technique, reworking of shell deposits and the small number of sites where these data are available. The record of human settlement in the region during this period can be gauged from the cave sedimentary records from Border Cave near

Geological history

Ingwavuma (Bird, 2003) and the Sibudu Shelter near Tongaat (Jacobs et al., 2008) where human occupancy in the coastal zone was influenced by sea level fluctuations and the regional environmental impact of climate-linked degradational cycles in the coastal hinterland. The short-term changes in the system that have occurred since the impact of anthropogenic change can be reconstructed accurately from the records of fluvio-lacustrine sedimentation that extend back thousands of years.

Palynological records derived from wetland sediment cores from Lake Eteza on the lower Mfolozi River, Lake Sibayi and the Mfabeni wetland (Grundling et al., 1998; Finch and Hill, 2008) preserve evidence of vegetation response to Holocene environmental change (Scott and Steenkamp, 1996; Neumann et al., 2008, 2010). The anthropogenic influences on the catchment can be gauged from the uppermost wetland deposits and the earliest topographic maps and aerial photography after 1937.

3.4 Evaluation of St Lucia sedimentation Martin and Flemming (1988) compared modern sediment yields from KwaZulu-Natal rivers to the palaeoyield derived from offshore sedimentation in the Natal valley of the Indian Ocean. Seismic profiles showed the modern sediment yield of 322 t km2 yr 1 to be 12–22 times higher than long-term rates averaged over 5 Ma and 130 Ma (14–27 t km2 yr 1). Flemming and Hay (1988) described modern sediment yield as exceeding long-term averages by a factor of between 12 and 30 but cautioned the extrapolation of modern sediment yield data backwards in time. Estuaries along the subtropical KwaZulu-Natal coast are generally linked to large flood-prone catchments that regularly scour and erode estuarine deposits. The St Lucia basin is fed by five rivers that drain catchments with differing terrain morphology, geology and human disturbance patterns. The open water body is subject to wind-generated waves and currents that redistribute the predominantly finegrained sediment deposited by the rivers. Fortuin (1992) found that the mud fraction is the most abundant sediment in the lagoon and dominates where water depths exceed 1 m. He suggested that this depth was the wind-wave base and therefore the mud was not reworked deeper down. Contrary to this Hobday (1976) assumed that the entire lake is now above the wave base and that wave-generated currents and direct wave action result in a

continuous erosion, transportation and redeposition of mud along the bottom of the lake. The restricted marine link results in much of the fluvial sediment entering the system being retained within the lake. The sedimentation history of the Lake St Lucia basin is similar in many respects to other estuaries along the coast where bedrock valleys incised during multiple marine regression/ transgression cycles linked to glacio-eustatic sea level fluctuations during the Quaternary period were infilled during the subsequent Holocene marine transgression (Orme, 1975; Maud, 2000). Of specific relevance to this investigation are dated lacustrine deposits derived from Lake Eteza, a dammed tributary of the lower Mfolozi River (Scott and Steenkamp, 1996; Neumann et al., 2010), Lake Futululu on the lower Mfolozi floodplain (Grenfell et al., 2010), the lower Mkhuze Swamps and dammed tributary valley (Ellery et al., 2003a; Turner and Plater, 2004), and Lake Sibaya (Neumann et al., 2008). Cross-sections depicting the valley infill associated with many rivers along the KwaZuluNatal coast, including the St Lucia lake and its tributaries, have been published by Orme (1974). The marine, lacustrine and fluvial channel infill history of St Lucia has many elements in common with the Kosi lake system (Wright et al., 2000; Wright, 2002) and Lake Sibaya (Miller, 2001). The organic-enriched lacustrine sediment and molluscan shells that were

57

58

Greg A. Botha et al.

dated are prone to erosional reworking and redeposition within channels and may not provide accurate sea level index points (see Kriel, 1965) within the alluvial valley fill or lacustrine deposits (Wright, 1995). A holistic understanding of the prehistoric influences on this ecosystem is necessary to predict the likely effects of future changes on the St Lucia catchment system and the impact on lake sedimentation. The closest high-resolution sedimentation and palaeoenvironmental record is from the alluvium-dammed tributary of the Mfolozi floodplain, Lake Eteza (Scott and Steenkamp, 1996; Neumann et al., 2010). The dearth of measured bedload data for South African rivers reduces the accuracy of any sedimentation rate predictions that could be applied to St Lucia. The general rule is to add 25% to allow for bedload and non-uniform sediment transport to fine suspended sediment samples from a river in flood. The influence of large flood events skews any long-term averages. Beck et al. (2004) cite sediment transport during a large flood observed on the Pongola and Thukela rivers as 8 to 13 times the mean annual sediment yield but add that the combined transported sediment from smaller, more frequent floods represents a large proportion of the total sediment load and is as important as the large resetting floods. Dollar and Rowntree (2003) cite evidence of slack water sediments in the Mfolozi River that indicate a flood of 28 000 m3 s 1 in the late Holocene, nearly twice the size of the largest recorded flood of 16 000 m3 s 1 in 1984.

3.4.1 St Lucia sedimentation The most comprehensive spatial assessment of lakebed sediments was conducted by Fortuin (1992) who described the sediments and their distribution, assessed relationships between sediment type and lake depth and used grain-size distribution curves to interpret the sedimentary processes active in the lake.

In general the eastern part is characterized by sandy substrate and the water is less turbid than the western and deeper parts of the lakes. Fortuin (1992) showed that mud substrate is more abundant in the deeper parts of the lake where it is not reworked by oscillatory wind-wave currents. Silt-dominated grain-size fractions occur in pockets but are not aligned with any physiographic features. Fine-grained sand is predominantly along the Eastern Shores area from Bird Island to Selley’s Lakes and the coarser sand substrate predominates in the shallow eastern shoreline which is less turbid than the western part. Most accounts of catchment erosion and sediment dynamics in the system refer to the problem of lake ‘siltation’. Interestingly, Fortuin’s (1992) study revealed an absence of substantial material in the silt-sized range in the lake sediments. In the north lakes, uniform, pelagic sedimentation occurs whereas the shallow shorelines around the southern lakes are characterized by ‘traction’ sediment transport with patches of suspension sedimentation around the islands. The present lake-bed substrate conceals a succession of sedimentary deposits that accumulated during the post-glacial marine transgression during the terminal Pleistocene to mid Holocene (Flandrian). Hill (1975a) contoured the depth of soft sediments in the lake which show an infilled trough reaching 15m depth extending between Lane Island and the Sengwana spit. A 40 m deep infilled channel extending from Hell’s Gate into the northern lake (Figure 3.3) could be the incised river channel that drained the basin during the marine regression leading to the LGM lowstand (Sydow and van Heerden, 1988). The description of the St Lucia lake environment by Kriel (1965), published in the comprehensive report by the Commission of Enquiry (Kriel et al., 1966) assessed the soft sediment deposits beneath the lake bed as representing 9–10 times the volume of the present lake storage capacity (lake and Narrows channel to the Mpate River).

Geological history

3.4.2 Sedimentation rate The lake bed is underlain by an estimated 2376 × 106 m3 of soft mud with a further 739 × 106 m3 of sediment beneath the swamps. Assuming that sedimentation started at 5000 yr BP these deposits represent an accumulation rate of 263 × 106 m3 yr 1 (Kriel, 1965; Fortuin, 1992). An assessment by Kriel et al. (1966) estimated that it took 1500 years to deposit the volume of sediments in the lake and marshes. This estimate excludes marine sediment influx and did not consider the erosion of lake banks along the Eastern Shores or current transport of bedload around the lake margins. The Kriel (1965) assessment predicted that the present annual sediment load of rivers entering the lake is 2.25 to 3.3 times higher than the average annual rate of deposition in the lake over the past 5000 years. At this rate of sediment accumulation the lake was predicted to fill to sea level within 115 years and reach the high water mark in about 250 years. The impact of catchment soil erosion led Blok (1976) to speculate that the lake could be filled up by sediments to sea level within the next 100 to 200 years. Begg (1978) noted that the present sedimentation rate to be 2–3 times greater than the mean rate over the past 5000 years. Annual silt accumulation has been estimated at 0.98 × 106 m3 yr 1 to 2.0 × 106 m3 yr 1. The accumulation of sediment and shallowing of the lake would affect lake salinities due to the unfavourable depth:area ratio. McCarthy and Hancox (2000) described the aggradation of Mkhuze River sediment in the Greater Mkhuze Wetland System (GMWS). East of the Lebombo Mountains the sandy valley infill deposits have formed a delta that progrades towards the east during large flood events. Most coarse sediment is deposited in upper reaches of the floodplain reach from the gorge to Mtundane Pan, the remainder of fine sediment being deposited before Lower Mkhuze Bridge (Goodman, 1993). The Mkhuze Swamps attenuated the Domoina flood discharge by 60–100%.

The suspended sediment load that enters the floodplain was estimated at 50 kg s 1 during an average discharge event (Linton et al., 1998 cited by McCarthy and Hancox, 2000). The sedimentary fill of the GMWS by the prograding Mkhuze delta has advanced about 25 km during the past 6500 years, an average of 3 m yr 1. Sedimentation rates along the lower Mkhuze River floodplain were calculated using 210Pb and 137Cs activities by Humphries (2008). Sedimentation rates of 0.24 cm yr 1 and 0.26 cm yr 1 were derived from cores and the mean sedimentation rates on the floodplain were found to be 0.25 to 0.35 cm yr 1 over the last 100–150 years. Over a longer period, sedimentation at the Lower Mkhuze Bridge over 6500 years occurred at 0.65 cm. yr 1 (McCarthy and Hancox, 2000). Long-term average sedimentation rates in the Mkhuze floodplain have decreased in response to valley infilling and decline in the valley cross-sectional area. Sediment deposition was found to be spatially variable with higher rates closer to the channel and lower accretion rates in distal floodplain margin environments leading to the development of a channel mound confined by raised levees. A study of sediment accumulation in the Mdlanzi Swamp, an interdune wetland that feeds the Mkhuze Swamps, was conducted by Turner and Plater (2004). Infilling of the interdune valley by fine-grained minerogenic and organic sediment after 1450 ± 40 yr BP indicates that early Iron Age landscape disturbance accelerated sedimentation. The Mdlanzi organic sediments show a very low rate of peat accumulation at c. 2–3 mm yr 1 whereas Thamm et al. (1996) report peat accumulation rates of up to19 mm yr 1. The palynological interpretation of a sediment core from Lake Eteza in the lower Mfolozi catchment led Scott and Steenkamp (1996) to the conclusion that the sedimentation rate has accelerated by 18 times since pine trees were introduced in the late 1920s. A more detailed resampling of the core by Neumann et al. (2010) confirms dramatic variations in sediment deposition rates. The uppermost deposit, attributed to sedimentation since Cyclone Domoina in 1994

59

60

Greg A. Botha et al.

showed fine sediment accumulation at 4.13 cm yr 1 between 1984 and 1992. Sediment accumulated at an average rate of 1.06 cm yr 1 during the periods AD 1984 to 394 cal. yr BP, 4066 to 4146 cal. yr BP and 6506 to 6602 cal. yr BP. Early sedimentation occurred at an average of 0.32 cm yr 1 (7768–9196 cal. yr BP). During the remaining periods sedimentation occurred at 0.13 cm yr 1. Grenfell et al. (2010) showed that Lake Futululu was formed by Mfolozi River floodplain aggradation where the sediment supply exceeded the rate of sedimentation in tributary valleys. The Mfolozi floodplain morphology was described by Grenfell et al. (2009) as an alluvial fan at the head with a transition to a flat middle reach, possibly controlled by faulting of the underlying rock, and the lower floodplain that is controlled by sea level. The tributary lake had been impounded by 1813 BC (cal. yr BP) and the peat within the lake accumulated at 0.13 cm yr 1. Sediment transport in the Mfolozi channel has resulted in the accretion of an alluvial ridge that rises some 2.5 m above the surrounding floodplain. The average suspended sediment yield is 61 t km 2 yr 1 with suspended sediment transport calculated at 6.8 × 105 t yr 1 (Grenfell and Ellery, 2009). An estimate of catchment sediment yields by Taylor et al. (2006a) was based on the 155 t km 2 yr 1 average sediment yield calculated by Midgley et al. (1994). The 100 year sediment mass deposited in the lakes would amount to 1.4 × 108 tons or a thickness of 0.2 m, taking into account the larger surface area due to sea level rise. This could be offset by the increased sediment-trapping potential of the Mkhuze Swamps as sea level rises. The wide range of sediment accumulation volumes and rates within the Lake St Lucia environment reflect the lack of essential, long-term monitoring data on the river catchments. Until detailed river channel mapping and sediment inventory studies are combined with hydrological modelling, it will be impossible to make meaningful improvements to the predictions of sedimentation impacts on the St Lucia system.

3.4.3 Estuarine processes and future sediment dynamics A compilation of historical records and maps of the St Lucia estuarine system and Mfolozi River mouth from 1883 (Taylor, 2006, 2011a) shows how the linkages between the lower Mfolozi channel and St Lucia Mouth have changed due to flood scouring, sediment movement in response to the tidal prism, sedimentation during droughts and accretion of the beach spit north of Maphelane. To reduce the impact on sugar cane plantations upstream, canalization of the lower Mfolozi channel, the construction of a flood diversion weir and periodic breaching of the closed river mouth have been undertaken (Taylor, 2011a). Excavation of a new Mfolozi River mouth and separate St Lucia Mouth after 1956 was engineered to reduce the perceived accelerated sedimentation of the lower St Lucia estuarine system. Dredging of the mouth and channel was undertaken in order to maintain an open mouth for better water circulation. The extreme flooding associated with the Cyclone Domoina rainfall event caused considerable scouring, and sediment deposition in the Mfolozi and St Lucia system was compounded by the upstream river diversion structures and results of mouth dredging operations. Van Heerden and Swart (1986) described the long-term control on estuary mouth migration as being the constraining Pleistocene aeolianite outcrops at Maphelane and First Rocks. They noted that during and immediately after floods, large volumes of sediment accumulate in the nearshore and the mouth migrates further north due to the influence of oblique southerly swells combined with the large littoral zone sediment budget. A study of the sediment dynamics of the St Lucia and Mfolozi estuary mouths by Wright (1995) during 1988/1989, following the 1984 and 1987 flood events, showed how sediment shoaling in the estuary mouth had continued despite dredging and deepening. Wright (1995) regarded the best method of interpreting the geological record from sediment cores as being the recognition of sediment

Geological history

characteristics associated with present-day sedimentary environments and facies. This approach was used by van Heerden (1975) to differentiate sediment sources in the lake-bed cores (Hobday, 1976). The mouth comprises distinct beach barrier, abandoned channel and estuarine/lagoonal environments. Marine sand advances up the estuary driven by barrier washover sedimentation and tidal effects to form a series of flood-tidal deltas. Gravel lag deposits are concentrated near the estuary mouth and on Honeymoon Bend. Sand-dominated sediment is also concentrated near the mouth and bend but extends upstream within the channel at some times of year. Generally the coarsest sand is near the

mouth, fining upstream. The provenance of the sand fraction can be differentiated into catchment derived, aeolian (fine grained, well sorted) and marine (medium grained, moderately sorted). Marinederived mud is negligible. During high-rainfall years the mouth discharge is able to maintain an open outlet to the sea, but as lake level drops and flow declines shoaling takes place. The understanding of long-term sediment dynamics in response to changing sea level, catchment provenance and Mfolozi discharge influences must draw from the evaluation of ancient sediment records as well as present sedimentation trends operating over a range of timescales.

3.5 Conclusions The present morphology of the St Lucia estuarine lake was achieved only within the past 1000 years. Marine inundation of the lake basin occurred for short periods during the Last Interglacial and the mid Holocene around 125 ka and 6 ka respectively. The St Lucia/False Bay and lower Mfolozi and Mkhuze floodplains were inundated during the Last Interglacial when coral reefs occurred in False Bay. During this period the proto-coastal barrier south of Cape Vidal formed an island similar to Bazaruto. During the subsequent marine regression sea level was lowered and by ~20 000 cal. yr BP sea level was drawn down to 130 m, exposing the continental shelf to dune accretion. It is likely that the coastal plain surrounding the St Lucia estuarine system has been a dry valley for longer periods over the past 2.6 million years during the Quaternary Period than it has been inundated as an estuarine-influenced lake basin. The present lake morphology is due to marine sediment influx during the period of increased

marine connection and tidal prism associated with the early Holocene. A series of raised beach ridges trace the falling level in the estuarine lake as the marine connection was closed by coastal barrier dune accretion. The sedimentation balance shifted towards fluvial input and lacustrine sedimentation. There are few accurate records of sediment accumulation rates against which future sedimentation trends in the estuarine lake can be assessed. The lower tributary rivers all show dramatically increased sedimentation rates during the late Holocene. Reduced fluvial discharge could reduce the sedimentation trend in the main lake although catastrophic flood events contribute significant loads that confound any long-term average estimates. During the Anthropocene, the legacy of human activities in the catchments over a hundred years and global change influences have altered the sedimentation patterns to which the lake environment had equilibrated over the preceding thousands of years.

61

Chapter contents 4.1 Introduction 4.2 The Agulhas Current 4.3 Inshore waters and currents 4.4 The Central Shelf 4.5 Marine influences on the St Lucia Mouth condition 4.6 Coastal productivity, nutrients, zooplankton and fish 4.7 Adult fish migrations offshore in relation to St Lucia Estuary 4.8 Offshore canyons and their potential impact on St Lucia Estuary 4.9 Coastal biogeography

Sea sunrise off the St Lucia Mouth. (Photo: Ricky H. Taylor, February 2012.)

4

The marine environment Allan D. Connell and Sean N. Porter

4.1 Introduction The marine environment adjacent to the St Lucia Estuary, on the east coast of South Africa, is characterized by three important features. Firstly, the offshore environment is dominated by one of the biggest western boundary currents in the world, which, adjacent to Lake St Lucia, is stabilized by a narrow shelf, less than 10 km wide (Figure 4.1). While this does not directly impact St Lucia Estuary, it strongly influences recruitment of marine organisms utilizing the estuary and lake system. Secondly, the coastline has a north–south orientation, and lies in a biogeographic mixing region, between tropical

western Indian Ocean biota, and those of more temperate regions to the south. The narrow shelf and offshore Agulhas Current play a major role in positioning this biogeographic split adjacent to St Lucia Estuary. Thirdly, waters of the KwaZulu-Natal (KZN) coastal region are essentially oligotrophic, supporting a diverse but impoverished marine community, and low-yielding exploitable resources. The role of terrestrial runoff, in coastal productivity, is elevated in such circumstances, adding another important dimension to the management of coastal ecosystems in this type of environment.

4.2 The Agulhas Current The dominant oceanographic feature along the entire KZN coastline is the Agulhas Current, one of the major western boundary currents of the world. Reviews of the coastal oceanography of the area can be found in Schumann (1988) and Lutjeharms (2006). Surface water temperature in the current core rises to about 28  C in summer, and drops to about 23  C in winter, while surface water inshore of the current is generally about 2  C cooler. The Agulhas Current mainly comprises water from the South Western Indian Ocean Subgyre (SWIOS), with smaller contributions from the South Equatorial Current, which drives some water into the northern end of the Mozambique Channel around the northern

tip of Madagascar, and onto the east coast of Madagascar, creating a small western boundary current running south along the coast (Lutjeharms, 2006). In the past this East Madagascar Current was assumed to round the southern tip of Madagascar, and merge with the Mozambique Current, to form the Agulhas Current in the region of 27 S. Satellite imagery, however, appears to indicate that, at times, it retroflects south of Madagascar, when no longer confined by the Madagascan shelf, and is thus not a regular stream eastwards, contributing to the Agulhas Current (Lutjeharms, 1988). Another interpretation suggests that these are simply anticyclonic eddies pulling cooler chlorophyll-rich

Ecology and Conservation of Estuarine Ecosystems: Lake St Lucia as a Global Model, eds. Renzo Perissinotto, Derek D. Stretch and Ricky H. Taylor. Published by Cambridge University Press. © Cambridge University Press 2013.

64

Allan D. Connell and Sean N. Porter

FIGURE 4.1 The KwaZulu-Natal coastline, showing the shelf break at 200 m.

inshore water offshore and eastward (Quartly et al., 2006). The retroflective behaviour of the southern limb of the East Madagascar Current therefore remains an enigma (Lutjeharms, 2006). There is also considerable evidence of eddies moving west (Quartly and Srokosz, 2004), below the southern tip of Madagascar, to join the Agulhas Current in the region of 27 S, about 180 km north of the St Lucia estuarine system, though some are deflected south by the Mozambique Plateau (Gru¨ndlingh, 1984). These eddies, along with the Delagoa eddy and its possible system of smaller eddies off Maputo, disturbing flows out of the southern end of the Mozambique Channel, may explain the 100 km bedload parting zone north of Cape Vidal, identified by Flemming and Hay (1988), where bedforms periodically switch direction. Off Durban the total Agulhas Current flow is estimated at 73 × 106 m3 s 1, of which the Mozambique Channel flow contributes about 18 × 106 m3 s 1 (Donohue and Toole, 2003), and the eddies around southern Madagascar, about 8 × 106 m3 s 1 (Quartly and Srokosz, 2004). The balance comes from the SWIOS, which not only contributes in the region of 27 S, but all along the entire KZN and Eastern Cape coasts, at a rate estimated at 6 × 106 m3 s 1 100 km 1 of coastline (Lutjeharms, 2006). The flow south via the Mozambique Channel, particularly in the northern section of the channel, is not a classic western boundary stream, but comprises cyclonic and anticyclonic eddies (de Ruijter et al., 2002; Ridderinkhof and de Ruijter, 2003), possibly caused by the turbulence created by the northern tip of Madagascar, to water being driven towards Africa by both the South Equatorial Current and Rossby waves, moving west across the subtropical Indian Ocean (Lutjeharms, 2006). No evidence of a seasonal variation in the volume flow has been found (Pearce and Gru¨ndlingh, 1982). Off Durban, the core of the Agulhas Current is located 40–60 km offshore, with an intense cyclonic shear on the inshore side, and a mean width of about 100 km (Schumann, 1988), but the core can change its position by 30 km and more in a day (Pearce, 1977a). Due to narrowing of the shelf off Port

The marine environment

Edward, Schumann (1988) reported that the Agulhas Current is usually so close inshore that the inside edge, indicated by a sudden temperature rise of about 2  C, is seldom discernable. The core was found to be

about 10 km offshore, with increased current speed, compared with off Durban. A similar situation probably occurs off the St Lucia Estuary, where the shelf is equally narrow.

4.3 Inshore waters and currents Being a western boundary current, travelling south in the southern hemisphere, Ekman veering causes Agulhas Current bottom water to slide onto the shelf all year round (Pearce, 1978; Schumann, 1988). This water is reported by Pearce (1978) to come from 40 to 60 m in the Agulhas Current, and is, on average, about 1.4  C cooler than Agulhas Current surface water. In summer the surface of this cooler inshore water is rapidly warmed by the sun, resulting in a marked difference ( 4  C, according to Pearce, 1977b) between surface and bottom temperature in shelf water of 30–50 m depth. Data from a thermister string deployed in 31 m off Amanzimtoti illustrate this (Figure 4.2). In winter, the weaker sun, radiation loss, and the nightly cooling effect of the winter land breeze, prevent this temperature difference from developing. Thus, in winter, inshore shelf water temperatures are generally within about 1  C from surface to bottom, at approximately 21  C (Figure 4.2). Satellite images confirm the presence of cooler water inshore of the Agulhas Current, both in

FIGURE 4.2 Temperature difference between surface and bottom, in 30 m water depth, off Amanzimtoti, from June to December 2005.

summer and winter, as shown in the two examples in Figure 4.3. Inshore currents are predominantly wind driven (Schumann, 1988), outside of the influence of topographically driven features such as the Durban– Richards Bay area, known as the Natal Bight, and the Port St Johns Bight. Although caused by relatively small kinks in the coastline, these are the sites of semi-permanent eddies of expanded cooler inshore water (Figures 4.3 and 4.4), with predominantly north-going currents inshore. The Durban Shelf area is described by Schumann (1988) as ‘a transition region between the wind-dominated shelf to the north, and the Agulhas Current dominated shelf to the south’, with evidence that the area between Durban and Park Rynie (his Mzinto) often shares the same water in a recirculating eddy. To the north of the St Lucia estuarine system, there is a much larger cyclonic eddy caused by the substantial offset of the coastline, to the west, from Inhambane down to Delagoa Bay (Schumann, 1998; Lutjeharms, 2006), which Schumann describes as having possible systems of smaller eddies. A further disrupting feature to inshore currents is the passage of a ‘Natal pulse’, essentially a moving gyre of water trapped inside the Agulhas Current. A ‘pulse’ appears to be generated when a topographically generated gyre becomes unstable and breaks free to move down the coast (or a large inshore deflection of the Agulhas Current becoming isolated), with a rapid clockwise rotation (when viewed from above), generating unusually powerful north-going currents inshore (Lutjeharms and Connell, 2000). The big eddy south of Port Edward in Figure 4.5 is probably an example of this feature.

65

66

Allan D. Connell and Sean N. Porter

FIGURE 4.3 Sea surface temperature images of summer (left) and winter (right), illustrating the presence of a band of cooler inshore water throughout the year along the KwaZulu-Natal coast. Side bar is sea surface temperature in  C.

FIGURE 4.5 A well-developed ‘Natal pulse’, partly obscured by cloud, south of Port Edward on 2 February 2006. Side bar is sea surface temperature in  C. FIGURE 4.4 Features inshore of the Agulhas Current on the KwaZulu-Natal coast: 31 May 2005, with the Durban (A) and Port St Johns (B) bights highlighted. Both show evidence of clockwise rotation. Side bar is sea surface temperature in  C.

A ‘pulse’ moves down the coast at a speed of about 22 km per day (Lutjeharms and Roberts, 1988), and, due to its size and depth, is often associated with

unusual, deep-water fauna being pulled up onto the shelf, such as squid egg balloons (Roberts et al., 2011), and deep-water copepods. Most studies suggest that these ‘pulses’ are relatively infrequent, with only three to four passing through each year (Lutjeharms, 2006).

The marine environment

4.4 The Central Shelf Figure 4.4 shows a band of cooler inshore water occupying the entire bight from Durban to just south of Richards Bay, often referred to as the Natal Bight, caused by the Agulhas Current being confined further offshore by the straight shelf break (Figure 4.1), with the coastline cutting away to the west, immediately south of Richards Bay. Flemming and Hay (1988) showed how bottom currents inferred from sediment dispersal and bedform patterns revealed a complex flow under this band of cool water. These confirmed a closed eddy system off the Thukela Mouth, and an accompanying depocentre for fine silts, as well as a

semi-permanent (also cyclonic) eddy just north of Durban. A similar eddy can be seen, centred just south of Port St Johns, extending north to where the 30 E meridian intersects the coast, about 30 km south of Port Edward (Figure 4.4). The unusual width of this feature on the day this image was taken (31 May 2005) suggests a Natal pulse may also be present off Port St Johns (see also Figure 4.3). For fishes moving up the coast on annual migrations, these two semi-permanent eddies form important ‘stepping stones’, since their inshore edges have north-going currents (Connell, 2006; Roberts et al., 2010).

4.5 Marine influences on the St Lucia Mouth condition Tide range is at most about 2.1 m at equinox spring tide, down to about 0.5 m at neap tides. The orientation of the coast at the St Lucia Estuary is roughly 28 (E of true N). The area is dominated by north-east and south-west winds in almost equal proportions, with north-east winds dominant in summer (Hunter, 1988). These two winds account for a wave climate dominated by two distinct populations, one with waves originating from around 30 , the other with waves from approximately 180 (van Heerden and Swart, 1986). Based on a perpendicular to the coast of 118 , the south-ofnormal component of waves is dominant at 47%, the eastern component comprising 26%, and the balance directly onshore. The southerly component also comprises higher wave heights and longer wave periods. Corbella and Stretch (2012) analyzed 18 years of Waverider buoy data from Richards Bay and Durban, and found that the two sets of data were complementary in wave heights, with some difference in wave direction due to local wind conditions. They found a mean significant wave height of 1.65 m, mean period of 10.0 seconds, and a mean wave direction of 121 . Satellite-based

measurements confirm that the average wave height between Richards Bay and St Lucia is essentially the same (Porter, 2009). Consequent to the average wave direction from the south-east, a net north-bound sediment transport occurs in the surf zone, at an average of between 0.8–1.0 × 106 m3 yr 1 (van Heerden and Swart, 1986). Under periods of low flow, or reduced tidal prism in the estuary, this will lead to northward migration of the mouth, and eventually contribute to its closure. The area is also influenced by tropical cyclones, which occasionally move across northern KwaZuluNatal, generating winds, and subsequently swells, offshore that can erode coastlines in unexpected ways, due to their being both of exceptional height and with an incidence to the coast that is generally north of perpendicular. While the western Indian Ocean usually experiences six cyclones per year, on average, less than one per year will move as far south as the northern KwaZulu-Natal coast (Hunter, 1988). The year 1984 was exceptional, when two, named Domoina and Imboa, moved across the area, dropping 300–400 mm of rain over wide areas, and

67

68

Allan D. Connell and Sean N. Porter

generating 10 m waves at Richards Bay (Hunter, 1988). Due to the north-of-perpendicular angle to the coast of these waves, it was estimated that net sand movement that year was to the south. The associated rainfall was sufficient to scour a deep mouth at St Lucia (van Heerden and Swart, 1986). In March 2007, after a prolonged period of mouth closure, an

unusual combination of a cold front developing into a cut-off low which remained stable for three days, some 700 km south-east of Durban, generated massive seas and a powerful storm surge (Hunter, 2007). These, combined with an equinox spring tide, smashed open the St Lucia Mouth, without the aid of exceptional rainfall.

4.6 Coastal productivity, nutrients, zooplankton and fish The Agulhas Current, being the dominant feature of the coastal oceanography of this region, plays a major role in the year-long nutrient loading and productivity of the KwaZulu-Natal coastal marine environment. Studies have shown that there is a negative gradient in nutrient concentration with distance offshore as far as the Agulhas Current core (Oliff, 1973; Pearce, 1977a; Carter and d’Aubrey, 1988), as well as a positive gradient with depth in the Agulhas Current itself. The continuous Ekman veering of Agulhas Current bottom water up onto the shelf results in the inshore waters being cooler, and slightly elevated in the inorganic nutrients N, P and Si, compared with the Agulhas Current core surface

water. In KZN inshore waters, there is thus a strong correlation between the cooler inshore water and chlorophyll-a production (Figure 4.6), and the chlorophyll-a levels are similar in summer and winter (Figure 4.7). Nutrient concentrations are generally low; at 10 m depth across the shelf at Richards Bay, Oliff (1973) measured nitrates of 2.0–2.7 μM l 1, phosphates of 0.98–1.1μM l 1, and silicates of 3.6–4.4 μM l 1. These compare well to concentrations of these nutrients measured by Porter (2009) in inshore waters of his Subtropical Natal Bioregion (region to the south of the St Lucia estuarine system). Both sets of concentrations are, however, significantly higher

FIGURE 4.6 Satellite imagery of 31 May 2005, demonstrating the strong relationship between the cool inshore water (left) and chlorophyll-a concentration (right), along the KwaZulu-Natal coast. Side bars are sea surface temperature in  C (left) and chlorophyll-a in mg m 3 (right).

The marine environment

Table 4.1. Comparison among nutrient concentrations (μm l 1) between the Delagoa and subtropical Natal regions Nutrient species Region NO2

NO3

PO4

SiO2

Delagoa

0.05

0.20

0.31

2.28

Subtropical Natal

0.14

4.26

1.11

5.45

–

2.70a

1.10a

4.38a

Data from Porter (2009). a Data from Carter and d’Aubrey (1988).

FIGURE 4.7 Comparison of chlorophyll-a concentration in summer (left) and winter (right), along the KwaZulu-Natal coast. Side bar is chlorophyll-a in mg m 3.

than Porter (2009) measured in inshore waters of his Delagoa Bioregion Region, between St Lucia Estuary and Kosi Bay, due to the lack of river discharges and very narrow shelf in this region (Table 4.1). Zooplankton have not been extensively studied on the KZN coast, and nothing has been added since the review of Carter and Schleyer (1988), apart from papers on fish larvae, which are not relevant to the present discussion. By far the dominant animals in most near-surface zooplankton communities are the copepods, and they are also a key component in the

food web from phytoplankton to fishes. Transects across the Agulhas Current have shown a higher biomass inshore of the current (Carter, 1977), and the bulk of the biomass (70%) in the upper 100 m (Carter and Schleyer, 1988). Carter (1977) found that the dominant copepod species in the core of the current was Paracalanus parvus, but on the shelf, inside the current, Centropages chierchiae and Calanoides carinatus were dominant (Carter, 1977). He recorded a shelf biomass minimum in the Durban area, in the late summer (March), and a winter biomass bloom dominated by the large calanoid C. carinatus. This is

69

70

Allan D. Connell and Sean N. Porter

a key species in the diet of sardine; thus a winter bloom would be an important food source for this species during its annual migration to KZN waters. During a long-term study of fish spawning patterns within the Aliwal Shoal Marine Protected Area off Scottburgh on the south coast of KZN, inferred from fish eggs collected in a surface-towed plankton net, Connell (2006) found Calanus agulhensis to be a dominant species, also important in the diet of sardine. Productive food chains, which in coastal seas lead to productive fisheries, are obviously dependent on sustained high primary production (phytoplankton), driven by nutrients, which are usually brought to the surface in cold upwelled water. This supports a rich zooplankton population, feeding on the phytoplankton bloom, which in turn supports fastgrowing, pelagic, shoaling fishes such as sardines and anchovies. This is clearly not the case in KZN, due to the oligotrophic nature of Agulhas Current waters, despite the Eckman-veer driven upwelling. The appropriate species of grazing fishes are present: Sardinops sagax, Etrumeus teres and Scomber japonicus eggs were the most abundant of all species, in a 15-year study of pelagic fish eggs off Park Rynie on the south coast of KwaZulu-Natal (Connell, 2001); they are just not present in high

biomass. The point is best illustrated by comparing the coastal fisheries around the South African coast. The KZN shelf does not support any year-round pelagic fishery, and only yields between 1500 and 2000 tonnes of linefish and about 500 tonnes of shoaling sardine-like pelagics, these latter associated with the annual winter ‘sardine-run’ up the east coast from Cape waters (van der Elst, 1988). By contrast, the southern Cape coast fishery yields in the region of 100 000 tonnes annually, and the Western Cape fishery in the region of 600 000 tonnes annually. Clearly the east coast does not support a significant fishery. The real value of the KZN fishery, in the local context, is the quality of the fish harvested (van der Elst, 1988), which finds its way into local markets via a small but active commercial fishery and an ever growing recreational fishery. Many of the fish species are of tropical origin; the warm, powerful Agulhas Current allows tropical species to extend their range further south than would otherwise be the case. But the area also exhibits a high degree of endemism (van der Elst, 1988; Turpie et al., 2000), due, again, in no small measure to the powerful Agulhas Current, which, along with the vast and open southern Indian Ocean, forms an effective barrier to coastal species migration, northeastwards, away from the southern tip of Africa.

4.7 Adult fish migrations offshore in relation to St Lucia Estuary A major feature of spawning patterns of fishes on the shelf of KZN is that some of the key spawning species move into these waters, from the south in winter, on spawning migrations (van der Elst, 1988; Hutchings et al., 2003). Of these, however, only kob Argyrosomus japonicus, Cape stumpnose Rhabdosargus holubi and garrick Lichia amia larvae actively recruit to estuaries, and the former two have been commonly found in the St Lucia estuarine system in the past (Wallace, 1975a; Harris and Cyrus, 1995, 1996). The garrick appears to spawn in the shallows of the Natal Bight, between Durban and the

Thukela River (Connell, 2006), but larvae and juveniles are rare in KZN estuaries (Wallace and van der Elst, 1975), evidently moving south into the estuaries of the Transkei and Eastern Cape (Wallace and van der Elst, 1975; Harrison, 2003). An examination of all fishes recorded from St Lucia in the past, including the very small species recorded by Harris and Cyrus (1995, 1996), reveals that all are Indo-Pacific or western Indian Ocean species, with the possible exception of R. holubi and Psammogobius knysnaensis, but even these two species are known to occur in Kosi Bay, to the north

The marine environment

of St Lucia (Whitfield, 1980a; Harris et al., 1995), providing a potential source of larvae to the St Lucia estuarine system. Note, however, that both Durban Harbour and Richards Bay have the same suite of species as St Lucia, with the exception of the endemic blenny Omobranchus woodi, which has not been recorded north of Durban, according to South African fishes collection records (E. Heemstra, pers. comm.). While Durban and Richards Bay at the south and north ends of the Natal Bight respectively, are likely to receive larvae from adult spawning both to the north and south of them, the St Lucia estuarine system is likely to be predominantly fed by recruits from the north, due to the extreme narrowing of the shelf north of Richards Bay, creating a barrier to northward movement along the shelf by larvae and juveniles of estuary-related species (but see note below regarding prawn postlarval recruitment). The St Lucia estuarine system was, in the past, associated with migrations of adult fishes, the most spectacular of which were the mullet species Mugil cephalus, Liza macrolepis and L. dumerilii, and the grunter Pomadasys commersonnii (Wallace, 1975a), the former because they attracted sharks and crocodiles to the area, and the latter because they were targeted by anglers. The mullet migration within the estuary was a seaward spawning migration of sexually mature fish, but Wallace (1975a) attributed the sudden appearance of spectacular numbers of fish in the estuary mouth area to arrivals from both the lake and the sea. Garratt (1993), while studying spawning behaviour of the riverbream Acanthopagrus berda at the mouth of the Kosi Estuary, about 170 km north of St Lucia Estuary, found similar congregations of the mullet L. macrolepis, where they were shown to spawn, just inside the mouth, along with Acanthopagrus, on the outgoing tide at night. The arrival of mullet shoals at St Lucia Mouth during April and May each year probably represents a similar spawning aggregation, and coincides with the April to November peak of mugilid eggs in the mouth of Durban Harbour (Connell, 2006). Wallace (1975a)

showed that few adult fish returned to the lake after spawning at sea. The annual ‘grunter run’ was marked by the arrival of fish from offshore in August, and movement of some spent fish into the lake from then until November. It appeared, however, that the majority of fish remained at sea, and few adult fish remained in the St Lucia estuarine system during winter (Wallace, 1975a). Less obvious migrations into and out of St Lucia include several species of penaeid prawns, the most abundant of which is Penaeus indicus (see Chapter 14), larvae of crabs such as Scylla and Varuna, and numerous juvenile fish, the latter comprising many fish species that utilize the system as a nursery area (see Chapter 15). With regard to the source of penaeid prawn entering St Lucia Estuary, Forbes and Demetriades (2005) found that their flood-tide study of postlarval recruitment to Kosi Bay (the first large estuarine system south of Maputo Bay) revealed a catch dominated by P. japonicus, to the virtual exclusion of other penaeid species. By contrast, recruitment to St Lucia, Richards Bay and Durban Harbour had a much wider variety of prawn species, suggesting recruitment from stocks on the Thukela Banks, located in the northern section of the Natal Bight, adjacent to the Thukela River. Given that the Agulhas Current between Kosi Bay and St Lucia is close inshore (Figure 4.1), it seems logical to assume that a similar P. japonicus dominated recruitment would be found in the St Lucia estuarine system, unless prawn larvae from elsewhere are being added to the mix entering St Lucia. While Richards Bay and Durban are at opposite ends of the Natal Bight, where south-going offshore currents and predominantly north-going inshore currents would aid recruitment of larvae to both these systems from the Thukela Banks, St Lucia Estuary lies north of the Cape St Lucia bottleneck, which should theoretically prevent prawn postlarvae from Thukela Banks reaching St Lucia. Although the studies of prawn recruitment at the four sites (Kosi Bay, St Lucia estuarine system, Richards Bay and Durban)

71

72

Allan D. Connell and Sean N. Porter

were conducted in different years, and no data are available on the diversity of prawn larvae in the nearshore marine environment off Kosi Bay at the time of the study (to eliminate the possibility that other species of prawn larvae were available, but disinclined to recruit to Kosi Bay Estuary), the data of Forbes and Demetriades (2005) suggest there may be a link between the St Lucia estuarine system and the coastal marine environment to the south. The St Lucia estuarine system provides an important nursery ground for juveniles of many marine species and is considered to be one of the most important estuarine systems along the east coast of South Africa for juvenile marine recruits (Cyrus and Vivier, 2006a; Vivier et al., 2010a; Bate et al., 2011). Begg (1978) has estimated that St Lucia

comprises 80% of the total estuarine habitat in KwaZulu-Natal. During times of prolonged mouth closure, significant impacts are expected on estuarine-associated marine species (Cyrus and Vivier, 2006a). Prior to the opening of the estuary mouth in March 2007 for example (Cyrus et al., 2011), a significant reduction in catch per unit effort was recorded for the Natal stumpnose, Rhabdosargus sarba, which makes extensive use of St Lucia (Mann and Pradervand, 2007). Furthermore, Vivier et al. (2010a) found that 79% of species recorded in the estuary spanning a period when the mouth opened were those that utilized the estuary for spawning or nursery grounds, and that recruitment of postlarvae of 20 marine species occurred following the mouth opening in March 2007.

4.8 Offshore canyons and their potential impact on St Lucia Estuary Six large (and some smaller) submarine canyons are located on the northern KwaZulu-Natal continental margin, the most southern of which is off Leven Point, due east of the northern edge of the main lake of Lake St Lucia. The most northern canyon is off Mabibi, about 80 km to the north. These canyons originate on an unusually steep continental slope at depths of 100–300 m, resulting from a late Pliocene influx of sediment (Green et al., 2008), and are essentially massive erosion gullies, eroding upslope to form cauliflower-shaped heads, and downslope, where they form discrete channels, generally straight or slightly meandering, and extending into water depths in excess of 500 m. Downslope gradients on the unstable shelf-edge wedge vary between 6 and 10.5 , dropping to 3–5 below 300 m (Green et al., 2007). Instability of the shelf-edge wedge is considered the cause of canyon inception. Although trending at right angles to the coast, offshore canyon locations do not correspond to any terrestrial hydrological features such as rivers, with the exception of Wright Canyon, opposite Lake Sibaya, which is linked by an old river course (Green,

2009). Nevertheless the canyon heads are sufficiently close to shore that the 50 m depth contour is less than 2 km offshore at Leven and Diepgat Canyons, closer than anywhere else on the KwaZulu-Natal coast. Old sea level notches, on erosion faces, indicate that the canyons are currently inactive, suggesting that canyon growth is linked to periods of lowered sea level associated with ice ages (Green and Uken, 2005). Postdating the last two glacial-induced low sea levels, at 18 000 and 180 000 years BP, when sea level dropped to about 120–125 m below present, has been a period of sediment starvation, which has further inhibited canyon development (Ramsay, 1994; Green, 2011). These canyons have attracted close attention, following the recent discovery of coelacanths resident in caves on the canyon walls (Venter et al., 2000). Little is known, however, about their local ecological impact. No published information is available to indicate whether they funnel cold water up onto the shelf, as is for example known to occur over the Monterey Bay Canyon (Ryan et al., 2005). A series of Acoustic Doppler Current Profiler (ADCP)

The marine environment

temperature recordings taken amongst canyons, and others simultaneously recording at the same depths away from canyons, indicated no discernable difference in temperature profiles (M. J. Roberts, pers. comm.). Examination of numerous satellite imagery indicates very little cooler water inshore of the Agulhas Current in the Sodwana Bay to Cape Vidal

area, apart from the regular Ekman-veering intrusion of bottom water onto the shelf, discussed earlier. The presence of coral reefs at Sodwana Bay and Leadsman Shoal, both located within the stretch of coast containing the canyons, would also indicate canyon-induced cold water intrusion is not a common occurrence.

4.9 Coastal biogeography The marine ecology in the vicinity of St Lucia Estuary is significant and encompasses an important area biogeographically on the East African coast. At a regional scale, the St Lucia estuarine system is an area where many taxonomic groups show a rapid transition between tropical and temperate water organisms, and is the location of biogeographic discontinuities for a suite of species in the western Indian Ocean (Bolton et al., 2004; Sink et al., 2005; Porter, 2009). Many taxa, particularly seaweeds (Bolton et al., 2004; Evans, 2005), echinoderms, brachyurans, molluscs, acarians (Awad et al., 2001) and fishes (Whitfield, 1980a; Turpie et al., 2000) also show relatively high levels of diversity compared with other sections of the southern African coast, demonstrating the fusion of tropical, subtropical and temperate water species that is a characteristic of the marine environment of St Lucia. Early marine biogeographic analyses on the east coast of Africa focused on intertidal invertebrate assemblages in southern Africa and described St Lucia as being subtropical (Stephenson, 1948; Jackson, 1976; Brown and Jarman, 1978; Emanuel et al., 1992; Bustamante and Branch, 1996). The coastal biogeographic region encompassing St Lucia was termed the Subtropical Natal Province, also referred to as the Subtropical East Coast Province, with its northern boundary declared to lie in the vicinity of Maputo in Mozambique and its southern boundary at Durban (Stephenson, 1948). At a regional scale of thousands of kilometres, the Subtropical Natal Province is situated between the

vast Tropical Indo-West Pacific Province to the north and the Warm Temperate South Coast or Agulhas Province to the south (Stephenson, 1948; Briggs, 1974; Lu¨ning, 1990; Emanuel et al., 1992; Bustamante and Branch, 1996). Recent analyses and intensive sampling have been conducted north of the St Lucia estuarine system and have refined the positions and boundaries of various biogeographic regions (Bolton et al., 2004; Evans, 2005; Lawrence, 2005; Sink et al., 2005; Porter, 2009). Phytogeographically, the St Lucia/ Cape Vidal area is where the fastest changeover of tropical Indian Ocean seaweed to temperate southern African species occurs (Bolton et al., 2004; De Clerk et al., 2005). South of St Lucia for several hundred kilometres the flora becomes an equal mix of tropical and temperate species and is therefore considered to be a true region of overlap by phytogeographers, not a separate biogeographic province, comprising temperate Agulhas and tropical Indo-West Pacific floras (Bolton and Anderson, 1997; Bolton et al., 2004). Zoogeographers, however, consider the area south of and adjacent to St Lucia Estuary as a distinct marine province (i.e. Subtropical Natal Province) (Emanuel et al., 1992; Bustamante and Branch, 1996; Turpie et al., 2000; Sink et al., 2005; Porter, 2009) with its northern border in the vicinity of Cape Vidal (a few kilometres north of St Lucia Mouth) (Sink et al., 2005; Porter, 2009). North of Cape Vidal, analyses of reef invertebrate fauna and flora show that the area should be considered an overlap region where species

73

74

Allan D. Connell and Sean N. Porter

from the Subtropical Natal Province transition with those from the Tropical Indo-West Pacific Province (Sink et al., 2005; Porter, 2009). This area of overlap is known as the Delagoa Bioregion (Spalding et al., 2007; Porter, 2009) and many of the reefs there support Africa’s southernmost coral communities (Riegl et al., 1995; Schleyer, 2000). These biogeographic patterns are generally mirrored by many other marine and estuarine taxa, although the precise location of boundaries among provinces may vary (Thandar, 1989; Williams, 1992; Turpie et al., 2000; Harrison, 2002; Acuna and Griffiths, 2004; Primo and Va´squez, 2004; Thandar and Samyn, 2004; Samaai, 2006). The relationship and relevance of coastal biogeography to St Lucia Estuary and the adjacent Mfolozi River becomes apparent when one considers the ecological forces driving these biogeographic patterns and the species that typify the adjoining marine environment. The Subtropical Natal Province is characterized by a high biomass of filter feeders relative to the Tropical Indo-West Pacific Province to the north of St Lucia Estuary (Porter, 2009). Sink et al. (2005) found that the filter-feeding bivalve Perna perna was the most important species distinguishing the Subtropical Natal Province from intertidal rocky shores north of Cape Vidal. On average, P. perna cover was only 12% on rocky shores north of Cape Vidal whereas it comprised 51% of shores further south. The dominance of P. perna is demonstrated locally near the mouth of the St Lucia Estuary on the intertidal rocky shores of First Rocks to the north and Maphelane to the south (Tomalin and Kyle, 1998). A similar pattern of high filterfeeder abundance has been found on shallow subtidal reefs. Filter-feeder biomass is three and a half times higher on reefs in the Subtropical Natal Province compared with those that lie north of St Lucia Estuary (Porter, 2009). Two dominant abiotic factors attributable in part to the St Lucia Estuary and Mfolozi River appear to be driving these patterns. Turbidity is significantly greater in the Subtropical Natal Province compared with the region north of St Lucia Estuary due to the

FIGURE 4.8 Contrasting levels of turbidity north and south of St Lucia for the 3-year Period 2003–2005. Sidebar is turbidity in inverse metres (m 1).

multitude of rivers that exude into the nearshore of this area (Sink, 2001; Porter, 2009) (Figure 4.8). In addition to increasing turbidity, estuaries and rivers input terrestrial and freshwater particulate organic matter (POM) into the inshore zone (Schleyer, 1981; Porter, 2009; Vorwerk and Froneman, 2009). The enhanced levels of turbidity and POM resulting from estuaries and rivers are likely to favour the increased biomass of filter feeders in the region. This is especially probable when viewed in the context of the coastal waters being oligotrophic (see Carter and d’Aubrey, 1988; Carter and Schleyer, 1988; Porter, 2009), comprising a high proportion of seaweeds that are coralline and hence unpalatable in nature, and that there is an absence of large productive systems such as kelp beds to form the basis of the food chain (Evans, 2005; Lawrence, 2005; Porter, 2009). Evidence from isotope-trophic analyses support this, and reveal that inshore marine POM can comprise 20% terrestrial/freshwater components near the mouths of the Mfolozi River and St Lucia Estuary (Porter, 2009). Rivers in the Subtropical Natal Province contribute between 9% and 33% to filterfeeder community diets and four species of filter feeders in the vicinity of the Mfolozi and St Lucia

The marine environment

Mouths, specifically, have been shown to assimilate between 10% and 20% terrestrial-freshwater POM (Porter, 2009). The assimilation of river-derived POM has also recently been established in filter feeders from the Eastern Cape of South Africa (Vorwerk and Froneman, 2009) and further afield in the Pacific Northwest (Tallis, 2009). Furthermore, levels of turbidity have been found to correlate with shallow-subtidal reef

biogeography at a regional scale (hundreds of kilometres) and among different reefs within the Subtropical Natal Province, suggesting a strong association with rivers and estuaries in general (Porter et al., 2007; Porter, 2009). Turbidity levels are significantly different and twice as high in the coastal waters of the Subtropical Natal Province than they are north of St Lucia (Porter, 2009) (Figure 4.8).

Acknowledgements We thank the Active Archive Center at the NASA Goddard Space Flight Centre, Greenbelt, Maryland, for allowing us access to the Level-2 MODIS data.

We are also grateful to Tioxide Southern Africa, Huntsman Pigments Division, for access to the water temperature data used in Figure 4.2.

75

Chapter contents 5.1 Introduction 5.2 Hydrological context 5.3 Annual average runoff and sediment yields 5.4 Modelling the impact of land-use changes in the Mfolozi catchment 5.5 Management implications 5.6 Conclusions

Mkhuze River exiting the floodplain into St Lucia. (Photo: Ricky H. Taylor, May 2010.)

5

Catchment hydrology Derek D. Stretch and Andrew Z. Maro

5.1 Introduction Lake St Lucia is supplied by five rivers, the catchments of which lie outside the boundaries of the Park. North to south these are the Mkhuze, Mzinene, Hluhluwe, Nyalazi and Mpate. The Mfolozi and Msunduze rivers in the south enter the sea together close to the mouth of Lake St Lucia. The largest rivers, the Mkhuze and Mfolozi, have little of their alluvial lower reaches in the Park. The rivers are seasonal, flowing during the wet summer months and reduced to isolated pools and seepage through bed sediments in winter. High sediment loads from the Mkhuze river which drains the Lubombo mountains fill its mouth, forming meandering distributaries, levees and pans with swamp and riverine forest. As a result, the waters of Lake St Lucia are turbid. (UNEP, 2008) Most of the iSimangaliso Wetland Park surrounding the St Lucia estuarine system is pristine and the park management philosophy aims for minimal intervention and the restoration or mimicking of natural processes where they have been disrupted. Nevertheless, as in many protected areas in southern Africa, the rivers that flow into the system drain from significantly disturbed catchments. Perhaps the most important aspect of this is the changing land-use patterns that have evolved over the last century. There are perceptions that these have led to significant increases in sediment yields, particularly from the Mfolozi catchment. This perception has been and continues to be a major driver of

management interventions including the separation of the St Lucia and Mfolozi mouths in 1952. The management issues are reviewed and discussed in Chapter 2. Hutchison and Pitman (1973, 1977) carried out the first detailed analysis of the regional hydrology of the St Lucia and Mfolozi system. The work was part of a larger study of the water and salt balance of Lake St Lucia (Hutchison, 1976; Hutchison and Midgley, 1978). There has apparently been no comprehensive hydrological study undertaken since that time, although several ad hoc studies have been carried out with specific (more limited) objectives, but without publication of the results in the primary scientific literature (e.g. Weston et al., 1995; Quibell, 1996; Van Niekerk, 2004). More recently, Grenfell et al. (2009) and Grenfell and Ellery (2009) investigated the hydrology, sediment dynamics and geomorphology of the Mfolozi River. The important role of the Mfolozi River (and the sediment yield issue) in the management of the St Lucia system adds to the significance of these specialist studies (refer to Chapter 2 and Chapter 7) and is discussed further in Section 5.5. The standard data source for water resources management in South Africa is the publication Surface Water Resources of South Africa (Midgley et al., 1994) widely known as WR90 and updated as WR2005 (Middleton and Bailey, 2008). The WR90/ 2005 compendium contains quaternary level

Ecology and Conservation of Estuarine Ecosystems: Lake St Lucia as a Global Model, eds. Renzo Perissinotto, Derek D. Stretch and Ricky H. Taylor. Published by Cambridge University Press. © Cambridge University Press 2013.

78

Derek D. Stretch and Andrew Z. Maro

catchment information (areas ~500 km2) together with monthly averaged data for rainfall, evaporation, streamflow and sediment yields. The study by Rooseboom (1992) provided the sediment yields for WR90/2005. The recent updates by Msadala et al. (2010) will presumably be incorporated into future releases of the WR90/2005 series. Data on dams, irrigation abstractions and other hydrologically relevant information (land cover, water use, wetlands, afforestation, etc.) are also included in WR90/2005. WR90 was intended as an updated survey of an earlier study conducted in 1981 by the Hydrological Research Unit (HRU) at the University of Witwatersrand (HRU, 1981). The objectives of WR90/ 2005 are to provide a basis for preliminary assessment and planning of water resources developments. The monthly streamflow data provided in WR90/ 2005 are simulated using the WRSM model based on a rainfall-runoff module developed by Pitman (1973). The WRSM model was calibrated against

observed streamflow records. ‘Naturalized’ streamflow was simulated by setting all upstream land-use components to ‘virgin’ conditions and adding back any land-water-use effects. The Pitman model was also used for the previous study of the St Lucia estuarine system by Hutchison and Pitman (1973, 1977). In that case they estimated that the Mfolozi River mean annual runoff (MAR) was reduced by about 2% from virgin conditions, while the MAR for all other rivers discharging into the St Lucia estuarine system had reduced by 18% relative to virgin conditions. The recent studies by Lawrie and Stretch (2011a, 2011b) were based on the same hydrological information as that used for the original work of Hutchison and Pitman (1973, 1977) but streamflow simulations were extended from 1972 to 2010. In this chapter we summarize information on the hydrology of the St Lucia/Mfolozi catchment system, and then focus on recent new results by Maro (2012) concerning the effects of land-use changes on water and sediment yields.

5.2 Hydrological context 5.2.1 Overview The catchments of the St Lucia estuarine system are mapped in Figure 5.1 while some of the key physical features such as topography, soil types and land-use patterns are shown in Figures 5.2–5.5. The southern half of the catchment system shown in Figure 5.1 is drained by the White and Black Mfolozi Rivers with a confluence about 50 km from the sea. After the confluence the Mfolozi flows through a lowlying coastal floodplain that has an area of about 150 km2. The river discharges into the sea between the Maphelane bluff and the St Lucia Narrows (about 2 km to the north). The Msunduzi River is a short channel that drains a small subcatchment around the lower Mfolozi floodplain and swamp area (Figure 5.1) and merges with the Mfolozi just before the mouth. The Mfolozi catchment has been divided into 26 quaternary subcatchments for water resources

planning purposes (WR90/2005) with a total area of 10 085 km2. The low-lying coastal floodplain and wetland/swamp near the mouth of the Mfolozi River has historically played an important role in the management of St Lucia because of human impacts that have interfered with its function as a sediment filter. A detailed review of the connectivity between these wetlands and Lake St Lucia is discussed in Chapter 6. The northern half of the catchment system shown in Figure 5.1 is drained by the Mkhuze River and several smaller rivers that discharge directly into the lake and Narrows. From north to south they are: Mzinene, Hluhluwe, Nyalazi, Mpate and other smaller streams. The sandy Eastern Shores of Lake St Lucia also drain into the lake system by way of small streams and groundwater flows (refer to Chapter 8 and Været et al. (2009) for details).

Catchment hydrology

Mkuze W3H011

ene Mzin

Klipfontein W2H030 W2H009 W2H007

Secondary Rivers

W3H012

W2H008

Primary Rivers

W2H028

DWA Weir Legend Impoundments

W3H015

Black Mfolozi

W2H006

Quaternary Catchment

W3H022

Hluhluwe W3H013

White Mfo

lozi W2H002

W3H014 ate Mp

W2H005

W2H010

Mfolozi W2H032

Msunduzi

N E

W

0

20

40

60

80

100 Kilometers S

FIGURE 5.1 Catchments of the St Lucia estuarine system. The locations of stream gauging sites are shown on the map.

The Hutchison and Pitman (1973) estimate of 18% reduction in total MAR (relative to virgin conditions) for the Lake St Lucia catchments comprised 16% reduction for the Mkhuze, 43% for the Hluhluwe (due mainly to a dam and associated abstractions), 25% for the Nyalazi, 8% for the Mzinene, and 31% for the Mpate. A reanalysis of this issue is clearly required in order to update these estimates since there have been considerable land-use changes during the period since 1973. In particular, abstractions from the Mfolozi for irrigation and dune-mining operations may have substantially increased since that time, as well as farming activities in the Mkhuze catchment system. The total catchment area draining directly into Lake St Lucia is estimated to be 7575 km2 (Hutchison and Pitman, 1973). This comprises 19 quaternary catchments delineated in WR90/2005 but excludes a substantial endorheic area to the north of the Mkhuze Swamps. The Mkhuze Swamps is a hydrologically important part of the system (refer to Chapter 6) and comprises an area of about 150 km2 to the north of Lake St Lucia. There are several low-lying pans and wetlands in the area north of the lake.

5.2.2 Catchment topography The topography of the Mfolozi and St Lucia catchment system is shown mapped in Figure 5.2. Elevations rise from sea level in the east to about 1600 m in the west. Gradients are small on the wide (~30 km) coastal plain but elevations increase to ~1000 m in the next ~120 km.

5.2.3 Soil types Soil types in the catchment are shown in Figure 5.3 and vary from loam soils on the higher elevations in the west to sandy soils on the lowlying eastern half of the catchment (NLC, 2005; Schulze and Horan, 2008; Schulze et al., 2010).

5.2.4 Land-use patterns Natural land cover (veld types) prior to significant human interference are shown mapped in Figure 5.4 (Acocks, 1988). Comparison with Figure 5.2 shows a strong correlation with altitude.

79

Derek D. Stretch and Andrew Z. Maro

FIGURE 5.2 Basin elevations (in metres above MSL) and primary contributing rivers to the St Lucia estuarine system.

Mkuze

Klipfontein Dam Legend

Black Mfolozi

Hluhluwe Dam

Ny ala

zi

White Mfolozi

Mfolozi

Elevation [m] 0 - 99 100 - 199 200 - 399 400 - 599 600 - 799 800 - 999 1000 - 1199 1200 - 1600 N

Msunduzi

E

W 0

20

60

40

80

100 Kilometers S

FIGURE 5.3 Soil types of the Mfolozi and St Lucia catchments (NLC, 2005; Schulze and Horan, 2008; Schulze et al., 2010). Mkuze

Black Mfolozi

Legend Mfolozi

zi

White

Ny ala

80

Mfolozi

Soil Type loam loamy sand sandy clay loam

Msunduzi

N W

0

20

40

60

80

E

100 Kilometers S

A current land-use map is shown in Figure 5.5 (NLC, 2005; see also Fairbanks et al., 2000) and shows significant changes in the land use. Tabulated values for the different land-use classes are shown in Table 5.1.

5.2.5 Climate context The St Lucia system is situated in the north-eastern coastal region of South Africa and is a summer rainfall area with strong east–west climate gradients.

The regional climate is strongly influenced by the warm energetic Agulhas Current (refer to Chapter 4) and is also vulnerable to tropical cyclones that have proven catastrophic in the past (e.g. Reason and Keible, 2004). Interannual variability in the region is high (Mason and Jury, 1997) with quasi-periodic drought/ flood periods that have been linked to southward migrations of the Intertropical Convergence Zone and to the El Nin˜o Southern Oscillation (ENSO) (Rautenbach and Smith, 2001). Hewitson et al. (2005) note that the regional oceans play an important role

Catchment hydrology

FIGURE 5.4 Veld types in the St Lucia catchment system. (Data source: Acocks, 1988.)

Legend Acocks Veld Types ARID LOWVELD COASTAL FOREST & THORNVELD HIGHLAND AND DOHNE SOURVELD LOWVELD NATAL SOUR SANDVELD NGONGONI VELD - ZULULAND NORTH-EASTERN MOUNTAIN SOURVELD NORTH-EASTERN SANDYHIGHVELD PIET RETIEF SOURVELD THE NORTHERN TALLGRASSVELD THE SOUTHERN TALLGRASSVELD THEMEDA-FESTUCA ALPINE VELD ZULULAND THORNVELD

N

0 10 20 30 40 50 Kilometers

W

E S

legend Current land-use Forest (Indigenous) Woodland Thicket/Bushland Unimproved natural grassland Improved grassland

FIGURE 5.5 Current land-use patterns in the St Lucia catchment system. (Data source: NLC, 2005.)

Forest plantations Waterbodies Wetlands Bare rock/erosion/gullies Degraded woodland/grassland Cultivated (irrigated) Cultivated (dryland) Cultivated (sugarcane) Urban Mines/quarries N W

0

20

40

60

80

100 Kilometers

in modulating large-scale circulations and the moisture availability that leads to aridity in the west and wetter conditions in the east of the subcontinent. Mid-latitude depressions to the south of the country are associated with most rainfall, and cut-off lows in particular have been associated with major floods in the past.

5.2.6 Floods Some of the major floods that occurred during the period 1918–2010 were: • March 1925 flood: A severe cyclone led to the highest rainfall on record for the period – 1400 mm over 11 days recorded at Riverview near Mtubatuba.

E S

• July 1963 flood: Associated with a cut-off low weather system. On 3 to 4 July 1963 large sections of the contributing catchments received exceptionally high 24-hour rainfall during which recorded rainfall exceeded 100 mm on either 3 or 4 July 1963. Moderate rain was measured on the days before and after the event. In some areas the daily rainfall was estimated to exceed 500 mm in 24 hours. • January/February 1984 flood: Cyclone Domoina at end of January 1984 followed by Cyclone Imboa 11 to 20 February 1984. Most of the subcatchment’s rainfall exceeded 100 mm for more than 2 or 3 days consecutively, with consecutive rainfall for 4 to 5 days in places.

81

82

Derek D. Stretch and Andrew Z. Maro

Table 5.1. Details of current land cover classes within the catchments of (a) the Lake St Lucia system catchments, and (b) within the Mfolozi catchment system (a) Current land cover (all)

Area (km2)

%

Forests: indigenous

363

2

Forests: plantations

894

5

Trees and grass mix: woodland

4 770

26

Trees and grass mix: wood/grass combination

2 798

15

491

3

–

–

6 369

35

Grass: degraded

812

4

Cultivation: commercial; dry-land

215

1

Cultivation: subsistence; dry-land

1 562

9

–

–

18 274

100

Natural

6 700

67

Agriculture

1 800

18

Degraded

1 400

14

Urban

300 N

Average annual sediment yield: 146 t/sq km/a [Data source: WR2005] 0 10 20 30 40 50 Kilometers

W

E S

Catchment hydrology

FIGURE 5.7 Average runoff coefficients for the St Lucia quaternary catchments. (Data source: WR2005: Middleton and Bailey, 2008.)

5.2% 4.1% 11.3% 3.9% 11.6% 3.7%

12.0%

11.5%

8.7%

3.2% 4.2%

4.5%

14.8%

9.0%

18.1%

8.4% 8.4% 8.8%

4.1%

17.0%

8.4%

8.5% 11.0%

12.1% 12.5%

11.0% [Data source: WR2005] 0 20 40 60

8.3%

7.4% 10.6%

10.2%

9.5%

8.2%

5.1%

12.0% 11.8%

10.9%

12.2% 7.4% 7.8%

12.7%

9.9% 10.8%

13.6%

13.5%

16.1%

N E

W

80

100 Kilometers

Daily maximum and minimum temperatures for the quaternary catchments were estimated from data at reference stations in the area together with assumed adiabatic lapse rates to correct for altitude changes (Schulze and Maharaj, 2004; Smithers and Schulze, 2005). The daily temperature series was then used to generate estimates of reference potential evaporation since there is no measured data. The Hargreaves and Samani (1985) method was used to estimate reference potential evaporation since it is a simple and practical method of estimating plant water requirements using a minimum of climatological data; that is, only daily minimum and maximum temperature data.

5.4.4 Soil types The soil information required for ACRU to simulate the impacts of land-use changes include: thicknesses of soil horizons, surface properties affecting infiltration, percentage of clay/sand/silt within the soil horizons, the relationship to permeability and hydraulic conductivity, water retention properties of the soil (i.e. wilting point, field capacity and total

S

porosity) and soil erodability. The soil properties were sourced from the Institute of Soil, Climate and Water (NLC, 2005). This database identifies nine broad categories of soil types in South Africa, of which three are present in the Mfolozi catchment (refer to Figure 5.3), with soil horizon depths from 0.2 m to 0.8 m and erodibility factors from 0.2 to 0.6.

5.4.5 Reference state simulations The baseline or reference state land cover used by Maro (2012) was the Acocks (1988) veld types that have been frequently used in hydrological studies. This served as the benchmark for assessing the hydrological impacts of changes in land-use and management practices. The Acocks veld types within the study area and schematics of the river flow systems are shown in Figure 5.4.

5.4.6 Current land use The South African National Land Cover database (NLC, 2005) indicates 42 distinct land-use classes within the study area – they were mapped using

85

86

Derek D. Stretch and Andrew Z. Maro

satellite imagery. This is summarized in Table 5.1 and shown in Figure 5.5. Land-use classes were overlain onto the Mfolozi quinary catchments. Each quinary catchment was then delineated into distinct land-use classes, and each land-use class was allocated a unique veld type. Maro’s (2012) simulations take into account wetlands, natural water bodies and impervious areas, but do not account for the effects of irrigated areas.

5.4.7 Water budget in ACRU Soil water deficit within ACRU is determined through the multilayer soil water budget after defining a critical soil depth. The critical soil depth takes into account different runoff-generating mechanisms as a result of different land-use or climate conditions. For simulations within this study, this critical soil depth was limited to the thickness of the top horizon soil. Schulze (1995) describes the way that ACRU models the timing of stormflow releases from subcatchments, including the effects of slope and size.

5.4.8 Baseflow Baseflow is the fraction of streamflow that originates from accumulated deep subsurface flow or intermediate/groundwater store. This is computed in ACRU exclusively from recharged soil water stored in the intermediate/groundwater zone (Schulze, 1995). This recharging effect is a result of rainfall events that have been redistributed through soil horizons into the intermediate/groundwater store when the deepest soil horizon’s water content exceeds field capacity. A release coefficient determines the rate of release of groundwater into the stream. This coefficient is dependent on catchment characteristics such as slope, area and geology. The same release

coefficient was applied to all subcatchments within the Mfolozi catchment.

5.4.9 Calibration of streamflow simulations After setting up the ACRU model, streamflow was simulated and accumulated on monthly time steps and verified against the Department of Water Affairs flow gauges (see Figure 5.1). The calibration process followed the steps outlined by Smithers and Schulze (2005).

5.4.10 Sediment yields in ACRU Sediment yield modelling in ACRU utilizes the Revised Universal Soil Loss Equation (RUSLE). Empirical coefficients in the RUSLE model are generally catchment specific but Maro (2012) used values recommended by Schulze (1995) based on research cited therein. The RUSLE model requires the peak discharge per stormflow event for each quaternary catchment. ACRU uses a peak discharge equation suggested by Schulze (1995) and based on methods developed by the US Soil Conservation Service (SCS). Other required input data for the sediment yield module in ACRU are: • Average catchment slope per quinary catchment – from a 20 m digital elevation model. • Stormflow volume per land-use management scenario. • Soil erodibility factor from NLC (2005). • Slope length and gradient function – internally calculated in ACRU using catchment slope. • Soil erodibility cover and management (Wischmeier and Smith, 1978; Schulze, 1995).

5.4.11 Calibration of sediment yields The validation of sediment yield simulations began with field methods of water collection for unobstructed flow using pop-bottles. In the

Catchment hydrology

laboratory, turbidity (NTU) readings of samples were taken before filtering to determine sediment concentration in the form of total suspended solids (TSS). A linear relationship was established between TSS and NTU. This allowed for conversion of the 10-year record of daily turbidity readings recorded at the Mtubatuba Water Treatment Works into daily sediment concentrations. Using these concentrations and the product of daily average streamflow measurements from weir W2H032 (Figure 5.1) daily suspended sediment loads were determined. Accumulated monthly suspended solids were increased by 20% to account for bedload. This factor was estimated by averaging measured ratios of suspended load to bedload obtained by Grenfell and Ellery (2009). Finally, the simulated annual average sediment yields for each quaternary catchment were compared with WR90/2005 data and were found to agree to within about 20%. The accumulated yield from the whole catchment agreed to within 5%.

5.4.12 Results – the effect of land use on runoff Maps of the simulated quaternary catchment MAR are shown in Figure 5.8 for both the Acocks natural veld types and current land use. The ACRU simulations for Acocks veld types are consistent (within 20%) with WR2005 naturalized streamflow (Figure 5.7). Compared with current land use they suggest a decrease of about 40% in the mean annual runoff from the Mfolozi catchments due to the changes. This change is much larger than has previously been assumed for the Mfolozi. Recall that Hutchison and Pitman (1973) previously estimated the change as about 2% for the period prior to 1972, but using a different approach where water utilization for specific land uses was

added back to estimate naturalized flows. The ACRU model uses a more integrated approach where the hydrological responses to the changes are parameterized in the model. If the Maro (2012) results are extrapolated to the other catchments in the St Lucia system, they suggest that overall reductions in inflows due to land-use changes may be a significant factor in the water balance of the lake that needs to be further investigated.

5.4.13 Results – the effect of land use on sediment yields Maps of the simulated quaternary catchment sediment yields are shown in Figure 5.9 for both the Acocks natural veld types and current land use. The ACRU simulations for current land use are broadly consistent with WR2005 data (Figure 5.6). They suggest an increase of about 50% in the annual total sediment yield from the Mfolozi catchments due to the land-use changes. The changes in average runoff and sediment yields for the Mfolozi system are summarized in Figures 5.10 and 5.11. Further insight into the nature of the sediment dynamics is evident in the simulated time series of annual average yields shown in Figure 5.12 that shows the relatively large contributions that are made by episodic flood events. Figure 5.13 shows simulated time series of monthly flows together with suspended sediment concentrations. The highly variable and seasonal characteristics of suspended sediment loads is evident, as well as the strong reduction in sediment yields since the onset of drought conditions after 2001. According to the simulations, peak monthly averaged sediment concentrations during flood events can exceed 10 g l 1 (with associated turbidity values of order 10 000 NTU.

87

88

Derek D. Stretch and Andrew Z. Maro

Mean Annual Runoff (MAR) Mfolozi Catchment (Acocks Veld Types)

23.76

Legend MAR [Mcm]

41.24

< 10

73.54

10–29

25.50 40.47

55.88 18.83

15.75

51.81 40.04

60–89 90–119

32.54

120–149

61.78 24.95

30–59

67.39

150–179

115.05

180–209

35.84

41.24

47.22 73.04

44.50

> 210

50.39

49.20

25.04

30.65 50.67 N

Average MAR: 1183 Mcm 20

0

40

60

80

W

100 Kilometers

E S

Mean Annual Runoff (MAR) Mfolozi Catchment (Current Land-use)

17.13

Legend MAR [Mcm]

26.53

28.65

10–29

18.47

21.16 37.49

19.52 34.74

24.92

< 10

33.68

13.47

60–89 90–119

19.57

120–149

25.82 22.44

33.09

30–59

28.17 50.46

150–179 180–209

40.59

21.53 54.11

32.93

35.71

> 210

29.90 14.08

18.88 23.57 N

Average MAR: 727 Mcm 0

20

40

60

80

100 Kilometers

W

E S

FIGURE 5.8 The effect of land-use changes on the mean annual runoff of the Mfolozi catchments (Maro, 2012).

Catchment hydrology

FIGURE 5.9 The effect of land-use changes on the mean annual sediment yields of the Mfolozi catchments (Maro, 2012).

89

90

Derek D. Stretch and Andrew Z. Maro

–27.9% –30.5%

–42.7%

–54.2%

–27.6%

–59.2%

–37.8% 32.3%

–14.5%

–58.2%

–39.9% –58.2%

–10.1%

–19.8%

–6.4%

–51.8%

–56.1%

13.3%

–54.4% –25.9%

–26.0%

–27.4%

–40.7% –43.8%

–38.4% –53.5% N

0

20

40

60

80

100 Kilometers

E

W S

FIGURE 5.10 Changes in the mean annual runoff for each quaternary subcatchment of the Mfolozi due to land-use changes (Maro, 2012).

6

FIGURE 5.11 Change factors for the mean annual sediment yields of each quaternary subcatchment of the Mfolozi due to land-use changes (Maro, 2012).

1400 1200

Current Land Use

Acocks Veld Types

1000 800 600 400 200 0 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Sediment Yield (t/Km2/a)

Catchment hydrology

Year

12 10 8 6 4

Discharge m3/s

1979

1978

1977

1976

1975

1974

1973

1972

1971

1970

1969

1968

1967

1966

1965

0

Sediment concentration

12

800 700 600 500 400 300 200 100 0

10 8 6 4 2

Discharge m3/s

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1991

1990

1989

1988

1987

1986

1985

1984

1983

1982

1981

0 1980

Discharge (m3/s)

Long-term monthly average variation in discharge and sediment concentration (1980-2010)

Sediment Concentration (Kg/m3)

1964

1963

1962

1961

1960

1959

1958

1957

1956

1955

1954

1953

1952

2 1951

800 700 600 500 400 300 200 100 0

1950

Discharge (m3/s)

Long-term monthly average variation in discharge and sediment concentration (1950-1979)

Sediment Concentration (Kg/m3)

FIGURE 5.12 Time series of annual average sediment yields from 1950 to 2010 for the Mfolozi catchment illustrating the relatively large contribution from significant flood events in 1963, 1984, and 1987 during the period (Maro, 2012).

Sediment concentration

FIGURE 5.13 Time series of monthly discharge and suspended sediment concentrations for the period 1950–2010 for the Mfolozi catchment. Note the reduced discharges and sediment loads during the period following the onset of the drought from 2001 onwards (Maro, 2012).

91

92

Derek D. Stretch and Andrew Z. Maro

5.5 Management implications The sediments yield of the Mfolozi catchment and the possible sedimentation impacts on the St Lucia Narrows and St Lucia Lake have been an important and recurrent management issue concerning the St Lucia/Mfolozi system. The threat of sedimentation has been cited as the main justification for the artificial separation of the Mfolozi from St Lucia in 1952 (refer to Chapter 2). In Chapter 7 it was argued that this intervention has been the most important factor in changing the overall functioning of the system. The ACRU modelling results of Maro (2012) indicate significant changes in the water and sediment yields of the catchments as a result of landuse changes from the original undisturbed Acocks veld types. About one third of the Mfolozi watershed has an altered land use (Table 5.1). This seems to be the first detailed modelling study investigating this issue, although it has been widely claimed that increases in sediment yields caused the sedimentation problems that led to management intervention in 1952. It is also noteworthy that the modelling also suggests significant reductions in runoff due to the land-use changes. Grenfell et al. (2009) estimated the sediment yields for the Mfolozi using the available data for suspended sediment concentrations, but their study was confined to the period after 2000. It is evident from the longer-term simulations shown in

Figure 5.13 that estimates based on data for the post-2000 period would have significantly underestimated the sediment yields. This explains why their estimate was a factor of 2 to 3 lower than that of Rooseboom (1992) and that published in WR90/2005. Previous ACRU model studies of the effects of land-use changes on runoff in South Africa have been reported by Kienzle et al. (1997) and Schulze (2000). They found an overall 23% reduction in MAR for the Mgeni catchment (4387 km2), which has about 50% natural vegetation. The highest MAR reductions were 50% to 60% in highly afforested or heavily cultivated subcatchments. Some of the reductions were also attributed to abstractions. The higher overall reduction obtained for the Mfolozi system suggests further investigation is required to understand the cause of its greater sensitivity to land cover changes, at least as predicted by the ACRU model. Several aspects of the sediment yields require further investigation. The analysis of Maro (2012) does not account for sediment routing effects such as the trapping of sediment within small farm dams or in low gradient parts of the river channel. The contribution of erosion gullies (Liggitt and Fincham, 1989; Watson et al., 1996; Le Roux et al., 2010) to sediment yields of the Mfolozi is also a factor that requires further investigation.

5.6 Conclusions In this chapter we have reviewed the hydrology of the catchments that form an integral part of the St Lucia estuarine system. A key focus has been on simulating the effects of land-use changes on the water and sediment yields from the catchments. The results suggest that the water and sediment yields have changed significantly as a result of shifts in land use – more than previously assumed.

This issue needs to be further investigated and incorporated into management plans for the sustainability of the St Lucia estuarine system. Future climate changes may also play a significant role in shifting the water and sediment yields of the St Lucia catchment system. This issue is discussed in Chapter 21.

Catchment hydrology

Acknowledgements Aspects of the research reported here were funded by grants from the National Research Foundation (NRF, Pretoria), Marine and Coastal Management (MCM, Cape Town), the World Wide Fund for Nature (WWF-SA, Stellenbosch) and the South Africa–Netherlands

Research Programme on Alternatives in Development (SANPAD). We are also grateful to the management and staff of the iSimangaliso Wetland Park and Ezemvelo KZN Wildlife for providing administrative, logistical and operational support during the study.

93

Chapter contents 6.1 Introduction 6.2 Connectivity of coastal plain wetlands and Lake St Lucia 6.3 Diversity and hydrogeomorphology of freshwater wetlands on the coastal plain 6.4 Floodplains of the coastal plain and their relationship with Lake St Lucia 6.5 Valley-bottom wetlands: blocked valley lakes and wetlands 6.6 Depression wetlands on the coastal plain 6.7 Connectivity revisited: a framework for examining artificial impacts to wetlands 6.8 Future landscape development 6.9 Conclusion

Freshwater wetland (left) and the Lake (right) showing the intimate juxtaposition of the freshwater wetlands of the Eastern Shores and the estuary. (Photo: Ricky H. Taylor, March, 2012.)

6

The wetlands William N. Ellery, Suzanne E. Grenfell, Michael C. Grenfell, Marc S. Humphries and Kirsten B. Barnes

6.1 Introduction The Maputaland coastal plain is host to extensive wetlands that are remarkable in their diversity and importance internationally, being home to five of South Africa’s 16 Wetlands of International Importance (McCarthy and Hancox, 2000). Kosi Bay and Lake St Lucia are the two largest estuarine wetland systems in southern Africa, and the Mfabeni Mire is one of the oldest (45 000 years) and best known mires in the world (Smuts, 1992). The Maputaland coastal plain comprises an important component of the Natal Mire Complex that extends from Richards Bay to Maputo and hosts two of the country’s largest floodplain wetland complexes: the

Mkhuze and the Mfolozi. This chapter reviews the research on wetlands on the coastal plain that are related hydrologically or geomorphologically to Lake St Lucia and therefore influence its structure and function. As a large dune-bounded estuary, Lake St Lucia is dependent on the inflow of water from marine, fluvial and groundwater sources. This chapter focuses on fluvial connectivity of streams that flow into the lake, and considers these as regulators of water and sediment inputs. Understanding connectivity helps appreciate how human interventions in catchments influence lake structure and function.

6.2 Connectivity of coastal plain wetlands and Lake St Lucia Connectivity is a major research theme in ecology, hydrology and geomorphology. The term connectivity in ecology refers to the degree to which the landscape facilitates or impedes movement of organisms among resource patches (Taylor et al., 1993). Within a hydrological and geomorphological context, connectivity considers the transfer of energy and matter between two landscape compartments or within a system as a whole (Chorley and Kennedy, 1971). Therefore, the term connectivity has different meanings in different disciplines. In this chapter connectivity is used in a geomorphological sense as

this constitutes a more appropriate way of examining the relationship of wetlands on the coastal plain to Lake St Lucia. Two aspects of connectivity are important for the freshwater streams entering Lake St Lucia: (a) connectivity of water flow, and (b) connectivity of sediment transport. Both these types of connectivity depend to the greatest extent on the continuity of streamflow within a catchment, so that those streams with a continuous channel connection (spatially and temporally) with Lake St Lucia have the greatest potential to supply both water and sediment.

Ecology and Conservation of Estuarine Ecosystems: Lake St Lucia as a Global Model, eds. Renzo Perissinotto, Derek D. Stretch and Ricky H. Taylor. Published by Cambridge University Press. © Cambridge University Press 2013.

96

William N. Ellery et al.

In contrast, those streams that do not connect directly with Lake St Lucia supply a disproportionately low volume of water and/or sediment. Given that Lake St Lucia is a large estuary, the contribution of surface water inputs is vital for the functioning of this ecosystem. However, sediment input to Lake St Lucia from continental catchments is viewed as a threat to Lake St Lucia through its effects on lake depth, lake size, turbidity and connectivity with the ocean. Herein lies one of the conundrums facing lake managers; is it possible to reduce river sediment supply without reducing river water supply? In respect of sediment generation, transfer and deposition, it is important to distinguish clastic and dissolved sediment because they behave in different ways within the fluvial system. Clastic sediment refers to the minerogenic material transported by streams, including clay, silt, sand, gravel, cobbles and pebbles that are moved downstream by fluvial systems. Such material may be moved in suspension, contributing to the muddy character of streams (suspended sediment), or it may be moved as bedload by being rolled or bounced along the stream bed (bedload sediment). Dissolved sediment is transported in solution and it may precipitate from solution as a consequence of a range of biogeochemical processes, giving rise to chemical sedimentation. Deposition of either clastic or dissolved sediment causes aggradation, which is the modification of the earth’s surface by deposition to increase uniformity of grade. Without wishing to oversimplify clastic sediment flux in streams, the sediment load that a stream can carry is primarily related to stream velocity and discharge. Provided sediment is available a stream

will carry sediment in proportion to its velocity and discharge. The amount of sediment a stream can carry based on these factors is referred to as its capacity (Ellery et al., 2009). However, the amount of sediment being transported by a stream is referred to as its load. When the capacity is greater than the load, erosion will occur, but when the load is greater than the capacity, deposition takes place. The deposition of clastic sediment is therefore usually a response to local scale variation in hydraulic characteristics that influence velocity and/or discharge, typically taking place within or immediately adjacent to the stream (Ellery et al., 2009). In contrast, the deposition of dissolved sediment takes place in backswamp environments. The prolonged residence of water in backswamp areas allows extensive water loss to the atmosphere by evapotranspiration, resulting in the precipitation of dissolved sediments in the soil, creating local relief. Thus, clastic and chemical sedimentation are spatially separated such that they interact to promote uniformity of grade within wetland environments (Ellery et al., 2009). Organic sedimentation is a third type of sedimentation that takes place in wetland environments. The accumulation of peat is not particularly well understood and is generally attributed to low temperatures, low pH, the presence of permanent flooding, the absence of clastic sediment input and an elevated base level. However, Ellery et al. (2012) argue that all that is needed for peat accumulation is an elevated base level such that permanent flooding is inevitable, combined with low inputs of clastic sediment.

6.3 Diversity and hydrogeomorphology of freshwater wetlands on the coastal plain The coastal plain of Maputaland contains a large number and wide variety of wetlands (McCarthy and Hancox, 2000), including depressions, unchannelled and channelled valley-bottom systems, floodplains

and lacustrine wetlands (for a preliminary classification of South African wetland hydrogeomorphic types see SANBI, 2009). Depression wetlands tend to be small (< 100 ha) with

The wetlands

catchments on the coastal plain and with substrates that are either sandy or organic (Goge, 2003). Channelled and unchannelled valley bottom wetlands tend to be much larger (hundreds to approximately 10 000 hectares), once again with catchments on the coastal plain and with sandy or (typically) organic sediments (Schoultz, 2001; Goge, 2003; Ellery et al., 2012). Given that depression and valley-bottom wetlands have catchments on the coastal plain underlain by reworked marine sediments, there is very little weathering to produce suspended clastic or dissolved sediment, such that these wetlands are characterized by very low clastic and dissolved sediment loads. In contrast, floodplain wetlands tend to be large (thousands to tens of thousands of hectares) with catchments that typically extend west of the coastal plain, and therefore have high clastic and dissolved sediment inputs produced from weathering bedrock in their catchments (Stormanns, 1987; Neal, 2001;

Barnes et al., 2002). Lacustrine wetlands occur on the fringe of lakes where water depth is relatively constant and sufficiently shallow for the establishment of emergent macrophytes. The distribution of the main wetlands associated with Lake St Lucia is indicated in Figure 6.1. Strictly, bodies of open water at the foot of tributary valleys of the Mkhuze floodplain that are colloquially referred to as ‘pans’, such as Muzi Pan and Yengweni Pan, are not depression wetlands but instead are blocked valley lakes that occur at the interface of aggrading floodplains. Their formation is described in Section 6.5. The following sections consider the structure and function of floodplains, valley-bottom and depression wetlands and their relationship with Lake St Lucia. A fourth wetland type does exist along the shoreline of Lake St Lucia, but this is simply a response to variation in lake water level and is not considered further in this chapter.

6.4 Floodplains of the coastal plain and their relationship with Lake St Lucia 6.4.1 Fluvial styles Four rivers enter the main water body of Lake St Lucia: the Mkhuze, Mzinene, Hluhluwe and Nyalazi rivers (Figure 6.2). The Mfolozi River intermittently shares its mouth with that of Lake St Lucia, and its connectivity with the lake is therefore somewhat different from that of the other four rivers. The Mzinene, Hluhluwe and Nyalazi rivers arise on, or a short distance inland of, the coastal plain and therefore have small catchments with limited topographic relief. These streams are characterized by highly variable flow and narrow downstream to terminate in lacustrine deltaic wetlands. The Mkhuze and Mfolozi rivers arise within the semi-arid northern KwaZulu-Natal Drakensberg Foothills, with the Mkhuze entering Lake St Lucia from the north and the Mfolozi from the south. A downstream reduction in channel dimensions and capacity is

characteristic of the Mkhuze River, where flow variability and transmission losses due to overbank flooding and losses from the stream to groundwater limit channel continuity (Humphries et al., 2010; Ellery et al., 2012). Only the Mfolozi River, with about double the catchment area of the Mkhuze, maintains a continuous channel through the length of its coastal floodplain (Grenfell et al., 2009). The area of the Mfolozi catchment is approximately 11 300 km2, while that of the Mkhuze is approximately 6080 km2 (Begg, 1978). Furthermore, only the Mfolozi floodplain has scroll bars (Grenfell et al., 2009), which are indicative of active meander migration and reflect the ability of the fluvial system to distribute sediment internally in features known as point bars. Towards the floodplain–lake confluence, the Mkhuze River narrows and ultimately terminates in

97

98

William N. Ellery et al.

FIGURE 6.1 Pan-sharpened Landsat-7 image showing the main wetland systems that interact with Lake St Lucia.

The wetlands

FIGURE 6.2 Wetlands in the immediate vicinity of Lake St Lucia. (Adapted from Wright et al., 2000.)

99

100

William N. Ellery et al.

lacustrine deltaic wetlands. This morphology is symptomatic of a disequilibrium flow-sediment feed (Kleinhans and van den Berg, 2011) that is manifest in a river environment characterized by high rates of channel bed aggradation relative to rates of lateral channel migration. This results in the channel occupying a sandy alluvial ridge that is elevated above the surrounding floodplain. The height of the alluvial ridge above the surrounding floodplain is related to the length of time that the channel has occupied its current position. River avulsion is an expected outcome of such conditions (Jerolmack and Mohrig, 2007), and is a naturally recurrent process within the Mkhuze floodplain system (Ellery et al., 2003a). These processes are described in more detail in Section 6.4.6.

6.4.2 Trunk stream clastic sedimentation It seems that floodplains and their associated wetlands on the coastal plain have formed over the last 20 000 years since the Last Glacial Maximum, during which time the valleys in which they occur must have undergone incision due to a lowering of sea level by about 120 m below present sea level (Ramsay, 1995; Ramsay and Cooper, 2002). Although no direct evidence for the onset of sedimentation in these systems exists, it is likely that it started about 6500 BP when sea level reached its present position (McCarthy and Hancox, 2000). Floodplains of the Maputaland coastal plain are dominated by clastic sediment (primarily sand and silt) on the channel belt and overbank reaches. An annual accretion rate of between 0.25 and 0.50 cm yr 1 has been calculated by Humphries et al. (2010) for the current Mkhuze River alluvial ridge using 210Pb (useful for the last 150 years), which is consistent with long-term aggradation rates estimated from valley fill sequences (Ellery et al., 2012). Sedimentation along the Mkhuze and Mfolozi trunk streams blocks flow along incoming tributary valleys, leading to sedimentation in tributary streams that will be described in Section 6.5.

6.4.3 Accumulation of dissolved sediment Although seasonal flooding is dominated by clastic sedimentation, residence of water in floodplain depressions may be important for solute precipitation (chemical sedimentation). Fluvial inflow into the Mkhuze floodplain with electrical conductivities between 300 and 1300 μS cm 1 (Stormanns, 1987) represents a source of naturally occurring solutes such as Ca, Mg, Si, K, Na and Cl (Barnes et al., 2002). Solutes accumulate on the Mkhuze floodplain as a consequence of evapotranspirational water loss from backswamp areas (Barnes et al., 2002; Humphries et al., 2010). Within zones of groundwater recharge within the Mkhuze floodplain, solutes entering the wetland become increasingly concentrated in the groundwater due to the influence of transpiration by trees, such as Acacia xanthophloea, which appears to be particularly effective in selectively removing water and causing the progressive concentration of groundwater solutes (Humphries et al., 2011a). This results in the development of saline groundwater, which is associated with the saturation, precipitation and accumulation of minerals such as CaCO3 and SiO2 (Humphries et al., 2011b). Within the Mkhuze floodplain, these processes occur mainly along the margins of the upper blocked tributary valley lakes (especially Muzi Pan) as well as on the lower floodplain upstream of its confluence with the Mbazwana Swamp. Although our understanding of such processes is limited to a few case studies in the Maputaland region, the importance of chemical precipitation and solute retention within wetland sediments has been shown to be significant in comparable studies in the Okavango Delta, Botswana, where chemical accumulation is seen as a major driver of ecosystem structure and function (McCarthy and Ellery, 1998). Mineral precipitation causes volume increase and expansion in the soil and plays a role in modifying landscape topography, while the development of highly saline groundwater results in marked vegetation zonation (McCarthy and Metcalfe, 1990; Ellery et al., 1993; McCarthy et al., 1993). Although

The wetlands

the Mkhuze system is dominated by clastic inputs, such evapotranspiration-driven processes are likely to be widespread in floodplain wetlands on the coastal plain of KwaZulu-Natal, promoting uniformity of grade of such floodplain environments. These floodplains thus represent dynamic systems that continue to aggrade as a result of the interplay between clastic and chemical sedimentation.

6.4.4 Floodplain ecosystem structure Floodplain ecosystem structure is determined largely by hydrological and geomorphological processes, such as levees and alluvial ridge development along the trunk stream as a result of overbank flooding during flood events. The elevated alluvial ridges flanking the trunk stream are dominated by coarse sediment and are flooded infrequently, which creates an environment suitable for the establishment of riparian forest, which is typically dominated by Ficus sycomorus (Patrick and Ellery, 2007). Bearing in mind that overbank flooding and clastic sediment deposition is concentrated in upper floodplain reaches, riparian forest is naturally concentrated on the channel margin in upper floodplain zones. Fine sediment not deposited in close proximity of the channel finds its way onto the upper floodplain surface. Areas of active deposition of silt are dominated by the strongly rhizomatous, tall, coarse grass, Echinochloa pyramidalis, in largely monospecific stands (Patrick and Ellery, 2007). Distal reaches of floodplains, where deposition of clay is widespread, are dominated by Phragmites mauritianus, while areas that receive water but little clastic sediment, where chemical sedimentation is most widespread, are dominated by Cynodon dactylon (Neal, 2001). Acacia xanthophloea is widespread and dominant in areas of the floodplain margin characterized by groundwater recharge, where groundwater solute loads are high due to evapotranspirational water loss (Humphries et al., 2011a). The only tree species able to tolerate extremely high groundwater solute concentrations is the tamboti Spirostachys africana.

6.4.5 Floodplain ecosystem functions – flood attenuation, sediment trapping The functions provided by floodplains are related to their hydrogeomorphic setting, which accounts for the ways that water enters, flows through and leaves a wetland (Kotze et al., 2008). Floodplains behave as streams during low flows, although there may be transmission losses of water and solutes downstream within the floodplain during such periods. However, during flood events, floodwaters are dispersed over a large area such that flow is much slower than in the stream itself, and floodwaters fill depressions and recharge regional groundwater, such that flood attenuation is inevitable. At the same time, deposition of clastic sediment takes place, particularly in proximal reaches, such that sediment trapping is also an important function. As mentioned earlier, chemical sedimentation associated with substantial water loss to the atmosphere by evapotranspiration is also an important function performed by floodplains.

6.4.6 Human impacts on floodplains and their effect on connectivity Due to the presence of clay and silt soils with a high exchange capacity and therefore fertility, floodplains serve as the most suitable sites on the otherwise dystrophic coastal plain for commercial and subsistence agriculture. Furthermore, their hydrological and sedimentological connectivity to Lake St Lucia has meant that large floodplains have been modified by humans to ensure that, under stress, the functioning of Lake St Lucia has been maintained. For these reasons the Mfolozi and Mkhuze floodplains have been subjected to particularly high levels of transformation and impacts are reported here in some detail.

Human activities on the Mfolozi floodplain Human activities on the Mfolozi floodplain have been studied and described by Grenfell et al. (2009). The earliest aerial photography of the Mfolozi

101

102

William N. Ellery et al.

• increase connectivity of the backswamp with downstream fluvial systems in order to drain the backswamp and make it suitable for farming.

FIGURE 6.3 Major features of the Mfolozi floodplain.

floodplain was taken in 1937, when sugar cane cultivation was restricted to the upper part of the floodplain, occupying about 30% of the floodplain. The remainder of the floodplain had indigenous vegetation. A number of artificial drains had also been excavated in the area under cultivation at that time. Lake Futululu was the largest floodplain margin lake on the north of the floodplain, while Lake Eteza occupied the Msunduzi drainage line entering the Mfolozi floodplain from the south-west (Figure 6.3). Commercial sugar production expanded rapidly on the floodplain such that by 1960 more than 50% of the floodplain was under cultivation. Expansion of agriculture on the floodplain was accompanied by modification of streams to increase hydraulic efficiency along the trunk stream and by excavation of an increased number of artificial drains on the floodplain itself. Flow along the Mfolozi River was accelerated by straightening and thus shortening the Mfolozi River course south of Lake Futululu (removal of the ‘Uloa Loop’ by straightening a number of meander bends; Figure 6.3). This trend of increased expansion of agriculture and associated hydraulic modification continued eastwards until 1970. The intentions of these activities were to: • decrease connectivity of the trunk stream with the floodplain and therefore increase longitudinal connectivity of the trunk stream with the ocean;

Two of the largest floods on record were experienced in January 1984 (Cyclone Domoina), and in September 1987 as a result of a cut-off low pressure system in the region. Discharge in the Mfolozi River during Cyclone Domoina peaked at approximately 16 000 m3 s 1, which was estimated to be three times the 100-year return flood (Travers, 2006). Sediment deposited during Cyclone Domoina produced a depositional feature about 10 km long and 3 km wide on the upper floodplain south of the Mfolozi River and north of the Msunduzi River to the west of the Uloa Loop. The cyclone led to the avulsion of the Mfolozi River towards the south (van Heerden, 1984). However, the Mfolozi River was artificially returned to its former course after Cyclone Domoina. There are no large impoundments on the Mfolozi River upstream of the floodplain but approximately 50% of the wetlands in the catchment have been transformed (Begg, 1988), potentially leading to substantial impacts on the hydrological and sedimentological characteristics of the Mfolozi River. However, such impacts have not been studied or documented, and are poorly understood.

Management of the Mfolozi for the benefit of Lake St Lucia During periodic droughts in the region, Lake St Lucia becomes hypersaline (Taylor et al., 2006a) and nature conservation authorities look for ways to introduce fresh water into the system. One such measure involved the excavation of a canal (the Mfolozi ‘Link Canal’) between the Mfolozi River and Lake St Lucia in the late 1970s (Collings, 2009). The design of the canal was such that sediment flux between the river and the lake would be small, but the intervention was unsuccessful due to a combination of its small size and low gradient (Grenfell et al., 2009). However, when Lake St Lucia is not hypersaline, authorities separate the mouths of the Mfolozi River and the lake in order to limit sediment influx into the latter.

The wetlands

Such interventions are most effective if sediment dynamics of fluvial systems are well understood. As mentioned earlier, the primary mechanism by which sediment is exchanged between the channel and floodplain in meandering rivers, such as in the Mfolozi floodplain, is through the process of lateral meander migration (Lauer and Parker, 2008). During lateral migration, sediment is eroded at the outer bank of meander bends, while sediment is deposited on point bars at the inner bank. In meandering river floodplains that are not undergoing long-term aggradation or incision, channel bank erosion during meander migration reintroduces sediment into the channel, but the added sediment is typically balanced by sequestration in point bars, and overbank deposition on floodplains and in off-river water bodies such as oxbow lakes (Lauer and Parker, 2008). However, the Mfolozi floodplain has been substantially transformed and its ability to sequester sediment in point bar accretion has been compromised. Rehabilitation of the Mfolozi floodplain such as proposed by Collings (2009) would need to address restoration of the meandering fluvial system if sediment flux into Lake St Lucia is to be limited in the desired way. Furthermore, any sustainable long-term solution to reducing sediment addition to the lake in a Mfolozi–Lake St Lucia jointmouth situation must consider the catchment of the Mfolozi River. This is a disheartening proposition, given the scale of disturbance in the Mfolozi catchment through (inter alia) careless dirt-road construction and drainage, gully erosion, and land clearing for agriculture.

Human activities on the Mkhuze floodplain Human activities in the Mkhuze floodplain have also been quite extensive and had some unexpected outcomes. Agriculture on the floodplain has increased exponentially from none in 1937 to 1385 ha (approximately 30% of the wetland area) in 1996 (Neal, 2001). Unlike the Mfolozi floodplain where cultivation started at the head of the floodplain due to the low-lying nature of the lower floodplain and proximity to the sea, which made drainage difficult,

cultivation in the Mkhuze floodplain started in the mid to lower floodplain in close proximity to human settlements along the western fringe of the wetland. Cultivation was also concentrated on the high-lying alluvial ridge and involved gradual clearing of riparian forest. Today, much of the floodplain is under cultivation, although mostly in the form of subsistence agriculture with small fields that produce a wide range of crops for local consumption. The most notable human impacts on the Mkhuze floodplain have been associated with diversion of water, initially to increase fluvial connectivity of the Mkhuze River and Lake St Lucia to increase the supply of fresh water to the lake during a hypersaline phase in the late 1960s and early 1970s (Ellery et al., 2003a). Under these conditions evaporation dominates the water balance of the lake. In the absence of freshwater inputs the lake then becomes hypersaline and toxic to freshwater life (Taylor et al., 2006a). In the early 1970s the Natal Parks Board in association with the St Lucia Scientific Advisory Committee made the decision to excavate a canal to shorten the route of freshwater inflow into Lake St Lucia by circumventing the north-eastern portion of the Mkhuze floodplain. This artificial canal, known as the Mpempe Canal, was excavated over a distance of about 15 km from the Mkhuze River to the Mpempe Pan, to re-enter the Mkhuze River north of the Demezane Pan (Figure 6.4). From Demezane Pan the bed of the Mkhuze River was cleared artificially for a few additional kilometres. The scale of the project was large in that heavy machinery was used for excavation and dynamite was used to blast logjams from the river course. The view was expressed that the canal was poorly designed as it was not of a suitable size or grade, with negative to extremely negative effects (Taylor, 1986, 1993; Goodman, 1987; Stormanns, 1987; Watkeys et al., 1993). Despite this attempt to artificially increase fluvial connectivity of the Mkhuze River and Lake St Lucia, the construction of the canal did little to reduce salinity in Lake St Lucia (Ellery et al., 2003a).

103

104

William N. Ellery et al.

About 10 km upstream of the point where the Mpempe Canal diverted water from the Mkhuze River, a commercial farmer cleared a large tract of riparian forest and in 1986 manually excavated a small canal, the Tshanetshe Canal, by deepening an existing hippo trail through the Mkhuze River levee to Tshanetshe Pan to the south (Figure 6.4). This minor canal was subject to rapid erosion through the levee and downstream of Tshanetshe Pan, thus connecting the Mkhuze River north of Tshanetshe Pan to the Mpempe Canal (Figure 6.4). This erosion led to channel avulsion and altered the course of the Mkhuze River such that the section of the original river course between Tshanetshe Pan and the head of the Mpempe Canal receives flow only during unusually high flows (Ellery et al., 2003a). Subsequent to the 1986 excavation much of the recent and on-going rehabilitation work in the Mkhuze floodplain has focused on attempting to redivert the Mkhuze River to a pre-avulsion course, and recent attempts to block the new channel course with large soilcrete structures have failed. However, Ellery et al. (2003a) demonstrated with a Differential GPS (DGPS) data set that the Mkhuze River channel was ‘primed’ for avulsion prior to excavation of the

Tshanetshe Canal, which precipitated channel abandonment; in this case the lateral slope from channel to floodplain in the vicinity of the breach was over five times the longitudinal channel slope. They therefore argued that the diversion of the Mkhuze River into the Tshanetshe Canal would most likely have happened naturally at some stage. Rather than attempting to maintain the Mkhuze floodplain in its pre-avulsion state it is argued that natural processes such as avulsions need to be encouraged in order to maintain spatial heterogeneity and allow renewal of the system in a way that is consistent with the natural dynamic (Ellery et al., 2003a; see also Ellery and McCarthy, 1994, 1996). Furthermore, such rehabilitation cannot demonstrably improve the ecology of the floodplain, or contribute to resolving the hydrological stresses facing Lake St Lucia. In addition, the lacustrine deltaic wetland in which the Mkhuze River terminates would constitute an effective fluvial disconnect and substantial sediment trap, and this river is not expected to contribute significantly to lake sediment supply. In summary, acting to ‘repair’ avulsions carries high financial risk and offers low ecological reward.

6.5 Valley-bottom wetlands: blocked valley lakes and wetlands Channelled and unchannelled wetlands form where former valleys are blocked by localized aggradation at the toe of the valley. Aggradation may be the result of deposition of clastic sediment along floodplains or by the formation of beach ridges (cheniers) along the edge of Lake St Lucia.

6.5.1 Geomorphology and hydrology The floodplains surrounding Lake St Lucia abound in blocked valley lakes and wetlands (Grenfell et al., 2010; Ellery et al., 2012), where tributary valleys abutting trunk channel floodplains are permanently flooded. These blocked valley lakes occur along both

the Mfolozi and Mkhuze floodplains (Grenfell et al., 2010; Ellery et al., 2012). For the purposes of this chapter, only the Mkhuze floodplain tributaries will be described. As aggradation took place along the Mkhuze River trunk stream, sediment transfer from the local catchments along the tributary valleys was progressively impeded and organic sedimentation became the dominant form of sedimentation in the tributary valleys, such as is taking place in the Totweni drainage line at present (Figure 6.4). Continued trunk stream aggradation eventually increased trunk stream elevation above the tributary valley, at which time a lake formed at the toe of the tributary valley, such as Mpanza and Mdlanzi

The wetlands

FIGURE 6.4 Major features of the Mkhuze wetland system.

lakes at the toe of the Totweni drainage line. As trunk aggradation continued, clastic sediment progressively entered the tributary valley such that organic sediments were eventually buried, such as for the Yengweni and Muzi Pans (Ellery et al., 2012). The

progressive blocking of tributary streams by the trunk stream caused tributary valley lakes to lengthen over time and the southern margin of lakes to migrate progressively up the tributary valley, which has been the case for all of the lakes in

105

106

William N. Ellery et al.

tributary valleys entering the Mkhuze floodplain from the north (Muzi, Yengweni, Mpanza, Mdlanzi; Ellery et al., 2012). At present, the Mkhuze floodplain is prograding into the Mbazwana Swamp, and will eventually block the valley, at which time a lake will form north of the floodplain. The sequence of tributary blocking, peat accumulation and input of clastic sediment into tributary valleys is clearly evident in the Mkhuze floodplain as a space-for-time substitution from west (oldest blocked tributary valley) to east (most recently blocked tributary valley) along the eastward-oriented floodplain system. As described above, these wetlands form when tributary valleys become impounded by floodplain aggradation of a river with a larger sediment load and catchment than the tributary stream. In these floodplains, the elevation of the trunk stream in the upper floodplain is greater than the adjacent tributary valley lake such that during a flood, water and clastic sediment readily enter the tributary from the trunk stream. Tributary valley lakes that are at a lower elevation than the trunk stream (such as Muzi and Yengweni Pans) receive the bulk of their water supply from the Mkhuze River during floods and are characterized by groundwater recharge (Barnes, 2009). However, the elevation of the trunk stream relative to the tributary valley becomes progressively lower downstream such that during trunk stream flood events, floodwaters do not flow into tributary valley lakes on the lower floodplain. Therefore, the tributary valleys that abut the lower floodplain are sustained by groundwater inputs from the coastal plain (McCarthy and Hancox, 2000), and are characterized by groundwater discharge. South of the confluence of the Mkhuze floodplain and Mbazwana Swamp, Lake St Lucia forms the base level to which streams and wetlands have adjusted their grade. It is likely that north-west to south-east trending beach ridges along the northern margin of the lake have influenced wetland development along the Mbazwana, by isolating the Mbazwana Stream from marine water inputs and therefore allowing freshwater wetlands to develop. A second blocked

valley wetland that results from separation of freshwater and marine water inputs is the Mfabeni Mire on the Eastern Shores of Lake St Lucia. Beach ridges with a similar orientation to those of the northern lake have formed along the northern shore of Catalina Bay, leading to separation of the Mfabeni Mire and Lake Bhangazi South from Lake St Lucia. Once peat formation was initiated in the Mbazwana and Mfabeni valleys, the flow of fresh water into the lake was impeded due to reduced hydraulic efficiency, which raised the regional water table and allowed further peat formation.

6.5.2 Ecosystem structure In describing ecosystem structure of blocked valleys it is important to distinguish between those that are characterized by: • groundwater recharge and abut floodplains near the head of the floodplain such as Muzi and Yengweni Pans on the Mkhuze floodplain, which are dominated by fine clastic sediment inputs; • groundwater discharge and abut floodplains near the toe of the floodplain such as Futululu (Mfolozi floodplain) and Mpanza and Mdlanzi (Mkhuze floodplain) lakes, which have limited input of fine clastic sediment near the toe of the blocked valley; and • groundwater discharge that arise on the coastal plain such as the Mbazwana Swamp upstream of the Mkhuze confluence, and the Mfabeni Mire on the Eastern Shores, which have no fine clastic sediment inputs at all due to an absence of fine sediment in their catchments.

Groundwater recharge valley-bottom wetlands Blocked tributary valley wetlands that are associated with floodplains and are characterized by groundwater recharge receive sediment from the adjacent floodplain during flood events. The grass Echinochloa pyramidalis is dominant on the floodplain as far as the lake margin, giving way to

The wetlands

open water. Groundwater recharge around the lake, particularly in the vicinity of the floodplain margin, becomes a dominant process. Acacia xanthophloea and Spirostachys africana are dominant here.

Groundwater discharge valley-bottom wetlands with limited input of fine clastic sediment The lower tributary valley in the vicinity of the floodplain typically contains a lake and is dominated by Cyperus papyrus floating marsh (Figure 6.5), a vegetation type restricted to permanently inundated settings where there is limited input of clastic sediment during flood events. Papyrus is a well-known peat-forming species (Ellery et al., 2003b). The margin of the basin upstream of the lake along the tributary valley, at the transition zone of peat to dune sand, is characterized by stands of Phragmites australis marsh, which suggests that the head of the basin is semi-permanently flooded. The valley margin is dominated by broadleaved evergreen trees typical of terrestrial settings. However, as floodplain sedimentation continues and slopes along the valley decrease upstream, the P. australis margin is increasingly flooded, and the entire wetland system migrates upstream, until eventually the valley is drowned in clastic sediment originating from overbank flooding from the trunk stream. At this point in time, C. papyrus disappears and Echinochloa pyramidalis becomes dominant on the edge of the lake that abuts the floodplain. Gradually, the broadleaved evergreen fringe disappears to be replaced by Acacia xanthophloea and Spirostachys africana.

Groundwater discharge valley-bottom wetlands that arise on the coastal plain The Mbazwana and Mfabeni mires are good examples of this wetland type, with substantial peat deposits and a combination of herbaceous vegetation in the eastern parts of these wetlands and swamp forest typically along the western margin (Schoultz, 2001; Venter, 2003). The herbaceous swamps are generally dominated by short sedges and other hydrophytes,

including Sphagnum (Venter, 2003), while the swamp forest communities are diverse but include Barringtonia racemosa, Syzygium cordatum and Ficus trichopoda dominated vegetation types (Venter, 2003). The distribution of the herbaceous and swamp forest vegetation types within the mires is poorly understood, although Schoultz (2001) suggests that it is related to interactions between topography and wind. Warm berg winds dry out herbaceous vegetation and render it highly combustible. Fire-protected sites will enable establishment of swamp forest. Berg winds typically blow from the west to the north-west and swamp forest typically occurs below steep east to south-east oriented slopes that are sheltered from such berg winds.

6.5.3 Ecosystem function In respect of the ecosystem services provided by these wetlands, they serve a very important role in surface water–groundwater interactions, occurring in zones of groundwater discharge and recharge. The Mfabeni Mire (and possibly the Mbazwana Swamp) has lower hydraulic conductivity than the sands of the coastal plain. As such, eastward draining groundwater that is intercepted by these southwardoriented mires is impeded, raising groundwater elevation that sustains inundation of the mire and therefore promotes peat formation. Currently, much of the Mfabeni Mire is elevated about 10 m above sea level by this set of processes. The role of Mfabeni Mire in steadily providing fresh water to Lake St Lucia, especially during periods when the mouth of the lake is blocked and the lake becomes hypersaline, has been well recognized (Taylor et al., 2006b). Although these wetlands are typically very effective in respect of both nutrient and detrimental solute trapping, in the region of Lake St Lucia they do not receive high solute loads and are therefore not particularly important in this respect, perhaps with the exception of those systems that abut the Mfolozi floodplain where sugar cane in their catchments is prevalent.

107

Dominant vegetation type

William N. Ellery et al.

sand Phragmites forest mauritianus marsh

Cyperus papyrus floating marsh

Echinochloa pyramidalis seasonal marsh

Riparian forest

River channel

Dominant sediment type

Blocked-valley lake or wetland

Palaeo-valley floor (coastal dune sands)

Organic sediment

Clastic floodplain sediment (sand & silt)

(peat)

FIGURE 6.5 Longitudinal section through a blocked-valley lake and wetland on the Maputaland coastal plain showing typical sediment and vegetation distributions. (Adapted from Grenfell et al., 2010.)

estimated bedrock valley floor - exact shape unknown

Aerial views looking downstream along the Futululu drainage line (photographs taken 15 August 2007)

108

A particularly important function of these wetlands relates to carbon sequestration since peat accumulation rates are high. Ellery et al. (2012) estimate that the active peatlands in the Totweni drainage system accumulate 1.3 tonnes of organic material per hectare per annum. Provided that the ecosystems in which peat accumulation is taking place are not damaged, these are important sites of carbon sequestration given their aerial extent. Furthermore, trunk aggradation and the eventual deposition of clastic sediment in tributary valleys, which buries peat for long periods of time, makes these systems particularly important long-term carbon sinks. These systems are likely to release carbon into the atmosphere during periods of cooling such as ice ages, when sea level drops and systems such as these degrade, thereby playing a role in global temperature regulation. Although there is no direct influence of carbon sequestration per se on the

Mfolozi River commercial sugar cane cultivation Riparian forest Echinochloa pyramidalis seasonal marsh

Cyperus papyrus floating marsh Lake Futululu

Lake Futululu Cyperus papyrus floating marsh Phragmites mauritianus marsh

functioning of Lake St Lucia, these processes highlight the global long-term importance of these wetlands in biogeochemical cycles.

6.5.4 Human impacts In general, valley-bottom wetlands that are related to Lake St Lucia, both those linked with floodplains as well as those that arise entirely on the coastal plain, have not been impacted excessively by human activity. There is a limited amount of agriculture in the head of the Mbazwana system in close proximity to the town of Mbazwana, and this is increasing over time. There is also harvesting of fibre and timber in the Mbazwana. The valley-bottom wetlands that abut the Mkhuze floodplain have very few impacts apart from harvesting of natural fibre for craft and construction purposes. The valley-bottom wetlands that abut the Mfolozi floodplain are subject

The wetlands

to invasion by alien plants, and the catchments adjacent to these wetlands are used for commercial timber and sugar production. These activities are

likely to affect water inputs to the wetlands, and also lead to an increase in solute loads – especially of plant nutrients from use of commercial fertilizers.

6.6 Depression wetlands on the coastal plain Many small depression wetlands exist on the coastal plain that reflect the coincidence of the land surface with the groundwater rest level (Watkeys et al., 1993; McCarthy and Hancox, 2000). The land surface is spatially heterogeneous due to small variations in elevation (metres to tens of metres) that reflect the interaction of hydrology and topography (Botha et al., 2003; Vrdoljak and Hart, 2007). As these wetlands are groundwater driven, wetness reflects the recharge state of the large unconfined aquifer that covers the area (Kelbe and Rawlins, 1992). Given rainfall variability that characterizes the area, these wetlands are dynamic and many are wet for

extremely variable periods, such that organisms that occupy them need to be adapted to tolerate variability (Vrdoljak and Hart, 2007). As a result of their dynamic nature in time, the concentration of dissolved solutes is always low ( +1.0 m

8% (9%)

9% (10%)

(a) Mouth state statistics

(b) Salinity statistics

(c) Water level statistics

Water levels eventually increase until the frontal berm is rebreached to restore a connection to the sea. The summary in Table 7.2 shows that on average the Mfolozi contributes about 20% of the total freshwater supply of the lake (~ 150 Mm3 yr
1) during the intermittent mouth closures. The mouth remains closed for 2–3 years at a time (Table 7.3) – a three-year return period flood usually suffices to cause breaching of the berm under these conditions.

7.2.4 Overview of historical human impacts Anthropogenic activities have reduced the inflows to the lake by up to 20% or 60 Mm3 per year (refer to Chapter 5 for details). Simulation results using ‘naturalized’ flows are shown in Figure 7.4 and illustrate the impact of these inflow changes. Comparing these results with those shown in Figure 7.3 it is evident that the higher

Estuary and lake hydrodynamics

FIGURE 7.4 Simulated monthly average lake salinities and mouth states using ‘natural’ flows (prior to any losses due to anthropogenic activities) for (a) scenario 1, (b) scenario 2 and (c) scenario 3. These results should be compared with those shown in Figure 7.3a, b, c respectively.

Monthly average lake salinity

(a)

sea water

simulated

closed mouth state

120 100 80 60 40 20 0 Monthly average water levels (mEMWL) breaching water level

simulated

3 2 1 0 –1 –2 1910

1920

1930

1940

1950

1960

1970

1980

1990

2000

2010

1990

2000

2010

1990

2000

2010

Monthly average lake salinity

(b)

simulated

sea water

closed mouth state

120 100 80 60 40 20 0 Monthly average water levels (mEMWL) breaching water level

simulated

3 2 1 0 –1 –2 1910

1920

1930

1940

1950

1960

1970

1980

Monthly average lake salinity

(c)

sea water

simulated

closed mouth state

120 100 80 60 40 20 0 Monthly average water levels (mEMWL) breaching water level

simulated

3 2 1 –1 0 –2 1910

1920

1930

1940

1950

1960

1970

1980

inflows significantly reduce salinities during dry periods (see also Table 7.3). Other changes are slightly longer open mouth conditions and increased water levels in scenario 2 during droughts. While desiccation of the lake is a regular feature of scenario 2 simulations under the reduced inflow regime (Figure 7.3b), it is less likely to occur under ‘naturalized’ inflow conditions. However, it

is worth emphasizing that the model predicts that the mouth state in scenario 2 (i.e. with inlet separated from Mfolozi) remains predominantly closed even with the increased inflows that existed under natural conditions in the past. The simulations further elucidate how the artificial separation of the Mfolozi from St Lucia has had a

121

122

Derek D. Stretch et al.

major impact on the natural functioning of the system. In particular, scenario 3 shows that prior to 1952 the system mostly remained connected to the sea via a combined inlet and only closed for short periods during droughts. Salinities were highly variable but water levels mostly remained near EMWL. Hypersaline conditions occurred during dry conditions with the mouth open, but when the mouth closed salinities were diluted by Mfolozi inflows. In contrast, scenario 2 simulations suggest that the post-2002 management strategy, if extended over the long term, results in a predominantly closed system that is only connected to the sea for brief periods during wet conditions; that is, with high freshwater inflows. Salt loadings would be low because of the limited seawater influxes, but water levels would be very variable and can fall to low levels during droughts with desiccation of large portions of the lake. In terms of impacts on the water level/salinity regime of the system, the simulations clearly show that the change in mouth dynamics due to separation from the Mfolozi is much more significant than the reduced terrestrial inflows into the lake. The reason for the change in the mouth dynamics of St Lucia when separated from the Mfolozi is clear from the water balance components summarized in Table 7.2. Total terrestrial outflows through a combined inlet are on average about six times higher than those through a separate St Lucia inlet (about 900 Mm3 per year versus about 150 Mm3). Higher outflows maintain the inlet in an open state for longer by flushing sediments out of the mouth area. The anthropogenic impacts highlighted here are similar to those for other major systems worldwide (see e.g. McLusky and Elliott, 2004; Wolanski, 2007). Two well-known examples are the Coorong estuary in Australia (Webster, 2010) and the Chilika lagoon in India (Ghosh et al., 2006). The Coorong is managed using a strategy similar to scenario 1 – the mouth is artificially maintained in an open state by dredging because terrestrial flows have been substantially reduced by agricultural abstractions. Like St Lucia, the system now experiences hypersalinity during low flow conditions. In contrast, the Chilika lagoon is similar to St Lucia under

scenario 2 – inflow changes due to abstractions and impoundments, together with mouth sedimentation and closures, have led to reduced salinities and lower water levels. In both examples the ecological implications have been significant (Ghosh et al., 2006; Lester and Fairweather, 2009; Webster, 2010).

7.2.5 The effects of severe drought conditions Extensive desiccation of the lake has not been recorded prior to 2006. A question that may be asked is Is this a rare but natural event or is it attributable to anthropogenic impacts on the system? The scenario 2 simulations (Figure 7.3b) show that desiccation of the lake is a consequence of the management strategy of allowing the mouth to close while artificially maintaining the separation from the Mfolozi – indeed it is a typical response to severe drought conditions in this scenario. The simulations show that the system would have responded differently to the post-2002 drought under different management scenarios. In the case of scenario 1 (artificially open mouth) Figure 7.3a shows that hypersaline conditions would have occurred but water levels would have been maintained near EMWL. In the case of scenario 3 (combined inlet with no management intervention) Figure 7.3c shows that the mouth would have stayed open for longer after the onset of the drought while salinities would have increased during that period. However, the mouth would then have closed and salinities reduced due to dilution by Mfolozi inflows. Water levels would also not have fallen substantially below EMWL and may even have increased.

7.2.6 Trajectories through the salinity/ water level state space Phase plots showing the trajectories through the water level/salinity state space for each scenario are shown in Figures 7.5 and 7.6. There are two main trajectories – one associated with open mouth and the other with

Estuary and lake hydrodynamics

100

100 open mouth

Salinity

90

100

open mouth

90

80

80

80

70

70

70

60

60

60

50

50

50

40

40

40

30

30

30

20

20

20

10

10

10

0

0 –2

–1

0

1

2

3

closed mouth

0 –2

4

open mouth

90

closed mouth

–1

0

1

2

3

4

–2

–1

0

1

2

3

4

Water level (mEMWL)

FIGURE 7.5 Long-term average lake water level versus salinity for (left) scenario 1, (middle) scenario 2 and (right) scenario 3. Open symbols denote open-mouth states while solid symbols denote closed-mouth states. 100

100 open mouth

Salinity

90

100 closed mouth

90

80

80

80

70

70

70

60

60

50

50

40

40

30

30

20

20

10

10

0

Average duration: closed -10 years open -18 months

–1

0

1

2

3

4

open mouth

mouth closes

Average duration: closed -3 years open -6 years

60 50 40 30

mouth breaches

mouth closes

mouth breaches

20 10

0 –2

closed mouth

90

open mouth

0 –2

–1

0

1

2

3

4

–2

–1

0

1

2

3

4

Water level (mEMWL)

FIGURE 7.6 Schematic depiction of the main trajectories in the salinity/water level state space for (left) scenario 1, (middle) scenario 2 and (right) scenario 3.

closed mouth conditions. With an open mouth there is a linear bidirectional trajectory in the salinity/water level state space that characterizes scenario 1 but also occurs in other scenarios when the mouth is open. With a closed mouth there are curved trajectories in the salinity/water level state space, each associated with a specific salt loading. These trajectories are bidirectional in scenario 2 and unidirectional in scenario 3. In scenario 2 the closed mouth trajectories extend to more extreme values than in scenario 3. In extreme conditions the system is prone to desiccation and hypersalinity in this scenario. The combination of low lake levels and high salinities is unique to scenario 2 and reflects the actual post-2002 conditions in the lake.

7.2.7 Occurrence and persistence statistics of water level and salinity states Occurrence probabilities and persistence times for six water level/salinity states and for the three management scenarios are summarized in Table 7.4 and shown as bubble plots in Figure 7.7. Scenario 1 has the highest salinity variability of the three scenarios. Hypersaline conditions are also more probable in this case. Salinities exceed that of seawater about 45% of the time and extreme hypersalinity occurs for 17% of the time. Fresh conditions only occur 6% of the time. Persistence times are limited to a few months in each state and there are frequent state

123

124

Derek D. Stretch et al.

Table 7.4. Occurrence probabilities for various salinity and water level states and for the three management scenarios States

Scenario 1

Scenario 2

Scenario 3

6%

41%

14%

2. Brackish (5–12)

13%

28%

26%

3. Low estuarine (13–25)

21%

16%

27%

4. High estuarine (26–45)

29%

7%

24%

5. Hypersaline (46–65)

15%

2%

7%

6. Extreme hypersaline (> 65)

17%

6%

2%

1. 
1.0 (Note 1)

0%

18%

0%

2.
1.0 !
0.5 (Note 2)

0%

15%

1%

3.
0.5 ! 0

60%

22%

41%

4. 0 ! +0.5

35%

26%

40%

5. + 0.5 ! +1.0

3%

11%

13%

6. ≥ +1.0

1%

8%

5%

Average salinities 1. Fresh ( 4)

Water level ranges (m)

Note 1: > 50% of the average lake area desiccated. Note 2: 25–50% of the average lake area desiccated.

changes. However, salinities above 45 can occur for durations longer than a year. In contrast, scenario 1 has the most stable water levels of the three scenarios – they remain within 0.5 m of EMWL for 95% of the time. This is due to the open mouth conditions, which allow seawater to flow into the lake when water levels drop. In scenario 2 the lake is predominantly fresh with a 41% probability that salinities remain below 4 (Table 7.4) and only a 10% chance that (average) salinities exceed that of seawater. Hypersaline conditions are associated with very low water levels and desiccation (Figure 7.7). Water levels fall below
0.5 m EMWL for about 30% of the time (> 25% desiccation) and below
1.0 m EMWL for 18% of the time (> 50% desiccation of the lake area). Extreme hypersalinity and desiccation

can persist for longer than a year at a time during dry periods. During wet periods high water levels with low salinity occur for similar durations. The highest persistence times are in extreme conditions when the mouth is closed or when the mouth is open, otherwise the system changes state every few months. In scenario 3 salinities are variable but remain below 45 for 90% of the time and extreme hypersalinity is rare. The risks of low water levels and significant desiccation are negligible in this scenario. Persistence times are generally shorter than scenarios 1 and 2. During dry periods, salinities > 45 can occur for about a year at a time but Mfolozi inflows prevent extreme hypersaline conditions from persisting for more than a few months.

Salinity states

Estuary and lake hydrodynamics

6

15

2

6

6

6

2

5

12

3

5

2

5

7

4

20

8

4

3

2

1

4

11

11

3

3

10

10

1

3

4

6

3

4

3

12

10

3

1

2

4

8

1

2

2

5

11

9

2

7

10

5

3

4

1

1

2

7

14

1

3

9

2

1

Salinity states

1

2 3 4 5 Water level states

6

1

11

8

2 3 4 5 Water level states

6

6

7

2

6

15

5

4

2

5

2

1

4

4

2

1

1

4

3

3

3

3

3

3

1

2

3

4

4

3

13

2

2

3

1

1

2

4

6

7

6

3

1

1

2

3

1

4

3

5

7

10

3

4

5

2 3 4 5 Water level states

6

1

2

1 6

1

1

6

2 3 4 5 Water level states

1

5

6

4 5

3

4

2

4

4

4

2

3

4

3

3

3

3

2

3

4

4

3

1

2

4

2

2

3

4

5

6

1

2

FIGURE 7.7 Bubble plots showing (top row) percentage time spent in each salinity/water level state combination and (bottom row) average persistence times (in months) for each salinity/water level state combination. In both rows the management scenarios are (left to right) scenarios 1, 2 and 3 respectively. The area of each bubble is proportional to the magnitudes of its associated variable.

In summary, scenario 1 has the highest occurrence of hypersaline conditions and scenario 2 the highest occurrence of fresh conditions. Scenario 3 incorporates aspects of both scenario 1 and 2 but without the concomitant risks of extreme hypersalinity and desiccation respectively. Scenario 2 is the most stable of the three scenarios in terms of longer average persistence times, but the longest persistence times are generally in extreme conditions.

7.2.8 Basin-scale heterogeneities Measured water levels and salinities for North Lake, False Bay and South Lake for the period 1963–2002 are shown in Figure 7.8. Measured salinities are also plotted against average lake salinities to show the spatial divergence from the mean under different conditions. When average lake salinities are below

35 there is a normal salinity gradient; that is, low salinities in the northern reaches of the lake increase toward that of seawater near the mouth. This occurs during wet periods when freshwater inflows are higher. As average salinities increase above 35 a reversed salinity gradient develops due to seawater inflows that compensate for evaporative losses during dry conditions. Note that extreme hypersalinity (> 65) occurred 15% of the time in North Lake and 7% in South Lake from 1963 to 2002; that is, under the scenario 1 management strategy. Spatial heterogeneities are most evident in extreme dry conditions. Before 2002 (i.e. scenario 1) salinity gradients of up to 50 were recorded during dry conditions. South Lake was found to provide a refuge for fauna that could not tolerate the extreme conditions in North Lake (Forbes and Cyrus, 1993). After the mouth was left

125

Derek D. Stretch et al.

North Lake

False Bay

South Lake

100

100

90

90

80

80

70

70

60

60

Salinity

Salinity

126

50

50

40

40

30

30

20

20

10

10 0

0 –2

–1

0

1

2

3

4

0

10

20

30

40

50

60

70

80

90 100

Average lake salinity

Water level (mEMWL)

FIGURE 7.8 (Left) Measured water levels versus salinities from 1963 until 2002 (open-mouth conditions) in North Lake, False Bay and South Lake, and (right) the measured salinity of each station versus the average lake salinity.

to close in 2002 (i.e. the management strategy changed to scenario 2), salinities of 60 were measured in South Lake while salinities of 200 or higher occurred in North Lake/False Bay. In addition, as water levels dropped, the different parts of the lake began to separate out. The spatial heterogeneity created during dry conditions in both

of these scenarios inevitably affects the biota of the system creating distinct biological responses in different parts of the lake (Taylor, 2006). However, it is worth emphasizing that these specific extreme conditions occur only when dry conditions are combined with a scenario 2 management strategy.

7.3 Wind-driven flows, waves and mixing The prevailing wind directions for the St Lucia Lake system are from the north-east or southwest (Figure 7.9). The wind roses shown are derived from weather stations near the St Lucia Mouth and from False Bay in the north (refer to Figure 7.1). The strongest winds range from 7 to 19 m s
1.

7.3.1 Wind set-up in the lake Hutchison (1976) estimated wind set-up for the St Lucia Lake using a simplified one-dimensional hydro- dynamic model. The wind speeds considered were 5, 10 and 15 m s
1 blowing either from the north or the south direction for an average water

depth of 1 m. The time it took for equilibrium set-up to be reached was estimated to be 24, 60 and 70 hrs respectively. The wind rose suggests that it would be unlikely for a wind speed of 15 m s
1 to blow continuously for 70 hrs. For a 10 m s
1 wind blowing from the north the estimated wind set-up magnitude was 0.38 m after 60 hrs. The currents induced from this wind set-up would not be strong enough to resuspend sediment from the lake bed since the induced bed shear stresses are too low. However, these currents would contribute to horizontal advection and diffusion of suspended sediments throughout the lake basins (Luettich et al., 1990).

Estuary and lake hydrodynamics

Wind Rose Lister’s Point (annual)

Wind Rose Estuary Mouth (annual)

N

N NNW NW

NNW

NNE

20%

NE

15%

25%

NNE

20%

NW

NE 15%

10%

WNW

ENE

ENE

10%

WNW > 19

5%

W

> 19 5%

12 - 19

E

0%

7 - 12

W

4-7

2-4

2-4 0-2

WSW

ESE

ESE

SE

SW

SE

SW

SSE

SSW

7 - 12

E

4-7

0-2

WSW

12 - 19

0%

SSW

S

SSE S

FIGURE 7.9 Wind roses in the St Lucia Lake area (left) at the estuary mouth, and (right) at Lister’s Point in False Bay. 9

1200

Turbidity (NTU)

7 6

800

5

600

4 3

400

2

200

Wind Speed (m/s)

8

1000

1 0

13

:3

0

3 :5 12

2 :2 12

6 :4 11

9 :1 11

10

:4

9

6 :1 10

27 9:

00 9:

00 8:

00 7:

6:

00

0

Time (hr:min)

FIGURE 7.10 Measured turbidity and wind speed over time within South Lake. The dashed grey line is the turbidity and the dashed black line is the wind speed.

7.3.2 Wind waves and sediment resuspension In shallow lakes sediment resuspension is driven mainly by wind-driven waves (Luettich et al., 1990). When the wind is strong enough windgenerated waves drive an oscillatory benthic boundary layer. Sediment concentrations within the water column increase approximately in proportion to the induced bed shear stresses. Turbulent mixing causes the water column to become well mixed. Figure 7.10 shows observed concentrations of sediment within South Lake as the wind speed increases.

Concentrations increase to an equilibrium value where the deposition flux equals the erosion flux. There was no significant vertical gradient in turbidity over most of the water column. Wind-induced wave growth depends on the wind speed, fetch and duration that the wind blows (Holthuijsen, 2007). The shallow nature of the St Lucia Lake complicates the wind-induced wave growth by the addition of the water depth as a limiting factor for wave growth. Waves that are initially formed by the wind (usually a few centimetres in height and wavelength) are in deep

127

Derek D. Stretch et al.

(a)

10–3 1

NE Wind Direction

0.3 g2 E U4

128

0.2 10–4

='

0.1

10–5 102

103 xg X= U2

104

FIGURE 7.11 Wave growth in finite depth water after Young and Verhagen (1996). The plot shows wave energy versus wind fetch for different depths (in non-dimensional form).

water since the ratio of the water depth to wavelength is large. At this stage the growth is dependent on the fetch and wind speed. As the waves propagate through the water they grow, assuming there is further energy transfer from the wind. The growth continues until the waves begin to ‘feel’ the lake bottom. It is at this point when the wave growth is said to be depth limited. A schematic depiction of the wave growth in finite water depth is shown in Figure 7.11 (after Young and Verhagen, 1996). Initially, for a short fetch, the wave growth is in accordance with deep-water relations. As the fetch increases the effects of the water depth become more important and the wave growth deviates from the deep-water behaviour and wave heights asymptote to a constant value that depends on the water depth. Using the model of Young and Verhagen (1996) together with bathymetric data, it is possible to estimate the areas of St Lucia Lake where wave growth is in depth-limited conditions. Figure 7.12 shows estimates for South Lake and for a wind speed of 8m s
1 from the north-east and south-west. The average water depth was assumed to be 0.8 m for these estimates. The shaded region indicates the area of the lake where wave

(b) SW Wind Direction

FIGURE 7.12 Figure showing the area of South Lake where wave growth is in depth-limited conditions (shaded) for (a) a north-east wind and (b) a south-west wind blowing at 8 m s
1 with an average water depth of 0.8 m. Arrows indicate wind direction.

Estuary and lake hydrodynamics

1200

1000

Turbidity (NTU)

800

600

400

200

0 0.0

0.1

0.2 0.3 0.4 Bed shear stress (Pa)

0.5

0.6

FIGURE 7.13 Measured turbidity in South Lake plotted against the bed shear stress calculated from measured wave heights and periods. Water depths at the sampling locations were in the range 0.5–1.2 m. Note the transition for shear stresses above 0.1–0.2 Pa and the approximately linear increase in turbidity above that range of shear stresses.

growth is depth limited. For the north-easterly wind approximately 55% of South Lake is in depth-limited conditions. For the south-westerly wind approximately 65% of South Lake is in depth-limited conditions. The differences can be attributed to the bathymetry of the lake. The results indicate that wave growth within South Lake can be expected to be predominantly depth-limited when water levels fall below EMWL, e.g. during drought conditions. Figure 7.13 shows measured turbidities in South Lake plotted against the bed shear stress. The shear stress was inferred from measured wave heights and periods. A critical value of the shear stress in the range 0.1–0.2 Pa is evident from the data, which is consistent with soft fluid mud (Partheniades, 2009). Above the ‘critical’ conditions the turbidity increases approximately linearly with shear stress. Figure 7.14 shows typical diurnal variations in turbidity that are expected to occur in the lake as a result of wind speed variations. These results were

Turbidity (NTU)

1000

500

0

1

2

3

4

5

1

2

3

4

5

6 Day

7

8

9

10

11

12

6

7

8

9

10

11

12

Turbidity (NTU)

1000

500

0

Day

FIGURE 7.14 A sample of simulated daily turbidity variations during summer (top) and winter (bottom) of 2002 (Pringle, 2011). Note the typical diurnal variations and the generally higher values during summer (Feb) due to the windier conditions.

129

Derek D. Stretch et al.

generated using a model calibrated against the data shown in Figure 7.13, and are shown for typical summer and winter periods. Summer wind speeds are on average higher than winter at this

location. It is evident that very high wind-driven turbidities (~700 NTU) can be expected to occur regularly in the lake, particularly when the water depths are shallow.

7.4 Tidal inlet hydrodynamics and morphodynamics Intermittent mouth closure of TOCEs is usually linked to seasonal changes in terrestrial runoff and other factors such as energetic wave action, small tidal range and high rates of longshore and cross-shore sediment transport (Ranasinghe et al., 1999; Ranasinghe and Pattiaratchi, 2003; Stretch and Parkinson, 2006). However, inlet processes for TOCEs (and tidal inlets generally) can be complex and site specific. The energetic wave climate (average Hs ~ 2 m) on this coastline plays a significant role in the inlet dynamics as it gives rise to an average northerly longshore sediment transport of approximately 850 000 m3 yr
1 (~2300 m3 day
1) but with large interannual variations (Schoonees, 2000). There is a dominant southerly swell on the eastern coast of South Africa (Rossouw, 1984). Ocean tide characteristics in the region are semidiurnal (M2 dominant with a period of 12.42 hours)

and microtidal with a spring tide range of about 2.0 m and neap tide range of 0.5 m.

7.4.1 Tide-driven exchange flows The available information regarding the tidal characteristics of the St Lucia and Mfolozi systems, either as linked or as separate systems, is very limited. The period 1990 to 2007 contains several short intervals when the Mfolozi and St Lucia were combined. The Mfolozi inlet naturally migrates northward due to the prevailing longshore transport regime. Occasionally this migration has resulted in (brief) periods of a recombined inlet before the systems were again artificially separated by management interventions. Taylor (2006) has documented a record of these ad hoc changes in inlet configuration. There is a water level recorder in the St Lucia Estuary located about 2 km from the mouth (see Figure 7.1).

1.50 Flood

Ebb

1.00 Water level (m)

130

Estuary water level Mean estuary level

0.50

Super elevation

Mean sea level 0.00

HE – Estuary range Ocean water level

–0.50 Hs – Sea range –1.00 0:00

2:00

4:00

6:00

8:00

10:00

12:00 Time

14:00

16:00

18:00

20:00

22:00

0:00

FIGURE 7.15 Sample of recorded water level variations in the sea and estuary to illustrate the tidal characteristics discussed in the text.

Estuary and lake hydrodynamics

Measure bridge data

(a)

Measured tidal data

Simulated levels

1.2 0.8 0.4 0 -0.4 -0.8 11-Jul-72

12-Jul-72

13-Jul-72

14-Jul-72

15-Jul-72

(b)

3-Jun-00

7-Jun-00

11-Jun-00

15-Jun-00

(c) 1.2 0.8 0.4 0 -0.4 -0.8 5-May-07

(d)

8-May-07

11-May-07

14-May-07

1.2 0.8 0.4 0

-0.4 -0.8 23-Mar-10

27-Mar-10

31-Mar-10

4-Apr-10

FIGURE 7.16 Comparison of simulated water levels and measured gauge data at the St Lucia Estuary Bridge for: (a) Hutchison’s 1972 measurements for a separate St Lucia system, (b) a combined Mfolozi and St Lucia system in 2000, (c) separate St Lucia system in 2007, (d) a separate Mfolozi system. Neap to spring periods are shown based on data availability.

These water level recordings give an indication of the tidal characteristics in the lower reaches of both the St Lucia and Mfolozi Estuaries. Figures 7.15–7.17 show some of the main features of the tidal characteristics, which may be summarized as follows: • Tidal energy is strongly attenuated as it propagates through the inlet. The ratio of the water level range

in the estuary to that in the sea, HE/HS, is typically less than 0.5 for spring tides and indicates strong inlet constriction. As the constriction increases, the HE/HS ratio decreases. The values are consistent with Keulegan’s K values (or repletion coefficients) of 0.5 or less (Keulegan, 1967). • The degree of inlet constriction, as reflected in the HE/HS ratio, varies with terrestrial flows and through the wet/dry seasons.

131

132

Derek D. Stretch et al.

Measured Flows

100.0

Simulated Flows

80.0 60.0 40.0 20.0 0.0 –20.0 –40.0 25-May-07

27-May-07

29-May-07

31-May-07

200 150 100 50 0 –50 –100 –150 11-Jul-72

12-Jul-72

13-Jul-72

14-Jul-72

15-Jul-72

3
1

FIGURE 7.17 Comparison of simulated and measured flows (m s ) at the St Lucia Estuary Bridge for (top) separate St Lucia system in 2007, (bottom) separate St Lucia system in 1972. (Data from Chrystal, 2012 and Hutchison, 1976.)

• Tidal prisms have been measured for the separated St Lucia Estuary (Table 7.5) and range from 0.3 to 0.6 Mm3 for neap tides and 1.0 to 2.0 Mm3 for spring tides (Chrystal, 2012). The volumes change with the degree of mouth constriction, as does the HE/HS ratio. • Tidal prisms for a combined system are less well documented, but Lindsay et al. (1996) estimated a tide prism of 0.32 Mm3 for the Mfolozi/ Msunduzi channels during neap tide and 0.65 Mm3 during spring tide. The St Lucia and Mfolozi systems had a common inlet at the time of their field study and the reported HE/HS ratio of 0.3–0.4 suggests that the spring tide prism on the St Lucia side would have been about 1.0–1.5 Mm3 (Table 7.5) giving a combined total of about 2.0 Mm3. • A strongly constricted inlet can control (limit) the tidal prism irrespective of changes in the tidal channel network. For example, the increased effective tidal area for a combined Mfolozi/St

Lucia system can result in a reduced HE/HS ratio with only small changes in the tidal prism – e.g. see Figure 7.16b, c. • There is a strong temporal asymmetry in the estuary tidal signal. The flood tide duration is typically about 5 hours while ebb tides can exceed 7 hours, i.e. the estuary tide is flood dominant with average flow rates higher during flood tides than ebb tides. • Spectral analysis reveals that tidal asymmetry is associated with the generation of an M4 over-tide with an M4/M2 amplitude ratio of 10–30%. The transfer of energy from M2 to M4 is a further indication of strong non-linear damping in the tidal transmission through the inlet. • The average water level in the estuary is significantly superelevated above the mean sea level. In particular the low tide levels in the estuary are well above the low tide levels in the sea during spring tides. Superelevation is caused by two main mechanisms: (1) wave radiation stress and

Estuary and lake hydrodynamics

Table 7.5. Measurements of the tidal prism for the St Lucia Estuary from 1972 (Hutchison, 1976) and 2007 (Chrystal, 2012). During 2007 the lake water levels were well below mean sea level at the time of mouth opening, and the flood tide prism was larger than the ebb, resulting in a net inflow into the lake

Date

Tidal stage

07/04/20

Spring

07/05/25

Peak flows (m3 s
1)

Volume (m3)

Flood

Flood

Ebb

Ebb

Net volume (m3)

Tidal range (m) Sea

Estuary

HE/HS

131.0

68.0

1 600 000

1 333 668

266 332

1.80

0.70

0.39

Neap

48.7

24.8

851 397

332 996

518 401

0.53

0.23

0.44

07/05/28

Mid

57.8

24.4

822 980

445 099

377 881

1.19

0.28

0.23

07/05/31

Spring

63.9

25.6

965 566

454 445

511 121

1.45

0.32

0.22

07/06/22

Neap

49.7

30.9

495 003

594 964


99 960

0.80

0.27

0.34

72/07/11

Spring

187.0

115.0

2 067 179

1 829 419

237 760

1.67

0.87

0.52

72/07/12

Mid

176.0

100.0

1 887 813

1 586 904

300 909

1.74

0.85

0.49

72/07/13

Mid

150.0

91.0

1 388 517

1 495 033


106 516

1.63

0.72

0.44

72/07/14

Mid

105.0

147.0

1 198 524

2 984 691


1 786 167

1.49

0.63

0.42

(2) inlet constriction, particularly the shallowness of the inlet during ebb tides that retards the outflows, as is evident in Figures 7.15 to 7.17 (Hanslow and Nielsen, 1992; Hanslow et al., 1996; Tanaka and Nguyen, 2008; Malhadas et al., 2009). Wind effects and strong terrestrial flows can also contribute to superelevation.

7.4.2 Inlet stability The stability of tidal inlets generally refers to whether they can maintain an approximate equilibrium in terms of their cross-sectional area. It can also refer to location stability, which concerns whether the inlet location changes or migrates. Considerable research has been done on navigable inlets in tide-dominated situations (e.g. Bruun, 1978) but the stability of inlets in wave-dominated environments (such as in South Africa) is less well understood. Energetic waves suspend sediments in the littoral zone, which can then be carried by the flood tide into the estuary

and be deposited in a flood delta. The growth of that delta starts constricting the flows and can eventually lead to closure of the mouth. In tide-dominated locations the ratio of the tidal prism P to the annual longshore sediment transport volumes M has been found to be a good overall indicator of inlet stability with values less than about 20, typically associated with constricted or unstable inlets that tend to close (Bruun, 1978). In the case of the wave-dominated St Lucia/Mfolozi system (with either separate or combined inlets) the ratio is very small, P/M < 5, and suggests that the inlet(s) will not stay open due to tidal flushing. This is not a surprising observation since there have been reports of intermittent mouth closure at St Lucia that go back to the early nineteenth century and continue to the present (see Table 7.9). Furthermore, it is well known that the P/M ratio alone is not a good indicator of inlet stability in this region. Mouth closures typically occur during persistent low flow periods. The important implication of this fact is that terrestrial flows are

133

134

Derek D. Stretch et al.

(a)

(c)

(b)

(d)

FIGURE 7.18 Photographs showing the progressive closure of the St Lucia mouth following the breaching event in March 2007. (a) and (b) are from 25 May 2007 while (c) and (d) are from 2 August 2007. Note the advancing flood delta as marine sediments are imported into the system. (Photographs: Ricky Taylor.)

required to maintain an open inlet at St Lucia. When the St Lucia/Mfolozi system discharges through a combined inlet, the average annual outflow through the inlet is about 900 Mm3 (~30 m3 s
1), whereas the average outflow from a separate St Lucia inlet is only 150 Mm3 (~5 m3 s
1). This 85% reduction is the reason that artificial intervention is required to keep a separated St Lucia inlet open, except perhaps during very wet conditions (Lawrie and Stretch, 2011a; C. P. Chrystal and D. D. Stretch, unpubl.).

Figure 7.18 shows the progressive re-closure of the St Lucia inlet following a breaching event in 2007. The northward migration of the spit from the southern side of the inlet is evident in the time sequence where it is also clear that it is driven by northerly longshore sand transport. Furthermore the progressive infilling of the flood delta as waves and tides carry marine sediment into the inlet is also evident, and ultimately led to re-closure of the mouth.

Estuary and lake hydrodynamics

7.5 Sediment dynamics in the St Lucia/Mfolozi complex An important and recurrent management issue concerning the St Lucia/Mfolozi system concerns the fate and transport of suspended sediments carried by the Mfolozi. Since the 1940s this issue has been used to justify artificial separation of the Mfolozi from St Lucia. As noted earlier in this chapter, this intervention has been the most important factor in changing the overall functioning of the system. Despite the importance of the sedimentation issue, the details are poorly understood and there remain many unanswered questions such as: • What are the sediment yields from the Mfolozi and the other catchments of the system and have they increased due to human activities over the last century? • What is the fate and transport of suspended sediments in the combined Mfolozi/St Lucia system? • Have sediments from the Mfolozi been transported in significant quantities from the mouth into the lake basin? • What are the effects of tidal flows on the sedimentation processes? • What are the effects of breaching events on the sedimentation processes? In this section we start addressing these questions by outlining what is known about sediment yields in the area and by showing some results obtained by extending the water/salt budget model to include suspended sediments. The focus here is on fine suspended sediments in the clay/silt particle size range (< 63 μm) because of their perceived importance in the sedimentation issue and impact on biological systems. The deposition of large quantities of fine sediments can smother the benthos and adversely affect benthic fauna (Wolanski, 2007; Cyrus et al., 2010a). Furthermore, the resuspension of fine sediments due to wind-generated waves or currents increases the turbidity of the water column, which in turn affects primary production through

reduced light availability. High turbidity can also negatively impact filter feeders (Carrasco et al., 2010). While the focus here is on fine suspended sediments, the influx of coarse beach sediments via the tidal inlet plays a key role in the mouth dynamics and can therefore indirectly also have important impacts on the biological functioning. However, these sediments have relatively high settling velocities and are therefore not transported far from the energetic near-mouth region.

7.5.1 Sediment yields and characteristics Estimated sediment yields for the Mfolozi catchment range from 0.68 to 2.4 Mtons yr
1 or 68 to 240 tons km
2 yr
1 (e.g. Grenfell et al., 2009, and references cited therein). These estimates are from semiempirical models and/or measurements of suspended sediment concentrations. Other regional estimates of 5–723 tons km
2 yr
1 suggest that the Mfolozi values are close to the median (Rooseboom, 1992). For example, estimated yields for the nearby Thukela and Pongola rivers are about twice as high as for the Mfolozi at 400–600 tons km
2 yr
1. In a global context the data given by Milliman and Meade (1983) suggest that the Mfolozi can be considered to have medium yields (normalized on catchment area). An extreme example is the Tana River in Kenya, which has a yield of about 1000 tons km
2 yr
1 associated with a very degraded catchment due to overgrazing and poor farming practices. The monthly distribution of average flows and suspended sediment yields for the Mfolozi are shown in Figure 7.19 and were obtained from direct turbidity measurements by the Mtubatuba water works. It is evident that for typical runoff conditions, the sediment yield is highest during the months November to April, with much lower values during the dry season from May to October. Furthermore higher yields occur at the start of the rainy season,

135

Derek D. Stretch et al.

Sediment Load (ktons)

7.5.2 Flocculation dynamics of fine sediments

Flow (Mm3)

180 160 Mass/Vol (Ktons/Mm3)

136

140 120 100 80 60 40 20 0 Oct Nov Dec Jan Feb Mar Apr May Jun

Jul

Aug Sep

FIGURE 7.19 Monthly averaged flows and suspended sediment loads in the Mfolozi based on measured suspended sediment concentrations. (Data from Maro, 2012.)

which suggests that the supply of sediment is a limiting factor (Basson, 2008). A distinguishing feature of the Mfolozi is its variable streamflow (coefficient of variation CV of 60–80%) that in global terms is high (Grenfell et al., 2009). This has implications for its sediment transport regime, which has significant contributions from large episodic flood events. Suspended sediment loads in the St Lucia/ Mfolozi system are dominated by particles less than 60 μm in size (Maine, 2011). Recent particle size analysis on samples from the Mfolozi River and South Lake showed that particles less than 60 μm constitute 97% of total sample volumes (Maine, 2011). These particles are predominantly silt-sized, but also contain a small clay-sized fraction and particulate organic matter. Fine suspended sediments such as these are typically referred to as cohesive sediments and tend to coagulate to form aggregates or flocs, a process known as flocculation. Flocculation affects the fate and transport of cohesive sediments by its effects on settling velocities. Small microflocs are typically 10–20 μm in diameter and are resistant to break-up by turbulence. Large macroflocs in excess of 100 μm have the potential to form depending on the intensity of the turbulence and other factors (Van Leussen, 1994; Wolanski, 2007; Mikes and Manning, 2010).

There are several factors which influence flocculation: turbulence, suspended sediment concentration, salinity, organic matter content, pH, the material composition of the sediments and the time duration for which optimal conditions are present (Van Leussen, 1994; Mikes et al., 2002; Maggi, 2006; Verney et al., 2009; Mikes and Manning, 2010). The cohesive properties of fine sediments are controlled by the clay fraction. The negative surface charge of clay particles results in the formation of an ionic double layer around the particle that forms an energy barrier preventing particle contact (Van Leussen, 1994; Maggi, 2006, Mietta et al., 2009). In saline conditions the size of the double layer reduces, increasing the number of effective collisions between particles, resulting in the formation of flocs (Van Leussen, 1994; Mietta et al., 2009). Turbulent mixing can stimulate flocculation by increasing the collision frequency between particles (Van Leussen, 1994; Maggi, 2006). However, high turbulence shear stresses also result in the destruction of flocs by pulling them apart. Collision frequency increases with suspended sediment concentration and organic matter generally enhances flocculation (Van Leussen, 1994; Mietta et al., 2009; Verney et al., 2009). Flocculation is a dynamic process controlled by the strengths of the drivers listed above. It generally occurs in estuaries where salinity and low turbulence provide suitable conditions for the formation of larger flocs (Mikes and Manning, 2010). This accelerates sediment deposition. The flocculation dynamics and settling velocity of St Lucia and Mfolozi sediments were investigated using a simple laboratory flocculator and digital imaging techniques (Maine, 2011). The results showed that floc growth occurs at low turbulence in saline conditions. Floc settling velocity was directly proportional to the floc diameter, but significantly lower than that of equivalent-sized quartz particles. Figure 7.20 shows floc size distributions and settling velocities of Mfolozi sediments under different turbulence intensities. The settling velocities of flocs

40

1.6

35

1.4 Settling velocity (mm/s)

% of total volume

Estuary and lake hydrodynamics

30 25 20 15 10

1.2 1 0.8 0.6 0.4 0.2

5

0

0 0

100

200

300

400

Equivalent diameter (µm)

0

100

200

300

400

Equivalent diameter (µm)

FIGURE 7.20 (Left) Size distribution of Molozi flocs at high (red) and low (black) shear rates. (Right) The relationship between floc size and settling velocity corresponding to the floc size distributions shown left. (Data from Maine, 2011.)

varied from 2.5 mm s
1 for the largest flocs (400 μm in size) to less than 0.01mm s
1 for fine unflocculated sediments. The effective density of flocs decreases as the size of flocs increases. Macroflocs typically have densities of 1050–1100 kg m
3 (Maine, 2011). Macroflocs are easily detected and tracked in the laboratory using digital imaging. These flocs form a small portion of the population but a larger portion of the suspended mass (Maine, 2011). A significant proportion of Mfolozi and St Lucia sediments were too fine to detect using imaging methods and exhibited settling velocities below 0.01 mm s
1 (Maine, 2011). The flocculation tests showed bimodal behaviour – large aggregates settled rapidly at approximately 0.1 mm s
1while the remaining population of fine flocs and unflocculated materials settled slowly. This is evident in Figure 7.21 where flocculated materials settled rapidly during the first 15 minutes of a jar test while a low background concentration of fine flocs remained for hours afterwards. The proportion of flocculated and unflocculated materials varied between tests, however: in certain tests only 50% of fine Mfolozi sediments flocculated. By Stokes Law, flocs with a settling velocity of 0.01 mm s
1 and effective density of 1600 kg m
3 will have a diameter of 5.2 μm.

Figure 7.22 shows that the proportion of Mfolozi sediment less than 5 μm and 10 μm is 60% and 80% respectively. At higher sediment concentrations the amount of unflocculated sediment will increase. This has significant implications for sediment behaviour in the field. It suggests that a portion of Mfolozi sediments may be transported significant distances with limited settlement occurring, particularly when conditions do not stimulate flocculation. Flocs that form and settle are vulnerable to resuspension and break-up in turbulent conditions. This is frequently observed in Lake St Lucia where wind-driven waves resuspend lake sediments (as previously discussed in Section 7.3). In summary, fine Mfolozi/St Lucia sediments have the potential to flocculate and accelerate settlement. However, the extent to which this occurs depends on the ambient conditions (salinity, turbulence, etc.) and the finest sediments may remain in suspension for long times.

7.5.3 Simplified budget for suspended sediments Insight into the fate and transport of suspended sediments from the Mfolozi can be gained by incorporating these sediments in the water/salt

137

Derek D. Stretch et al.

FIGURE 7.21 Jar test observations of the settlement of Mfolozi sediment under quiescent conditions. (Source: Maine, 2011.) Cumulative % volume

138

100 80

•

60 40 20 0 0.1

1

10

100

1000

Particle size (mm)

•

FIGURE 7.22 Cumulative particle size distribution of fine Molozi sediment (June 2011). (Data from Maine, 2011.)

• budget model of Lawrie and Stretch (2011a). The simplifying assumptions required to do this are as follows: • The monthly sediment loads in the Mfolozi were based on the average monthly sediment concentrations measured at the Mtubatuba water works (see Figure 7.19). • It is assumed that when the inlets are combined and the mouth is closed all the Mfolozi flow and its

•

suspended sediments are diverted into the Narrows (see Figure 7.1). The sediments are assumed to settle in the Narrows. This assumption is based on observed trapping efficiencies of more than 90% during recent periods when the Mfolozi closed and water was fed through to St Lucia via the Back Channel (Kelbe and Taylor, 2011; van Alphen et al., 2011). The lake is a large storage area and is therefore assumed to have a 100% trapping efficiency with respect to suspended sediments that are brought in from its northern catchments. When the mouth is open, all the Mfolozi sediments are discharged to the sea – mixing between the Mfolozi and St Lucia systems is assumed to be small and the proportion of re-entrant sediments is neglected. Tide-driven flows give rise to a tidal prism of about 0.5 Mm3 at neap and 1.0–2.0 Mm3 at spring (depending on the mouth constriction). The total volume exchanged per month is therefore 45–75 Mm3 per month. The influx of fine suspended sediments from the sea on the flood tide is assumed

Estuary and lake hydrodynamics

Annual suspended sediment budget (Ktons) 1500 1000 500 0 –500 –1000 –1500 1910

1920

1930

1940

1950

1960

1970

1980

1990

2000

2010

1990

2000

2010

1990

2000

2010

Monthly suspended sediment budget (Ktons) 1000 500 0 –500 –1000 1910

1920

1930

1940

1950

1960

1970

1980

Accumulated suspended sediment gain/loss (Mtons) 6.0 4.0 2.0 0.0 –2.0 –4.0 –6.0 1910

1920

1930

1940

1950

1960

1970

1980

FIGURE 7.23 Simulated suspended sediment loads (in kilotons) in the St Lucia estuary when linked to the Mfolozi (scenario 3). Positive values are gains due to inflows from the Mfolozi (when the mouth is closed) while negative values are losses from discharges to the sea. Shaded regions indicate closed-mouth conditions. During open-mouth conditions the sediment transport is driven by tidal flows. This budget considers only fine suspended sediments in the clay/silt size ranges (< 60 μm) and ignores the coarse marine sediments that are transported in the mouth region.

negligible, but there is a net efflux of suspended sediments on the ebb tide with concentrations in the range 100–300 mg l
1 (based on measurements by Cyrus and Blaber, 1988). • Breaching events occur at the end of each closed mouth period and they resuspend sediments and discharge them to sea. Flow rates increase with the breaching water level and are based on results from a hydraulic model (Lawrie and Stretch, 2011a; Chrystal, 2012). No direct measurements of the sediment concentrations for breach outflows are available and they must be estimated from published data and/or semi-empirical models (e.g. Partheniades, 2009). Results for a representative 90-year simulation of the suspended sediment budget for the Narrows are shown in Figure 7.23. Table 7.6 compiles some of the relevant statistics and how they change as a function of the breaching level. It can be seen that during the

closed periods there are accumulations of sediment in the system, which are driven by small inflow pulses from the Mfolozi during the wet seasons. Larger flood events also contribute but tend to occur at the end of the closed periods since they typically breach the system. It is also evident that the breach outflows remove large amounts of suspended sediments from the system. The outflows created by the mouth breaching events can have magnitudes similar to those of significant floods (e.g. Parkinson and Stretch, 2007). These flows can therefore result in substantial scouring of accumulated sediments from the Narrows and St Lucia Mouth, and play an important role in the sediment budget. Note, however, that increased sediment outflows associated with higher breaching levels are counterbalanced by increased influx of sediments due to longer closed periods. During open mouth periods the tidal flows also remove resuspended fine sediments from the system

139

140

Derek D. Stretch et al.

Table 7.6. Simulated suspended sediment loads (in kilotons) in the St Lucia Estuary when linked to the Mfolozi (scenario 3), for breaching levels ranging from 1.0 m to 2.5 m GMSL. Positive values are gains due to inflows from the Mfolozi (when the mouth is closed) while negative values are losses from discharges to the sea driven by high lake levels and from tidal flows Suspended sediment budget for St Lucia Breaching level (mGMSL)

2.5 m

2.0 m

1.5 m

1.0 m

Open %

65%

67%

73%

82%

Closed %

35%

33%

27%

18%

203

181

149

100

Loss (ktons yr )


267


230


204


180

Net gain/loss (ktons yr
1)


65


49


56


80

Mfolozi inflow (Mm3 yr
1)

160

142

115

76

No. of breachings

10

10

11

14

Gain (ktons yr
1)
1

although their contribution is smaller than for breaching outflows. Overall the model results suggest that when viewed over decadal or longer timescales, sedimentation is a

natural part of the open/closed cycles that are a feature of this system. A long-term balance between gains and losses is possible depending on the details of the physical processes involved and climatic factors.

7.6 Modelling the outcomes of management interventions Management interventions in the functioning of St Lucia over the last 50 years have focused on changes to the inlet configuration, manipulation of the mouth state through artificial breaching and/or dredging, general sedimentation control through dredging, and development of schemes to increase freshwater inputs. Management schemes to supplement the freshwater supplies to the lake by transfers from nearby catchments have also been considered in the past. For example, Hutchison (1976) identified and evaluated various options. Some were subsequently partially implemented but were not sustainable (Whitfield and Taylor, 2009). A more recent evaluation of this option is given by Lawrie and Stretch (2010) and included the effects of the water transfers on the mouth dynamics.

It has already been stated that without the Mfolozi the system is susceptible to hypersalinity and/or desiccation during dry periods. Combining the St Lucia/Mfolozi mouths would significantly decrease these risks. However, the risks of high sediment inflows from the Mfolozi during the wet seasons, and when the mouth is closed, remain a management concern. One management option is to use controlled mouth breaching to avoid excessive suspended sediment loads from entering St Lucia during closed mouth periods. The water balance model provides a means to evaluate the outcomes of this type of option. With a combined St Lucia/Mfolozi mouth, the Mfolozi contributes a significant amount of fresh water during closed mouth conditions – about half the amount needed to re-breach the system.

Estuary and lake hydrodynamics

(a)

(b)

FIGURE 7.24 Simulated monthly average lake salinities and mouth states for (a) scenario 3A and (b) scenario 3B. Detailed water balance results are shown in Table 7.7.

Evaporative losses, direct rainfall inputs, and inflows from the lake catchments also make significant contributions to the water balance (Lawrie and Stretch, 2011a). Large flows occur in the Mfolozi River during summer months and also carry the highest suspended

sediment loads (see Figure 7.19). The water and salt budget model was used to explore the impact of artificially breaching the St Lucia/Mfolozi mouth before the start of the rainy season. The following modifications of the scenario 3 simulations were investigated:

141

142

Derek D. Stretch et al.

Table 7.7. Water balance simulations for various artificial breaching scenarios described in the text Management scenarios

Scenario 3

Scenario 3A

Scenario 3B

Scenario 3C

Scenario 3D

Annual average water balance contributions (Mm3) Mfolozi inflows

141

56

80

95

106

Rainfall onto lake

296

281

287

287

290

Catchment inflows

279

279

279

279

279

Evaporation losses


473


450


460


460


465

Water balance contributions during closed mouth conditions (Mm3) Mfolozi inflows

1330

339

787

740

1081

Rainfall onto lake

956

280

634

497

765

Catchment inflows

379

87

223

197

289

Evaporation losses


1586


525


1137


859


1321

37

15

29

23

33

Statistics Closed mouth (months) Salinity > 45 ppt

9%

18%

12%

13%

7%

Water level <
0.25 mEMWL

6%

9%

6%

7%

6%

• 3A: Breach the mouth in November if it has been closed for 1 year or more. • 3B: Breach the mouth in November if it has been closed for 1 year or more, but only if the water level is above 0.55 mEMWL (1.0 mGMSL). • 3C: Breach the mouth in November if it has been closed for 2 years or more. • 3D: Breach the mouth in November if it has been closed for 2 years or more, but only if the water level is above 0.55 mEMWL (1.0 mGMSL). November was chosen as the ‘decision month’ as flows and sediment loads begin to increase with the onset of the rainy season (Figure 7.19). Samples of the simulation results are shown in Figure 7.24 and tabulated in Table 7.7.

For scenario 3A (Figure 7.24a – unconditional breach after 1 year) the inflows from the Mfolozi would decrease by 75%. In scenario 3B (Figure 7.24b – breaching conditional on water level) the Mfolozi contribution is decreased by 50%. Reduced Mfolozi inflows mean a decrease in the suspended sediment load transferred into the St Lucia Estuary. The impact of both these interventions on the average salinity, water levels and mouth state are also shown in Figure 7.24 and Table 7.7. Note that in scenarios 3A and 3B the occurrence of low water levels is not significantly changed although the probability of high salinities is increased. In scenario 3C (unconditional breach after 2 years) the Mfolozi inflow volumes are reduced by 41% while in scenario 3D (conditional breach after 2 years) the reduction is 20% (Table 7.7).

Estuary and lake hydrodynamics

The simulations suggest that during the post-2002 dry period, conditions would have been less severe

and persistent with these management interventions (compare Figure 7.24 with Figure 7.3).

7.7 Links between physical and biological functioning

L|H

M|H

H|H

L|M

M|M

H|M

L|L

M|L

H|L

(–0.5m) MEDIUM (+0.5m) AVERAGE WATER LEVEL

HIGH

(>+1m)

LOW

(12) MEDIUM (45) AVERAGE SALINITY

HIGH

(>65)

(>65)

LOW

(0)

The timescales of variations in salinity and water levels are also important since biological systems take time to adjust to new conditions. Therefore both the occurrence and persistence of salinity and water level states are important in driving the biological responses of the system. Lawrie and Stretch (2011b) provided a broad overview of the biological responses that may be

HIGH

• submerged macrophytes may cause frictional damping of wind-driven waves and thereby reduce turbidity of the water column; • filter feeders can remove suspended sediments from the water and thereby affect suspended sediment concentrations.

HIGH

AVERAGE SALINITY (12) MEDIUM (45)

There are also feedback mechanisms from biological systems to the physical components such as:

AVERAGE WATER LEVEL (–0.5m) MEDIUM (+0.5m)

( 3000 mg l 1 (Groundwater Development Services, 1995; Dennis and Dennis, 2009). The groundwater quality generally decreases (i.e. increasing TDS) downstream in the secondary porosity aquifers. In contrast, the overall quality of the groundwater in the primary porosity aquifer is good with TDS values generally in the range 0 to < 1000 mg l 1 (Meyer and Godfrey, 1995). Hydrochemical assessment on the main lithological units within the primary porosity aquifer just south of St Lucia has been presented by Hattingh (1998). Baseline studies of groundwater quality on the Eastern Shores of the St Lucia catchment by Bjørkenes et al. (2004) and Bjørkenes et al. (2006) show two distinct and spatially separated groundwater types, one dominated by sodium and chloride and the other by calcium and carbonate (Figure 8.3). Comparison between the hydrochemistry and geochemistry has been presented by Mkhwanazi (2010) from samples in the primary porosity aquifer across the Zululand Coastal Plain. The groundwater quality is generally poorest in the lower rainfall regions of the secondary porosity

Groundwater hydrology

SO42−

80

20 40

Mg2+

Groundwater chemistry Eastern Shores: 1: Western area 2: Eastern area +: Precipitation 60

60

2

60

20

80

40

40

40

60

2 80

2 1

HCO3− 20

40

+ 60

+ 80

20 +

+

20

Cl− Ca2+

A. Anions

1 80

60

2

40

80 + +

Na++K+

20

B. Cations

FIGURE 8.3 Trilinear Piper diagram showing the general grouping of the principal anions and cations for samples of water quality taken from two areas on the Eastern Shores of Lake St Lucia (from Simonsen, 2003).

aquifer and often has TDS concentrations that exceed 1000 mg l 1 (Dennis and Dennis, 2009). These authors have also shown that the TDS of the groundwater in the primary porosity aquifers is seldom > 200 mg l 1.

8.4.3 Groundwater modelling The historical water level data sets suffered from large inaccuracies in the specification of the site position and elevation that has hindered geohydrological mapping. This has improved recently with the advent of GPS for locating the more recent monitoring points. The inaccuracies in locating spatial data and the lack of transient measurements in groundwater monitoring in the St Lucia catchment’s primary aquifer have restricted the conventional methods of evaluating the role of groundwater in maintaining the mass balance of the St Lucia estuarine system. Consequently, the assessments of land-use impacts on the groundwater have mainly been inferred from studies done in similar environments in other places (Gush et al., 2002) or extrapolated from the site-specific studies (Rawlins, 1992; Været et al., 2009; Clulow et al., 2011).

The paucity of data in the St Lucia catchment has necessitated a more pragmatic approach to mapping groundwater and examining its role in the St Lucia catchment through the development and application of groundwater models. These modelling studies have helped to understand the groundwater system and its role in catchment management. The early lumped mass-balance models for St Lucia and its catchments developed by Hutchison (1976) and Pitman (1980) provided an estimate of the groundwater contribution to the overall water needs for the lake. Subsequently, Kelbe and Rawlins (1992) and then Kelbe et al. (1995) attempted to determine the groundwater dynamics of the Eastern and Western Shores primary aquifers using 3-D numerical models of the groundwater system. These studies have provided some estimates of the seepage rates along the lake and ocean shorelines. Subsequent model upgrades (Wejden, 2003; Været et al., 2009) were used to evaluate the impact of climate change and changes in land use on the groundwater dynamics and interlinked systems. Several other researchers have estimated the seepage rates in the greater Zululand Coastal Plain from groundwater gradients and hydraulic properties along the boundaries of the aquifers (Meyer and Godfrey, 1995; Parsons and Associates, 2009).

157

158

Bruce E. Kelbe et al.

FIGURE 8.4 Simulated depth (m) to the regional water table for the St Lucia catchment showing the areas with shallow groundwater that are vulnerable to land-use impacts.

A catchment-scale groundwater model has been developed and continually upgraded to support spatial mapping of the water table (Figure 8.2), depth to the water table (Figure 8.4) and management of the groundwater within the catchment. Similar models have been constructed for the primary aquifer in the northern coastal plain of Maputaland to evaluate the formation and distribution of wetlands, and these studies were presented by

Grundlingh (2011). The depth to the water table, as illustrated in Figure 8.4, is crucial in identifying areas that are prone to the impacts from land use on the groundwater and its subsequent release into the rivers and wetlands. The primary porosity aquifers have very shallow water tables, creating the extensive array of wetlands that are the primary physical feature of the ecotourism industry in Zululand.

8.5 Groundwater recharge Recharge to the groundwater in the St Lucia catchments occurs primarily through percolation from the unsaturated zone. The available soil water in

the unsaturated zone is dependent on the rainfall rate and land cover. The rainfall in this region is generally highest in summer and lowest in winter. In summer

Groundwater hydrology

the rainfall is derived mainly from convective storms in conjunction with frontal systems, which produce short duration high-intensity events. Interception loss from the vegetation during these events is unlikely to be a significant proportion of the rainfall reaching the land surface. During winter the rainfall from the frontal systems is generally of longer duration and lower intensity, which can lead to significantly greater interception losses. There is a marked difference in the infiltration rates between the deep, highly permeable coastalplain soils and the shallower less-permeable soils of the large inland river catchments. There is little evidence of overland flow on the coastal plains, which indicates that most of the rainfall that reaches the ground infiltrates the soils. Even in extreme events the excess rainfall (above infiltration rate) will often pond until it has time to move through the soil except when there is direct access to the drainage lines. Measurement of these recharge processes is difficult and consequently many geohydrologists assume a simple relationship between mean annual precipitation (MAP) and groundwater recharge to account for the various intermediary losses. Others prefer to use a proportion of the mean annual runoff (MAR). However, these generalizations use simple numerical factors that can lead to large differences in recharge estimates and so they need to be treated with caution.

8.5.1 Secondary porosity aquifers Early estimates of the recharge by Groundwater Development Services, (1995) based on baseflow in the main river systems assumed that groundwater storage in the secondary porosity aquifer is constant over the period of consideration. Their recharge estimates varied from 0.5% to 9.5% of the mean annual runoff (MAR). More recent studies by Dennis and Dennis (2009) derived estimates of the recharge rates for some of the upper sections of the St Lucia catchments that are given in Table 8.3.

Table 8.3. Recharge rates for the resource units in the Mkhuze and Hluhluwe catchments Characteristics

Recharge (mm yr 1)

Head waters of Mkhuze River

43

Lower Mkhuze River

40

Mouth of Mkhuze River

39–43

Catchment of Hluhluwe River

41

Adapted from Dennis and Dennis (2009).

8.5.2 Primary porosity aquifers Meyer and Godfrey (1995) estimated that the recharge across the Zululand Coastal Plain varied from about 17% of MAP at the coast, declining to about 5% at a distance of 60 km inland. These early estimates of recharge are assumed to account for all the losses between incident rainfall and groundwater runoff (baseflow) including evapotranspiration. However, they are based on a large-scale model and do not take into account the depth to the water table, landform and other features that will influence the local-scale recharge rate. On a local scale, Rawlins (1991, 1992) and Rawlins and Kelbe (1991) observed a simple relationship between the direct groundwater response and rainfall events in a shallow aquifer on the Eastern Shores. They observed that all cumulative (5-day) rainfall events in excess of 10 mm induced a rise in the shallow water table that is directly proportional to the magnitude of the event. Kelbe and Shezi (2010) observed a relationship between rainfall and groundwater levels that varied with depth to the water table at three hydrologically similar sites on the coastal plain from Zululand to northern Mozambique. This relation is illustrated in groundwater response measurements at Richards Bay on the southern end of the Zululand Coastal Plain (Figure 8.5) for water level (pressure) measurements at three depths in a coastal environment similar to the St Lucia coastal plain. The shallow water table shows an immediate response to

159

Bruce E. Kelbe et al.

90.0

33.6 33.5

BH1

BH2

BH3

BH1

BH2

BH3

Rainfall

BH1

BH2

BH3

33.4 80.0

33.3 33.2 33.1

70.0

33.0 32.9 32.8

60.0

32.7 32.6 32.5

50.0

32.4 32.3 32.2

Rainfall (mm/hr)

Water Level (mMSL)

160

Kwambonambi Formation

40.0

32.1 8m

32.0 31.9

30.0

31.8

Kosi/Port Durnford Formation

31.7 31.6

20.0

24m

31.5 31.4 31.3

10.0

Uloa Formation

31.2 31.1 31.0

0.0

36m

361 352 344 335 327 319 310 302 294 285 277 269 260 252 244 235 227 218 210 202 193 185 177 168 160 152 143 135 127 118 110 102 93 85 76 68 60 51 43 35 26 18 10 1 Time from DOY 1(01H00) 2002

FIGURE 8.5 Measured groundwater response (mMSL) at 3 m (BH3), 26 m (BH2) and 36 m (BH1) depths and the corresponding rainfall events in Richards Bay (Kelbe and Germishuyse, 2001).

the rainfall events, with the peak water level occurring within hours of the rainfall event. This response is progressively delayed with increasing depth to the groundwater table, with a peak occurring several days after a major storm event at depth in excess of 36 m. The difference in water levels between the three HSUs is also indicative of the leaky nature of these aquifers.

To overcome difficulties in deriving the recharge, many of the numerical models described above have included the different hydrological processes that link the rainfall to the recharge. This is imperative when numerical models attempt to simulate the impact of land use on the groundwater (Rawlins and Kelbe, 1998; Været et al., 2009).

8.6 Groundwater discharge The paucity of geohydrological data in the St Lucia system requires that the results from studies in similar hydrogeological settings along the Zululand Coastal Plain are used to infer the geohydrological conditions in the St Lucia system. The measured water table fluctuation in Richards Bay (Figure 8.5) shows three recession rates that are similar to characteristic storm hydrographs in rivers (Beven, 2001). There is an initial short rapid decline

in head followed by a period of slower decline over several days before reaching the long-term slow recession rate, which is assumed to be related to the discharge of groundwater (baseflow) into the draining streams. The cause of these different recession rates in the groundwater is uncertain. Simultaneous high frequency measurements of rainfall and water level fluctuations (Figure 8.5) can be used to estimate the recharge rate. The highly

Groundwater hydrology

10,000

Culvert(m3/h)

Streamflow (m3/hr)

33.5

WTElvtn (mMSL)

1X) = 31.559 × 0.014 R = 0.722

33.4

33.9

33.3

WL in domain (m)

33.2 33.1

33.7

33.0 32.9 32.8

3

Hourly Streamflow (m )

32.6 32.5 0

20

40

60

80

100

120

140

WL in Borefide (m)

1,000

33.3

33.1

Water Table Elevation (m)

33.5 32.7

Water Table Elevation (mMSL)

32.9

32.7

32.5 02/10/31 00:00

02/10/29 00:00

02/10/27 00:00

02/10/25 00:00

02/10/23 00:00

02/10/21 00:00

02/10/19 00:00

02/10/17 00:00

02/10/15 00:00

02/10/13 00:00

02/10/11 00:00

02/10/09 00:00

02/10/07 00:00

02/10/05 00:00

02/10/03 00:00

02/10/01 00:00

02/09/29 00:00

02/09/27 00:00

02/09/25 00:00

02/09/23 00:00

02/09/21 00:00

02/09/19 00:00

02/09/17 00:00

02/09/15 00:00

02/09/13 00:00

02/09/11 00:00

02/09/09 00:00

02/09/07 00:00

02/09/05 00:00

02/09/03 00:00

02/09/01 00:00

02/08/30 00:00

02/08/28 00:00

02/08/26 00:00

02/08/24 00:00

02/08/22 00:00

02/08/20 00:00

02/08/18 00:00

02/08/16 00:00

02/08/14 00:00

02/08/12 00:00

02/08/10 00:00

02/08/08 00:00

02/08/06 00:00

02/08/04 00:00

02/08/02 00:00

02/07/31 00:00

02/07/29 00:00

02/07/27 00:00

02/07/25 00:00

02/07/23 00:00

02/07/21 00:00

02/07/19 00:00

02/07/17 00:00

02/07/15 00:00

02/07/13 00:00

02/07/11 00:00

100

FIGURE 8.6 Simultaneous measurement of streamflow and groundwater response in a shallow primary porosity aquifer in Richards Bay (after Kelbe and Germishyse, 2010). The source of the noise in the streamflow measurements is unknown.

dynamic nature of the groundwater on the coastal plain, as illustrated in Figure 8.5, is not captured with current monitoring programmes in St Lucia, where most measurements are recorded at monthly intervals. The simultaneous measurement of rainfall and groundwater response could be most useful when establishing a reliable model of recharge in this area. The National Department of Water Affairs has initiated a programme to increase the frequency of water level measurements to capture the dynamic nature of the aquifers.

remarkable similarity in the shallow (2–3 m) water table fluctuations and the stream discharge. The correlation between the two data series (inset in Figure 8.6) indicates that the relationship is a function of the baseflow and stormflow components. A similar relationship is expected for the Nkazana and other perennial streams draining the Eastern Shores. It is also expected that this relationship can be extended to the numerous exposed surfaces of the groundwater forming seepage zones into the wetlands in the coastal plain.

8.6.1 Groundwater discharge into streams and rivers (baseflow)

8.6.2 Groundwater discharge into Lake St Lucia

The streams that drain the primary aquifer are expected to behave in a similar manner to other streams in similar hydrological settings along the coastal plain. The streamflow and groundwater response to incident rainfall in the Richards Bay primary porosity aquifer is shown in Figure 8.6 together with the water table response. There is

The contribution of baseflow from the secondary porosity aquifer to the overall lake water balance in Lake St Lucia is considered negligible except in extreme drought conditions when it does contribute an increasingly significant proportion of the freshwater inflow into the lake. In conjunction with incident rainfall this is likely to have an influence on

161

162

Bruce E. Kelbe et al.

the salinity (Kelbe and Taylor, 2011). However, the biggest impact of groundwater on the lake is from direct seepage along the lake shore and through the

small streams draining the primary porosity aquifer (Kelbe et al., 2003, 2004) as discussed in the next sections.

8.7 Groundwater seepage No direct measurements have been made of the groundwater seepage rates through the shorelines of the lake or ocean. Most estimates of the groundwater seepage have been derived from numerical simulations using groundwater models (e.g. Kelbe et al., 1995; Wejden, 2003; Været et al., 2009). This seepage of groundwater provides a steady supply of fresh water along the shoreline. Any changes in the elevation of the drainage boundary (shoreline) or water table profile will create changes in the water table gradient that will ultimately change the water table elevation in the surrounding aquifer. The change in the water table elevation profile will have a direct impact on the groundwater-dependent wetlands in the adjacent terrestrial area close to the shoreline. This together with other possible influences on the shoreline from linking the Mfolozi to St Lucia Estuary (Lawrie and Stretch, 2011a) could also affect the adjacent Eastern Shores area by the maintenance of a relatively higher water table. These different issues are discussed below.

8.7.1 Changes in the drainage boundaries Changes in the sea level (Været et al., 2009), groundwater profile and the influence of the Mfolozi/ Msunduzi linkage (Lawrie and Stretch, 2011a) will influence the lake level that will induce other impacts on the aquatic system that are discussed below.

8.7.2 Groundwater seepage with open mouth In the period from 1956 until 2002 the St Lucia Mouth was maintained open for most of the time. As a result the water level of the lake–estuarine system remained consistently close to mean sea level. Estimates of the

groundwater seepage into the lake were derived by several researchers using numerical simulation models when the lake level condition was kept at a constant level. They used calibrated finite-difference models with spatial resolutions of the order of 100 m by 100 m. While the lack of adequate exploration and monitoring data limits their reliability, these models have provided the best estimate of the groundwater dynamics and head profile in relation to changing land use and climate. These numerical models have been used to estimate the groundwater fluxes under different land-use and climate change scenarios: 1. Kelbe et al. (1995) and Nomquphu (1995) simulated the groundwater fluxes into the lake along the eastern and western shoreline under different land-use scenarios. Their study indicated a 15–75% reduction in groundwater flux into the lake along the Western Shores due to commercial afforestation (pine). 2. Været et al. (2009) examined the impact of deforestation on the groundwater regime on the Eastern Shores in comparison with predicted impact of climate change. They concluded that the impact of land-use change, mainly due to the pine plantations, was significantly greater than the impact of the predicted climate change scenarios.

8.7.3 Observed groundwater seepage during drought conditions with mouth closed and no connection with the Mfolozi River During 2002 the mouth management changed and the lake became isolated from the ocean. At the same time a severe drought started to develop. The lack of river inflow and seawater influx and the high evaporation from the lake caused the lake level to

Groundwater hydrology

FIGURE 8.7 Groundwater seepage along retreating lake shoreline in Catalina Bay. The orange line shows the salinity front between fresh (EC < 5 mS cm 1) and brackish (EC > 40 mS cm 1) water.

decline to a point where it formed several small isolated shallow impoundments in the main bays, with extensive beaches where the groundwater seepage became evident. This provided an opportunity to investigate the shoreline seepage along the lake (Kelbe et al., 2004). The influence of the shoreline seepage into the lake was mapped near the Jetty in Catalina Bay by measuring the boundary between fresh and brackish water (Figure 8.7). Along the Eastern Shores, groundwater seepage was also identified from the presence/absence of groundwater-dependent vegetation (Taylor et al., 2006b). In some places seepage zones converged to form small rivulets. Although the actual seepage rate could not be determined, these observations were useful to verify the numerical model simulations described above. In several places along the Eastern Shores shoreline, groundwater seepage converges to form concentrated flow such as the Nkazana Stream. In one site where the shoreline had been eroded (probably by hippopotamus tracks into the wetland), a small channel had formed. Here the flow was measured at 10 m3 hr 1 in May 2003 (Figure 8.7), long after the beach had receded to the level shown in Figure 8.7. Along the Western Shores only a few minor seepage zones were identified. The lack of seepage faces is here attributed to the very shallow

FIGURE 8.8 An illustration of the groundwater response to changes in lake level and the ecological system that would be affected. The low water level in 2004 caused freshwater seepage along the expanding beach compared with the higher lake levels in 2001. A further rise in the lake level would induce higher groundwater elevation profiles that would impact on the shoreline vegetation.

nature of the primary porosity aquifer and the high elevation of the Cretaceous bedrock (aquiclude) in this region. The low discharge along the shoreline is exacerbated by the extensive afforestation which was estimated to reduce the groundwater seepage by up to 50% (Kelbe et al., 1995) along sections of the shoreline. Only one significant seepage zone was found in the northern sector of False Bay and a few smaller seepage zones are associated with sandy deltas along the western shoreline of False Bay. These only became evident when the lake level dropped. They are associated with upstream gullies, and the groundwater from them promoted the growth of reeds and sedges in the adjacent dry lake bed.

8.7.4 Groundwater seepage during drought conditions with mouth closed and the Mfolozi diverted into St Lucia The condition of a closed mouth and the Mfolozi flowing into the St Lucia Estuary has not been observed since the 1940s. A prolonged period when the Mfolozi and St Lucia mouths are closed and water from the Mfolozi River is diverted directly into Lake St Lucia would raise the lake level to a metre or more

163

164

Bruce E. Kelbe et al.

above mean sea level (Lawrie and Stretch, 2011b). As this would occur after a period of lower water levels in the lake, we can speculate that it could raise the elevation of the groundwater boundary – and hence reduce the groundwater table gradient resulting in an increase in the groundwater storage. The effects of the two recorded scenarios on the shoreline ecohydrology or hydroecology are

illustrated in Figure 8.8. The stable lake level shown is that measured in 2001. The scenario where the water table drops to a very low level is what was measured in 2004. It is speculated that the third scenario (not shown in Figure 8.8) – with very high water levels above the 2001 level – could shift the seepage zone upward and kill the swamp forest by inundation.

8.8 Evaporation and land use One of the most difficult processes to measure and the most important in terms of water loss from the St Lucia catchment is evapotranspiration. Simple water balance studies (Clulow et al., 2011) show that evapotranspiration from the Mfabeni Mire on the Eastern Shores accounts for over 90% of the water loss. However, there are spatial variations because the evapotranspiration rate is dependent on the rooting depth, depth to the water table and the vegetation type. Lindley and Scott (1987), using the data obtained from the series of monitoring boreholes established on the Eastern Shores in 1973, estimated that the presence of pine plantations increased the evapotranspiration losses by about 210 mm yr 1(i.e. by about 15–20%). The indigenous

vegetation does not have the same detrimental impact on the groundwater as commercial forestry (Rawlins, 1992; Været et al., 2009). Clulow et al. (2011) found a large variability in the evapotranspiration between the different indigenous vegetation types on the Eastern Shores. The swamp forest trees tend to lose water at a high rate constantly, while the trees of the drier areas have a rapid loss of water only immediately after a rainfall event. A. D. Clulow (pers. comm., 2011) suggests that the evaporative losses from the dune forest could be about 60% of the atmospheric demand (potential evapotranspiration). This contrasts strikingly with the losses projected for the pine and eucalyptus forest that can use up 300% of the atmospheric demand.

8.9 The dependence of Lake St Lucia on groundwater The changing groundwater profile during the various stages of the lake conditions due to prevailing weather systems and mouth dynamics can have a significant impact on the hydroecological conditions that are examined for the following scenarios.

8.9.1 Groundwater contribution to the water balance of Lake St Lucia and its ecological importance The persistent ‘base’ flow between storm events in rivers has been equated to the groundwater contribution. Ephemeral streams indicate a limited

availability of groundwater due to low groundwater storage, while perennial streams are indicative of significant groundwater storage. The St Lucia rivers with large catchments are generally perennial although they have dried up completely during drought periods. This is indicative of relatively low groundwater storage when compared with the streams in the coastal plain with much smaller catchments that seldom if ever stop flowing. Two estimates of the various components which illustrate the limited contribution that groundwater provides to the overall water balance of the lake have been summarized in Table 8.4. While these estimates vary

Groundwater hydrology

1200

Table 8.4. Water balance studies

Annual runoff

Annual runoff

Component

Hutchison, 1976 (Mm3 yr 1)

Direct precipitation

268

Average river runoff

249

Groundwater seepage

46

Seawater inflow

70

Van Niekerk, 2004 (Mm3 yr 1) (Virgin flow estimates in brackets)a

Inflow rates(106M3)

1000 Groundwater

800 Baseflow Total Ground Water (?)

600 400

Averaged groundwater inflow

200 0

397

Outflow through mouth

236

a

10

20 30 40 50 60 70 80 90 Frequency of flow rates exceeded (%)

100

FIGURE 8.9 Depth duration curve for the rainfall, runoff and groundwater.

Evaporation

Total

23.1 (45.5)

0

633

362 (418)

‘Virgin’ refers to natural land-use conditions in the catchments.

widely, they both indicate that the groundwater provides about 6–7% of the total volume. The groundwater contribution to the water balance of Lake St Lucia is derived from the baseflow in the rivers and from seepage faces along the lake shoreline. The storm and groundwater component of river runoff from the major rivers feeding into St Lucia has been described in Chapter 5. The direct seepage contribution has been estimated independently for the Eastern and Western Shores using numerical models by Kelbe et al. (1995); Taylor et al. (2006b) and Været et al. (2009). The contribution from the Mkhuze catchment is unknown. The estimated seepage rate into the lake for a 30-year simulation period (Kelbe et al., 1995) is shown in comparison with the estimated runoff from the rivers (including baseflow) in Figure 8.9. Although the groundwater only makes up a small

component of total hydrology under average conditions, this does change in times of drought when groundwater forms a much greater proportion of the total fresh water that enters the lake. Analysis of a small stream in Richards Bay (Figure 8.6) with a similar hydrological setting to the Nkazana Stream has shown that as much as 80% of the flow is derived from groundwater seepage and the remaining 20% from storm flow (Kelbe and Germishuyse, 2010). In these hydrological settings the groundwater storage is the dominant mechanism leading to highly persistent perennial streams.

8.9.2 Importance of groundwater in the maintenance of groundwater-dependent habitats along the shoreline of Lake St Lucia Possibly the greatest importance of the groundwater is in the maintenance of stable habitats along the lake shoreline. The persistent groundwater seepage creates stable and consistent habitats for freshwaterdependent ecosystems (see Chapter 11). These habitats form refugia for populations of certain species during periods of lake hypersalinity – and nodes from which recolonizsation of the lake can occur once conditions have improved. This is unusual in an estuarine system characterized by extreme variability in abiotic conditions.

165

166

Bruce E. Kelbe et al.

FIGURE 8.10 The last remaining pool for hippos in Tewate Bay during the drought of 2006. Hell’s Gate is visible across the dry lake bed. (Photo: Bruce E. Kelbe.)

There is little understanding of the upwelling of groundwater through the lake-bed sediments in nearshoreline sites. Observations indicate that emergent plants such as Schoenoplectus scirpoideus and Phragmites australis can grow standing in water that has a salinity that exceeds the tolerance level of the plant if upwelling occurs and hence the plant roots are in low-salinity water. Due to low water levels induced by the 2002 to 2011 drought, the groundwater seepage along the Eastern Shores created several ponds which provided refuge for several animal species during the harsh drought conditions when the lake levels were almost completely depleted. Almost a third (more than 300) of the entire hippo population resided in the groundwater seepage pond at Tewate Bay during the extreme dry conditions (Figure 8.10). North of Tewate Bay is an area with little groundwater seepage along the shoreline. Here drinking water has not been available during

droughts and many of the game animals from there have moved southwards to where water is available. These shoreline seepage ponds create the only wetland habitat in which several species of fish have been recorded (Vrdoljak and Hart, 2007). One of these shoreline groundwater-dependent sites yielded five fish species previously unrecorded in St Lucia on the Eastern Shores. The concentrated seepage zones such as the Nkazana Stream and Tewate Bay can provide freshwater refugia sites during very harsh lake conditions with salinities in excess of seawater (Vrdoljak and Hart, 2007). These water bodies have also proved to be of vital importance to the crocodiles of the St Lucia estuarine system during hypersaline times (see Chapter 17). However, a raised lake level will flood seepage sites – destroying the stable nature of these habitats in a similar concept to the shoreline as illustrated in Figure 8.8.

8.10 Mpate River catchment From the Western Shores, the Mpate River flow was incorporated into the general runoff estimates described earlier in the chapter. However, this

catchment has a much greater groundwater storage capacity than the large catchments to its north. It feeds directly into the Narrows where its contribution

Groundwater hydrology

is a significant proportion of the Narrows water balance when the St Lucia Mouth is closed. The change in water quality in the Narrows was simulated with the current levels of groundwater contribution to the Narrows including the Mpate (Kelbe and Taylor, 2005). The impact of halving and doubling the groundwater contribution showed significant changes in the peak salinity profiles.

The measured salinity peak prior to overtopping from the Mfolozi in November 2003 would have been significantly reduced if the groundwater contribution could have been doubled by changes in land-use management. A study is being undertaken to quantify the impact of deforestation due to a fire that destroyed a large section of the Mpate catchment.

8.11 Lessons learnt about groundwater which can be applied to other systems • In secondary porosity catchments the base flow of rivers is fed by groundwater. Groundwater depletion in these aquifers will cause rivers to dry up more frequently, affecting downstream users and affecting river ecosystems and the downstream estuary ecosystem. In the river catchments afforestation, the increased evaporation losses due to farm dams, the abstractions of water for irrigation and the increased transpiration losses of water due to woody plant encroachment all affect the groundwater. • In primary porosity catchments, the lowering of the groundwater table alters groundwaterdependent habitats as well as drying areas used for subsistence agriculture and making it necessary for wells that supply drinking water to be deeper. The main impacts are due to afforestation and the large-scale abstractions of groundwater. • The quality of the groundwater is affected by the addition of nutrients and the introduction of pollutants. The quality is affected by agricultural runoff and the leachate from mine dumps. In the sand aquifers the introduction of sewage into the groundwater from pit latrines and French drains is likely to affect its nutrient status. • The South African National Water Act recognizes groundwater as an important resource, and

provides powerful legislation for its control and use. However, there is a severe shortage of skills and capacity needed to monitor and manage groundwater as specified in the Act. The national government is under increasing pressure from interest groups for the abstraction of groundwater in various forms that will impact on the surrounding communities. A ten-year moratorium on the issuance of forestry licence applications is currently under review and if revoked could lead to large-scale forestry that would have major impacts on the surrounding communities and environment if the groundwater is significantly impacted. • Global change will affect the groundwater. Changing climate will alter rainfall regimes affecting the patterns and amount of groundwater recharge and discharge. Close to the coast, rising sea level affects the boundary conditions and hence the level of the groundwater. This affects coastal water bodies. • The work done at St Lucia highlights the importance of having long-term monitoring of the groundwater, the need to understand the groundwater systems through the development of conceptual models and the development of more formal numerical models.

167

Chapter contents 9.1 Introduction 9.2 Water depths and system partitioning 9.3 Salinity 9.4 Temperature 9.5 Turbidity 9.6 Irradiance 9.7 pH 9.8 Dissolved oxygen 9.9 Nutrient concentration 9.10 Sediment composition and size structure 9.11 Temporal variability and states of the St Lucia system

Wind patterns on the North Lake. Wind stirring prevents vertical stratification and sets up currents. (Photo: Ricky H. Taylor, April 2010.)

9

Physico-chemical environment Renzo Perissinotto, Nicola K. Carrasco and Ricky H. Taylor

9.1 Introduction The physico-chemical environment of the St Lucia Estuary is one of the most extreme of all estuaries in South Africa, at times driven by cyclone-induced floods and at others by a closed mouth for years as a result of prolonged drought. In the past, floods have occurred in 1955, 1963, 1975, 1984 and 1987. These have alternated with droughts from 1949 to 1951, from 1968 to 1972, and the most recent drought phase, which started in 2002. It has been suggested that these cyclical changes dictate the basic physico-chemical characteristics of the system (Begg, 1978) and freshwater input is, therefore, the single most important driving factor as it has the potential to influence an array of other environmental variables including salinity, temperature, water depth and mouth state. These variables, in turn, drive the biotic communities and determine the distribution of species and habitats within the entire ecosystem. In aquatic ecosystems, it is important to consider the physico-chemical aspects at several scales so that the impacts of local and global changes on the biota can be interpreted. The smallest is the scale at which chemical reactions take place (not considered here), then there is the organismal scale representing the physico-chemical environment in which an animal or plant is immersed. This includes the immediate salinity, temperature, light regime, turbidity, etc. to which an organism is exposed. The largest scale is that of the entire system and represents the physico-chemical patterns that govern

species distribution, such as the spatial pattern of salinity gradient from one end of the system to the other. The St Lucia system is also characterized by a marked spatial variability between its different basin components. It is at this level that complex interactions between physico-chemical parameters result in each major portion of the system assuming its unique identity. Of special relevance are contrasts between marine versus freshwater inflows, with periodical formation of reversed salinity gradients (Begg, 1978; Whitfield and Taylor, 2009). Contrasts between shallow versus deep areas are also important in determining light penetration, salinity and temperature, while differences between sheltered and exposed basins are more important in affecting turbidity levels and oxygen concentration, for instance. Temporal variability is also important at all spatial scales, as often the limitations are not only determined by the range of the conditions that change, but also by the rate at which the change takes place. In the highly variable estuarine habitat, the physico-chemical conditions should at times be stable enough to provide suitable conditions over a long enough period for populations to recruit, grow and reproduce, as well as for successions of species to take place. It is then necessary to determine which changes are anthropogenic in nature, so that management can be implemented to counter those that are natural. Temporal changes in the St Lucia estuarine system may take place over long time

Ecology and Conservation of Estuarine Ecosystems: Lake St Lucia as a Global Model, eds. Renzo Perissinotto, Derek D. Stretch and Ricky H. Taylor. Published by Cambridge University Press. © Cambridge University Press 2013.

170

Renzo Perissinotto et al.

periods, such as the cyclical wet and dry phases that are experienced in the region, or on shorter timescales where seasonality influences parameters such as temperature and salinity. The St Lucia estuarine system, therefore, exhibits an extreme spatial and temporal variability, with virtually all its physico-chemical properties shifting over a wide range in response to the alternation of dry and wet phases in the region. Such fluctuations can result in the system becoming at times ‘as saline as the Dead Sea’ and at other times completely limnetic (Begg, 1978). The extent of the most recent drought cycle (2002–2011) has been

magnified by a range of anthropogenic interventions which have been undertaken during the last century. In particular, the closed connection between the Mfolozi and the St Lucia Estuary has meant additional freshwater deprivation from the historic main water supply. The condition of the system is, therefore, much more severe and knowledge of the physico-chemical environment can be used to explain the distribution of the fauna and flora, as well as monitor the trends in the parameters to determine the effect of different driver changes, whether they be natural or artificial.

9.2 Water depths and system partitioning The St Lucia estuarine system is very shallow, having a mean depth of less than 1 metre (Hutchison and Midgley, 1978), and hence has a very large surface area to volume ratio that is particularly important in determining the balance between evaporative water losses and fresh/sea water inflows (Chapter 7) (Figure 9.1). This means that the system is very sensitive to changes in the balance between gains and losses of fresh water. Freshwater inflow (Chapter 5) plays a significant role in determining the state and functioning of St Lucia and the physicochemical conditions prevailing in the system. During periods of high freshwater inflow (wet phases), the system tends to form one homogeneous entity, with environmental parameters across the lake being fairly similar. During periods of low freshwater inflow (dry phases), in contrast, salinity and water depth are strong determinants sculpting the distribution of different habitats within the system. While periods of drought are not uncommon to St Lucia, the current state of Lake St Lucia is unprecedented (Cyrus et al., 2010a) and a direct consequence of the management decision to allow the mouth of the St Lucia Estuary to close, while artificially maintaining its separation from the Mfolozi (Lawrie and Stretch, 2010).

It is evident that under different management strategies, the system would have responded differently to the most recent drought (Chapter 7 and references therein). Nevertheless, under the current conditions, the shallowness of the system leads to poor water circulation and mixing, and reductions in water depth during periods of freshwater deprivation may lead to discontinuous water flow between the different basins (Figure 9.1). Under extreme conditions this results in different parts of the system being fragmented and physically separated from each other (Pillay and Perissinotto, 2008). During the peak of the most recent drought (2006), the South Lake was completely cut off from the Narrows and North Lake (Taylor, 2006), creating discrete habitats. Additionally, as precipitation rates decrease and evaporation rates increase, salinity levels rise, resulting in hypersaline conditions developing throughout the lake region of the system (Cyrus and Vivier, 2006b: Taylor, 2006). While the lower reaches of the estuary are relatively protected from the drought due to freshwater input from the Mpate and Mfolozi rivers (Whitfield and Taylor, 2009), hypersalinity and low water levels become increasingly more severe towards the north. This reversed salinity gradient and depth profile has

Physico-chemical environment

Rank variables Resemblance: D1 Euclidean distance

25

Distance

20 15 10 5

1

3

4

Stream outlet into False Bay

Oxbow

Mkhuze Mouth

Esengeni

Public Jetty

Mouth

Bridge

Dead Tree Bay

Lister’s Point Pond

Charter’s Creek

Catalina Bay 2

Caspian Point

Hell’s Gate

Makakatana

Fani’s Island

Lister’s Point

0

5

FIGURE 9.2 Bray–Curtis similarity dendrogram showing the subdivision of the St Lucia estuarine system into different basins, using physico-chemical properties measured during the period February 2005 to May 2011.

FIGURE 9.1 Surface area of the St Lucia Estuary under different climatic regimes. Light blue: after Cyclone Domoina (February 1984); dark blue: at average water levels (January 2001); black: during severe drought conditions (January 2005).

fractionated the system into a variety of different habitats in relatively close range to one another. During the most recent freshwater deprivation crisis that began in 2002, 14 stations were sampled regularly every 3 months, from August 2004 to October 2005, by researchers at the University of KwaZulu-Natal (Pillay and Perissinotto, 2008; Perissinotto et al., 2010a). After this initial phase of wide spatial cover, it was decided, on the basis of similarity analysis, that representative stations be chosen for future monitoring. Physico-chemical data (temperature, salinity, water depth, turbidity, pH, dissolved oxygen

and nutrients) collected during 2005 at each of the 14 sampled stations were averaged across seasons and ranked. An Euclidean distance matrix was then calculated and cluster analysis (group-averaged) was used to visually assess spatial differences in physicochemical parameters throughout the estuarine system. The outcome shows a clear subdivision of the 14 stations into five broad groups (Figure 9.2). Group 1 consists of Lister’s Point, Fani’s Island and Hell’s Gate. Both Lister’s Point and Hell’s Gate are situated in False Bay, while Fani’s Island is situated off North Lake. These stations are all characterized by high salinity and turbidity levels as well as low water depths. Group 2 includes Makakatana and Catalina Bay (Old Jetty). Both stations are located in the South Lake and are characterized by brackish salinities and relatively low turbidity levels. Group 3 is made up of Caspian Point, Charter’s Creek and Dead Tree Bay, which are all situated in the South Lake. Salinity levels recorded here are slightly higher than those recorded at the stations in the previous group and dissolved oxygen levels are also relatively low in these regions. The mouth, Bridge and Public Jetty stations all group together (Group 4), showing a high degree of similarity (Figure 9.2). These stations are generally characterized by low salinity and turbidity levels, as well as greater water depths. Esengeni and Oxbow, both situated in the

171

172

Renzo Perissinotto et al.

Narrows, cluster together in Group 5. Environmental parameters at these stations are similar to those recorded at the mouth, Bridge and Public Jetty, with the exception of grain size, which is much finer in the stations of the Narrows. Finally, Lister’s Point Pond, Mkhuze Mouth and the Stream Outlet into False Bay are all outliers, separated from the other stations because all three are freshwater bodies.

The above groupings illustrate the degree of fragmentation between the different basins within the lake system. On the basis of these clusters, one representative station was chosen from each group for future sampling and monitoring. These include: Lister’s Point in False Bay; Charter’s Creek and Catalina Bay on the west and east coast of South Lake, respectively; Esengeni in the Narrows; and the St Lucia Mouth.

9.3 Salinity Salinity is one of the most important factors in estuarine ecosystems, because it determines the osmotic physiology of organisms (McLachlan and Erasmus, 1974; Matthews and Fairweather, 2004). Virtually every aquatic organism is affected by changes in salinity, with stenohaline species capable of tolerating only a narrow range and euryhaline species a much broader one. At St Lucia, salinity concentrations have been measured for the past 60 years, providing a unique record of the range and spatial patterns in the system. Generally, salinity levels above 60 and below 15 are associated with mass mortality of estuary-dependent marine and oceanic species (Schmidt-Nielsen, 1997; Carrasco and Perissinotto, 2011a). Typical estuarine species, however, exhibit substantially larger thresholds. For instance, Grindley (1982) reported that all dominant copepods found in the St Lucia system in March 1970 were surviving at salinities of about 75–78 and 2–7, on the opposite ends of the range (Chapter 13). In contrast, most submerged and riparian vegetation appears to be suppressed already at salinity levels above 40 (Taylor et al., 2006a; Bate and Taylor, 2007). See Chapter 11 for salinity tolerance ranges. Historically, the St Lucia estuarine system has experienced quasi-decadal cycles in salinity, with development of hypersaline conditions alternating with return to limnetic states at regular intervals. In the lakes region, major peaks in hypersalinity

(salinity > 100) have been recorded during the periods 1970–1971, 1983–1984, 2003–2006 and 2009–2010 (Whitfield and Taylor, 2009; Cyrus et al., 2011). Conversely, oligohaline to limnetic conditions (salinity < 5) have occurred in 1964, 1976–1978, 1984–1986, 1988–1992, and 2000–2001 (Whitfield and Taylor, 2009; Chapter 7). During periods of open mouth conditions, salinities throughout the mouth region, the Narrows and the southern part of South Lake are essentially marine, because of the tidal influence. This influence may eventually extend to the northern lakes during dry periods, when no fresh water enters the system from the tributaries (Carrasco et al., 2010). Conversely, the whole system may become less saline during periods of heavy rain and freshwater runoff, even if tidal action persists. Sustained outflows, such as those observed in the wake of Cyclone Domoina in January 1984, may cause salinity drops by as much as 60% in a few days, with the entire system becoming dominated by limnetic conditions for extended periods of time (Cyrus, 1988b; Forbes and Hay, 1988). When the system, however, closes and becomes isolated from the ocean, a reversed salinity gradient is established. The St Lucia Estuary receives freshwater input through the Mpate and Mfolozi rivers, which enter the estuary in the region of the mouth and Narrows. Salinity levels during closed mouth conditions can, therefore, vary from near

Physico-chemical environment

fresh water at the mouth to > 200 at times in the northern regions. This situation has effectively persisted during the past 10 years (2005–2011), with only a brief period of interruption between March and August 2007. Superimposed on the effects of mouth state are also marked and often rapid changes in salinity related to seasonality. During the rainy season, which spans approximately from October to April, a sustained inflow of freshwater runoff from the catchment areas results in conditions remaining approximately meso/oligohaline (closed mouth) or euhaline/marine (open mouth) at the Mouth and the Narrows. In the lake basins, however, hypersaline levels can predominate even at this stage, especially in False Bay. During the rest of the year, when rainfall in the area virtually ceases, evaporative water losses far exceed the residual inflow of fresh water from the tributary rivers and hypersaline conditions in the upper reaches may attain extreme levels, recently in excess of 200 (Whitfield and Taylor, 2009; Figure 9.3A) under closed mouth states. During extreme hypersaline conditions (≥ 300), there is a sequence of precipitation of salts, for example gypsum salts drop out before sodium chloride; so there is a change in the composition of ions of the different elements as well as their concentration. During a recent monitoring study of a briefly interrupted closed phase, undertaken from February 2005 until May 2011 (Figure 9.3A), salinity values at the Mouth and Narrows have ranged from 1.8 to 36.1, while in the South Lake and North Lake/False Bay regions the range has varied from 10 to 81.2 and from 18.3 to 216, respectively – and then drying and depositing crystallized salt. Cases of extreme hypersalinity were prominent in the lakes region during this period, especially between August 2004 and March 2007 and again between February 2008 and November 2010. Salinity levels were much less extreme during the remainder of the survey, with fewer records of hypersaline conditions reported between April 2007 and November 2007, a period

coinciding with the anomalous breach of the mouth from the ocean side. During this latter period, at the mouth and in the Narrows, salinity ranged from 1.8 to 36.4, whereas in the South Lake and False Bay regions it ranged from 10.6 to 33.4 and from 18.3 to 42.9, respectively (Figure 9.3A). Sediment salts and salinity gradients across the sediment–water interface are also important properties affecting ecosystem function (Little, 2003). In the St Lucia estuarine system, exchanges between the two compartments are likely to be driven mainly by water-column turbulence, bioperturbation and diffusion processes. Measurements of salt load in the sediment of St Lucia were carried out in 2006, at the peak of the recent drought, when over 70% of the total lake surface was desiccated (Bate and Taylor, 2008). It was estimated that over two million tonnes of salt was contained in the upper 20 cm layer. By comparison, after the March 2007 mouth-breaching event, about 12 million tonnes of salt entered the estuarine lake when it was filled with seawater (Bate and Taylor, 2008). These authors estimated that prior to the closure of the mouth in 2002, the system contained a total of about five million tonnes of salt. Thus, there would have been a net loss of approximately three million tonnes between the closure of the estuary mouth and its subsequent large-scale desiccation. Through measurements of redissolved sediment salts, collected both within the lake surface and in dust blown ashore by the wind, it was established that the balance of salt could be accounted for by a combination of three main factors. These include: (a) losses via wind-blown dust (containing about 4% salt); (b) burial into the lake sediment (salt concentrations of 15–20 found in cores down to at least 35 m depth, according to unpublished work by Vogel and van Urk (1975); and (c) failure of some crystalline components to redissolve into the water column (insoluble portion, possibly including calcium and magnesium salts) (Bate and Taylor, 2008).

173

174

Renzo Perissinotto et al.

FIGURE 9.3 Variation in salinity (A), temperature (B), light extinction (C), turbidity (D), pH (E) and dissolved oxygen (F) at False Bay (grey symbols) and the mouth (black symbols) of the St Lucia system during the period August 2004 to May 2011.

Physico-chemical environment

9.4 Temperature Temperature not only determines the metabolic activity of virtually every aquatic organism, but also affects critical abiotic processes, such as the solubility of gases and solids and water density. The temperature range for most aquatic life is relatively narrow, particularly in tropical and subtropical regions (Longhurst, 1998). Although a few invertebrates can live at about 50
C, some cyanobacteria above 70
C and thermophilic bacteria and archaea even above the boiling point of water, most aquatic organisms generally are unable to tolerate temperatures in excess of 48
C (SchmidtNielsen, 1997). Similarly, at the lower end of the range a few specially adapted organisms can tolerate subfreezing temperatures, but most of them will die at temperatures far above the freezing point (Schmidt-Nielsen, 1997). In the St Lucia estuarine system, for instance, mass mortality of the cichlid fish Oreochromis mossambicus was recorded during a cold snap caused by a winter front passing through the area in July 2009 (R. H. Taylor, pers. obs.). A similar event was observed earlier in the nearby freshwater Lake Bhangazi South, with the same species dying in numbers at temperatures of between 10 and 15
C (Bruton and Taylor, 1979; Whitfield, 1998). During the earliest modern scientific survey undertaken in St Lucia, between July 1948 and July 1951, Day et al. (1954) reported that with an open mouth waters inside the estuarine system were slightly cooler than in the adjacent ocean in winter, but warmer than this in the summer. The annual average for the entire period was 23.9
C, with minimum and maximum values recorded of 17.5 and 32
C, respectively. As expected, greater variability and wider ranges were observed in the shallow lakes, where warming and cooling processes are rapid because of the larger surface:volume ratio than in the Narrows and mouth. Ranges at the time varied from 18.8 to 25.6
C at the mouth and from 20 to 30.2
C

in the northern lakes (Day et al., 1954). Overall cooler temperatures were reported for the period July 1964 to January 1965 by Millard and Broekhuysen (1970), with an average of 21.5
C and extremes of 13.5 and 38.5
C. The lowest temperature value available for the system appears to be 12
C, recorded at False Bay and South Lake in June 1976 (Blaber and Whitfield, 1976). Begg (1978) reported an average temperature of ± 29
C for mid-summer and ± 17
C for the mid-winter period, yielding a similar annual average of 23
C. He also observed that thermal stratification does not normally occur within the estuarine lake because of effective wind mixing in the lakes and tidal action in the lower reaches, at least under open mouth conditions. Forbes and Benfield (1986) reported the formation of a thermocline during flood tides in the lower part of the Narrows. However, differences between surface and bottom temperatures were never found to exceed about 2
C. It is possible though that some degree of stratification may occur in the deeper and more sheltered areas when the mouth of the estuary is closed. Stratification can also occur where mats of surface macrophytes develop, effectively preventing mixing. During the 2004–2011 monitoring period, temperature has ranged from 15.2 to 41.2
C, with the lowest values generally recorded from June to August (20.8 ± 0.25
C, SE) and the highest from November to February (28.4 ± 0.43
C, SE) (Figure 9.3B). Because of the shallowness of its water column, the lake region consistently exhibited the widest temperature ranges (15.2 to 41.2
C), while variations were more restricted in the deeper Narrows and mouth areas (16.2 to 32.3
C) (Figure 9.3B). No significant differences were ever recorded between surface and bottom temperatures in the lake basin during this period. However, in the Narrows and at the mouth differences in excess of 1
C between surface and bottom waters were observed on five occasions, with the largest of 1.83
C recorded at the mouth in February 2007.

175

176

Renzo Perissinotto et al.

9.5 Turbidity Turbidity is a measure of the degree to which the water loses its transparency due to the presence of suspended particulates. The greater the concentration of total suspended solids (i.e. phytoplankton, sediments from erosion, resuspended sediments from the bottom, waste discharge, algal growth, urban runoff) suspended in the water, the murkier it appears and the higher the turbidity. Particle size, shape and density play a key role in settling and flocculation rates and, therefore, together with turbulence, are critical factors controlling the turbidity of estuarine waters (McLusky and Elliott, 2004). Because of the size and shallow nature of St Lucia, wave action is affected by the fetch as well as water depth. As the system is shallow, the waves stir up sediment and the system is generally very turbid. Parts of the lake may, however, be sheltered. Submerged macrophytes play an important role in dampening the effects of waves and the emergent vegetation fringing the margins protects the shorelines from wave erosion (Chapter 7). As the orientation of the lake is in the direction of the prevailing winds, a change in wind direction can alter water level at either end of the lake by up to 50 cm within several hours (Taylor, 2006). This action causes the flooding and exposure of mudflats. In the St Lucia estuarine system, turbidity patterns were investigated in detail by Cyrus and Blaber (1988) during the period from January 1980 to June 1983, under open mouth conditions and tidal penetration. During this period, turbidity ranged from 2 to 568 NTU (Nephelometric Turbidity Units), averaging 43.2 ± 64 (SD) NTU. The major determinant of turbidity across the estuarine lake was found to be the wind, and turbidity gradients were observed stretching as far as 17.2 km from the mouth during flood tidal phases (Cyrus and Blaber, 1988). The same authors obtained a linear regression equation for the relationship between sediment load (SL, mg l1) and turbidity (NTU): SL ¼ 16.47 + 1.18

NTU; r2 ¼ 0.93, n ¼ 12. A classification of waters according to a turbidity scale was also proposed: (a) < 10 NTU ¼ clear water; (b) 10–50 NTU ¼ semi-turbid; (c) > 50 NTU ¼ turbid (Cyrus and Blaber, 1987a; Cyrus, 1988a). Fielding et al. (1991) described the St Lucia system as characterized by high turbidity levels, with measurements showing maxima of 250 NTU. This was attributed to silt-laden freshwater inflow from the tributary rivers and the ability of any windinduced turbulence to resuspend sediment over the shallow lake areas. It was also observed that turbidity was generally higher on the Western than on the Eastern Shores of the system and in the Narrows, because of the much higher contribution of silt to the total sediment composition in these areas. Carrasco et al. (2007) derived an empirical relationship between turbidity (NTU) and silt concentration (SC, g L1) using natural water and sediment from the St Lucia system. The linear regression obtained was: NTU ¼ 421.97 SC (g L1); r2 ¼ 0.98, p ¼ 0.001. Turbidity readings recorded at the St Lucia mouth and Mfolozi River were used as an indication of the range of silt concentrations that were to be used in these experiments. Measurements were taken in April 2006 over a period of about 3 h. Values ranged from 10.23 NTU ± 1.19 SD at the St Lucia Mouth to 2588 NTU ± 141.53 SD at the Mfolozi Mouth. During the period August 2004 to May 2011, turbidity in the St Lucia system ranged from 1 NTU at Catalina Bay to 951 NTU at Charter’s Creek (Figure 9.3D). The highest turbidities were generally recorded at Esengeni, Charter’s Creek and Lister’s Point (Carrasco et al., 2010). Turbidity values reported by MacKay et al. (2010) for the different parts of the estuarine lake range from 0.7 to 340 for North Lake, from 0.7 to 1350 for South Lake and 5.9 to 227 for the Narrows and mouth.

Physico-chemical environment

9.6 Irradiance Visible light, or photosynthetically available radiation (PAR, 400–750 nm), is essential for estuarine primary production. Before crossing the air–water interface light availability depends on factors such as seasonal and annual solar radiation, time of day and weather conditions. Within the water column, photon flux decays exponentially with depth, because of absorption and scattering by the water itself, but also by particles suspended within the medium. Algal photosynthesis is confined to the areas of water and sediment that experience a photon flux of at least 1% (down to 0.1% in highly adapted species) of the total available at the surface (Kirk, 1994). On the opposite end of the range, photoinhibition may occur when algae adapted to low irradiances are suddenly exposed to unusually high photon fluxes (Platt et al., 1980). The first measurements of light penetration in St Lucia waters were taken using a Secchi disc of approximately 15 cm diameter. For the period July 1948 to April 1951, Day et al. (1954) reported that the Secchi depth in the estuarine lake was in the range of 15 to 140 cm (average: 48.3 ± 40.6) during the winter (June–August; dry season) and 8 to 51 cm (average: 25.4 ± 15.9) during the summer months (December– February; rainy season). The significant difference between the two extreme seasons was attributed to the inflow of silt-laden waters during the rainy season, particularly from the Mfolozi River. At that stage, the region from the mouth (open) to the current road bridge (Narrows) exhibited the lowest

light penetration values observed within the entire system, apparently as a result of resuspension of fine sediments by tidal action and again resulting from the high silt load of the Mfolozi water (Day et al., 1954). Actual measurements of photon flux through the water column of the St Lucia estuarine system were taken at key stations during the recent investigations by the University of KwaZulu-Natal, from August 2004 to May 2011 (Perissinotto et al., 2010a). Within the estuarine lake, surface irradiance varied between 3.59 and 4173 μmol m2 s1, while bottom irradiance levels ranged between 0 and 2628 μmol m2 s1, depending on water depth, turbidity and cloud cover. Spatially, the amount of irradiance reaching the bottom of the water column was highest in the lake region, with average values of 744 ± 78.6 μmol m2 s1 (¼ 62% of surface irradiance) recorded in South Lake and 762 ± 102 μmol m2 s1 (¼ 56% of surface irradiance) in North Lake/False Bay. In the deeper regions of the St Lucia Mouth and the Narrows, bottom irradiance averaged 341 ± 52.9 μmol m2 s1 (¼ 23% of surface irradiance) and 39.2 ± 20.4 μmol m2 s1 (¼ 3.61% of surface irradiance), respectively. Light attenuation coefficients (Kd, m1) ranged from 0.01 to 51.9 m1, peaking in the South Lake region of the system (Figure 9.3C). Coefficients varied between 4.21 and 14.8 at the mouth and Narrows, between 0.01 and 51.9 in South Lake and between 0.9 and 33.2 in the North Lake/False Bay area.

9.7 pH In estuarine/marine waters, the pH generally ranges between 7.5 and 8.5 and changes in its value are related to the balance in the carbonate–bicarbonate system, which is to a large extent regulated by photosynthesis and respiration (Parsons et al., 1984).

The carbonate–bicarbonate system keeps waters slightly alkaline through the process of buffering. Buffering also ensures that waters do not become too acidic or too alkaline, through reversibility in the reactions. So, when the pH of the water increases the

177

178

Renzo Perissinotto et al.

reaction shifts in a direction that allows carbonic acid to release hydrogen ions, thereby causing a drop in pH. Conversely, when the water pH starts decreasing bicarbonate recombines with hydrogen ions, causing pH to rise. These relatively rapid adjustments are normally enough to maintain a very limited range in pH, to within ± 1 unit (Schmidt-Nielsen, 1997). The recent global escalation in CO2 emission into the atmosphere is, however, also causing a net increase in the amount of gas entering seawater, with a consequent shift of the reaction towards the release of further hydrogen ions and, therefore, increase in acidity (Hall-Spencer et al., 2008; Marshall et al., 2008; Chapter 21). A whole range of estuarine biological processes is affected by shifts in pH values. For instance, shell formation in organisms that utilize calcium carbonate as a protective hard structure is negatively impacted by water becoming more acidic, because buffering reactions deprive the medium of carbonate ions that would otherwise be available to combine with calcium ions to form the calcium carbonate (CaCO3) precipitate (Marshall et al., 2008). Also, photosynthetic and respiratory processes can in turn have major effects on pH values in estuarine and coastal waters, particularly in shallow and isolated pools. Tidal rock pools, for instance, experience marked diel cycles in pH changes in response to shifts between photosynthesis and respiration in their algal communities (Dejours, 1981). The two opposing

physiological processes cause alternation of CO2 (mainly removed from bicarbonate) depletion in the water during daytime, with consequent alkaline shift in the pH to over 10 and CO2 release at night leading to drastic drops in pH to about 7 (Dejours, 1981). Very few measurements of pH have been reported in the literature from St Lucia so far and little attention has been given to pH changes within the system. In their early work at St Lucia, Millard and Broekhuysen (1970) reported that waters within the estuarine lake were generally alkaline, ranging from an average of about 7 in summer to 9 in winter. During the recent monitoring survey from August 2004 to May 2011, pH values at the mouth and Narrows have ranged from 6.63 to 8.99 (average 8.21 ± 0.43 SD), while in the South Lake and North Lake/ False Bay regions the range has varied from 7.45 to 9.2 (average 8.45 ± 0.37 SD and from 5.04 to 9.03 (average 8.18 ± 0.63), respectively, remaining generally within the range of 6.9 to 9.2 (Figure 9.3E). Significant exceptions to this pattern were recorded in May 2007 at Lister’s Point (False Bay), during the brief open phase, when pH values reached the lowest observed during the entire study, at 5.04 (Figure 9.3E). The causes for this acidic event may have been related to resuspension of hydrogen sulphide from the highly anoxic sediments characteristic of the False Bay area (Taylor, 2006).

9.8 Dissolved oxygen Well-mixed, shallow estuarine waters are generally saturated with oxygen. Cold water holds more, however, and in deep pools and channels, hypoxic to anoxic conditions (< 1 mg O2 l1) may develop regularly, particularly in the presence of vertical stratification and accumulation of organic material with a high biological oxygen demand (BOD). Anthropogenic eutrophication of the water column, through massive nitrogen and phosphorus inputs,

often leads to anoxia, as phytoplankton blooms strip the water of nutrients causing a rapid escalation in respiratory oxygen requirements at night (Livingston, 2001). This is often compounded by the subsequent collapse of the ageing bloom, which sinks to the bottom of the estuary where its decomposition drastically increases the BOD of the system. This process is particularly rapid and detrimental when the mouth of an estuary is closed. As water

Physico-chemical environment

circulation becomes restricted, water-column stability is enhanced and residence time is sufficiently large to maximize nutrient uptake by algae (Thomas et al., 2005; Gale et al., 2006; Perissinotto et al., 2010a). Gas exchanges at the water–air interface are expected to be generally effective and large within Lake St Lucia, because of the large surface:volume ratio and the regular wind-induced turbulence that characterize the area. However, in the deeper and sheltered parts of the estuary, such as the Narrows and the mouth, gas exchanges may become periodically reduced and oxygen concentrations may drop well below saturation levels (Figure 9.3F). High temperatures may also contribute to suboptimal oxygen concentrations in the shallowest areas and during the hottest part of the year, as gas solubility decreases with increasing water temperature. This can contribute to fish die-offs. Since August 2004, absolute and relative oxygen content in the water column has been highly variable, ranging from 0.06 to 13.4 mg l1 and 0.9% to 212% saturation, respectively, over the survey period. The average values for the estuarine lake

during the period August 2004 to May 2011 are 7.19 ± 2.78 (SD) mg l1 and 97 ± 34.6 (SD) % saturation, indicating a generally well aerated water column. However, hypoxic to anoxic conditions were recorded at Lister’s Point (False Bay) in November 2004, February 2005 and November 2006 as well as Hell’s Gate (North Lake) in February 2005 (Figure 9.3F). These are areas that prior to the measurements had experienced mass mortality in both benthic and pelagic communities, as a result of the extreme hypersaline conditions prevailing at the time. In the two extreme basins, oxygen levels during the monitoring period ranged between 1.9 and 13.4 mg l1 (average: 7.47 ± 2.02 (SD) mg l1) at the mouth and between 0.5 and 10.8 mg l1 (average: 6.68 ± 2.55 (SD) mg l1) in the False Bay region (Figure 9.3F). While occasional differences were obtained at the mouth and the Narrows, with O2 values decreasing from surface to bottom, concentrations were normally uniform through the shallow water column of the lakes region. It would be interesting in the future to take measurements of oxygen concentration within the sediment and within submerged macrophyte beds.

9.9 Nutrient concentration Dissolved inorganic nitrogen (DIN: sum of nitrates, nitrites, ammonia and urea) and phosphorus (DIP: mainly orthophosphates) are the key macronutrients controlling primary production in subtropical estuaries (Perissinotto et al., 2010a). Estuarine productivity can experience limitation either from lack of nitrogen or phosphorus, or even co-limitation of the two elements (Conley et al., 2009). Estuarine waters normally receive large inputs of nutrients from processes such as catchment runoff, detrital remineralization within bottom sediments and tidal exchange with the ocean. In addition to this, anthropogenic eutrophication during the past few decades has contributed dramatically to the increase in nutrient loading of estuarine waters, mainly through use of fertilizers in

cultivated floodplains and residential and industrial wastewater disposal (Conley et al., 2009). Nutrient recycling is rapid and efficient in estuaries, particularly under closed or constricted mouth conditions, and it is estimated that nutrients brought into the system are reused three to fourfold before being exported (McLusky and Elliott, 2004). Exchanges with the ocean are in most cases balanced, in the sense that most estuaries receive from tidal inflow as much as they lose during outflow. However, in the case of the St Lucia estuarine system, and similar estuaries located on the eastern seaboard of southern Africa, there is a net flow of nutrients from the estuary to the Indian Ocean, because the Agulhas Current is notably an oligotrophic system (Perissinotto et al., 2010a).

179

180

Renzo Perissinotto et al.

The nutrient load of False Bay and its tributary rivers was investigated by Johnson (1976) during a wholeyear survey in 1975. Prior to that only a few measurements had been taken at selected rivers (e.g. Nyalazi, Hluhluwe, Mkhuze) during a snapshot study of the National Institute for Water Research (NIWR, 1969). Results showed that in the tributaries DIP levels were low (< 0.65 μM) almost during the whole of 1975, with a mid-summer peak of 0.8–1.2 μM. Higher values were, however, recorded within False Bay, where maximum DIP concentrations reached 1.32 μM, thereby suggesting that sediments within the lake itself may have been releasing phosphorus at the time (Johnson, 1976). By comparison, levels recorded during the NIWR river survey in 1969 ranged from 4.52 μM in the Hluhluwe to 7.74 μM in the Nyalazi. DIN concentrations at the same time ranged between 3.86 and 14.3 μM within False Bay, reaching peak levels of 44.3 μM in the Nyalazi and 42.5 μM in the Hluhluwe River (Johnson, 1976). Interestingly, while lake nutrient values generally increased with increasing freshwater inflow from the rivers, in the rivers themselves concentrations were often highest under low flow rates. Using algal bioassays, Johnson (1976) concluded that DIP was the primary limiting nutrient between January and July 1975, while both DIN and DIP were co-limiting productivity between August and December. During the period August 2004 to May 2011, watercolumn DIN values at St Lucia ranged between 0.001 and 770 μM, while DIP ranged between 0.0001 and 15.14 μM (Figure 9.4A, B). The highest DIN concentrations occurred from November 2004 to February 2005 in the lake basins and particularly in False Bay, under the most severe hypersaline conditions and the lowest water levels recorded through the entire study. The lowest DIN levels were observed in August 2004 and again during May 2007 (open phase) throughout the lake system. Regarding DIP, highest levels were measured in November 2009 and May 2010 in the False Bay region, while the lowest were observed in August 2004 and March 2008 virtually throughout the estuarine lake. In contrast, sediment pore-water concentrations of nutrients were about fivefold higher than water-column

values, ranging between 0.23 and 1178 μM for DIN and 0.03 and 36.8 μM for DIP (Figure 9.4C, D). Highest DIN levels were observed in August 2005 at False Bay and in June 2006 in the Narrows, while DIP was consistently higher in South Lake, in August 2007 and May 2010 and again in March 2007. The lowest levels of pore-water DIN and DIP occurred throughout the system in December 2009 and November 2010 and in August 2005 and March 2008, respectively (Figure 9.4C, D). DIN:DIP molar ratios were above the Redfield optimum of 16:1 for 50% of the time, indicating a potential for P-limitation during a period mainly from June to November. The ratio was also below the Redfield value about 48% of the time, suggesting the potential for N-limitation mainly from February to May (Perissinotto et al., 2010a). During the 2006–2007 primary productivity study undertaken by van der Molen and Perissinotto (2011), DIN levels declined gradually over the investigation period, while DIP increased sharply after mouth opening in March 2007. The inflow of seawater into the estuary may have been the cause of major resuspension of sediment phosphorus into the water column, particularly at the mouth and in the Narrows that had previously maintained oligo- to mesohaline characteristics. Here, increased salinity levels would have provided sulphate for microbial reduction within the sediment, thereby releasing phosphorus that was previously stored in the benthic subsystem. Phosphorus previously adsorbed in clay particles would have also been released into the water column by the sudden intrusion of seawater (Conley et al., 2009). Although there are few indications of pollution at present, Mackay (1993) identified St Lucia as increasingly facing the threat of pollution and eutrophication. She conducted an audit of the various land uses in the catchment areas and the possible pollutants associated with each activity. Potential pollutants are heavy metals from mining, agricultural runoff (especially of pesticides and fertilizers) and the local impacts from boating and tourism. As little is known about this, Mackay (1993) has recommended that monitoring of water quality should be initiated.

Physico-chemical environment

FIGURE 9.4 Variation in (A) water-column DIN, (B) water-column DIP, (C) pore-water DIN, (D) pore-water DIP concentrations and (E) in water-column DIN:DIP ratio at False Bay (grey symbols) and the mouth (black symbols) of the St Lucia system during the period August 2004 to May 2011.

181

182

Renzo Perissinotto et al.

9.10 Sediment composition and size structure The structure and composition of bottom sediments play a crucial role in estuarine ecosystems, as they determine the availability of interstitial water, oxygen, nutrients and solid surfaces, among others (McLusky and Elliott, 2004). Sediments at the surface are generally well oxidated and white/yellow to brown in colour, while deep reduced sediments are normally black. Intermediate between these two extremes is a greyish layer, representing the point of rapid decrease in redox potential and known as redox potential discontinuity or RPD (Little, 2003). The depth of this RPD is determined by the penetration of freshly oxygenated water into the sediment. So, the more compact and fine the sediment is, the less the amount of oxygen capable of reaching the layers below the surface will be and, consequently, the closer to the surface the RPD will be. Conversely, coarse and porous sediments will be aerated more thoroughly and the RDP will, therefore, be positioned much deeper down (Little, 2003). In estuarine sediments the RDP effectively limits the benthic fauna, by excluding species that are unable to form burrows and obtain oxygen from overlying waters (McLusky and Elliott, 2004). Sediment characteristics also affect the availability of interstitial space for benthic organisms, which in turn determines the level of bioturbation and the degree to which the structure and chemistry of the sediment itself is changed by biotic factors. In their early investigation of the St Lucia system, Day et al. (1954) described it as characterized by drastically different sediment types dominating the various component regions. The mouth region (open to the ocean) exhibited mainly ‘clean, hard sand’, while further up towards the bridge the bottom was described as rapidly changing from sand to ‘soft slurry mud’ through ‘sandy mud’ (Day et al., 1954). Deposition of Mfolozi silt in the form of ‘soft glutinous mud’ in this area was reported as choking the channel, thereby ‘smothering the aquatic vegetation and bottom fauna and restricting the tidal flow which keeps the mouth scoured and open’ (Day et al., 1954). Above the bridge,

the bottom was described as dominated by muddy sand, as a result of decreasing influence of the Mfolozi waters and deposition of its mud. The Western Shores of the lake reportedly exhibited a varying bottom structure, from soft mud to hard sand with rocky outcrops interspersed throughout. Lister’s Point in False Bay was described as having a ‘soupy mud’ bottom. The Eastern Shores, conversely, were typified by a ‘sandy bottom’, with overlying clean waters (Day et al., 1954). This was confirmed by Fortuin (1992), who presented a detailed map of the sediments of St Lucia

FIGURE 9.5 Sediment grain size distribution pattern in the St Lucia estuarine system, as derived from the survey undertaken by Fortuin in 1992.

Physico-chemical environment

100

% contribution

80

60

40

20

0 Mouth

Esengeni

Catalina Bay Charter’s Creek Lister’s Point

Gravel

Very coarse sand

Coarse sand

Fine sand

Very fine sand

Mud

Medium sand

FIGURE 9.6 Sediment composition, as percentage dry mass of each fraction, in the different basins of the St Lucia estuarine system during July 2011. Gravel: > 2000 μm; very coarse sand: 2000–1000 μm; coarse sand: 1000–500 μm; medium sand: 500–250 μm; fine sand: 250–125 μm; very fine sand: 125–63 μm; mud (silt): < 63 μm.

(Figure 9.5). His report is based on a collection of 300 samples from throughout the system, except False Bay and the Narrows, and provides the grain size distribution patterns in the system.

More recently the sediment granulometry of the system has been reported by Pillay and Perissinotto (2008) and MacKay et al. (2010). Results show that sediments in the mouth region are composed of fine sand (125–250 μm), becoming more muddy in the northern parts of the Narrows. Sediments in the southern half of South Lake are a mixture of coarse (250–500 μm) and fine sand, but become finer further north into the system, being classified there as fine (< 63 μm) to coarse silt (63–125 μm). The latest granulometric analysis conducted in July 2011 shows a clear dominance of mud/silt in the sediment of the Narrows (Esengeni, 75% of total) and False Bay (Lister’s Point, 83% of total) (Figure 9.6). Sediment in the mouth area, as observed previously, is composed mainly of medium (63.8% of total) and fine sand (26.9% of total), while at Charter’s Creek and Catalina Bay this is codominated by medium and fine sand, with 39.8– 46.2% and 43–44% of the total, respectively. However, while at Charter’s Creek there are also significant amounts of mud (4.62%) and of coarse sand (5.16%), at Catalina Bay it is very fine sand that makes up virtually the full balance with 8.7% of the total (S. J. Bownes, unpubl. data).

9.11 Temporal variability and states of the St Lucia system Throughout its history, St Lucia has experienced alternating periods of wet and dry phases which last between 4 and 10 years at a time (Begg, 1978). During periods of high freshwater input, lake water levels increase and salinity levels decrease to near freshwater conditions, resulting in the estuary shifting towards a limnetic state. In contrast, periods of low freshwater input and high evaporation rates result in low lake levels when the mouth is closed and the estuary is separated from the Mfolozi (Chapter 7). During open-mouth conditions, seawater would then flow into the estuary, increasing the water and salinity levels to that of seawater, and becoming hypersaline in the north. Closed-mouth conditions,

however, result in the shift towards a hypersaline state as the salt is concentrated. Closed-mouth conditions are also characterized by a reversed salinity gradient as the system receives most of its freshwater input in the region of the mouth and Narrows. The high salinity conditions in the northern part of the system are also a result of the larger surface:volume ratio that prevails in that region, compared with South Lake. These processes are schematically shown in Figure 9.7. During the most recent drought (2002–2011), the estuary has experienced three different mouth states. A closed mouth state, which persisted from June 2002 to February 2007, an open mouth state from

183

184

Renzo Perissinotto et al.

Dry period

EXTREME drought or flood CONDITIONS

Wet period

Large scale mortality and loss of diversity

Increased freshwater inflow

Decreased freshwater inflow

Lower evaporation rates

Higher evaporation rates

Increased lake water levels

Decreased lake water levels

OPEN MOUTH

CLOSED MOUTH

Decreased seawater inflow

Freshwater surplus

Low salinities

Low salinities

Lake water level above sea level and rises to overtopping

Lake water level slightly above sea level

LIMNETIC STATE

OPEN MOUTH

CLOSED MOUTH

Increased seawater inflow

Desiccation

Reversed salinity gradient

Reversed salinity gradient

Lake water level slightly below sea level

Lake water level well below sea level

HYPERSALINE STATE

FIGURE 9.7 Conceptual diagram representing shifting states in response to the alternation of climatic cycles from drought to floods in St Lucia (revised after Begg, 1978). The desiccation branch in this figure is artificially created by management interventions, i.e. separating the Mfolozi from St Lucia, while allowing the mouth to close (see Chapter 7).

March to August 2007 and a re-closed mouth state from November 2007 up until current time (2012). Data from an on-going monitoring programme were used to identify the key determinants of these states and the dynamics of the shifts involved in the transition from one state to the other. Physicochemical data collected from February 2005 to May 2011 were subjected to principal component analysis (PCA), in order to get an indication of what environmental parameters are mainly responsible for shaping the different basin environments during different mouth states (i.e. closed, open and reclosed) (Figure 9.8). During the closed mouth state, the St Lucia Mouth and Esengeni show substantial overlap in the PCA plot, with salinity and water depth as the two main environmental parameters separating these stations from the others. The mouth and Esengeni are both characterized by deeper waters (~1 m) and salinity levels which approximate oligohaline to limnetic conditions. Catalina Bay, Charter’s Creek and Lister’s Point by contrast are all very shallow stations (~0.2 m) with high salinity levels. It is, therefore, not surprising that temperature and salinity are the two variables that best explain

the distribution of these stations. Not only are shallow waters more susceptible to heating, but increased evaporation leads to further salinity increases. These stations are also situated lower down in the plot, showing a closer association to turbidity than the mouth and Esengeni, reflecting the high turbidity levels experienced in the lake (Figure 9.8). The PCA plot of open mouth conditions does not show much grouping between stations. Instead, there appears to be an even distribution throughout the plot. This is typical of an open mouth state in the St Lucia Estuary (Chapter 7), as with increased seawater inflow, salinity levels across the lake approximate seawater conditions (~ 35). Water levels within the lakes will also increase during this state, decreasing the discrepancy between sites in terms of environmental characteristics (Figure 9.8). During ‘re-closed’ mouth conditions, PCA results showed strong groupings, particularly between the mouth and Esengeni as well as between Charter’s Creek and Lister’s Point. The environmental parameters that best explain the distribution at the mouth and Esengeni are water depth and dissolved oxygen. Lister’s Point and Charter’s Creek, in contrast, are situated quite far from these parameters, again highlighting the shallow water levels which are typical of these areas. Turbidity and salinity also stand out as factors separating these stations from the others. Lister’s Point and Charter’s Creek are both characterized by fine sandy substratum, which is easily resuspended by wind action resulting in higher turbidity levels. Catalina Bay is positioned roughly in the centre of all these stations. Salinity levels here are slightly lower than those observed in the rest of the lake. Catalina Bay also has a more coarse sediment composition, resulting in relatively low turbidity levels in comparison with those recorded at Charter’s Creek and Lister’s Point. Water depths are, however, usually very low (~ 0.2 m) here, explaining the greater distance from the depth parameter in the PCA plot. The lake stations again showed a greater association with temperature, reflecting the warmer water temperatures which are generally experienced here (Figure 9.8).

Physico-chemical environment

2

4

CLOSED

2

OPEN

Depth

0 0 Temperature

pH

Depth Salinity

DO

–2

DO Turbidity

–6

PC2

Salinity

–4

pH

–2

0 PCI

2

4

Turbidity

Depth

–2 Temperature

Turbidity

pH

–4 –4

Salinity DO

Temperature

PC2

PC2

Mouth Esengeni Catalina Bay Charter’s Creek Lister’s Point

2

0

–2

Station

RE-CLOSED

–4 –4

–2

0 PCI

2

4

–4

–2

0 PCI

2

4

FIGURE 9.8 Principal component analysis showing the contribution of the different environmental variables to the difference between representative stations of the St Lucia estuarine system during the current dry phase (February 2005–May 2011).

As is evident from the above, the St Lucia estuarine system is characterized by a marked spatial variability between its different basin components, which differ depending on the state of the system. At times this state is driven by cyclone-induced floods and at others by a closed mouth for years, as a result of prolonged drought. In recent years (2002–2012), however, drought conditions have been exacerbated by the artificial diversion of the Mfolozi waters, historically the main freshwater supply to St Lucia (see also Chapter 7 and references therein), resulting in the recent drought being the most severe. The range in values of the physico-chemical variables within as well as between the different basin environments are an expression of these extreme conditions. Most notable are the salinity levels, which may vary between fresh water and crystallizing brine from one end of the lake system to the other. While an inverse salinity gradient during closed mouth conditions is rare for estuaries, it is not unique to St Lucia alone. The Coorong in Australia is

a choked coastal lagoon and since most of the freshwater input occurs through the barrages at the same end as its connection to the sea, the Coorong acts as an inverse estuary. During periods of low flow, salinity levels in the lagoon can increase to well over 100 (Webster, 2010). As in St Lucia, recent decades have seen a decline in the biodiversity of the Coorong and it is thought that the alterations to its ecology are primarily due to the high salinity levels and changes to water level regime (Paton et al., 2009). Physico-chemical variables are known to drive the biotic communities and determine the distribution of species and habitats within the entire ecosystem. However, the extent of variation, especially on some of the short temporal timescales that are experienced in these systems, have the potential to exclude large components of the system biodiversity, as not only do the extreme conditions exceed the tolerance limits of certain organisms, but more importantly the timescale of change is often shorter than the rate of recovery.

Acknowledgements This study was funded using grants from the National Research Foundation (NRF, Pretoria), Marine & Coastal Management (MCM, Cape Town), the World Wide Fund for Nature (WWF-SA, Stellenbosch) and the South Africa–Netherlands Research Programme

on Alternatives in Development (SANPAD, Durban). We are very grateful to the management and staff of Ezemvelo KZN Wildlife and the iSimangaliso Wetland Park for providing logistical support throughout the study.

185

Chapter contents 10.1 Introduction 10.2 Diversity and community structure 10.3 Biomass 10.4 Productivity 10.5 Exceptional blooms

Hippo path across drought-exposed shoreline. Note the water coloured by microalgae. (Photo: Xander Combrink.)

10

Microalgae Renzo Perissinotto, Guy C. Bate and David G. Muir

10.1 Introduction Microalgae are key primary producers in shallow lakes, where they may alternate their dominance with macrophytes depending on the preponderance of turbid versus clear water conditions (Moss, 1994; van der Berg et al., 1998). Both water-column (i.e. phytoplankton) and sediment-associated microalgae (i.e. microphytobenthos) serve crucial ecological functions, by providing a link between inorganic compounds and organic matter available to higher trophic levels and top predators (Mortazavi et al., 2000). High benthic microalgal biomass with low primary production and low phytoplankton biomass, with high rates of pelagic primary production, have often been observed in shallow estuaries, particularly in those experiencing a regular opening and closing of the mouth (Perissinotto et al., 2003). Possible explanations for this have included: (a) more favourable light conditions in the water column compared with the sediment surface; (b) higher grazing impact by zooplankton in the water column; and (c) settling of phytoplankton cells on the sediment (Anandraj et al., 2007). Major changes in phytoplankton and microphytobenthic standing stocks, community structure and productivity are strongly related to the alternation of open and closed phases in estuaries (Perissinotto et al., 2002; Anandraj et al., 2007). For instance, in the East Kleinemonde Estuary (Eastern Cape, South Africa), the phytoplankton community may be dominated by diatoms during the closed phase, but by

dinoflagellates and cryptophytes during the subsequent open phase (Whitfield et al., 2008). Other studies have shown that nutrients might be the main limiting factor for phytoplankton growth during the closed phase, since light penetration is generally greatest during this period (Nozais et al., 2001). In contrast, low light might be limiting phytoplankton growth during the open phase, since at this stage turbidity and nutrient loading increase in response to inflow of fresh water from land and tidal exchange with the ocean (Thomas et al., 2005). Microphytobenthic standing stock also exhibits strong fluctuations, with the lowest values coinciding with the open phase (Nozais et al., 2001). However, unlike phytoplankton, in shallow systems benthic microalgal biomass does not appear to be inhibited by either low water-column nutrient concentration or light availability (Perissinotto et al., 2006). Despite the generally good level of knowledge available for the higher trophic levels of the St Lucia Estuary, very few studies have so far been conducted on the organisms that are at the base of its food webs. Phytoplankton studies that were undertaken in this estuary are limited to the taxonomic works of Cholnoky (1968), Grindley and Heydorn (1970) and Millard and Broekhuysen (1970). Johnson (1976, 1977) completed a study on pelagic cell volumes, while Fielding et al. (1991) undertook the first measurements of chlorophyll-a biomass. More recent, holistic investigations have included both benthic and pelagic microalgae, either in taxonomic

Ecology and Conservation of Estuarine Ecosystems: Lake St Lucia as a Global Model, eds. Renzo Perissinotto, Derek D. Stretch and Ricky H. Taylor. Published by Cambridge University Press. © Cambridge University Press 2013.

188

Renzo Perissinotto et al.

(Bate and Smailes, 2008), biomass/abundance (Perissinotto et al., 2010a; Muir and Perissinotto

2011) or productivity (van der Molen and Perissinotto, 2011) studies.

10.2 Diversity and community structure The diverse habitats of Lake St Lucia are subject to environmental change on varying temporal scales, and it might be expected that the diversity of microalgae would be correspondingly great. There are, however, few long-term quantitative studies for any group and none for many groups that may be important, particularly the Dinophyceae, Chlorophyceae and prokaryotes. This is despite the fact that cyanobacterial/diatom mats are a prominent feature of much of the benthos, particularly under hypersaline regimes. Table 10.1 presents all the recorded microalgal species in the St Lucia estuarine system, barring those recorded by Cholnoky (1968) for freshwater streams flowing into the lakes. What is immediately apparent is the paucity of data relating to groups other than the diatoms. Only recently has some attention been directed to the cyanobacteria and other groups.

10.2.1 Cyanobacteria Millard and Broekhuysen (1970) and Johnson (1976, 1977) noted cyanobacterial species in the system, identifying a few common benthic and planktonic species. Recently, this list has been expanded somewhat, but identifications presented here are tentative, and based on morphologies shown in Silva and Pienaar (2000). The shores and shallow benthos of South Lake, North Lake and False Bay often have extensive cyanobacterial mats with a complex associated eukaryotic community. These mats persist through periods of extreme environmental change, exposed to high temperatures, intense insolation and hypersalinity. At salinity up to 100, they are composed of several filamentous species such as Leptolyngbya sp., Anabaenopsis arnoldii, Lyngbya sp., and other members of the Oscillatoriales, commonly

mixed with colonial or solitary species such as Stanieria sp., Chroococcus sp., Gloeothece sp. and Aphanothece sp. The mats are often several millimetres thick, of a dense, gel-like consistency. Oxygen is often trapped within this matrix to the extent that portions of the mat lift and tear away, and are commonly seen floating on the water surface. Under extreme hypersaline conditions, such as those observed in North Lake and False Bay in 2010 (Chapters 7 and 9), observations suggest that the diversity of the mats decreased, and they became dominated by Leptolyngbya sp., with very few active diatoms or other eukaryotes persisting (D. G. Muir, pers. obs.). In 2011, when the system was flooded by heavy rains, the dense mats established during the preceding drought were stripped en masse from the substrate and cast up on the shore, where they dried and formed the basis for an explosion of beetle species (mainly the staphylinid Bledius pilicollis), the larvae of which fed on them (Carrasco and Perissinotto, 2012). In the plankton, particularly at salinities around that of seawater, Myxosarcina sp., Spirulina sp., Oscillatoria sp. and very small unicellular genera (suggestive of Synechococcus and Prochlorococcus) are commonly found and may persist as hypersaline conditions begin. During extreme hypersalinity (> 150), unicellular forms such as Synechococcus and Cyanothece predominate in the plankton and may reach bloom status (Muir and Perissinotto, 2011). Many of the cyanobacteria mentioned above are diazotrophic, and though their role in the nutrient dynamics of the St Lucia system is entirely unknown, it is fair to assume that they are of considerable importance, particularly as a source of new fixed nitrogen to the system.

Microalgae

Table 10.1. Microalgal species recorded in Lake St Lucia Group

Species

Habitat

Sample locality

Salinity range

Author

Cyanophyceae

Anabaenopsis arnoldii

Pl

FB, NL

33~39

IJ

Aphanothece sp.

B/Pl

FB, NL, SL

35~200

M*

Chroococcus sp.

Pl

FB

8 ~14

IJ

Cyanosarcina sp.

Pl

Ub

10–100

M*

Cyanothece sp.

Pl

FB, NL, SL

40~200

MP

Gloeothece sp.

Pl/B

Ub

20~40

M*

Leptolyngbya sp.

B

FB, NL, SL

35~200

M*

Lyngbya conferroides

B

Ch, SL

7.6–29

MB

Microcoleus chthonplastes

B

Ch, SL

7.6–29

MB

Spirulina sp.

B, Pl

Ub

10–150

M*

Stanieria sp.

B

FB, NL, SL

35~200

M*

Synechococcus bacillaris

Pl

FB, NL

8~40

IJ

Synechococcus leopoldensis

Pl

FB, NL, SL, M

4~50

IJ

Noctiluca scintillans

Pl

FB, NL

40~50

Gr

Prorocentrum micans

Pl

FB

42–48

Gr

Unidentified (probably several spp.)

Pl

FB, NL, SL

5~39

IJ

Chrysophysceae

Dictyocha fibula

Pl

M

30~34

IJ

Haptophyceae

Halopappus ? adriaticus

Pl

FB

32–44

IJ

Chlorophyceae

? Chlamydomonad

Pl

SL, NL, FB

20–100

M*

Euglenoids

Water’s edge aggregations

M

~10

M*

Nannochloris atomus

Pl

FB, NL, SL

35

IJ

Pediastrum, Cosmarium, Gonium

Pl, B

FB

1  105 cells ml 1) over a wide range of salinity values and were commonly seen in phytoplankton samples collected recently (D. G. Muir, pers. obs.). It is likely that several species occur, but these remain unidentified, despite their evident importance in the phytoplankton. There have, however, been occasions when dinoflagellate blooms have occurred, as described by Grindley and Heydorn (1970) and discussed below.

10.2.3 Haptophyceae and Chrysophyceae Johnson (1976) mentioned the occurrence of small populations of the marine coccolithophorid Halopappus adriaticus in False Bay (salinity: 32–42)

and the silicoflagellate Dictyocha fibula, limited to the immediate environs of the St Lucia Mouth. Both are marine species and their occurrence was probably fortuitous and resulting from prolonged mouth opening.

10.2.4 Chlorophyceae Phytoflagellates were noted throughout the St Lucia system and are frequently abundant in the phytoplankton (Johnson, 1976, 1977), but for the most part remain unidentified. Johnson (1977) tentatively identified only one (Nannochloris atomus), at salinity close to that of seawater, but tolerating slightly higher salinity as well, in South Lake, North Lake and False Bay. In more recent work, however, a larger, unidentified biflagellate chlorophyte (resembling Chlamydomonas sp.) with a

Microalgae

distinct cell wall was also often seen and has been observed to persist in salinity up to 100, and many other phytoflagellates of varying morphology are commonly seen, particularly in live samples (D. G. Muir, pers. obs.). At the mouth, transient patches of euglenoids often occur, forming brilliant green patches at the surface of the damp sand at the water edge. This may be due to the limnetic conditions prevailing in this area and throughout the Narrows during closed mouth states (Chapter 9). It is very likely that there are many other phytoflagellates that await identification.

10.2.5 Bacillariophyceae The diatoms have received the closest study of all the microalgal groups in St Lucia and reports have been made of the benthic (Cholnoky, 1968; Bate and Smailes, 2008), epiphytic (Gordon et al., 2008) and planktonic populations (Johnson, 1976, 1977; Bate and Smailes, 2008). Given their wide separation in time, the various studies provide some insight into changes in diatom populations under different environmental conditions, and in particular the impact of salinity on community structure, but sampling and analytical methods differ so greatly from one another that comparison is difficult. Cholnoky (1968) was the first to examine the planktonic and benthic diatoms of St Lucia, although the planktonic studies related specifically to the streams flowing into the lake. The salinity gradient in 1965, when Cholnoky was collecting his samples, was seawater (35) at the mouth dropping to c. 10 in the lakes. This represented a relatively normal situation during climatic conditions of average rainfall, when the mouth remained open most of the time. Cholnoky (1968) sampled 15 freshwater streams, eight sites in False Bay, six sites on the Eastern Shores affected by fresh groundwater seepage, three sites on the Western Shores, three in the Narrows and two at the St Lucia Mouth. Apart from the fresh stream water (which will not be considered further here), his sampling was of the benthos. He identified 60 common taxa in total for the whole estuary and

lake system (in addition to 52 which occurred extremely rarely). At the mouth, 26 common taxa were found, 26 in the Narrows, 32 in South Lake, 26 in the North Lake and 32 in False Bay. Of these, 13 taxa were ubiquitous, while others were clearly limited by their salinity tolerance, with distinct communities being particularly marked at the mouth and Narrows (and of marine origin) and others (presumably more stenohaline species) in the lakes area (Table 10.1). The next study to be conducted was the survey of the phytoplankton of St Lucia by Johnson (1976), which was published as an M.Sc. thesis in 1977. Immediately prior to her study, the estuary had been subject to a prolonged drought with associated hypersaline conditions, which ended as her sampling commenced (she measured salinities between 25 and 35). She recorded 47 taxa, but it should be noted that 39 of these are more properly regarded as benthic species. Particularly in the lake regions, the shallowness of the water and very active wind-driven water movement lead to constant resuspension of settled cells: only eight of her recorded taxa are truly planktonic. In 2004–2005, Bate and Smailes (2008) conducted a diatom survey during a drought, which had commenced in 2002. This had led to a salinity reversal, with hypersaline conditions (> 100) in False Bay and the North Lake, and < 12 at the mouth (Chapters 7 and 9). They recorded a total of 69 taxa. At the same time, Gordon et al. (2008) studied epiphytic communities on the predominant macrophytes, Stuckenia pectinata and Ruppia cirrhosa. What is remarkable about these studies is the dissimilarity of the communities they describe. Johnson (1976) recorded only eight taxa from Cholnoky’s study (1968), while Bate and Smailes (2008) recorded only two in common with Johnson (1976) and ten with Cholnoky (1968) (Table 10.1). Gordon et al. (2008) recorded 21 taxa, of which 12 are unique to their study, the others being benthic species also recorded by the other authors mentioned. Only one species (Nitzschia lorenziana) is recorded

199

200

Renzo Perissinotto et al.

ubiquitously throughout the system by all except the Gordon et al. (2008) study. It is evident that the structure of diatom communities in the estuarine system is highly mutable and that salinity is the principal driver of diatom population structure. Cholnoky (1968) believed that almost all diatoms in St Lucia are of marine origin. Of 79 species recorded in the streams supplying the lake, only four were recorded in the lake itself, these being identified as brackish water species. In the absence of mouth

opening, different communities may still survive conditions adverse to them, by burrowing or occupying refugia from which new populations are seeded under advantageous environmental conditions, adding to the complexity of succession patterns in the system (Bate and Smailes, 2008). From the foregoing, it appears that the diatoms found in Lake St Lucia are well adapted to the wide salinity variations that occur periodically as a result of weather extremes.

10.3 Biomass The first measurements of chlorophyll-based microalgal biomass for the St Lucia estuarine system were reported by Fielding et al. (1991) for the period 1987–1988, when there were differences in salinity between three sampling periods, but no great salinity gradient on each occasion. Chlorophyll-a (chl-a) levels obtained ranged from hardly detectable to just over 16 mg m–3, with the highest average recorded in February 1988 at False Bay and the lowest during September 1987 in the region of the Narrows (Fielding et al., 1991). On the basis of the methodology employed in the analysis (spectrophotometry), Fielding et al. (1991) suggested that their values could probably be regarded as overestimates, by comparison with results obtained using more accurate techniques, such as highperformance liquid chromatography (HPLC). Also, they could not detect any significant difference between chl-a levels on the turbid Western Shores and those of the much clearer waters on the Eastern Shores. Salinity could also not have been a factor determining chl-a variations in that study, as there was virtually no salinity gradient in the system at the time (Fielding et al., 1991). More recently, microalgal biomass in the water column (phytoplankton) and sediment (microphytobenthos) have been measured since August 2004, with a minimum of five and a maximum of 14 stations being sampled regularly every three months to identify the mechanisms

responsible for spatio-temporal patterns in microalgae across the whole system (Perissinotto et al., 2010a). During this period, St Lucia watercolumn chl-a concentrations have averaged 23.3 ± 39.1 mg m–3, with the highest value of 413 mg m–3 recorded at Lister’s Point in False Bay just prior to the mouth-breaching event of March 2007 (Perissinotto et al., 2010a) (Figure 10.1). The average value is 1.5-fold higher than the maximum chl-a value obtained by Fielding et al. (1991). Exceptional bloom levels (> 100 mg m–3) were observed on 11 occasions between August 2004 and May 2011. Within the sediment, the average microphytobenthic chl-a concentration measured at St Lucia during this study has been 147 ± 332 mg m–2, with the highest value of 2576 mg m–2 recorded at Charter’s Creek, also in March 2007 (Figure 10.1). Both pelagic and sediment maxima are among the highest values so far reported in the literature for any estuarine ecosystem, including those impacted by eutrophication (Lukatelich and McComb, 1986; Cloern, 1996; Adams et al., 1999; Underwood and Kromkamp, 1999; Livingston, 2001; McLusky and Elliott, 2004; Thomas et al., 2005; Perissinotto et al., 2006). These are by far the highest values reported for any South African estuary to date. The phytoplankton maximum is about four times greater than maximum values recorded for permanently open systems (Adams et al., 1999) and up to 50 times greater than

Microalgae

FIGURE 10.1 Surface phytoplankton (0–10 cm upper layer) and microphytobenthic (upper 1 cm of sediment) chlorophyll-a measured at the five representative stations of the St Lucia Estuary during the period August 2004 to May 2011. Vertical bars represent ± SE; continuous top horizontal line indicates period of mouth closure, while dotted part indicates open mouth conditions. PPL, phytoplankton; MPB, microphytobenthos.

values observed in smaller temporarily open/closed systems (Nozais et al., 2001; Perissinotto et al., 2002). Similarly, the highest microphytobenthic biomass values in the St Lucia Estuary were between five and seven times greater than those reported for temporarily open/closed estuaries (TOCEs) (Nozais et al., 2001; Perissinotto et al., 2002), and up to 11 times greater than reported for the permanently open

systems (Adams et al., 1999). The most probable reason for these high values is that, with St Lucia being an estuarine lake rather than a normal permanently or temporarily open system that is periodically flushed to a greater or lesser extent with river water, there is the possibility for microalgal cells to remain within the system. In addition, the nutrient concentrations were higher than those

201

202

Renzo Perissinotto et al.

normally found in other estuaries. A possible reason for the higher nutrient levels is that St Lucia is not a closed system with respect to nutrients, because large numbers of hippo feed on the shore at night and import ‘new’ nutrients into the water body during the day. In addition, the density of diazotrophic populations of cyanobacteria appears to have become of particular importance during the drought period, and this may be contributing to the high total dissolved nitrate, nitrite and ammonium concentrations observed (Chapter 9). Average values are well within the range of those reported from other parts of the world, in similar estuarine types. For instance, in the shallow PeelHarvey Estuary of Western Australia average benthic microalgal biomass concentrations of 100–200 mg m–2 have been recorded regularly (Lukatelich and McComb, 1986). Mean global estuarine values fall in a range from 10 to 500 mg m–2, with the highest of 500 mg m–2 reported from the Bay of Fundy, Canada (McLusky and Elliott, 2004). Values much higher than this maximum have often been obtained in the upper reaches of the St Lucia estuarine system in recent years. Apart from the absolute maximum measured at Charter’s Creek in March 2007, another 12 values exceeding 0.5 g m–2 were also recorded in the regions of False Bay, North and South lakes between August 2004 and March 2007 (Perissinotto et al., 2010a) (Figure 10.1). Measurements taken with a BBE Fluoroprobe (Moldaenke) have indicated that the bulk of this exceptional microphytobenthic chl-a concentration is related to the formation of thick mats of cyanobacteria, which are particularly well adapted to the hypersaline conditions that regularly occur in the area (e.g. Muir and Perissinotto, 2011). Regarding phytoplankton biomass, the values obtained at St Lucia are above those reported globally for estuaries (1.1–22 mg chl-a m–3, McLusky and Elliott, 2004), but maxima are within the range of values obtained under dense bloom and/or hypereutrophic conditions (Livingston, 2001; Allanson and Baird, 2008). When integrated over the entire water column, phytoplankton biomass levels ranged between 0.02 and 537 mg m–2 (average 31 ± 55 mg m–2), very similar to those of the Peel-Harvey

Estuary, which also occasionally attained maxima of about 600 mg m–2 (Lukatelich and McComb, 1986). As is often the case for shallow South African TOCEs (e.g. Nozais et al., 2001), biomass values in the water column at St Lucia were generally lower than on the sediment throughout the lakes region (False Bay, North and South lakes). However, in the deeper areas of the Narrows and the mouth, it was phytoplankton that provided the largest contribution to the total areal biomass of microalgae (Perissinotto et al., 2010a). This, however, may only be the case under closed mouth and low salinity conditions. In terms of size-class contribution, the phytoplankton assemblages of the estuary during the period of the investigation have been dominated by nanoplankton (2–20 μm), with an average of about 64% of total chl-a, followed by microplankton (> 20 μm) and picoplankton (< 2 μm), each with about 18% of the total (Perissinotto et al., 2010a). Clear patterns from ordinations and cluster analyses emerged in the study undertaken by Perissinotto et al. (2010a), showing that phytoplankton biomass was spatially distinct during the period August 2004 to August 2007, with the mouth, Narrows and South Lake forming a similar group and the North Lake and False Bay a second, separate entity. During the same period, microphytobenthic biomass at the mouth and Narrows (Cluster 1) was generally different to that observed in the other areas, with most of the South and North lakes forming a separate group (Cluster 2) and False Bay essentially standing on its own (Cluster 3; Perissinotto et al., 2010a). During the same study period, phytoplankton biomass showed positive and significant correlation with water level, oxygen content and surface irradiance. Negative correlations were significant with salinity, pH, turbidity and water-column DIP, probably indicating that phosphate was readily taken up by microalgal photosynthetic activity. Microphytobenthos, conversely, was positively related to salinity and water-column irradiance, but negatively to water depth. Water depth and turbidity are key factors in the control of microalgal productivity through their effect on light availability,

Microalgae

which is the real proxy factor (Kirk, 1994). While a shallow water depth enhances light penetration in parts of the lakes region, the large and exposed water surface results in high turbidity generated by the strong frontal winds and even by the normal breeze that blows during daytime from the ocean (Fielding et al., 1991; Chapter 9). Wind-driven, lateral water circulation is a regular occurrence in the shallow lake areas and appears to be responsible for some of the changes in microalgae biomass that have been measured and the patterns that are formed. Temperature and salinity also play a major role in the microalgal biomass distribution within the St Lucia Estuary. Temperature is mostly controlled by water depth, with deeper areas being least sensitive to fluctuations but shallowest parts reaching almost 50  C during the hottest summer days (Chapter 9). Because of the extreme large water surface:volume ratio characterizing the shallow lake basins (depth < 0.2 m, during the study period), hypersaline conditions prevail regularly during droughts when evaporation exceeds precipitation. Both temperature and salinity can have serious negative impacts on microalgal biomass at such extreme values, but hypersaline levels of between 100 and 200 can, for instance, promote dense algal blooms of halotolerant/halophylic species, particularly cyanobacteria (Muir and Perissinotto, 2011). Grazing by invertebrates is regarded as the main top-down factor affecting microalgal biomass in estuaries and other aquatic ecosystems (Nichols, 1985; Kibirige and Perissinotto, 2003a). In some TOCEs, zooplankton can graze up to 70% of the available phytoplankton biomass and even in excess of 100% of its production (Kibirige and Perissinotto, 2003a). In the St Lucia estuarine system, both zooplankton and macrofauna abundance and

biomass have been found to be negatively correlated with microphytobenthic chl-a concentration on occasions (Pillay and Perissinotto, 2008; Carrasco et al., 2010; Chapters 12 and 13). The contribution of epiphytic microalgae to the overall biomass and food webs of the St Lucia estuarine system was investigated only recently, during a period of severe drought, from November 2004 to October 2005 (Gordon et al., 2008). Epiphytes play an important role in aquatic ecosystems, by providing an alternative linkage to free microalgae in food web structures (e.g. Winning et al., 1999). Their nutritional value is generally higher than that of the submerged macrophytes to which they are attached and they enter the food chain through selective feeding or accidental ingestion (Ziemann and Wetzel, 1980). They can also become part of the detrital pathway through the breakdown of the host plants. In the St Lucia estuarine system, at the time of the study, epiphytes were associated with the only two dominant macrophyte species remaining in the estuarine lake: Ruppia cirrhosa (mainly in South Lake) and Stuckenia pectinata (virtually restricted to the Narrows). Results show that epiphyte chl-a biomass was highest at the peak of the wet season, in February 2005 (88 ± 20 SE mg chl-a m 2 of macrophyte leaf area), and lowest at the end of dry season, in October 2005 (15 ± 0.9 SE mg chl-a m 2 of macrophyte leaf area) (Gordon et al., 2008). Marked spatial differences were also evident, with chl-a concentrations in South Lake exhibiting significantly higher values than in the Narrows. However, this appeared to have been mainly the result of prolific growth of epiphytic filamentous macroalgae (Cladophora and/or Ulva), rather than microalgae, on R. cirrhosa in the South Lake (Gordon et al., 2008).

10.4 Productivity Fielding et al. (1991) provided coarse estimates of phytoplankton production derived from the empirical equation of Cole and Cloern (1987). Values were

obtained from actual measurements of chl-a biomass, euphotic depth and photosynthetically available radiation and ranged from 218 to 252 mg C m 2 d 1

203

204

Renzo Perissinotto et al.

FIGURE 10.2 Temporal variation in water-column (A) and sediment surface (B) primary production at the five representative stations of the St Lucia Estuary during the period June 2006–May 2007. Vertical bars represent ± SE.

(Fielding et al., 1991), which is equivalent to approximately 18–21 mg C m 2 h 1 under the assumption of a 12 h daylight period. Recently though, van der Molen and Perissinotto (2011) have shown that there was no significant correlation between microalgal biomass and productivity during a drought year in the St Lucia Estuary, suggesting that the use of biomass data as a proxy for production may not generally produce accurate estimates of production for this estuarine system (Herlory et al., 2005). A lack of correlation between microalgal productivity and biomass may indicate either that biomass produced locally is rapidly consumed and incorporated into food webs (i.e. high productivity, low biomass), or alternatively that such biomass is not utilized while accumulating (i.e. low productivity, high biomass). Direct, in situ primary productivity measurements at St Lucia were undertaken during the study of van der Molen and Perissinotto (2011) in June 2006 to May 2007. Values obtained in this study ranged from 0 to 180 mg C m 2 h 1 in the pelagic, and from 0 to 34 mg

C m 2 h 1 in the benthic subsystem (Figure 10.2). These rates are comparable to the values obtained recently from small temporarily open/closed estuaries (TOCEs) in the region (Anandraj et al., 2007), except that pelagic values seem to be lower at St Lucia than in TOCEs, while the opposite applies to benthic production. For instance, in the Mpenjati Estuary on the KwaZulu-Natal (KZN) South Coast, pelagic primary production ranged from 51 to 216 mg C m 2 h 1 and in the Mdloti Estuary, on the KZN North Coast, from 1 to 340 mg C m 2 h 1 (Anandraj et al., 2007). However, benthic microalgal productivity was in the range of 0.2–6 mg C m 2 h 1 for the Mpenjati and 0–16 mg C m 2 h 1 for the Mdloti (Anandraj et al., 2007). In the global context, maximum daily pelagic production measured at St Lucia (about 2.16 g C m 2, assuming a 12:12 light:dark ratio) would be on the lower range of the maxima recorded for estuarine bays and lakes in North America and Europe (0.55–9 g C m 2; McLusky and Elliott, 2004). The highest benthic microalgal production rates are, however, similar to those measured in other systems around

Microalgae

the world, including those reported from Langebaan Lagoon, on the Cape West Coast (Fielding et al., 1988; McLusky and Elliott, 2004). In terms of average gross production rates, microphytobenthic values reported so far from St Lucia are in the order of 10.3 ± 9.9 mg C m 2 h 1. This is close to the levels obtained from shallow coastal sediments in several systems worldwide (Cibic et al., 2008), but lower than rates obtained from tidal and exposed mudflats in western Europe, which are generally > 20 with peaks of about 200–300 mg C m 2 h 1 (Spilmont et al., 2005; Forster et al., 2006). Van der Molen and Perissinotto (2011) showed that salinity, irradiance and temperature largely influence the rate of microalgal production in the St Lucia estuarine system. This could indicate a seasonal pattern in the productivity of the water column, although the system appears to be more affected by the alternation of dry and wet cycles rather than seasonal changes (Chapter 9). A significant impact, although limited to benthic microalgal productivity, was detected after the seaward breaching event experienced by the St Lucia Mouth in March 2007. After the berm at the mouth was overtopped during a severe storm and seawater flooded in, the water level in the lakes rose rapidly, with most of the previously exposed mudflats becoming submerged under at least 10 cm of water. These changes were promptly followed by a several-fold increase in benthic productivity, from 4.13 ± 5.44 mg C m 2 h 1 prior to the breach to 15.4 ± 12.6 mg C m 2 h 1 one month after the event (Figure 10.2), while pelagic productivity remained virtually unchanged (van der Molen and Perissinotto, 2011). Despite the drought conditions prevailing during the primary production study, nutrient concentrations in the St Lucia estuarine system were generally well above the limiting thresholds of about

2 and 0.2 μM for dissolved inorganic nitrogen (DIN) and phosphorus (DIP), respectively (Fisher et al., 1992). However, DIN:DIP ratios were greater than the Redfield optimum of 16 about 60% of the time and 27% of the time lower than this value. Hence, there is a potential for microalgal production in St Lucia becoming limited by availability of either P or N for a substantial part of the year. However, during the period August 2005 to August 2007, threshold values capable of limiting production were observed only 15% (N) to 33% (P) of the time in the water column and 1.5% and 5.7% of the time in the pore water (Perissinotto et al., 2010a). Primary production measurements suggest that P-limitation may occur predominantly from June to November and Nlimitation from February to May (van der Molen and Perissinotto, 2011). A summer productivity peak would thus occur during the period of transition from P- to N-limitation (Figure 10.2). Productivity-irradiance models specific for this geographical region still need to be generated and applied, in order to estimate production for the entire estuarine system over larger timescales, involving both dry and wet climatic cycles (Meyercordt et al., 1999). Also, in a shallow aquatic system such as St Lucia, wind-induced turbulence causes large and constant exchanges of microalgal biomass between benthic and pelagic compartments (Scheffer, 1998). It remains to be clarified, therefore, to what extent primary productivity measured in the water column is actually due to phytoplankton, rather than to resuspended benthic microalgae. Taxonomic studies on the composition of the microalgae in each of the two subsystems could be instrumental in defining the relative contribution of the phytoplankton and microphytobenthos functional groups to the total production of the system.

10.5 Exceptional blooms Nuisance blooms of microalgae have been identified worldwide as an increasingly common phenomenon, particularly in enclosed water bodies and usually

associated with exogenous anthropogenic input, but also with the effects of salinity increases and global climate change (Paerl and Huisman, 2009).

205

206

Renzo Perissinotto et al.

FIGURE 10.3 Aerial view of the Cyanothece bloom recorded in the North Island Channel, near the Mkhuze Mouth, on 20 July 2009. (Photo: courtesy of R. H. Taylor.)

There are three such blooms on record in St Lucia, all occurring in the False Bay/North Lake area. A red water bloom, attributed to the marine dinoflagellate Noctiluca scintillans, was reported by Grindley and Heydorn (1970) during July and August 1969. It is probable that the seed population was marine, brought into the lake during prolonged mouth opening. The cause of the bloom was not clear, but water was hypersaline in the three months immediately prior to the bloom (max salinity: 89 in False Bay; 71 in South Lake) but fell to 42 as the bloom appeared. In April, ‘exceptional bioluminescence’ was noted in the South Lake, but may have been due to another smaller organism, since Noctiluca was not noted in net tows. In July 1969, a three week period of unusually low winds occurred and was accompanied by the appearance of pink streaks of aggregated Noctiluca and bioluminescence at night in North Lake and False Bay. The onset of south-westerly winds moved the blooms northward and concentrated them before they decayed and dispersed, only to reappear briefly after a second windless period in August and then finally dissipate as the salinity rose from 48 to 70 by

December. The bloom was accompanied by a minor fish kill, attributed to deoxygenation of the water by the decaying cells (Grindley and Heydorn, 1970). Another bloom, extending throughout the North Lake and False Bay, was reported by Ezemvelo KZN Wildlife staff in July 2009. This bloom had a characteristic red/orange coloration and was thus easily observed (Figure 10.3). It remained in False Bay and North Lake until February 2011. The bloom was associated with exceptionally high salinity throughout the northern reaches as a result of the extended drought period. From July 2008 to December 2010, for example, the salinity never fell below 60 and on two occasions rose above 200. The water level throughout this period fell considerably and the intrinsic development of the bloom by growth was exacerbated by its concentration as the volume of water fell. The organism responsible was identified, by both morphological and 16S rRNA sequencing, as the cyanobacterium Cyanothece sp. (Muir and Perissinotto, 2011; Figure 10.4). The genus is commonly associated with hypersaline situations such as saltworks (Roussomoustakaki and Anagnostides, 1991), salt pans (De Philippis et al.,

Microalgae

FIGURE 10.4 Phase contrast micrograph of bloom-forming Cyanothece cell, with typical star-shaped granular cytoplasm concentrated around the centre of the cell and vacuoles at the periphery.

1993) and solar lakes (Margheri et al., 1999) but has also been found in marine waters (Welsh et al., 2008). In this instance, our own research suggests that its highest growth rate occurs when the salinity is close to that of seawater and that it is truly extremely halotolerant, rather than halophilic, and probably of marine origin. The Cyanothece bloom persisted

throughout the period of hypersalinity in the northern reaches, until exceptionally heavy rains at the end of 2010 led to a fall in salinity to the level of seawater once more, at which point the bloom dissipated: it constituted one of the longest cyanobacterial blooms on record worldwide (Muir and Perissinotto, 2011). At the time of writing (October 2011), a third bloom has been identified at Lister’s Point, False Bay. This is of a very small cyanobacterium which forms a 2.0 μm (approx.) coccobacillus immediately before division, with newly divided cells being 1.0 μm diameter cocci (D. G. Muir, pers. obs.). It is likely that this is Synechococcus sp., which has been recorded worldwide as a bloom species in lagoons and estuaries (Philips et al., 1999; Alvarez et al., 2006; Schapira et al., 2010) and was also noted by Johnson (1976). Its current population is about 2.0  105 ml 1. It has been suggested that hypersaline conditions trigger the appearance of these blooms, and that the driver for their development is the reduction of populations of grazers such as zooplankton and ciliates (Muir and Perissinotto, 2011). It may be that there is also a requirement for nutrient augmentation to move the bloom organism to exponential growth, but to date knowledge about nutrient dynamics of the lake systems is limited and remains the subject of on-going research (Chapter 9).

Acknowledgements This study was funded using grants from the National Research Foundation (NRF, Pretoria), Marine & Coastal Management (MCM, Cape Town), the World Wide Fund for Nature (WWF-SA, Stellenbosch) and the South Africa–Netherlands Research Programme

on Alternatives in Development (SANPAD, Durban). We are very grateful to the management and staff of the iSimangaliso Wetland Park and Ezemvelo KZN Wildlife, for providing administrative, logistical and operational support during the study.

207

Chapter contents 11.1 Introduction 11.2 Distribution of macrophytes 11.3 Abiotic factors and macrophyte response 11.4 Recent studies on response to drought and mouth closure 11.5 Summary of changes in macrophyte habitats 11.6 Conservation and management 11.7 Future research

Mflolozi mangrove forest with bed of pneumatophores in foreground. (Photo: Lynette Clennell, July 2012.)

11

Macrophytes Janine B. Adams, Sibulele Nondoda and Ricky H. Taylor

11.1 Introduction The St Lucia estuarine system has six major macrophyte habitats: mangroves, fringing reeds and sedges, saline lawns, succulent salt marsh and submerged plants. All of these are important as primary producers; they produce organic material detritus, modify the physical environment and create a variety of habitats for other estuarine biota. Submerged macrophytes provide a substratum for epiphytes, which in turn provide food for invertebrate fauna and refuge for juvenile fish. The extensive reed and sedge habitats serve as breeding areas for waterbirds. These plants are highly effective at absorbing or removing large quantities of nutrients from the water, which then return to the system as detritus or are exported by smoke when they burn.

Unique crabs and invertebrates are associated with the mangrove habitat, which also serves as a nursery area for juvenile fish and crustaceans. Macrophytes play an important role in carbon sequestration, wave attenuation, bank stabilization, shoreline protection, sediment trapping, turbidity reduction, nutrient cycling and nutrient export (Adams et al., 1999). The salt marsh grasses, Paspalum vaginatum Swartz and Sporobolus virginicus (L.) Kunth. are grazed by a variety of animals such as hippos and ducks. Macrophytes in the St Lucia estuarine system are important indicators of ecosystem health. The knowledge on their growth and distribution in response to changes in environmental conditions is presented in this chapter.

11.2 Distribution of macrophytes The first references to macrophytes in Lake St Lucia were made by Von Bonde (1940) and Day (1948), when they referred to the beds of submerged macrophytes in the Brodie’s Crossing area as the ‘weed problem’. These were likely pondweed, Stuckenia pectinata (L.) Bo¨erner, which was considered to hinder the migration of fish (Day et al., 1954). Also some mention was made of mangroves. Ward (1976) was the first to describe the St Lucia system vegetation in detail and this covered the period 1962 to 1976. Details on the ecology and distribution of submerged

macrophytes and shoreline vegetation were outlined. Taylor et al. (2006a) described four primary habitats: (a) open water; (b) intertidal shorelines; (c) ‘dry’ shorelines and islands, where there is evaporation of saline water as well as exposure to desiccation; and (d) ‘wet’ shorelines, the groundwater-dependent habitat where the effects of high salinity are moderated by groundwater seepage. A recent mapping exercise used similar terminology (Nondoda, 2012). Aerial photographs (July and August 2008) from Land Resources International (Pty) Ltd were used

Ecology and Conservation of Estuarine Ecosystems: Lake St Lucia as a Global Model, eds. Renzo Perissinotto, Derek D. Stretch and Ricky H. Taylor. Published by Cambridge University Press. © Cambridge University Press 2013.

210

Janine B. Adams et al.

with ArcGIS 9.3 software to map the distribution of the different macrophyte habitats (Figure 11.1). Table 11.1 indicates the mapping units as well as the area covered by each habitat type. The boundary of the map for the north, east and western shores is the floodplain within the 5 m contour line, and to the south the boundary is the north bank of the Mfolozi River and floodplain. Mapping is influenced by the boundary used, but as long as this is clearly defined it can be revised by others in the future. Salt marsh with halophytic grasses and succulents covered close to 1000 ha, whereas mangroves occupied an area of only 305 ha. Grasses and shrubs covered 2300 ha, while the common reed Phragmites australis occurred in a variety of non-tidal, groundwaterdependent and Mkhuze Swamps habitats, with an approximate area cover of 6500 ha. Extending from this southern boundary, a map is available of the vegetation of the Mfolozi–Msunduzi floodplain (Nondoda et al., 2011), but is not dealt with in this chapter. .

11.2.1 Submerged macrophytes Submerged macrophytes are rooted in the substratum and appear to be confined to those portions of the estuary where there is sand or reasonably consolidated mud, and not in sediments that are too soft and unstable to anchor the plants (Taylor et al., 2006a). There are three main species of submerged macrophytes: Stuckenia pectinata, Ruppia cirrhosa Petagna (Grande) and Zostera capensis Setch. Stuckenia prefers salinity lower than seawater (< 20) and when conditions are favourable it spreads rapidly to form ‘doughnut’ circles (Figure 11.2). Since 2002, closed mouth and low salinity conditions in the Narrows have resulted in large beds forming in this area. When the lake is fresh, and where the rivers enter the estuary, R. maritima L., Najas marina subsp. armata (Lindb. f.) Horn, and Lamprothamnion papulosum (K. Wallroth) J. Groves are found. Ruppia maritima has been observed in the shallow parts of Catalina Bay and Charter’s Creek and in shallow pools on the islands. During Day

FIGURE 11.1 Distribution of the different macrophyte habitats mapped using ArcGIS 9.3 software on aerial photographs (July and August 2008) from Land Resources International (Pty) Ltd.

(1948) and Ward’s (1976) periods of study, no submerged macrophytes were observed south of The Forks, due to the high turbidity of the water originating from the Mfolozi River containing fine silt and clay suspended matter. Open mouth conditions, high flow velocities and wave action in the Narrows prevented the establishment of submerged macrophytes, because this area was characterized by soft sediments and high turbidity. Zostera capensis has occurred in the estuary from Fani’s Island and southwards, when the estuary has salinity close to that of seawater. This plant has not been observed in the estuary since 2005 and it is unknown whether the plant is now extinct from the system. Long-term monitoring of transects from 2005 to 2010 showed that R. cirrhosa was mostly present at Charter’s Creek, Makakatana and Catalina

Macrophytes

Table 11.1. Mapping units used for the 2008 map and area covered by the different habitat types Mapping unit Submerged macrophytes

Area cover (ha) 19.8

Description Characterized by Ruppia cirrhosa Petagna (Grande) and Stuckenia pectinata.

Mangroves

304.9

Mangroves (Avicennia marina (Forssk.) Vierh. and Bruguiera gymnorrhiza (L.) Lam.) mainly in the Narrows and mouth area.

Intertidal reeds

206.7

Phragmites australis (Cav.) Steud. at the margin edges, sometimes emergent shoreline vegetation. Mainly occurring in the Narrows and fringing Mpate River.

9.7

Schoenoplectus scirpoideus (Schrad.) Browning observed at sites with freshwater input at the margins, sometimes as emergent shoreline vegetation.

Sedges

Salt marsh

967.3

Dominant species Sarcocornia spp., Salicornia meyeriana Moss. and Atriplex patula L.

Juncus kraussii

196.8

Often occurring at the landward fringe of the mangroves.

Juncus kraussii and grass

50.6

Mosaic of Juncus kraussii Hochst. susp. kraussii and grass, on the Makakatana Peninsula.

Grass and shrubs

2279.6

Sedge, grass and shore slope lawn observed in areas where there is no freshwater input. Saline lawns characterized by Sporobolus virginicus, Paspalum vaginatum and Stenotaphrum secundatum (Walt.) Kuntze.

‘Dry’ Phragmites

5912.1

‘Dry’ Phragmites – pale brown in colour on aerial photographs (fed by occasional flooding of low saline water and by direct rainfall). Killed by drought, prolonged submergence and hypersaline water. Mainly observed at Mkhuze Swamps.

Swamp forest

486.3

Observed on the banks of Mfolozi Estuary, in the vicinity of the back channel and Narrows. Swamp forest in other areas along the Eastern Shores was not mapped because it did not fall within the 5 m contour line.

Groundwaterdependent communities

184.4

Characterized by freshwater sedge community consisting of Phragmites, Juncus and Schoenoplectus along the Eastern Shores.

Non-tidal reeds

426.9

Groundwater-dependent reed swamp monoculture, bright green in colour on aerial photographs, in proximity with swamp forest.

8.6

Observed at the margins, in close proximity to mangroves and intertidal reeds.

Hibiscus tiliaceus

Water column

30 457

Sand

578.2

Mud

1642.8

Open water habitat. Bare ground, exposed sand banks. Non-vegetated mudflats.

211

212

Janine B. Adams et al.

FIGURE 11.2 Aerial photo of the submerged macrophyte Stuckenia pectinata forming large circular beds in North Lake. (Photo: R. H. Taylor, April 2004.)

Bay. This plant thrives under the fluctuating salinity and water level conditions (Adams and Bate, 1994a, 1994b; Riddin and Adams, 2008). Analysis of the 2008 aerial photographs showed that submerged macrophytes occurred mainly at Brodie’s Shallows in close proximity to the Narrows and Makakatana. During a field survey in May 2010, beds of S. pectinata were observed at Makakatana, indicating fresh to brackish (< 15) conditions. Prior to 2002, when the mouth was open and this site was saline, large beds of the more marine species Z. capensis would be present.

11.2.2 Mangroves Mangroves occur in the intertidal habitat along the shoreline of the Narrows. The primary mangrove colonizer, Avicennia marina (Forssk.) Vierh., is the dominant species and occurs as far as The Forks area. Bruguiera gymnorrhiza L. Lam is dominant from The Forks to Potter’s Channel but thereafter it only occurs sporadically and then as single trees as far north as Fani’s Island. Both species occur in the Mpate River where it enters the Narrows. The red mangrove Rhizophora mucronata Lam. is absent from the St Lucia system, possibly because it requires more

waterlogged conditions and is usually associated with intertidal channel and creek habitats (Steinke, 1999; Hoppe-Speer et al., 2011).

11.2.3 Reeds, rushes and sedges This habitat covers the largest area and the dominant species are the common reed, Phragmites australis and the sedge Schoenoplectus scirpoideus. Reeds cannot tolerate high salinity (> 20) and therefore stands occur at sites provided with a constant supply of fresh water. ‘Dry’ reed beds are found where this constant groundwater input is lacking. The reeds in these areas brown off and die rapidly in response to relatively short periods (few months) of drought. These dry areas are mapped as ‘dry’ Phragmites (Table 11.1) and are common in North Lake. During the period 2008–2011, there was a thick band of reeds present at the water edge in front of the mangroves in the Narrows, as a result of fresh water flowing in from the Mfolozi River Back Channel and from Mpate River. Reeds and sedges fed by groundwater also occur along much of the eastern shoreline of the St Lucia system, where there is an abundant supply of fresh groundwater flowing in

Macrophytes

from the eastern dune cordon (Chapter 8). On the western shoreline and in False Bay, this habitat only occurs where incised drainage lines meet the estuary shoreline. It also occurs along the margins of the rivers entering False Bay. In the 2008 mapping exercise (Table 11.1) these are referred to as ‘intertidal’ reeds. Stands of the rush Juncus kraussii Hoschst are found in saline (35) but less frequently inundated sites, for example Potter’s Channel and along the Eastern Shores at Catalina Bay and Mamba Point. In areas that become fresh over time, for example on the main entrance road to the St Lucia Estuary, P. australis appears to be shading and hence replacing J. kraussii. The fern Acrostichum aureum L. is often found along the fringes of the Juncus community. J. kraussii, known as ‘incema’ in Zulu, is harvested and used in crafts such as baskets, place mats, sleeping mats, gifts for wedding ceremonies and strainers for traditional beer. In the 2009–2010 years, the biomass of J. kraussii harvested in the areas adjacent to the St Lucia Estuary is estimated to be 170 tonnes (wet weight), having a value of over ZAR 680 000 (
USD 85 000) while in the unprocessed form (Kyle, 2010). From the late 1970s, harvesting of J. kraussii has been controlled and permits are required for harvesting in the iSimangaliso Wetland Park.

11.2.4 Salt marsh On the Western Shores and along all of the False Bay shoreline, where there is no significant groundwater inflow, dry shoreline habitats are found. Typical is the succulent salt marsh with species such as Salicornia meyeriana Moss and Sarcocornia natalensis (Steud.) Dur and Schinz. There are also saline lawns with Sporobolus virginicus, Paspalum vaginatum and Stenotaphrum secundatum (Walt.) Kuntze and the ‘dry’ stands of Phragmites australis. Similar vegetation is also found on the floodplains of the rivers entering False Bay, on the islands, the peninsulas and along those parts of the eastern shoreline where the topography of the estuary edge acts as a barrier to freshwater inflow. Succulents colonize exposed saline soils but are not tolerant to inundation. These plants grow rapidly from a large seed bank. Species include S. meyeriana, an annual, and perennials S. natalensis and Atriplex patula L. Sporobolus virginicus is more tolerant to hypersaline conditions compared with P. vaginatum and occurs in sandy soil, whereas P. vaginatum occurs in clay soil, where hypersaline conditions do not occur. Stenotaphrum secundatum, in comparison with P. vaginatum and S. virginicus, is less tolerant of hypersaline conditions and prolonged inundation (Ward, 1976).

11.3 Abiotic factors and macrophyte response The St Lucia estuarine system has a variable, diverse environment and therefore knowledge of the response of the dominant macrophytes to abiotic conditions can be used to predict how the vegetation will respond to future changes. Taylor (2006) captured this understanding in an expert system which uses attributes of the physical environment to identify the potential for growth and colonization by macrophytes. Changes in macrophyte area have largely followed mouth manipulation, dredging, canalization of the Mfolozi floodplain and a reduction in freshwater input. These impacts have

accentuated the effects of natural salinity fluctuations that are unique to St Lucia.

11.3.1 Mouth condition and water level fluctuations Closed-mouth conditions and a high water level can influence the macrophytes. Flooding kills succulent salt marsh within months, but other macrophytes such as saline Sporobolus lawns are more resilient. Prolonged submergence of mangrove pneumatophores limits gas exchange and causes

213

214

Janine B. Adams et al.

die-back. In the early 1960s, dredging activities blocked off the Dukuduku Stream near the St Lucia Mouth and a large area of mangroves was killed by the ensuing flooding (Taylor, 2006). Past high lake levels in conjunction with hypersaline conditions have been particularly detrimental to the shoreline vegetation. In 1971, heavy rains followed a drought period. Raised water levels in the southern areas such as Potter’s Channel resulted in the inundation of peripheral communities with water of salinity of up to 50 (Ward, 1976). This resulted in the death of the protective fringe of reeds and erosion of the platform followed. Two cyclones in 1984 resulted in losses of vegetation. The cyclones, Domoina and Imboa, struck the St Lucia system in January/February 1984. Over 600 mm of rain fell during Domoina and over 150 mm during Imboa. The latter cyclone coincided with extremely high tides and rough seas. The main effect was the flooding of the Mfolozi River, which spilt into the St Lucia Estuary, causing flooding in the Narrows and up into the lakes (Steinke and Ward, 1989). The water level during these cyclones rose 2.5 m above mean summer levels and it was only by April 1984 that it returned to normal. There were direct losses of mangroves due to breakage, uprooting and the collapse of trees. The white mangrove, Avicennia marina, was removed where the Mfolozi River broke into the St Lucia estuarine system, along the northern bank of the lower estuary. Indirect losses were due to prolonged inundation of mangrove pneumatophores with turbid water, loss of leaf litter and loss of mangrove propagules. The layer of fine silt deposited on the pneumatophores by the floods decreased gaseous exchange causing stress and die-back of mangroves. Fringing vegetation (Phragmites australis and Hibiscus tiliaceus L.) was also adversely affected. In the Narrows, water level fluctuations (marine tide- and wind-influenced) maintain the mangroves and intertidal salt marsh. Extended periods of mouth closure with no water exchange can result in dry, saline, anoxic soil, which inhibits plant growth. An assessment of mangroves at four different sites in

2010 (S. Hoppe-Speer et al., unpubl. data) showed that freshwater seepage, wind movement and flow from the Mfolozi River through the Back Channel were important in maintaining the mangrove habitats when the St Lucia Mouth is closed. Despite the mouth being closed since 2002, recruitment of mangroves has been observed. Seedling (< 100 cm) and juvenile mangrove trees (101–150 cm) were found fringing the water column of the Narrows. However, closer to the terrestrial fringe, the habitat was dry and saline due to lack of tidal exchange. These mangroves also had higher ion concentrations in their leaves, indicating saline sediment and groundwater conditions. There was a large litter layer as tides no longer moved broken branches, leaves and other debris around (Figure 11.3). In some areas, there was competition from freshwater species such as P. australis because the Narrows has remained fresh during the extended closed-mouth conditions (2002 to 2012).

11.3.2 Salinity Historically, salinity has been an important factor influencing macrophyte distribution and growth in the St Lucia estuarine system. Marsh productivity decreases as salinity increases and the shoots of all rooted vegetation die in salinity above 55. If too much salt accumulates plants cannot survive; first the mangroves die, then succulent salt marsh plants, leaving barren mudflats. Freshwater abstraction and loss of freshwater inflow from the Mfolozi River has increased the rate of salinity change and, on average, the lake is more saline now than it was during the first half of the twentieth century. The rate of salinity change is rapid when the system changes from hypersaline to fresh, after rainfall and river inflow. During drought conditions, with a low lake level, wind can blow water over the exposed sediments, which results in rapid salinity changes from 25 to 55 within a few hours as sediment surface salt crystals are dissolved (Bate and Taylor, 2008). The plants growing in these areas are the succulents, which are able to survive these rapid salinity fluctuations.

Macrophytes

FIGURE 11.3 Photograph indicating large mangrove litter layer which accumulated because tides no longer moved broken branches, leaves and other debris around. (Photo: J. B. Adams, May 2010.)

The succulents are often bright red, due to the accumulation of anthocyanin pigments. These osmotic solutes protect the plants against drought and salinity stress. Anthocyanins typically accumulate in the photosynthetic tissues, where they act as antioxidants protecting photosystems against photo-oxidative stress (Steyn et al., 2002; Wang et al., 2003). The three main species of submerged macrophytes (Stuckenia pectinata, Ruppia cirrhosa and Zostera capensis) die off rapidly when salinity exceeds their tolerance ranges. Ruppia cirrhosa is excluded beyond 50 (Adams and Bate, 1994b). Zostera capensis is excluded above 45 and below 10 (Adams and Bate, 1994b). Stuckenia pectinata is excluded above 20, depending on the duration of exposure. Plants do not have time to colonize and grow if salinity fluctuations out of their tolerance ranges are too rapid. In the past, fluctuating salinity conditions have resulted in massive die-back of submerged macrophytes, with resulting detritus pulses to the estuary. During drought conditions, when the mouth was kept open, hypersaline conditions with salinity

up to 150 was recorded (Chapter 7). No submerged macrophytes would survive under these conditions. Under natural conditions, with inflows from the Mfolozi River when the mouth is closed, salinity has seldom risen higher than 45. The submerged macrophytes R. cirrhosa and Z. capensis survive under these conditions. There are large sections of the Eastern Shores of Lake St Lucia where groundwater seeps out of the sand aquifers into the edge of the lake. Vertical stratification may occur where fresh groundwater seeps through the sand of the estuary bed. This creates a stratum of low salinity water in the root zone, with more saline water below, which is important for the survival and biodiversity of a variety of sedges, grasses and shrubs. Transects sampled from 2005 to 2010 during drought conditions indicate that macrophyte species richness was significantly higher at Catalina Bay (27 species), compared with Makakatana (13 species), Charter’s Creek (11 species) and False Bay (3 species). This is essentially along a gradient from fresher to more saline conditions.

215

216

Janine B. Adams et al.

11.3.3 Life history traits Estuaries are dynamic environments and thus a flexible life-cycle pattern is necessary to ensure the survival of macrophytes. Although vegetative growth is the dominant reproductive strategy for most estuarine macrophytes, seed banks do play a critical role in their re-establishment when water level fluctuations result in local extirpation of species (Keddy and Reznicek, 1982; Welling et al., 1988; Adams and Riddin, 2007). Macrophytes therefore need to complete their life cycles in order to replenish seed stocks, otherwise their persistence may be compromised. This was studied in the East Kleinemonde Estuary, a small dynamic temporarily open/closed Eastern Cape (South Africa) estuary that experienced high water level when the mouth remained closed for 17 months due to drought (Vromans, 2011). The emergent macrophytes rapidly declined in cover and eventually died back. Once the water level dropped, they recovered rapidly from a persistent seed bank. The succulents Sarcocornia tegetaria S. Steffen, Mucina & G. Kadereit and Salicornia meyeriana responded quickly to relatively small water level fluctuations (11 to 20 cm), by germinating within one month after exposure when the estuary water level dropped from 2.3 to 2.1 m amsl. Both species were able to complete their life cycles and set viable seeds within three and four months respectively, thereby replenishing the seed bank (Vromans, 2011). This confirmed the suggestion by Riddin and Adams (2008, 2009) in the same estuary that a minimum period of four months would be required for salt marsh to develop 100% cover under ideal conditions. Although not quantified, similar rates of expansion are expected in the St Lucia system. A large seed bank clearly exists for the St Lucia system, as expansion of salt marsh has occurred in exposed areas. The submerged macrophytes Ruppia cirrhosa and Stuckenia pectinata could also endure fluctuating salinity and water level conditions, as they have such a seed bank. Ruppia cirrhosa has persisted in the St Lucia Estuary under drought conditions, as it has a large seed bank allowing it to germinate when

conditions are favourable. In contrast, Zostera capensis seldom sets seed and was last observed in the estuary in 2005. Ruppia cirrhosa can go from germination to flowering and seeding within a few months, and hence replenishes the seed bank rapidly.

11.3.4 Interactions with other biotic components Newly sprouted Phragmites australis shoots are heavily grazed by large herbivores after burning, and this combination of fire followed by grazing can change dry reed bed areas into saline-grass lawns. On the west bank of the Mkhuze River mouth, a combination of burning followed by cattle grazing has altered the vegetation. Phragmites areas have been changed to saline lawn dominated by Paspalum vaginatum and Sporobolus virginicus. Hippos and other herbivores, including domestic stock, readily graze on the lawns formed by the saline grasses. Hippos create natural pathways in the shoreline vegetation. These paths from the wallows to the lake edge can drain perched wetland areas. Such action has been seen to promote hygrophilous tree and shrub establishment where, prior to the hippo activity, edaphic conditions were unsuitable for woody plant establishment (Ward, 1976). Hippos can also potentially influence the general nutrient relationships, by releasing large quantities of plant material (detritus) into the lake in their faeces (Begg, 1978; Chapter 18). Recently, kudu have been observed browsing on mangrove trees in the Narrows area. The drought conditions have made these habitats more accessible because they have become drier when the mouth is closed and they are no longer intertidal. The dry mud has provided a firm surface to walk on. In some cases a browse line of up to 2.2 m is clearly visible. Shorter trees, especially of Bruguiera gymnorrhiza, have been killed by these browsing activities. Submerged macrophytes form an important substratum for epiphytes. Epiphytes associated with submerged macrophytes were sampled during a drought from November 2004 to October 2005 (Gordon et al., 2008; Chapter 10). Very low water

Macrophytes

levels and high, variable salinity characterized the estuary at the time. Stuckenia pectinata and Ruppia cirrhosa were the dominant submerged macrophytes throughout the estuary, with S. pectinata occurring in the Narrows at lower

salinity (~ 10) and R. cirrhosa in the southern lake at higher salinity (> 15). Epiphytic biomass varied greatly between sites and over time ranging from 11–72 mg Chl-a m2 for S. pectinata and 17–165 mg Chl-a m2 for R. cirrhosa (Gordon et al., 2008).

11.4 Recent studies on response to drought and mouth closure There has been long-term monitoring of macrophyte distribution along permanent transects in 2005, 2006 and 2010. These data were combined with an analysis of past aerial photographs to indicate changes in macrophyte distribution over time and in response to drought (Nondoda, 2012).

11.4.1 Mapping Images from 1960, 2001 and 2008 were digitized to compare vegetation distribution and cover in the Narrows, Makakatana and the Eastern Shores. The 2001 images were used to illustrate vegetation distribution prior to the drought, which started in June/July 2002. The 2008 images were used to illustrate the situation during the drought, after 6 years of mouth closure. The early images from 1960 also represented a drought condition which started in 1958; however, at that time the mouth was open. A receding shoreline was evident in the North and South Lake in 2008. Low water level during the drought since 2002 has resulted in the expansion of shoreline vegetation into the dry estuary bed. The dominant species were the salt-tolerant succulents (Salicornia meyeriana, Sarcocornia natalensis and S. tegetaria S. Steffen, Mucina and G. Kadereit) and grasses (Paspalum vaginatum, Sporobolus virginicus and Stenotaphrum secundatum). The largest areas of salt marsh developed at Makakatana (75 ha) and in the Brodie’s Shallows area (343 ha) (Table 11.2). In 2001 these areas were covered with water. In 2008 and 2011 clusters of succulent salt marsh, such as Sarcocornia spp. and S. meyeriana were observed under extreme low water level conditions and growing on exposed sediment at Brodie’s Shallows

(Figure 11.4). These plants can tolerate some rise and fall in water level in response to rainfall and wind. At Makakatana in 2008, salt marsh colonized areas where the water column had receded (Figure 11.5). At Fani’s Island there was a decrease from inundated Phragmites to ‘dry’ Phragmites, as the water level receded. In the Narrows exposed areas were colonized by Phragmites fringing the channel and in the vicinity of The Forks exposed sand was colonized by Juncus kraussii and by sedges. Creeks and streams were overgrown with vegetation; for example, the Oxbow area was colonized by mangroves while P. australis established itself in the Nkazana and Tewate streams. During the drought there was an expansion of vegetation along the Eastern Shores at sites of groundwater seepage, which persisted even during the worst of the drought. These are important refuge sites not only for plants but also as a source of fresh water for drinking by crocodiles, birds, terrestrial herbivores and hippos. Freshwater input from groundwater seepage from the Eastern Shores has promoted diversity, with a variety of brackish macrophyte species present from the families Cyperaceae, Juncaceae and Juncaginaceae. Species diversity was probably lower prior to the drought, when water level was high and this area was flooded.

11.4.2 Long-term transect monitoring Changes in sediment characteristics and macrophyte cover were assessed at four sites (Lister’s Point, Charter’s Creek, Makakatana, Catalina Bay) during

217

218

Janine B. Adams et al.

Table 11.2. A comparison of area covered by different habitat units at Brodie’s Shallows and Mkakatana Location

Habitat unit

Area (ha) 1960

Brodie’s Shallows

2001

2008

60.0

28.2

0

Intertidal Phragmites

0

0

2.7

Salt marsh

0

0

343.2

Sand

0

0

78.8

62.2

0

19.3

321.7

415.7

0

444

444

444

26.8

23.5

0

Salt marsh

0

0

75.1

Sand

0

0

39.0

87.3

90.5

0

114

114

114

Grass and shrubs

Submerged macrophytes Water Total Makakatana

Grass and shrubs

Water Total

FIGURE 11.4 In 2008 and 2011 clusters of succulent salt marsh plants, such as Sarcocornia natalensis and Salicornia meyeriana, were observed under extreme low water level conditions and growing on exposed sediment at Brodie’s Shallows. (Photo: R. H. Taylor, October 2010.)

Macrophytes

FIGURE 11.5 Aerial photograph and maps indicating the expansion of macrophytes due to the drop in water level at Makakatana in 2008. (Photo: Ricky H. Taylor.)

219

220

Janine B. Adams et al.

Table 11.3. A comparison of sediment characteristics for the different sites including all sampling dates (mean ± SE, n in brackets) Sediment characteristics

Lister’s Point

Catalina Bay

Charter’s Creek

Makakatana

Moisture content (%)

30.1 ± 1.9 (82)

14.3 ± 0.7 (72)

15.2 ± 1.0 (90)

12.8 ± 0.9 (107)

Organic content (%)

7.9 ± 1.0 (82)

2.8 ± 0.9 (72)

3.4 ± 0.7 (90)

1.7 ± 0.2 (107)

Electrical conductivity (mS)

44.0 ± 4.7 (48)

9.4 ± 2.1 (41)

25.3 ± 2.8 (54)

23.0 ± 2.4 (64)

% sand

22.4 ± 1.2 (25)

11.8 ± 1.3 (21)

12.0 ± 0.5 (27)

11.8 ± 0.4 (32)

% silt

52.2 ± 2.4 (25)

74.1 ± 2.9 (21)

75.1 ± 1 (27)

76.1 ± 0.9 (32)

% clay

25.4 ± 1.4 (25)

14.2 ± 1.6 (21)

13.0 ± 0.6 (27)

12.1 ± 0.5 (32)

2005, 2006 and 2010, along transects that extended from the terrestrial edge into the submerged habitat. Low rainfall and the exposure of sediment resulted in the accumulation of salt on the surface layer at Lister’s Point (False Bay). There was an increase in sediment salt content from the terrestrial fringe to the water edge for all sites except Catalina Bay. This was related to the freshwater seepage at the latter site, whereas the other sites reflected the high water column salinity conditions. A significant increase in the abundance of Sarcocornia natalensis between October 2005 and May 2010 (H ¼ 9.41, n ¼ 82, p < 0.05) was observed at Makakatana, possibly in response to stable exposed conditions and lower salinity.

Extreme drought conditions were buffered by groundwater discharge at Catalina Bay and freshwater inflow from the Narrows at Makakatana. However, at False Bay a salt crust on the surface sediment and highly saline, dry conditions restricted macrophyte growth (Table 11.3). This site had the lowest macrophyte cover (< 10%) and macrophyte species richness (three species). The dominant species were Sporobolus virginicus and Chenopodium album L., which are highly salt tolerant species. Catalina Bay had the highest species richness (18 to 27), as a result of fresh groundwater seepage from the sand dune aquifers on the Eastern Shores.

11.5 Summary of changes in macrophyte habitats Changes in the macrophytes over time can be summarized for three main time periods: (1) the estuary under natural conditions, when the mouth would have remained open for long periods of time with no articial manipulation; (2) the periods 1952–2000, when the mouth was artifically dredged to keep it open; and (3) the period 2002–2011, when the mouth was closed. Lawrie and Stretch (2011b) described this as scenario 1 (open mouth due to dredging), scenario 2 (closed mouth, no dredging) and scenario 3 (natural).

11.5.1 Natural conditions Under natural conditions the mouth of the St Lucia Estuary and the Mfolozi River would have been linked, the water in the estuary would have been fresher and the mouth would have remained open for long periods of time (Chapter 7). Lawrie and Stretch (2011b) describe this as scenario 3 (natural): the combined Mfolozi/St Lucia inlet with no active mouth manipulation – commonly the ‘natural’ state

Macrophytes

of the system. When the mouth closed, fresh water from the Mfolozi River entered the lake resulting in low salinity with a high water level. Such an event would have prevented the growth and expansion of mangroves in the Narrows, whenever there was a high water level. This is because when the pneumatophores are inundated for longer than three months, the trees die. The area covered by mangroves has doubled since 1937, when there were only 169 ha. By 2008 there were 304.9 ha (Table 11.1). This figure may be an overestimate due to the difficulty in delineating narrow mangrove areas on the aerial photographs. Between these years, there have been both losses and increases in mangroves due to manipulation of the St Lucia/Mfolozi mouth system, dredger activities, and periodic losses due to cyclones, hail and fire. In the late 1970s and early 1980s, Ward and Steinke (1982) calculated an area of 160 ha for the mangroves of St Lucia Estuary. There is no doubt that the mangrove area has expanded in the St Lucia Estuary during the time when there was artificial breaching of the mouth and fluctuating tidal conditions. Mangroves are very sensitive to fire, and during this time pine plantations were being established and a necessary management strategy was to protect the plantations from fire. This fire protection to the mangroves would have also facilitated mangrove expansion. However, under natural conditions mangrove extension would have been limited by periodic closed-mouth conditions that resulted in a raised water level. This backflooding would have killed many of the mangroves, particularly if water level was raised to above 1 m for a few months. The highest salinity in the ‘natural’ scenario would occur when the mouth was open during long periods (years) of dry conditions. Salinity > 45 can occur for about 10 months at a time, but Mfolozi River inflow prevents extreme hypersaline conditions from persisting for more than about 4 months (Lawrie and Stretch, 2011b). Under these conditions salt-tolerant succulents would have temporarily replaced reeds and sedges.

11.5.2 The period 1952–2002 From 1952 the mouth was artificially dredged to keep it open during drought periods. This resulted in the estuarine system experiencing saline conditions and the water level being maintained close to that of mean sea level. This represents scenario 1, as described by Lawrie and Stretch (2011b): separate Mfolozi River and St Lucia inlets with management interventions to keep the St Lucia Mouth open and maintain inlet separation. Although mangroves would have thrived in the Narrows, which remained tidal and with salinity close to that of seawater, no submerged macrophytes would have been present in the Narrows, due to water flow being too strong and the sediment too unstable. The small water level variations but highly variable salinity would favour the growth of salttolerant succulents and grasses. Extreme hypersaline conditions would have occurred for about 30% of the time in this scenario, and would have persisted for about a year (Lawrie and Stretch, 2011b; Chapter 7). Artificially maintaining an open mouth would decrease the chance of desiccation, but salinity would exceed 65 about 17% of the time. An analysis of the earliest available aerial photograph of the St Lucia system showed that reed and sedge swamps occupied an area of 6032 ha in 1937, whereas by 1996 they only covered 3789 ha (Riddin et al., 2000). This figure may be inaccurate due to the problems of delineating reed areas, shown on the poor quality 1937 aerial photographs. However, the artificial open-mouth conditions from 1952 and the higher overall salinity would definitely have resulted in the loss of reeds, particularly in the north-eastern shallows and in the Selley’s Lakes area. In 1970/1971, the salinity of the water in False Bay rose as high as 100 and large reed swamp areas were lost due to these hypersaline conditions. They were replaced by saline hygrophilous grasses. Taylor et al. (2006a) showed that the open-mouth hypersaline condition is the most detrimental to macrophytes, resulting in die-back of shoreline vegetation and the

221

222

Janine B. Adams et al.

FIGURE 11.6 Seepage from the Eastern Shores into Tewate Bay supporting a groundwater-fed macrophyte habitat. (Photo: R. H. Taylor, April 2011.)

subsequent wave-driven erosion of the lake shoreline. This understanding supported the management decision that resulted in the mouth being kept closed after 2002.

11.5.3 The period 2002–2011 The most recent conditions experienced in the St Lucia system have been that of severe drought. The mouth has remained closed and the estuary has been characterized by extremely low water levels. This represents scenario 2 (Lawrie and Stretch, 2011b): separated Mfolozi and St Lucia inlets but without any mouth manipulation to keep the St Lucia Mouth open. The model results showed that without the Mfolozi link desiccation of about 50% of the lake

area would result for 32% of the time over an average duration of 15 months. Extreme conditions of desiccation and high salinity have resulted in the expansion of succulent salt marsh species and grasses into the resulting bare areas. The drought and low water level since 2002 have resulted in the expansion of shoreline vegetation into the lake, especially along the Eastern Shores into the exposed sand and mudflat areas. The pine plantations along the Eastern Shores were removed by 2007, and this has increased freshwater seepage into the lake in this area (Været et al., 2009), and has been important in maintaining the brackish grass and sedge shoreline habitats. The most important seepage zones are at Catalina Bay south, Catalina Bay north, Brodie’s Crossing, Dead Tree Bay and Tewate Bay (Figure 11.6).

Macrophytes

While areas of the estuarine system have been hypersaline, the water in the Narrows has remained fresh/brackish as a result of river inflow and flow from the Mfolozi River via the Back Channel. The submerged macrophyte Stuckenia pectinata and reeds have flourished in the Narrows as a result, but the persistent low salinity (< 15) has resulted in reeds outcompeting mangroves along the Narrows. Succulent salt marsh plants and saline grasses have flourished where water level and salinity fluctuates, such as Brodie’s Shallows, Makakatana and Western Shores. In very saline areas, such as False Bay and at Lister’s Point, macrophyte cover is sparse. Unfortunately, Zostera capensis has been lost from the system in recent times. Macrophytes are remarkably adaptable and there has been an expansion in the exposed habitats along the shoreline, during drought conditions

characterized by low water level. Indeed, there is a higher richness of macrophyte species now than when these areas were covered by water during open mouth conditions; that is, greater than 10 emergent species versus 3 dominant submerged macrophyte species. While the Lawrie and Stretch (2011a, 2011b) models and simulations are useful, these do not depict the spatial variation in the St Lucia estuarine system. Aerial photograph and transect analyses showed that the macrophytes are spatially variable, with those habitats fed by fresh groundwater sources versus those with no groundwater inputs responding very differently to drought. Low water level conditions during the closed mouth drought period, together with low salinity conditions along the Eastern Shores, have resulted in high macrophyte species richness.

11.6 Conservation and management To ensure the health and survival of the macrophyte habitats in the St Lucia estuarine system, a strategic adaptive management approach is needed. It is necessary to monitor system changes that are due to management interventions and use the insights thus gained to adapt the management. In the past, and probably into the future, the system will continue to be characterized by extreme environmental conditions. This will be exacerbated by anticipated climate change, which suggests an increase in the frequency of droughts and floods and wetter conditions characterized by extreme rainfall events (Chapter 21). Such events would mean less time available for the system to recover (Lawrie and Stretch, 2011b), with a possible loss of biodiversity. A higher sea level may also mean that the St Lucia estuarine system becomes a deeper, and hence more stable, system. Macrophytes are resilient; however, saline conditions coupled with dry wind-blown sediment prevent germination and growth. Increases in salinity can decrease the reed and sedge habitat, whereas

rapid increases in water level can result in the dieback of salt marsh and mangroves. Fluctuating water levels and lack of intertidal conditions have resulted in the loss of the seagrass Zostera capensis from the St Lucia estuarine system. The current poor health of the mangroves is also a cause for concern. Healthy ecosystems are more resilient to change. A way to ensure resilience is the determination and implementation of the Estuarine Ecological Water Requirements (Reserve) (Van Niekerk and Turpie, 2011), and for the St Lucia system this means reestablishing the link between the Mfolozi River and the estuary (Chapter 2). The St Lucia Estuary represents more than 50% of the total estuarine area in South Africa and it is therefore essential that the current health status be improved. The processes underpinning the goods and services provided by estuaries, such as the assimilation and cycling of nutrients, also need to be protected if resilience is to be maintained. Globally, eutrophication in aquatic systems is a problem and this is possibly a future threat to the

223

224

Janine B. Adams et al.

St Lucia Estuary. In particular, inputs from the Mkhuze and Mfolozi rivers are a cause for concern, as they drain extensive agricultural areas. Contaminated agricultural runoff can introduce toxic substances (e.g. herbicides and pesticides) and excessive nutrients from fertilizers, and increased suspended solid loads enter as a result of soil erosion. Eutrophication results when excessive nutrients, mainly inorganic nitrogen (N) and phosphorus (P), discharge from agricultural, industrial and domestic sectors into a water body. These nutrients stimulate primary production which increases the organic load. Upon decomposition, the oxygen demand outweighs the oxygen supply and hypoxia or anoxia develops causing fish kills (Cloern, 2001; Heisler et al., 2008). Additional symptoms of eutrophication include changes in microalgal dominance from diatoms to flagellates, increase in nuisance phytoplankton taxa such as toxic cyanobacteria, decreased light

availability and high macroalgal growth rates. Excess nutrients induce the shift between two alternating states, by favouring the rapid growth of macroalgae or phytoplankton to exclude rooted submerged macrophytes (Flindt et al., 1999, Collie et al., 2004; Orfanidis et al., 2005; Viaroli et al., 2008). The increase in blooms of filamentous green macroalgae in South African estuaries is a typical response to eutrophication (Adams et al., 1999). Globally, similar systems, particularly those with intermittent connections to the sea, have also shown an increase in macroalgae, as a response to eutrophication (e.g. Waquoit Bay in the USA and the Peel-Harvey estuarine system in Australia; Kinney and Roman, 1998). The degradation of rooted submerged macrophytes will impact negatively on the associated fauna such as invertebrates, fish and bird populations that are dependent on them for food, shelter and breeding habitats.

11.7 Future research Macrophytes are good indicators of change. Because they are rooted, they can be used as a record of what conditions have been present at any particular site for the lifespan of the plants that occur there. Although we do have a considerable amount of knowledge about the tolerance ranges of the plants to various conditions, long-term ecological research is still needed to study the response of the macrophytes to fluctuating environmental conditions. This will improve our predictive abilities by linking biotic response to physical and chemical change. An understanding of the rates of recruitment, growth, expansion and die-back is needed for each dominant macrophyte species. This can then be applied in models to predict habitat availability under a given set of conditions that would occur given various management scenarios. The implications of loss of habitat, for example submerged macrophytes for faunal components, and the links between the primary and secondary

producers need investigation. Die-back of the submerged macrophytes and fringing vegetation can result in huge pulses of detritus into the estuarine system. The importance of this to the food webs and secondary production is unknown. Further research on the mangroves can contribute to a global understanding of mangrove response to stress. The current scenario of closed-mouth conditions, low salinity and low water level provides a unique situation for the study of mangrove responses because these are trees that are adapted to saline intertidal conditions. A preliminary study in 2010 (S. Hoppe-Speer et al., unpubl. data) showed that sediment characteristics are unfavourable for mangrove growth at sites now characterized by a lack of tidal flooding. Sites that were dry with high salinity had no mangrove seedlings/saplings, or they occurred in low density. In the Narrows there is competition from reeds and terrestrial vegetation, as a result of the low salinity conditions. Long-term

Macrophytes

data are needed to assess the influence of mouth closure on recruitment and survival of the mangrove forest at St Lucia. Reconnecting the Mfolozi with the St Lucia system will result in localized losses of mangroves due to an increase in flow and inundation. An assessment of the status of invasive plants in the St Lucia system catchment and riparian zone is

needed. At this stage, there are signs of the early stages of invasion of the mangrove and salt marsh habitats by the alien shrubs Schinus terebinthifolius Raddi and Tamarix ramosissima (L.) Karst. (Chapter 19). The threats of these and other invasive species on ecosystem function and value could be large and the management needed to control them expensive.

225

Chapter contents 12.1 Introduction 12.2 Benthic macrofauna 12.3 Meiofauna 12.4 Ecological importance of benthic fauna 12.5 Conclusion

Male of Neosarmatium smithi flushed out by floods at the St Lucia Mouth. (Photo: Lynette Clennell, March 2012.)

12

Benthic invertebrates Deena Pillay, Sarah J. Bownes and Holly A. Nel

12.1 Introduction Benthic fauna in the St Lucia Estuary have historically attracted significant research interest (Day et al., 1954; Millard and Broekhuysen, 1970; Boltt, 1975; Blaber et al., 1983; Owen and Forbes, 1997; Pillay and Perissinotto, 2008, 2009; MacKay et al., 2010). The system is highly variable in both space and time, and undergoes cycles of wet, normal and dry periods, which can last several years. Understanding the effects of these phases has been the major scientific motivation for ecological research on benthic fauna in the system. Benthic macrofauna, defined here as sediment-dwelling invertebrates larger than 500 mm, has been one of the most intensively studied faunistic groups in the system. Meiofauna, which are interstitial invertebrates between 63 and 500 mm in size, has been significantly under-researched in the estuary. In terms of the benthic macrofauna, some studies have

been autecological in nature, focusing on the abundance, distribution and physiology of the numerically or gravimetrically dominant species such as the mud crab (Scylla serrata), pencil bait (Solen cylindraceus) and the ocypodid crab Paratylodiplax blephariskios. Other studies have addressed spatio-temporal patterns in macrofaunal assemblages during wet, normal and dry phases. Recent research on the meiofauna has explored spatio-temporal patterns of assemblages under drought conditions. The aim of this chapter is to review existing knowledge on the benthic invertebrates of the St Lucia Estuary focusing on the autecology of dominant species and the role of environmental fluctuations, particularly the cycling of the system between drought, normal and wet phases, in structuring macro- and meiofaunal assemblages.

12.2 Benthic macrofauna The common benthic macrofauna recorded in the St Lucia Estuary are listed in Table 12.1 along with their distributions. It is not intended to provide a comprehensive list of all macrofauna recorded in the system but serves to highlight the commonly reported species and their distributions in the estuary. It must be borne in mind that elucidating more specific patterns in terms of species abundance in

space and time is difficult due to differences in taxonomic identifications between studies and the variability in data reporting among studies, which contain a mix of quantitative and semi-quantitative approaches. Polychaetes, crustaceans and molluscs are the main macrofaunal components in the St Lucia Estuary (Table 12.1). Most of the species listed are

Ecology and Conservation of Estuarine Ecosystems: Lake St Lucia as a Global Model, eds. Renzo Perissinotto, Derek D. Stretch and Ricky H. Taylor. Published by Cambridge University Press. © Cambridge University Press 2013.

228

Deena Pillay et al.

Table 12.1. Commonly reported macrofaunal taxa in the St Lucia Estuary (1954–2010) and their known distributions Taxa

Distribution

INSECTA Chironomidae (larvae)

Recorded in False Bay under hypersaline conditions.

ANNELIDA Polychaeta Dendronereis arborifera Dendronereides zululandica Marphysa macintoshi Ceratonereis keiskama

Glycera spp.

Prionospio sexoculata CRUSTACEA Amphipoda Grandidierella bonnieroides Eriopsia chilkensis

Mouth, Narrows and South Lake. Mouth, Narrows and South Lake. South and North Lakes, False Bay. Marphysa simplex reported by Day et al. (1954). South and North Lakes, False Bay. Ceratonereis hircinicola reported by Millard and Broekhuysen (1970) and Ceratonereis sp. reported by Blaber et al. (1983) and Pillay and Perissinotto (2008). Recorded at the mouth and South Lake. Glycera convoluta reported by Day et al. (1954), Millard and Brockhuysen (1970) and Owen and Forbes (1997) and Glycera natalensis by MacKay et al. (2010). Mouth, Narrows and South Lake.

Corophium triaenonyx Melita zeylanica

Narrows, South Lake and North Lake. Narrows, South and North Lake. Victoropsia chilkensis reported by Owen and Forbes (1997). South Lake. Present throughout the system.

Isopoda Cyathura estuaria Cirolana luciae

Narrows, South Lake and North Lake. South and North Lakes, False Bay.

Brachyura Hymenosoma orbiculare Paratylodiplax blephariskios Scylla serrata

Present throughout the system. Abundant in muddy areas in the Narrows. Narrows, South and North Lake. Greatest abundance in the Narrows.

Tanaidacea Apseudes digitalis

Narrows, South Lake northwards.

Mysidacea Mesopodopsis africana

Present throughout the system.

MOLLUSCA Gastropoda Assiminea sp.

Narrows, South and North Lakes. Assiminea bifasciata reported by Boltt (1975).

Benthic invertebrates

Table 12.1. (cont.) Taxa

Distribution

Nassarius kraussianus

South Lake.

Bivalvia Solen cylindraceus Brachidontes virgilae Salmacoma litoralis

Narrows, South Lake, southern parts of North Lake. Solen capensis reported by Day et al. (1954) and S. corneus by Millard and Broekhuysen (1970). Southern Narrows. Narrows and South Lake. Previously reported as Macoma litoralis.

euryhaline (Millard and Broekhuysen, 1970) and are widely distributed in the system. The polychaetes commonly encountered in the literature are dominated by errant species, and few sedentary ones have been recorded. Among the polychaetes, nereid forms such as Dendronereis arborifera, Dendronereides zululandica and Ceratonereis keiskama have been commonly recorded. Glycerids, capitellids and the eunicid Marphysa macintoshi are also well represented in the macrofaunal literature. Crustaceans are represented by diverse groups in the system, including amphipods (Grandidierella bonnieroides, Eriopisa chilkensis, Corophium triaenonyx, Melita zeylanica), isopods (Cyathura estuaria, Cirolana luciae), tanaids (Apseudes digitalis), mysids (Mesopodopsis africana) and brachyurans (Hymenosoma orbiculare, Paratylodiplax blephariskios, Scylla serrata). Bivalves and gastropods are the dominant mollusc classes recorded, with the latter being dominated by the microgastropod Assiminea sp. and the larger Nassarius kraussianus. Among the bivalves, Solen cylindraceus (razor clams or pencil bait), Brachidontes virgiliae and Salmacoma litoralis have been frequently reported.

12.2.1 Bivalves Bivalves are important components of the benthic macrofauna in the St Lucia Estuary. Recent assessments have identified 24 bivalve species in the system based on surveys undertaken from December

1982 to April 2011, and from specimens at the KwaZulu-Natal Museum (Pietermaritzburg) and the Iziko South African Museum (Cape Town) (Nel et al., 2012). This assessment indicated that only 12 of the 24 bivalve species identified had been previously reported in the published literature for this estuary (Table 12.2). Twelve species had not been reported previously in the literature from this estuary (Table 12.2; Nel et al., 2012). Single shells of another two previously unrecorded species, Anodontia edentula and Timoclea lavrani, were also found in the estuary, but these may have been introduced dead through tidal exchange. The species Brachidontes semistriatus, Saccostrea cucullata, Pitar abbreviatus, Hiatella arctica, Solen capensis, Tellina triliatera, Tivela compressa and T. natalensis, although having been mentioned previously in the literature for the St Lucia Estuary, are not considered to be part of the bivalve fauna of the estuarine lake, as they may represent either fortuitous records or erroneous identifications (Nel et al., 2012). One of the 12 bivalve species not previously recorded in the St Lucia Estuary is the Asiatic hard clam (or thick-shelled clam), Meretrix meretrix, has a general Indo-Pacific distribution and in East Africa has been reported as far south as Maputo Bay (Scarlet, 2005; Branch et al., 2010). According to Kilburn and Rippey (1982), does not typically occur south of Mozambique (see Chapter 19). Meretrix meretrix was first recorded in the St Lucia Estuary in July 2000, at Charter’s Creek (Nel et al., 2012). In 2005 and 2011, recently dead M. meretrix were found in abundance in

229

230

Deena Pillay et al.

Table 12.2. Bivalve species (in alphabetic order) reportedly found in the St Lucia Estuary during surveys undertaken from December 1982 to April 2011, and from specimens at the KwaZulu-Natal Museum and the lziko South African Museum Scientific names

Previously reported

Not previously reported

Anadara natalensis

X

Anomia achaeus

X

Arcuatula capensis

X

Barnea manilensis

X

Brachidontes virgiliae

X

Chambardia wahlbergi

X

Corbicula fluminalis

X

Dendostrea sandvichensis

X

Dosinia hepatica

X

Eumarcia paupercula

X

Fulvia fragilis

X

Irus irus

X

Macomopsis moluccensis

X

Mactra cuneata

X

Martesia striata

X

Meretrix meretrix

X

Saccostrea forskahlii

X

Salmacoma litoralis

X

Siliqua cf. polita

X

Solen cylindraceus

X

Soletellina lunulata

X

Tellina cf. rousi

X

Tellina s.l. bertini Theora lata

X X

both South and North Lakes. Specimens ranged in size from 0.5 cm to 7 cm, suggesting that a viable population was at some stage thriving within the estuary. The time and mode of its introduction into the

St Lucia Estuary remain unresolved. The absence of the species in the latest survey and the large number of dead M. meretrix suggest it may no longer be present in the system (Nel et al., 2012).

Benthic invertebrates

FIGURE 12.1 Lateral view of the bivalve Solen cylindraceus (pencil bait) with body withdrawn into shell (top) and foot (F) and siphon (S) extended (bottom). (Photo: Lynette Clennell.)

Millard and Broekhuysen (1970) reported the presence of dead rock piddocks, Barnea manilensis, in the St Lucia Estuary. The extremely large numbers (> 100 ind. m
2) of dead specimens observed in March/April 2011 along the western shoreline of the estuarine lake suggest that aggregations of this species were potentially important filter feeders and major role players in the settling of suspended silt within the system. This species was found burrowed in exposed cretaceous sandstone. Currently most of the filtering action within the lake appears to be fulfilled by S. cylindraceus. Apart from this species, Brachidontes virgiliae, Dosinia hepatica, Salmacoma litoralis, Macomopsis moluccensis and Tellina cf. rousi were the only bivalves, out of the 24 species, found alive in the estuarine lake during the latest survey in March 2011 (Nel et al., 2012). Solen cylindraceus (Figure 12.1), commonly referred to as eastern pencil bait or stick bait (Kilburn and Rippey, 1982; Branch et al., 2010), can reach a maximum size of 9.5 cm in length (Kilburn and Rippey, 1982). Despite being numerically and gravimetrically dominant in the benthos of the St Lucia Estuary, it seldom exceeds 5.0 cm in length under the current harsh conditions. Solen cylindraceus is exposed to marked fluctuations in physico-chemical parameters due to the cyclical changes from wet to dry periods (Nel et al., 2011).

This species has consistently been recorded in the literature on the St Lucia Estuary (Day et al., 1954; Millard and Broekhuysen, 1970; Boltt, 1975; Blaber et al., 1983; Owen and Forbes, 1997; Pillay and Perissinotto, 2008; MacKay et al., 2010). McLachlan and Erasmus (1974) and de Villiers et al. (1989) suggested a lethal temperature for S. cylindraceus would be 44–44.5  C and regarded this higher than any temperature the species is likely to be exposed to in its natural habitat. The St Lucia Estuary, however, does experience similar and even higher temperatures in shallow waters (Carrasco and Perissinotto, 2011a). Solen cylindraceus is considered a euryhaline osmoconformer (McLachlan and Erasmus, 1974; de Villiers and Allanson, 1989). Recent experiments have indicated that S. cylindraceus could tolerate salinities between 30 and 60 when exposed to sudden changes (1 to 5 min) and between 15 and 65 when exposed to gradual changes (Nel et al., 2011). Pillay and Perissinotto (2008) reported highest abundances of S. cylindraceus in the St Lucia Estuary at relatively stable marine salinities between 25 and 50, with abundances decreasing at low (< 10) and rapidly changing salinity. An increase in S. cylindraceus abundance was recorded in the South Lake during stable marine salinities of about 30 to 45 (Blaber et al., 1983; Forbes and Cyrus, 1993). MacKay et al. (2010) found S. cylindraceus in the field at salinities from 10 to 70. This bivalve has limited horizontal mobility and thus employs behavioural strategies, such as burrowing to cope with unfavourable environmental conditions, as it is incapable of complete valve closure (de Villiers and Allanson, 1989). In the field, S. cylindraceus may burrow up to 40 cm deep in order to avoid unfavourable salinities (McLachlan, 1974; de Villiers and Allanson, 1989). Even when completely shut, this bivalve is exposed at its anterior and posterior ends, which may result in a lower tolerance to rapidly changing salinities (McLachlan and Erasmus, 1974; Matthews and Fairweather, 2004). Solen cylindraceus is considered a fast-growing species, living to about five years (McLachlan, 1974).

231

232

Deena Pillay et al.

Growth rings, often used to determine the age of bivalves, are too indistinct and irregular to be of any use in the study of growth in S. cylindraceus. Therefore, in situ size frequency analyses or growth incubations are required to estimate its growth rates. Recent incubations undertaken at Charter’s Creek showed that spat (< 27 mm length) grow at an average rate of 0.053 cm per day. An interesting question pertinent to filter-feeding bivalves in the St Lucia Estuary is how some species are able to survive and feed in seemingly dry, exposed sediment. Dissected S. cylindraceus that were collected at Makakatana in October 2005 in dry sediment indicated the presence of gut chlorophyll-a pigments (1378 ng (chl-a) ind.
1 and 1628 ng (phaeopigment) ind.
1), suggesting that they were able to filter feed under these conditions (D. Pillay, unpubl. data). This may be facilitated by winds or light breezes blowing films of water over the dry surface, enabling S. cylindraceus to filter feed on microalgae periodically.

12.2.2 Scylla serrata Scylla serrata (Figure 12.2) is a large predatory portunid crab that is common in estuaries on the east coast of southern Africa (Day, 1981). Hill (1979) estimated the adult population size of S. serrata in the St Lucia Estuary to be roughly 179 000 individuals. While the population size was large, animal densities were low. In the St Lucia Estuary, the greatest density of S. serrata was one individual per 295 m2 (0.003 ind. m
2). Density of the same crab in the Kleinemond Estuary in the Eastern Cape of South Africa was one crab every 124 m2 (0.008 ind. m
2; Hill, 1975b). It was also shown that densities of S. serrata in the St Lucia Estuary decreased from the mouth to the northern parts of the system, being greatest in the Narrows, declining by half in South Lake and decreased further in North Lake. The presence of fine-grained sediment and complex root structures of mangroves and other plants were important factors

FIGURE 12.2 Front view of the mud crab Scylla serrata. (Photo: Charles Griffiths.).

governing the distribution of S. serrata in the St Lucia Estuary (Hill, 1979). Gut content analyses indicate that S. serrata is an important benthic predator in the St Lucia Estuary (Hill, 1979). Bivalve fragments, predominantly those of Lamya capensis (¼Arcuatula capensis) were recorded in 48% of sampled individuals. The remains of the crabs Hymenosoma orbiculare and Paratylodiplax blephariskios were found in 16% of the sample and 12% contained crushed crustacean remains. The gastropod Assiminea bifasciata occurred in 4% of the guts of S. serrata (Hill, 1979). Seylla serrata is euryhaline, being able to survive salinities up to 65 under laboratory conditions (Hill, 1979). Crabs could feed up to salinities of 56 but became moribund at salinities greater than 60. Such salinities occur frequently in the St Lucia Estuary and can persist for several years (Boltt, 1975; Hill, 1979). Under these conditions, S. serrata may be eliminated from significant portions of the estuary and restricted to parts of the Narrows and South Lake where salinities rarely become hypersaline (Hill, 1979). The Narrows and South Lake may thus function as ecological refuges for populations of S. serrata during extreme hypersalinity and as ‘reservoirs’ to colonize other parts of the system when salinities stabilize (Hill, 1979). Field evidence indicates that S. serrata may survive salinities down to 2 but cannot survive in completely fresh conditions (Macnae, 1968; Hill, 1975b, 1979).

Benthic invertebrates

12.2.3 Paratylodiplax blephariskios Paratylodiplax blephariskios is a small burrowing ocypodid crab that is endemic to the subtropical eastern coast of southern Africa (Manning and Holthuis, 1981). In the St Lucia Estuary, P. blephariskios occurs most commonly in the muddy Narrows, where densities of this species can be as high as 2000 ind. m
2. The distribution and abundance of P. blephariskios in the St Lucia Estuary seems to be determined primarily by proximity to the St Lucia Mouth, the mouth state and sediment grain size and degree of sorting. Owen et al. (2000) reported that within the Narrows, P. blephariskios density is an order of magnitude lower in the upper Narrows than the lower portion. A reduction in tidal action may restrict recruitment of P. blephariskios into the upper Narrows (Owen et al., 2000).

12.2.4 Benthic macrofauna between wet and dry cycles Early studies on macrofaunal assemblages of the St Lucia Estuary between low (Millard and Broekhuysen, 1970), marine (Blaber et al., 1983) and hypersaline conditions (Day et al., 1954; Boltt, 1975) have focused on the lake systems. Owen and Forbes (1997) later documented the effects of low salinity and hypersaline conditions on the Narrows over an 11-year period. The most recent studies have been made on the entire system under dry, hypersaline cycles (Pillay and Perissinotto, 2008; MacKay et al., 2010). A summary of key findings of these studies along with background ecological information is presented in Table 12.3. Macrofaunal assemblages under hypersaline conditions in the lake systems were investigated between 1972 and 1973 (Boltt, 1975). Salinities at the onset of the study ranged between 45 and 80 and chironomid larvae were the only organisms found in False Bay samples and ostracods dominated North Lake samples. These taxa appeared to be the only

macrofauna capable of surviving the excessive salinities. Over the next six months, salinities in the system dropped, resulting in the rapid colonization of False Bay and North Lake by fast-growing taxa with planktonic larvae and an overall increase in macrofaunal diversity. Boltt (1975) postulated that South Lake acted as a refuge for species during hypersaline conditions because freshwater inflow prevented the development of extreme salinities in this region. Macrofauna in South Lake may therefore act as a pool to recolonize parts of North Lake most affected by hypersaline conditions (Boltt, 1975). Hypersalinity has also been shown to influence benthic assemblages in the Narrows (Owen and Forbes, 1997). Densities of the crab Paratylodiplax blephariskios decreased by an order of magnitude under these conditions, whereas the polychaetes Prionospio sexoculata and Dendronereides zululandica together with the bivalves Dosinia hepatica and Eumarcia paupercula increased in abundance (Owen and Forbes, 1997). The bivalve Solen cylindraceus also reappeared in the Narrows during this hypersaline phase following a period of low salinities. Macrofaunal biomass declined in the lower and middle sections of the Narrows, which was predominantly due to reductions in densities of the crab P. blephariskios. Biomass in the upper Narrows increased under hypersaline conditions, mainly because of increased abundance of the bivalves D. hepatica and E. paupercula (Owen and Forbes, 1997). A recent assessment of the St Lucia macrofauna under hypersaline conditions has highlighted the sensitivity of the northern lakes to droughts (Pillay and Perissinotto, 2008). In this study, which was undertaken in 2005, hypersaline conditions were especially severe in the northern parts of the system, being as high as 125 in Hell’s Gate (Pillay and Perissinotto, 2008). In addition, water depths decreased drastically in the northern regions with significant parts of the estuary being lost due to

233

Table 12.3. Studies undertaken on the macrofauna of the St Lucia Estuary (1954–2010) and the major outcomes Study

Period

Location

Salinity

Major outcomes

Day et al. (1954)

1948–1951

Whole system

25–52.6

Identification of faunal assemblages associated with biotypes in the estuary. Negative effects of siltation, particularly smothering of sedentary species. High salinities and scarcity of benthic fauna and flora in False Bay and North Lake. Emphasis on the importance of maintaining openmouth conditions for fish and invertebrates.

Millard and Broekhuysen (1970)

1964–1965

Whole system

5.5–36.3

Identification of fauna associated with biotopes in the estuary. High proportion of euryhaline species in the St Lucia Estuary relative to Langebaan Lagoon and Knysna Estuary. Decrease in faunal richness from the lower to northern parts of the system. This was thought to be related to the extreme conditions experienced in the north and its distance from the mouth, which limited colonization from the sea. Decrease in abundance of species with a preference for marine salinities and increase in brackish species.

Boltt (1975)

1972–1973

False Bay, North and South Lake

45–80 at onset of study, dropping to ± 35 within the first 6 months

Impoverished benthic fauna under high salinities (45–80), with low numbers of chironomids surviving in False Bay. Rapid colonization of previously denuded areas by fast-growing species with planktonic larvae following shifts to marine salinity. Recolonization most likely occurred from isolated refuges with low salinity or South Lake where salinity was kept low by riverine inflow. Low standing stock of macrofauna in the system relative to northern hemisphere temperate lakes.

Blaber et al. (1983)

1981–1982

South Lake

± 35

Dominance of slow-growing species (Solen cylindraceus, Eumarcia paupercula, Dosinia hepatica) in response to stability in the ecosystem. Decline in abundance of smaller gastropods (Nassarius kraussianus and Assiminea globulus), polychaetes (Prionospio sexoculata), nematodes and harpacticoids. Increase in species richness associated with a drop in salinity from hypersaline to marine conditions. Increase in biomass from hypersaline to marine conditions by four times.

Cyrus (1988b)

Owen and Forbes (1997)

1981–1987

1983–1994

South Lake

The Narrows

± 35–45 at onset, dropping to 0.5 due to heavy rains and cyclones, stabilizing at ± 35

Reduction in macrofaunal density by ± 40% following cyclonic activity and heavy rainfall.

± 0–50. Study covered periods of flooding and hypersaline conditions

Numeric and gravimetric dominance of the benthos by Paratylodiplax blephariskios (Brachyura).

Shift in size structure of the bivalve Solen cylindraceus, with juveniles dominating after the cyclones and rains.

Decline in Marphysa macintoshi (Polychaeta) and increase in Apseudes digitalis (Tanaidacea) abundance following cyclonic activity. Solen cylindraceus (Bivalvia) appeared after the cyclones and then declined in abundance. Victoriopsia chilkensis (Amphipoda) density increased during stable conditions. Appearance of the bivalves Eumarcia paupercula and Dosinia hepatica during hypersaline periods. Greater macrofaunal diversity in upper Narrows with greater contribution of lake species. Pillay and Perissinotto (2008)

2005

Whole system

5–125 over the entire system. Hypersaline conditions common in North Lake and False Bay

Reduction in macrofaunal abundance, species richness and diversity in the northern parts of the system where hypersaline conditions were prominent and water levels were lowest. No macrofauna recorded in some of the northernmost sites, indicating lethal levels of high salinity and low water depth.

Table 12.3. (cont.) Study

Period

Location

Salinity

Major outcomes Rarity of K-selected species such as bivalves (specifically Solen cylindraceus) and larger macrofauna. Macrofaunal abundance, species richness and diversity declined with decreasing water depth, and negatively correlated with microphytobenthic biomass.

MacKay et al. (2010)

2004–2008

Whole system

4.9–90 over the system. Hypersaline conditions recorded in North and South Lake

Large-scale variability masked patterns in physical environment and macrofaunal assemblages. A core set of taxa exists under low depths and high salinities that do not depend on open-mouth conditions for recruitment. Identification of salinity:abundance relationships for macrofaunal taxa. Apseudes digitalis (Tanaidacea) widely distributed from 5 to 70, Solen cylindraceus (Bivalvia) most abundant between 40–60, Prionospio sexoculata (Polychaeta) abundant between 20–70, Ceratonereis keiskama (Polychaeta) and Corophium triaenonyx (Amphipoda) occurred most frequently at salinities < 40; C. triaenonyx abundance peaked at 20.

Benthic invertebrates

evaporation (Taylor, 2007). In general, density, richness and diversity of macrofauna decreased significantly from the Narrows to False Bay and North Lake. In the most extreme cases, no macrofauna were recorded in the False Bay and North Lake samples. Macrofaunal richness, diversity and abundance were positively influenced by water depth, indicating that decreasing water level was an important mechanism by which the drought affected macrofaunal assemblages (Pillay and Perissinotto, 2008). MacKay et al. (2010) also surveyed the macrofauna of the St Lucia Estuary during the latest drought phase between 2004 and 2008, and did not find obvious spatial and temporal patterns in assemblages, with little direct evidence linking drought-related factors to macrofaunal communities. It appeared that a highly resilient set of core species was present in the system but in different proportions in various parts of the system. These species most likely do not require open-mouth conditions for recruitment (MacKay et al., 2010). Differences in findings between the Pillay and Perissinotto (2008) and MacKay et al.(2010) studies most likely relate to differences in temporal scales between studies and specific habitat traits among the sites sampled. The Pillay and Perissinotto (2008) survey took place four times over one year while the survey of MacKay et al. (2010) took place either annually or biannually over five years. The lower temporal resolution and greater time span in the latter study may have increased levels of variability and masked ecological trends. In addition, Pillay and Perissinotto (2008) surveyed sites in the northern parts of the system that were shallower than those sampled by MacKay et al. (2010). In combination, the studies may indicate that the shallower fringing habitats of the St Lucia Estuary are more susceptible to drought conditions than deeper areas. There are important outcomes of studies of the St Lucia benthos under hypersaline conditions that are pertinent to understanding the ecology of the system and other similar systems globally. Various authors have highlighted the reversed salinity

gradient that persists in the system under drought conditions (Day et al., 1954; Pillay and Perissinotto, 2008; MacKay et al., 2010) and is consistent with findings on other systems worldwide under conditions of low freshwater inputs (Hastie and Smith, 2006; Webster, 2010; Kingsford et al., 2011). Extreme hypersalinity is also a common feature of the system during drought cycles, especially in the northernmost sites in North Lake and False Bay, where salinities greater than 70 have been recorded (Boltt, 1975; Pillay and Perissinotto, 2008; MacKay et al., 2010). In addition to experiencing hypersaline conditions, the northern parts can get extremely shallow ( 104 ind. m 3) in the mouth region of the St Lucia estuarine lake shortly after mouth re-closure in August 2007 was, therefore, a unique occurrence (Carrasco et al., 2010). Associated with the arrival of O. dioica was a shift in zooplankton community structure, from dominance by the mysid Mesopodopsis africana and the copepod Pseudodiaptomus stuhlmanni to dominance by O. dioica and the copepod Acartiella natalensis. The absence/scarcity of other planktonic grazers in swarms of planktonic tunicates is not uncommon (Fraser, 1962; Berner, 1967; Deibel, 1980), as they exhibit a number of properties which enable them to outcompete other grazers. Drastic changes in the physico-chemical environment that occurred in the wake of the re-closure of the estuary mouth in August 2007, such as decrease in turbidity and increase in water level, appear to have favoured O. dioica, to the exclusion of the other grazers. Also, given its efficient feeding mode (Alldredge, 1981; Deibel, 1998), O. dioica may have outcompeted P. stuhlmanni with the rapid consumption of food sources. The slower regeneration times of copepods such as P. stuhlmanni (e.g. Paracalanus crassirostris versus O. dioica ratio of  1/10, Hopcroft and Roff, 1995) may have further made this species vulnerable to predation, while O. dioica thrived. Indeed, both P. stuhlmanni and M. africana are major components in the diet of a number of zooplanktivorous fish in Lake St Lucia (Blaber, 1979). Oikopleura spp. are also known to be consumed preferentially by many fish species, because of their high carbon and nitrogen content, as well as their lack of a carapace and slow reaction times (Gorsky and Fenaux, 1998). However, they exhibit extremely fast generation rates (Deibel, 1998) and this advantage may have counterbalanced the effect of fish predation at the St Lucia Mouth.

259

260

Nicola K. Carrasco et al.

After 2 months, O. dioica disappeared completely from the estuary. Possible factors that may have contributed to its demise include: (a) a drop in salinity down to 23, which is outside its range of tolerance (Uye and Ichino, 1995); (b) the possibility of O. dioica having become the primary target of fish predation, in the absence of P. stuhlmanni and M. africana; and (c) copepods depressing O. dioica densities through their ability to ingest its eggs and juveniles (Lo´pez-Urrutia et al., 2004). Whichever combination of factors it may be, the zooplankton community at the mouth was left dominated by A. natalensis until October 2008. Only then did the zooplankton community at the mouth begin resembling that of its pre-breaching state. Shifts in zooplankton community structure, such as this opportunistic dominance of O. dioica under closed mouth conditions, emphasize the complexity and erratic nature of the system in response to environmental variability.

simple food web (Carrasco and Perissinotto, 2012). In June 2009, a bloom of an orange-pigmented cyanobacterium (Cyanothece sp.) was recorded in False Bay, persisting uninterruptedly for 18 months. This bloom was probably initiated by high levels of nutrients regenerated from decomposition of organisms, which had died because of their inability to cope with the harsh conditions prevailing in the region during the previous months (Muir and Perissinotto, 2011). Stable isotope analysis suggests that Cyanothece sp. was the main prey item of F. cf. salina from May to July 2010. The ciliate in turn became prey to the cyclopoid copepod A. cf. dengizicus, which was eventually presumably consumed by greater and lesser flamingos that flocked to the area when the copepods attained swarming densities (R. H. Taylor, pers. comm.) (Figure 13.3). Greater flamingos feed mostly on zooplankton, but lesser flamingos are also capable of consuming cyanobacteria directly, accumulating their pigments to impart the typical pink coloration to their feathers

13.5.4 Halotolerant communities During 2010, the Lister’s Point region of False Bay supported a small zooplanktonic community, as only four taxa were able to withstand the harsh environmental conditions prevailing at the time. These were the flatworm Macrostomum sp., the harpacticoid copepod Cletocamptus confluens, the cyclopoid copepod Apocyclops cf. dengizicus and the ciliate Fabrea cf. salina. Both A. cf. dengizicus and F. cf. salina are potentially undescribed species (F. Fiers, W. Petz, pers. comm.). Apocyclops cf. dengizicus was also recorded as a dominant zooplankton species during 2005 and 2006 in North Lake (Jerling et al., 2010a) at a time when the water level in North Lake was very low and parts of the lake occasionally dried up. This was the first record of the genus Apocyclops in southern Africa. All five taxa demonstrated remarkable salinity tolerances, only disappearing from the region once salinity levels exceeded 130. In addition to having some of the highest recorded salinity tolerances for invertebrates, two of these taxa were involved in a remarkably

FIGURE 13.3 Diagram representation of the simple food web observed at False Bay (Lister’s Point) from May to June 2010.

Zooplankton

(Warren, 2006). Once salinity levels exceeded 130, all zooplankton disappeared from the water column, presumably after producing spores or resting cysts capable of surviving unusually harsh conditions for long periods of time. The heavy summer rains which fell in December 2010/January 2011 alleviated some of the drought conditions, resulting in the reappearance of the aforementioned species, as well as high densities of the rotifer Brachionus rotundiformis and the cladocerans Diaphanosoma cf. excisum and Moina cf. micrura and the harpacticoid copepod Nitocra taylori, which is a species only recently described and possibly endemic to St Lucia (Go´mez et al., 2012). Similar events have been previously recorded in the northern lakes of St Lucia, with hypersaline conditions resulting in a number of changes in some of the basic trophic relations (see above). For instance, the environmental conditions

experienced in the early 1970s were similar to those experienced during 2010 (Chapters 7 and 9). Thus, on the latest occasion it was expected that the system would respond in a similar manner to that observed earlier. However, this was not the case. Not only were the organisms involved in this study completely different from those documented by Grindley and Heydorn (1970), but they had also not been previously recorded in the St Lucia Estuary by Grindley (1976, 1982). It is possible that the magnitude of the salinity increase in the two different occasions may have played a critical role in this discrepancy. The hypersaline conditions which were experienced in the early 1970s ranged from 70 to 90, whereas salinity levels during 2010 ranged from 100 to 200. This is an indication that the system has indeed changed significantly over time, with the magnitude and intensity of the events increasing in response to climate change.

13.6 Trophic interactions The study of food webs has been useful in explaining community dynamics in highly variable environments such as estuaries. Zooplankton, as secondary producers, play a vital role in linking primary and/or secondary production with higher trophic levels and as the dominant zooplankton in St Lucia, Mesopodopsis africana, Pseudodiaptomus stuhlmanni and Acartiella natalensis have the potential to structure estuarine communities. Mysids and copepods are important in the food webs of aquatic ecosystems as both consumers and producers (Mauchline, 1980). As consumers, mysids are generally considered omnivorous, feeding on a wide range of items. Diet may include phytobenthos, phytoplankton, detritus, sediment, microzooplankton, mesozooplankton and small benthic invertebrates (Wilhelm et al., 2002; Kibirige and Perissinotto, 2003a; Lehtiniemi and Nordstro¨m, 2008, Vilas et al., 2008). Similarly, copepods are capable of utilizing a wide range of diets (Kleppel, 1993) and, like mysids, may feed either by means of

suspension-feeding currents or by actively capturing moving zooplankton, thereby resorting to raptorial or ambush feeding (Jiang and Osborn, 2004). In the St Lucia estuarine system, Carrasco and Perissinotto (2010) demonstrated the ability of M. africana to modify its diet in response to shifting environmental conditions, on both short temporal and spatial scales. Isotopic analysis suggested that M. africana fed mainly on mesozooplankton. At the St Lucia Mouth and in the Narrows (Esengeni), copepods and particulate organic matter (POM) were consumed in the highest proportions. In contrast, on the Eastern Shores of South Lake (Catalina Bay) the main food source for M. africana was POM, presumably because copepods were virtually absent from this station at the time. This suggests that in the estuarine system zooplankton diet is strongly linked to food availability, a trend which has been widely documented elsewhere (Jerling and Wooldridge, 1995; Viherluoto et al., 2000; Winkler et al., 2007; Vilas et al., 2008). On the Western Shores of South

261

262

Nicola K. Carrasco et al.

Lake (Charter’s Creek), with the exception of the macroalga Cladophora sp., M. africana generally consumed all available carbon sources in relatively even proportions (Carrasco and Perissinotto, 2010). While no change occurred in the dominant food items between wet and dry seasons, the proportions of the different sources in the mysid diet varied. The copepod P. stuhlmanni fed proportionally more on POM than on the other sources at the mouth, both during wet and dry seasons. The limited data available for A. natalensis suggest that it consumed preferentially POM, with sedimentary organic matter and microphytobenthos playing a minor role in its diet. The opportunistic feeding habit observed in all three species, M. africana, P. stuhlmanni and A. natalensis, may help prevent the exhaustion of any of the primary food sources, but such non-selective feeding patterns may have also been a reflection of stressful trophic conditions. Feeding by both M. africana and P. stuhlmanni was more selective at the mouth than at Charter’s Creek, a finding consistent with the harsh environmental conditions present at Charter’s Creek at the time. The high turbidity and salinity levels frequently recorded at this station, combined with low water levels, may have placed additional stress on zooplankton communities, decreasing their ability to feed selectively. Studies have shown that, amongst the zooplankton of St Lucia, M. africana occupies the highest trophic level, followed by P. stuhlmanni and A. natalensis (Carrasco and Perissinotto, 2010, 2011b; Govender et al., 2011). These three species also fall prey to a number of estuarine resident and juvenile marine fish species in the St Lucia estuarine system (Blaber, 1979; Carrasco et al., 2012; Chapter 15). Blaber (1979) documented their importance in the diet of zooplanktivorous fish, such as Gilchristella aestuaria, Hilsa kelee and Thryssa vitrirostris in St Lucia using gut content analysis. In this assessment, P. stuhlmanni was first in importance in terms of calorific contribution, followed by M. africana (Blaber, 1979). In a later assessment by Carrasco et al. (2012) using stable isotope analysis,

M. africana was found to contribute a substantial proportion to the diet of fish species such as Leiognathus equula, Ambassis ambassis and Gerres acinaces. For G. aestuaria, however, P. stuhlmanni was of more dietary importance. Oreochromis mossambicus exhibited the most variable diet, with all prey items contributing relatively even proportions to the diet. This flexibility in its diet may be advantageous in surviving harsh conditions, such as those currently experienced through most of the lakes region of the estuary and may explain their dominance throughout the system during the current freshwater deprivation crisis (Carrasco et al., 2012). Zooplankton also constitute an important food resource for some bird species (Chapter 16). While the greater flamingo Phoenicopterus roseus is the only obligatory filter-feeding bird in Lake St Lucia (Hockey and Turpie, 1999), certain waders are known to prey on the larvae of chironomids and the adults are eaten by the flock-foraging whitewinged tern and red-winged pratincole, as well as by several species of swallows and martins (Hockey and Turpie, 1999). There are also a number of bird species that feed on aquatic invertebrates such as gastropods (Assiminea cf. ovata and Nassarius kraussianus), bivalves (Brachidontes virgiliae and Solen cylindraceus), amphipods (e.g. Melita zeylanica), isopods (e.g. Philoscia hirsuta), crabs (Hymenosoma orbiculare) and prawns (Penaeus indicus) (Whitfield and Cyrus, 1978; Whitfield and Blaber, 1979a; Nel et al., 2011). Although these taxa are not classified as zooplankton, they all at some point have a planktonic stage within their life cycle. Whitfield and Cyrus (1978) studied the feeding succession and zonation of aquatic birds at False Bay and documented the importance of this food group, as aquatic invertebrate-feeding birds exceeded piscivorous feeders both in terms of abundance and species richness. Some of the dominant invertebrate-feeding birds included little egret (Egretta garzetta), avocet (Recurvirostra avosetta) and spoonbill (Platalea alba)(Whitfield and Cyrus, 1978).

Zooplankton

13.7 System threats 13.7.1 Extended mouth closure Extended periods of mouth closure may affect the zooplankton community in a number of ways, largely depending on salinity changes that occur during such an event, but also related to exchanges between the estuary and the adjacent marine environment. During prolonged mouth closure, water may either become hyposaline relative to seawater, if enough freshwater inflow is maintained, or it may become hypersaline if evaporation exceeds freshwater inputs. In either case the composition of the zooplankton community will change to reflect the salinity conditions (Grindley, 1982). In terms of exchanges with the sea, open-mouth conditions will lead to a much higher diversity due to the presence of euryhaline and stenohaline taxa entering the system (refer to section on ‘Diversity and community structure’ above). Another important consequence of extended mouth closure relates to the prevention of movement of species between the sea and the estuary, not only for fish, but also for those invertebrates that have an obligatory marine phase as part of their life cycle, for example various estuarine brachyuran species. The effect of such a closure in St Lucia can be illustrated by using brachyuran meroplankton as an example. Brachyuran zoea larval stages were recorded in mesozooplankton samples collected between 2005 and 2008 in St Lucia (Jerling et al., 2010a). Further analysis of data from that study indicated that both Hymenosoma sp. and Paratylodiplax blephariskios were present in the plankton. Their distribution patterns were, however, different. While no zoeae of P. blephariskios were recorded during 2005 to 2007 when the mouth was closed, it was the dominant zoea type in the Estuary (mouth and Narrows) and even extended into South Lake after the mouth had opened and closed again during 2007 (Figure 13.4). Hymenosoma zoeae were present in the Estuary before the mouth opened in March 2007, but were not recorded in this part of St Lucia after the mouth had opened (Figure 13.4).

The crab P. blephariskios was reported as the numerically as well as gravimetrically dominant benthic species in the Narrows section of the St Lucia system (Owen and Forbes, 1997). Historical studies showed a decline in abundance of this crab after mouth closure, leading Owen and Forbes (1997) to speculate on reasons for this decline as a possible intolerance to hypersalinity, an obligatory marine phase in the crab’s life cycle, or some effect of salinity on aspects of the feeding and breeding biology under elevated salinities (Owen and Forbes, 2002). However, during the recent mouth closure between 2004 and 2007, when hypersalinities were not recorded in the Estuary, the crab was also absent from the benthos (MacKay et al., 2010), indicating that hypersalinity was probably not the cause of the crab decline. The mouth opened in March 2007 and by May 2007 P. blephariskios was recorded in the benthos (MacKay et al., 2010), but not

FIGURE 13.4 Mean abundance of brachyuran zoeae recorded in three sections (Estuary, South Lake and North Lake) of the St Lucia system during 2005–2008. Vertical bars indicate SD. The mouth was open between March 2007 and August 2007.

263

264

Nicola K. Carrasco et al.

in the plankton (Figure 13.4). This suggests that the zoeae recorded in the plankton during November 2007 were released by the newly established benthic P. blephariskios population in the St Lucia Estuary. Recruitment for this population probably resulted from larvae and subsequent megalopa stages from the adjacent Mfolozi/Msunduzi system (refer to section on ‘Mfolozi–St Lucia link’ above). Adults of P. blephariskios do not normally occur in the lake basins of St Lucia (Owen and Forbes, 1997). The zoeae of this species recorded in South Lake were probably released in the Estuary and migrated, or were washed, into South Lake with the influx of seawater. Peak P. blephariskios zoeae densities were recorded in May 2008 after the mouth had closed again in August 2007; by November 2008 densities had again declined. In a study on brachyuran larval history strategies in the Mlalazi Estuary, situated approximately 100 km south of the St Lucia system, Papadopoulos et al. (2002) suggested that most of the crab species resident in the Mlalazi Estuary (including P. blephariskios) complete development in the marine environment, with the possible exception of Hymenosoma orbiculare. This species occurs all along the southern African coastline and displays great morphological variation. Edkins et al. (2007) proposed that South African crown crabs in fact represent five distinct species, with three of them yet to be formally described, one of which is located in the region of Lake St Lucia. Although H. orbiculare may be regarded as an estuarine crab, it is known to complete its life cycle successfully in coastal water bodies isolated from the sea, such as Lake Sibaya situated to the north of St Lucia (Allanson et al., 1966). It is, therefore, not surprising that zoeae of this species were recorded in the plankton of the Estuary and South Lake during the time when the mouth had been closed for an extended period (Figure 13.4). Results from the St Lucia mesozooplankton study support the suggested obligatory marine phase of P. blephariskios and emphasize the importance of an open mouth in the life cycle of this species. Zoeae released in the estuary during extended mouth closure will not be able to complete their life cycles, leading to the eventual extinction of the local benthic P. blephariskios population.

13.7.2 Silt loading from the Mfolozi Despite the high conservation priority status which the St Lucia estuarine lake has, the system has been subjected to a number of anthropogenic stressors, including raised levels of sedimentation. The estuary is a naturally shallow system, with an average depth ranging from 0.96 m to about 4 m and the entire water column is regularly mixed by wind-driven action (Forbes et al., 1987). Combined with increased silt deposition, this results in high turbidity levels generally prevailing within the lake system. Suspended sediments impact a whole range of physico-chemical and biological properties in ecosystems (Hart, 1986). Turbidity can have both adverse and beneficial effects on zooplankton (Hart, 1988). On the positive side, suspended sediment may hamper zooplanktivorous fish in their visual location of prey (Vinyard and O’Brien, 1976; Gardner, 1981). Additionally, dissolved organic matter may adsorb to surface-charged sediments, providing an additional energy source to those zooplankton organisms that are capable of ingesting such particles (Arruda et al., 1983). On the negative side, phytoplankton productivity generally declines with increased turbidity, as a result of particles backscattering light and reduced irradiance penetration (Kirk, 1985). This net reduction in algal food is compounded by the effect of excess suspended silt on the filter-feeding efficiency of zooplankton (Arruda et al., 1983; McCabe and O’Brien, 1983). Indirect effects of suspended sediment on water temperature may also be of importance in special instances (Schiebe et al., 1975). Mesopodopsis africana is a key species in St Lucia and contributes substantially to the diets of organisms at higher trophic levels. The effect of higher turbidity levels on this species would, therefore, be of immense importance. Experiments conducted by Carrasco et al. (2007) showed that higher silt levels indeed have a significant negative effect on mysid health, as mortality increased significantly with increased silt concentration (Figure 13.5). Additionally, mysids in higher silt concentrations were more opaque and generally less responsive than those in lower silt

Zooplankton

FIGURE 13.5 Mortality of Mesopodopsis africana at increasing silt concentrations. Error bars show standard error.

concentrations. Possible causes for this could include damage from physical abrasion with silt particles, or alternatively a depression in feeding rate at higher silt concentrations. Results showed that silt concentrations of about 2.58 g litre 1 (1088 NTU) and above significantly affect the survival of M. africana. This experiment was only run for 12 hours and it is possible that using longer incubation times mortality would increase, lowering significantly the critical limit of tolerance.

13.7.3 Hypersaline conditions and salinity tolerance In the St Lucia estuarine lake, a reversed salinity gradient has persisted since 2002 (with the exception of the brief open mouth phase in 2007), with extreme values ranging from 1.8 in the Narrows to > 200 in the upper reaches of the lake (Chapters 7 and 9). Due to shallow lake levels, water temperature in summer can at times exceed 40  C. The biotic communities living here are, therefore, characterized by a strong spatial heterogeneity. Temperature and salinity are both considered dominant ‘ecological master factors’, which may act individually or in synergy to modify the structure, function and distribution of estuarine communities (Kinne, 1971; Alderdice, 1972; Dorgelo, 1976). The survival of zooplankton species and the success of their populations are partly dependent upon their ability to tolerate the highly variable

conditions that characterize this environment. Temperature and salinity are also at the mercy of global climate change. According to the latest IPCC report (2007), temperatures are rising by 0.2  C per decade and the frequency and intensity of extreme weather events is also increasing (Planton et al., 2008). In St Lucia, rising temperatures will not only push the thermal tolerance limits of zooplankton, but increased evaporation from the large surface area of this estuarine lake will lead to concomitant salinity increases, which would exacerbate the impact on the resident communities. De Deckker and Geddes (1980) studied the salinity ranges of species present in the ephemeral salt lakes near the Coorong Lagoon in South Australia. Some of these lakes are susceptible to desiccation in the summer months, due to evaporation and lowering of the water table. Salinity levels, therefore, have the potential to become hypersaline to the point where salt is saturated and salt crusts form. These authors found that species richness during the drying phase of the lakes could be maintained as most of the taxa are salt tolerant. However, above a salinity of 150, species richness fell sharply and only the brine shrimp Parartemia zietziana survived salinity levels above 200 (De Deckker and Geddes, 1980). Grindley (1976, 1982) recorded salinity ranges for many of the common zooplankton taxa inhabiting the estuary, and suggested that most of the estuarine plankton species could tolerate and breed at salinity levels above 40. Species that were found surviving above 60 included Pseudodiaptomus stuhlmanni, Acartiella natalensis, Halicyclops sp., Mesopodopsis africana, gastropod larvae, and fish eggs and larvae. A few species, including a second species of Halicyclops, chironomid larvae and two species of unidentified harpacticoids survived salinity levels between 70 and 80 (Grindley, 1976, 1982). Between 2005 and 2011, a few species with exceptionally high salinity tolerance were recorded in the St Lucia estuarine lake. The cyclopoid copepod Apocyclops cf. dengizicus, the flatworm Macrostomum sp. and the harpacticoid copepod Cletocamptus confluens were found at salinity levels up to 130. The ciliate Fabrea cf. salina was even able to tolerate salinity levels above this. Harpacticoid copepods were not identified

265

266

Nicola K. Carrasco et al.

FIGURE 13.6 Diagram of updated salinity tolerance for the main species (revised after Grindley, 1976, 1982). The asterisk (*) denotes taxa which were reported by Grindley (1976, 1982) but not confirmed in recent records.

down to species level in Grindley’s (1976, 1982) studies. It is, therefore, possible that C. confluens was also present during Grindley’s earlier surveys. However, F. cf. salina, Macrostomum sp. and A. cf. dengizicus have not been previously documented in the system. A diagram of updated salinity tolerances of the main zooplankton species (revised after Grindley, 1982) is provided in Figure 13.6. The above-mentioned deductions of salinity tolerance were based on the occurrence of individuals in their natural environment. Given the importance of M. africana in St Lucia, further work was conducted recently in order to determine its temperature and salinity tolerance, both through in situ studies and the use of laboratory experiments (Carrasco and Perissinotto, 2011a). Two types of experiments were conducted: mysids held at 14 (ambient salinity) were either directly transferred to target salinities (shock experiments); or were acclimated to these salinity levels by gradually increasing the salinity over 24 h until target salinity treatments were reached (acclimation experiments). Results indicate that M. africana is a broad euryhaline species and, on a global scale, the only other species confirmed to be as tolerant of hyperhaline conditions as M. africana is Neomysis intermedia, which is capable of surviving salinity levels from 0.06 to 45 (Murano, 1966; Roast et al., 2001). Mesopodopsis africana was especially tolerant of the

lower salinity levels used in experiments. In the 20  C acclimation experiment, the time taken for half of its population to die when exposed to salinities of 1 and 2.5 (LS50) ranged from 8 to > 168 h, respectively (Carrasco and Perissinotto, 2011a). Acclimation is an important factor in establishing salinity tolerance of organisms (Bhattacharya, 1982; Delisle and Roberts, 1986; Kefford et al., 2007), as stepwise transfer of individuals to different salinity levels may increase their tolerance (Bhattacharya, 1982; Baylon and Suzuki, 2007). In contrast, repeated salinity changes may also have a cumulative negative impact and increase mortality at salinity levels that are otherwise tolerated (Vilas et al., 2009). Carrasco and Perissinotto (2011a) showed that acclimation significantly increases mysid survival, with acclimation rate being of critical importance in this process. Mysids exposed to salinity levels of 50 and 60 did not survive past the first hour during the in vitro experiments, but survived at 64 during field acclimated conditions. Thus, it is possible that by using acclimation rates of weeks to months, experimental results may approach those obtained in the field. To our knowledge, this is the highest salinity-tolerance threshold obtained so far for any mysid species. However, on occasions when mysids were collected at 60 to 64 in the field, it was observed that they were physiologically stressed and probably

Zooplankton

experiencing sublethal effects, as they were highly inactive and weakly responsive to mechanical stimulus (N. K. Carrasco, pers. obs.). While M. africana exhibits some of the highest recorded upper salinity and temperature tolerances recorded for a mysid, temperatures above 31  C and salinities above 65 appear to exert a marked influence on the distribution of this species. Such high values are often recorded in the shallowest parts of the St Lucia estuarine system. Because of the importance of mysids in estuarine trophic webs, the understanding of their community ecology is crucial in aiding the development of ecosystem-based management plans for the St Lucia system.

13.7.4 Desiccation effects Due to the nature of the most recent freshwater deprivation crisis (2002–2011), desiccation of the system has become a serious threat. Evaporation of water not only means less habitat for zooplankton, but exponential salinity increases as well. Shallow water levels are also more susceptible to heating and high temperatures are known to negatively affect the physiology of many zooplankton species (e.g. Moore et al., 1996; Norberg and DeAngelis, 1997). During the current freshwater deprivation crisis, the northern lake basins have been extremely susceptible to desiccation. At the peak of the drought in 2005–2006, up to 70% of the lake bed was dry (Whitfield and Taylor, 2009). Groundwater seepage points along the Eastern Shores have been hypothesized to form microhabitats of reduced salinity, capable of acting as reservoirs during dry cycles and providing refugia for estuarine species, for later recolonization of the estuary (Taylor et al., 2006b). However, data from a study carried out in 2005 (Singh, 2009), indicate that the zooplankton

communities of these refugia are significantly different from those of neighbouring sites and, therefore, probably not the primary source of recolonization and recruitment. It is more probable that recolonization occurs from adjacent parts of the estuary that do not completely dry out during drought periods (e.g. the Estuary, South Lake or the Mfolozi system). It is also likely that some species of zooplankton are able to produce resting stages/eggs capable of surviving dry/hypersaline conditions for long periods of time (e.g. Uye, 1985; Williams-Howze, 1997; Engel, 2005; Moscatello and Belmonte, 2009). In the salt lakes near the Coorong Lagoon, South Australia, most of the cladocerans, calanoid copepods and ostracods are known to survive the dry phase of lakes as eggs (Belk and Cole, 1975). However, some others survive as immature or adult stages as certain cyclopoid and harpacticoid copepods are capable of undergoing encystment as adults (or near adults). Although encystment and diapauses at this stage are well known, there are few records of them surviving in dry lakes (De Deckker and Geddes, 1980). Evidence of this has recently been found in the St Lucia system, as numerous cysts of the ciliate Fabrea cf. salina were observed in samples collected at False Bay during extremely hypersaline conditions in 2010. While Lake St Lucia has shown a high degree of resilience, harbouring zooplankton species that are capable of enduring harsh environmental conditions both through salinity adaptations as well as the ability to form resting stages or spores, continued freshwater deprivation may jeopardize the unique biodiversity that the system harbours and it is uncertain at this point to what extent the system will recover, should freshwater deprivation be mitigated.

Acknowledgements Thanks are due to the iSimangaliso Wetland Park Authority and the staff and management of Ezemvelo KwaZulu-Natal Wildlife (EKZN Wildlife). Funding for this project was provided by the National Research Foundation (NRF, Pretoria),

Marine and Coastal Management (MCM, Cape Town), the World Wide Fund for Nature (WWF-SA, Stellenbosch) and the South Africa-Netherlands Research Programme on Alternatives in Development (SANPAD, Durban).

267

Chapter contents 14.1 Introduction 14.2 Penaeid diversity and distribution 14.3 Prawns in South Africa: species, distribution, life cycles and biology 14.4 Immigration and emigration: movement, dispersal and origin of South African prawns 14.5 Population dynamics 14.6 The estuarine bait fisheries 14.7 Ecological significance of prawns in the St Lucia estuarine system 14.8 Status and threats to the major nursery grounds and the offshore habitat: the future of South African prawn stocks

Penaeus indicus collected from poachers’ nets in False Bay. (Photo: Ricky H. Taylor, April 2012.)

14

Penaeid prawns Anthony T. Forbes and Nicolette T. Forbes

14.1 Introduction The St Lucia bait prawn fishery, which operated for approximately 45 years from the early 1950s to the mid 1990s when it was officially abandoned, embraced an era of major change in the life of the fishing village of St Lucia. Human manipulation of the river catchments, floodplains and mouth dynamics, the latter particularly in the case of the Mfolozi River, eventually demonstrated that it was possible to push the St Lucia estuarine system, despite its large size, into an unprecedented physico-chemical condition. This occurred in the early years of this century, when record high salinities were followed by the lake drying up for the first time in recorded history. Apart from various other impacts, the mouth management policies for the Mfolozi and St Lucia, exacerbated by catchment abstractions of one sort or another and further exacerbated by harbour developments at Richards Bay, combined to remove the regional prawn nursery function served by these two systems, culminating in the collapse of the inshore prawn trawl fishery. The history of the development, operation and final demise of the St Lucia bait prawn fishery is inextricably linked with the long-standing history of the St Lucia estuarine system as a fisherman’s mecca, home of the grunter run and fishing competitions, despite the lake falling into one of the oldest nature

conservation areas in the country. Angling requires bait and the long-standing provider of prawn bait for anglers was the then Natal Parks, Game and Fish Preservation Board. The subsequent shortening of this rather cumbersome title to the Natal Parks Board possibly reflects changing perspectives. Folklore of the time had it that successful fishing in the St Lucia estuarine system ideally required bait prawns from the St Lucia estuarine system, since prawns from elsewhere were less effective. This chapter will begin by describing the typical prawn life cycle and the biology of these organisms in the St Lucia estuarine system, followed by the nature of the bait fishery operation and finally the contribution of St Lucia to prawn population dynamics off the east coast of South Africa. Use was made of the published literature, unpublished personal field observations, catch data and the results of tagging programmes carried out by the authors in Richards Bay and St Lucia, unpublished catch data obtained from the records of the Chief Directorate: Marine and Coastal Management, minutes of the Prawn Fisheries and Development Association and inputs from various people who have been involved in different ways with the prawn resource in KwaZulu-Natal and elsewhere in the country. The latter are acknowledged at the end of the review.

Ecology and Conservation of Estuarine Ecosystems: Lake St Lucia as a Global Model, eds. Renzo Perissinotto, Derek D. Stretch and Ricky H. Taylor. Published by Cambridge University Press. © Cambridge University Press 2013.

270

Anthony T. Forbes and Nicolette T. Forbes

14.2 Penaeid diversity and distribution The penaeid prawns, or shrimps as they are referred to outside the Indian and western Pacific Oceans, have a world-wide distribution in tropical and subtropical seas within the 20
C isotherms. They are most varied in the Indo-Pacific, where there are roughly five times more species than in the Atlantic. The Indo-Pacific region, which stretches from eastern Australia and the Philippines to eastern and southern Africa, supports 125 species of which 124 are endemic. Diversity and endemism decrease in all directions from this subregion, declining to 16 species and one endemic in the southern African region, which stretches from Durban to Swakopmund (Dall et al., 1990). The life cycles in all known members of the family Penaeidae involve planktonic larvae with a variety of naupliar, protozoeal, mysis and postlarval stages, followed by juvenile and adult stages (Figure 14.1). The greatest differences between the species lie in the preferred habitats of postlarvae, juveniles and adults; that is, whether they are predominantly estuarine, inshore or offshore, and whether demersal or pelagic. Depending on these preferences, Dall et al. (1990) recognized four different types.

Type 1 species appear to be restricted to smaller members of the genus Metapenaeus and may have entirely estuarine life cycles in which the postlarvae move upstream into lower salinities while the juveniles subsequently return towards higher salinities in the lower estuary, mature and breed. No South African examples exist. The postlarvae of most Type 2 species prefer estuaries or estuarine-like environments. This preference is characteristic of most Penaeus and Metapenaeus species including all the commercial South African examples. This habitat preference demands a degree of euryhalinity because of the salinity fluctuations typical of estuaries, but this tolerance does vary from species to species. Speciesspecific substratum preferences are typical of these estuarine juvenile stages and heterogeneous estuarine environments incorporating a variety of habitats will support a greater variety of species. During maturation, the juveniles or subadults emigrate from the estuaries to the offshore adult breeding grounds. The size or degree of maturity at emigration varies from species to species. Substratum FIGURE 14.1 Life cycle of a typical Type 2 species. The blue line separates the planktonic from the benthic phases in the life cycle and illustrates the stages at which immigration occurs and emigration begins (Graphics: Marine & Estuarine Research, 2011).

Penaeid prawns

preferences characteristic of the juveniles tend to be maintained in the adult phases. The life span of most Type 2 species is generally accepted as about one year in the tropics, but up to two years in cooler, more temperate waters where growth rates would be slower. Laboratory development time from spawning to postlarva is about eight to 24 days, largely depending on temperature, while estimates of field durations are about 14–21 days. Recruitment of the postlarvae to the estuarine nursery grounds from the adult breeding grounds offshore is followed by a period of growth in the estuaries, after which the juveniles or subadults return to the marine environment and complete their growth phase. There is no known deviation from this cycle in the South African species.

Virtually wherever they occur, prawns with a Type 2 life cycle are exploited in both the juvenile, estuarine phase and in the offshore, adult breeding phase. As might be imagined, this situation can be a cause for conflict between users of the juvenile and adult resources with each group of fishers claiming that their particular ‘rights’ are being infringed. The larvae of Type 3 species are pelagic while the postlarvae prefer relatively high salinity, usually sheltered, inshore waters. The adults remain in the marine environment. The only South African example of this type is the endemic, cool water, noncommercial Macropetasma africanus. Type 4 species have an entirely offshore life cycle; some are harvested in South African waters. Neither type is relevant in the St Lucia context.

14.3 Prawns in South Africa: species, distribution, life cycles and biology Revisions by Perez-Farfante and Kensley (1997) of what were previously subgenera within the genus Penaeus resulted in several name changes at the generic level. Published references incorporating these name changes have been left as such. The recent studies of mitochondrial and nuclear genes (Ma et al., 2011) have, however, argued for the retention of the genus Penaeus and this convention will be followed in this chapter. The distribution of these basically tropical species on the east coast of South Africa is enhanced by the warm, southward-flowing Agulhas Current and results in a general extension of the tropical and subtropical fauna of the western Indian Ocean into the South African region (Macnae, 1962). There is a rapid attenuation of this fauna, including the penaeid prawns, towards the southern boundary of KwaZulu-Natal, although the seasonal occurrence of prawns in estuaries, at least as far south as Port Elizabeth (Figure 14.2), has long been appreciated by local anglers (Hughes, 1966). The seven species recorded in the region are listed in Table 14.1. Of these, P. canaliculatus (the striped prawn) and P. latisulcatus (the western king prawn or

brown prawn, as it is known in the South African region), although having very wide distributions in the Indian and western Pacific Oceans (Dall et al., 1990), are by far the least common and the least known in the region and can be disregarded in the St Lucia context. Identification of the prawns that might be encountered in the St Lucia estuarine system can be done using the rostral formula, which refers to the number of teeth on the upper and lower sides of the spiny projection between the eyes. The speckled prawn, Metapenaeus monoceros, has no teeth on the lower side and is the only species in the region with this feature. The ginger prawn, P. japonicus, has 9–11 teeth on the upper side and a single tooth on the lower side. It is extremely rare in the St Lucia estuarine system, as it prefers sandy substrata into which it burrows, emerging only at night. The green prawn, P. semisulcatus, has a rostral formula of 6–7 over 3. It is strongly associated with weed beds and is, accordingly, generally rare in St Lucia. The tiger prawn, P. monodon, has a 7 over 3 rostral formula. It is not as common as the speckled prawn, it is not dependent on weed beds and is far

271

272

Anthony T. Forbes and Nicolette T. Forbes

FIGURE 14.2 Localities on the South African coast mentioned in the text, the regional areas on the KwaZulu-Natal coast and details of the southern sections of the St Lucia system.

Penaeid prawns

Table 14.1. Southerly distribution records of penaeid prawns in South Africa (see Figure 14.2 for places and estuaries mentioned in the table) P. canaliculatus Olivier

P. indicus MilneEdwards

P. japonicus Bate

P. latisulcatus Kishinouye

P. monodon Fabricius

P. semisulcatus de Haan

M. monoceros Fabricius

Port Edward

Swartkops

Knysna

No record

Swartkops

No record

Nyara

Day et al., No record 1952

Knysna

Knysna

No record

No record

No record

No record

Hughes, 1966

Swartkops

Swartkops

Swartkops

Swartkops

No record

No record

de Freitas, Swartkops 1980

Knysna

Knysna

Durban Bay

Swartkops

Durban Bay

Durban Bay and off Port St Johns

Day, 1969 East London

Keiskama

Breede

No record

Bashee

Durban Bay

Keiskama

Branch East London et al., 1994

‘Transkei’

False Bay

No record

No record

Durban Bay

East London

Author Barnard, 1950

Swartkops

more common than the green prawn. As the name suggests, it tends to be striped and also grows to a larger size in estuaries than any of the other species. The Indian or white prawn, P. indicus, has a rostral formula of 7–8 over 4–5. It is typical of muddy bottoms and turbid waters, has very little coloration and is by far the most common species in the southeast African region. The southern limits of distribution of all the above species (Table 14.1) are indicative of their tropical affinities. The records mentioned in the table in some cases (Day et al., 1952; Hughes, 1970) reflect collections by the authors concerned, but in the others are based partly on museum specimens (Barnard, 1950) or literature records. Lack of agreement between the different authors is striking. It would presumably be expected that, over time, recorded southern limits for species such as these with their offshore planktonic larval stages are more likely to expand. However, in some cases the same author (Day et al., 1952; Day, 1974) extended a species range in one case (P. japonicus), but reduced

it in another (P. indicus). Overall, the number of species historically recorded increased from five at St Lucia (Joubert and Davies, 1966) to seven at Durban (Joubert, 1965), dropping again to five in the Swartkops Estuary at Port Elizabeth (Hughes, 1970). Only two species have been recorded beyond Port Elizabeth, namely P. indicus as far as Knysna, and P. japonicus from the Breede River (Day, 1974) or ‘False Bay’ (Branch et al., 1994) (Figure 14.2). Prawn benthic habitat preferences are typically well defined in the juvenile stages, which occur in estuarine nursery grounds, and in the adult stages which occur on the offshore, but generally shallow breeding habitats. In the southern African situation, more information is available regarding the juvenile than the adult preferences. There is no published information specifically on South African juvenile prawn habitat preferences, but as the local species also occur in Mozambique, information derived from Maputo Bay, which is a relatively large system with a variety of benthic habitats, is incorporated here. In an early paper, Hughes (1966) described the

273

274

Anthony T. Forbes and Nicolette T. Forbes

habitat preferences of P. monodon, P. semisulcatus, P. indicus and M. monoceros in Maputo Bay. De Freitas (1986) extended this investigation to include P. japonicus and M. stebbingi; the latter species has not been recorded south of Maputo Bay. De Freitas (1986) questioned Hughes’ identification of P. semisulcatus and accordingly de Freitas’ description of habitat preferences will be incorporated here. De Freitas (1986) selected 11 stations in Maputo Bay, which were sampled for any prawns present. Penaeus semisulcatus was obtained almost exclusively on intertidal flats covered with seagrasses. This accords with records from the St Lucia estuarine system, where this species was uncommon and only obtained in Zostera beds (pers. obs.). Penaeus japonicus in Maputo Bay was obtained only on intertidal flats in bare, sandy mud to muddy sand where it burrowed during daytime or during low tides. This behaviour was mirrored in Durban Bay, where push netters for many years collected ‘ginger shrimp’ at night over sandy sediments in shallow water during low tide periods (Joubert, 1965). De Freitas (1986) recorded P. indicus and P. monodon in similar habitat types, although

P. indicus was three times more abundant. Both species were most common in muddy areas within mangrove swamps. A similar situation prevailed in the major KwaZulu-Natal nursery area of St Lucia and Richards Bay, where the bait prawn fisheries operated in the muddy channels and backwaters of these two areas and produced primarily the white prawn, P. indicus, and very much smaller numbers of the tiger prawn, P. monodon, and the speckled prawn, M. monoceros. De Freitas (1986) described M. monoceros as occurring in a ‘more diverse number of habitats; from areas with submerged macrophytes to the deeper reaches of the mangrove swamps’ and most individuals obtained were caught in ‘a low salinity zone’, which was interpreted as demonstrating a greater tolerance of low salinities than the Penaeus species. Whether the occurrence of M. monoceros in low salinities represents a distinct preference is unknown. Unpublished observations of bait fishery catches and collections in St Lucia and Richards Bay supported the observations and conclusions reached in the Maputo Bay area.

14.4 Immigration and emigration: movement, dispersal and origin of South African prawns Day et al. (1954) as well as Millard and Harrison (1954) all referred to the uncertainty at that time of the nature of the shallow-water penaeid prawn life cycles in South Africa. As indicated earlier, the fundamental nature of these life cycles is now well known for most of the commoner species, including the South African examples, although there are clearly still real differences between the species in terms of aspects such as habitat preference, diel behaviour, dietary preferences and growth rates. Despite the extensive literature, last reviewed by Dall et al. (1990), there are still many unknowns or uncertainties regarding aspects such as

mechanisms of migration of larvae and postlarvae from marine breeding grounds into estuaries and subsequent emigration to adult habitats, and also the extent of dispersal of planktonic larval stages from the breeding grounds. It is also relevant that the combination of particular habitat requirements and availability of such habitats on the east coast of Africa due to the generally narrow continental shelf results in prawn habitats, and therefore fisheries, being localized into specific areas in the western Indian Ocean, typically associated with suitable estuarine nursery grounds.

Penaeid prawns

The life cycles of the local penaeid prawns (Figure 14.1) are punctuated by two major events, firstly the transition from a marine planktonic larva to a benthic estuarine juvenile, which obviously represents a major morphological, physiological and behavioural transformation, and secondly the return of the juveniles or subadults from the estuarine nursery grounds to the offshore adult breeding grounds. The analyses of offshore catch species composition referred to earlier, and the presence of tiger prawns, Penaeus monodon, only in the two largest commercial size classes, indicate a minimum threshold size before emigration in this species. The emigration of P. indicus is a more protracted process and involves a wider range of size classes beginning well before adult size is reached, as shown by the presence of individuals of this species in the smallest offshore commercial size classes. The mechanisms involved in migration from breeding grounds to estuarine nurseries in P. plebejus in Australia have been discussed by Rothlisberg et al. (1995), but not investigated experimentally. Available evidence indicates that larval migration in any of the Type 2 penaeid species does not involve active swimming up some sort of concentration gradient, but rather the selective use of currents based on active vertical migration. The details of the integration of this process with the accessing of suitable currents remains to be clarified. Subsequent movement and dispersal in estuaries represents a different process, as these occur in a tidal environment where currents flow in opposite directions at regular intervals. This latter phase coincides with the transformation of the planktonic, migratory postlarval stage into the benthic, substratum-specific juvenile. It was apparent from St Lucia studies that there was very little variation in the size of the postlarvae recruiting into that system in any one species. More than 80% of P. indicus, P. japonicus and P. semisulcatus postlarvae had carapace lengths between 1.5 and 2 mm; more than 90% of

P. monodon and 75% of Metapenaeus monoceros fell between 2 and 3 mm (Forbes and Benfield, 1986). The conclusion arising from this situation was that immigration and dispersal of postlarvae in estuarine environments involves a very specific set of behavioural cues and responses, which are associated with a particular period in the development of the prawn. The stages preceding the postlarvae are entirely planktonic, while the subsequent juvenile stages are essentially benthic. The postlarvae are intermediate, alternating between planktonic and benthic phases following a largely tidal, diel or lunar rhythm. The literature is replete with references to statements that postlarvae are most abundant in the estuarine water column over flood tide periods, particularly at night and often over particular moon phases (Dall et al., 1990). Possible cues to the above behaviour were investigated by Hughes (1969), who suggested that a salinity response was involved. He proposed and experimentally found support for a hypothesis that postlarvae on the bottom responded to increasing salinities associated with flood tides, by becoming more active and thereby being transported upstream. During subsequent slack water, the postlarvae sank back to the bottom. Falling salinities during ebb tides did not elicit any response, but the subsequent flood tide and increasing salinities again resulted in greater activity and further transport upstream. This process would be repeated and would result in net upstream transport and the possibility of the postlarva locating a suitable substratum on which to settle. The possibility of an additional mechanism was investigated by Forbes and Benfield (1986) in relation to the St Lucia estuarine system, where Champion (1976) noted that prawns were still present in the Narrows during periods of hypersalinity; that is, when salinities in the lake were greater than seawater. This situation was at odds with the Hughes (1969) model, which was predicated on a normal estuarine salinity gradient; that is, one in which

275

276

Anthony T. Forbes and Nicolette T. Forbes

salinities declined upstream. Wickham (1967) established the presence of a pressure response in P. duorarum juveniles in the USA, although this was discarded as a possible mechanism for migration by Penn (1975) in the case of P. latisulcatus in Shark Bay, Western Australia, because of the depth at which the postlarvae and juveniles occurred and the small local tidal range. Forbes and Benfield (1986) investigated a possible pressure response in the postlarvae recruiting into the St Lucia estuarine system, based on the hypothesis that a response by the postlarvae to increasing salinities during a flood tide in a normal estuary could be replaced by an ability to detect increasing hydrostatic pressures on the bottom, resulting from depth increases during flood tides. Conversely, hydrostatic pressures would decrease on the bottom during ebb tides. In an experimental laboratory situation, where postlarvae of P. japonicus were exposed to simulated pressure changes that would occur on the bottom in the Narrows, the postlarvae responded by becoming more active under increasing pressures, a response which would have allowed them to be transported up the Narrows in a manner corresponding to Hughes’ salinity response model. The pressure changes which generated increased activity were in the order of 1–2 kPa. Whether all species have a salinity or a pressure response, or both, and how either is selected is unknown. Emigration of juveniles or subadults from the estuarine nursery grounds to the marine environment has been suggested to be triggered by minor flood events which would be a feature of the summer rainfall areas occupied by the penaeid prawns in South Africa. In addition, the major South African species P. indicus is relatively sensitive to salinities < 10 and would not require major floods to be induced to leave the estuarine environment. At the same time it has been noted that, unlike P. monodon which only contributes to the largest offshore size categories, the smallest size classes in the offshore fishery include P. indicus and, while estuarine catch

records indicate that P. indicus seldom exceeds a carapace length of 25 mm in the estuaries under open mouth conditions, emigration begins well before this threshold is reached. The cues for this species are unknown. Tagging experiments using Australian streamer tags were carried out in successive years (January 1990 and February 1991) in Richards Bay (Figure 14.2) and St Lucia. Five thousand prawns were tagged at Richards Bay and 1550 at St Lucia. Fifty-four of the Richards Bay tagged animals were recovered from the Thukela Bank (Figure 14.2) by the prawn trawling fleet over a period between 44 and 135 days after release. Coordinates of the recovery sites indicated net movements of more than a kilometre per day in some cases. Fifteen of the prawns tagged at St Lucia in 1991 were captured by the prawn fleet on the Thukela Bank during the following season. In each case the recoveries amounted to c. 1% of the tags. The above operation was mounted in response to a user conflict situation between the then estuarine bait fisheries and the offshore trawl operation and a statement that the Thukela Bank prawns originated from Mozambique. While the tagging operation clearly demonstrated, as might have been expected on a geographical and proximity basis, that prawns from Richards Bay and St Lucia were contributing to the Thukela Bank catches, it obviously did not entirely resolve the question of input from Mozambique. Maputo Bay has historically supported a prawn fishery (de Freitas, 1980) based largely on P. indicus. Maputo Bay and St Lucia are about 260 km apart with only Kosi Bay and the small Mgobezeleni Estuary in between. Neither of these systems generate much silt due to the sandy nature of their catchments. The local neritic zone, and particularly the benthic habitat, are consequently totally unsuitable environments for the prawn species dominating the catch on the Thukela Bank. It therefore appears highly unlikely that juvenile prawns emigrating from the local Mozambican estuaries into Maputo Bay would be able to negotiate

Penaeid prawns

the intervening unfavourable environment and thereby contribute to any significant degree to the Thukela Bank population. The above argument suggests that prawn stocks on the Thukela Bank are derived largely from KwaZulu-Natal nursery grounds, and are dependent on juveniles emigrating particularly from the large systems of St Lucia and Richards Bay. A similar question arises in relation to the origin of postlarvae migrating into the KwaZulu-Natal nursery grounds: are they derived from a Thukela Bank breeding population, or is there an input from ‘further north’ as suggested by Macnae (1962)? Available genetic information on P. indicus (Querci, 2003) and P. monodon (Forbes et al., 1999) from Mozambique, Madagascar and South Africa does not indicate any distinction between populations from these areas. Penaeid prawns are, however, genetically highly conservative and this lack of any contrast cannot be taken as proof of regular genetic exchange between South African populations and those further north. As already mentioned, the habitat requirements and life cycles of the penaeid prawns result in the presence of isolated stocks in the entire western Indian Ocean region and the degree of exchange between them is unknown. In the absence of conclusive genetic

evidence, an attempt was made to resolve the question of local larval sources by using data on postlarval recruitment into KwaZulu-Natal estuaries comparing Kosi Bay, St Lucia, Richards Bay and Durban Bay (Figure 14.2). Forbes and Cyrus (1991) and Forbes et al. (1994) described the contrast in the species composition of immigrating postlarvae between Kosi Bay, which was dominated by P. japonicus to the virtual exclusion of any other species, and the three more southerly systems in closer proximity to the Thukela Bank where, in addition to P. japonicus, there was recruitment of postlarval P. indicus, P. monodon and M. monoceros. Assuming a ‘northern source’ and a south-flowing current, it would be expected that postlarval recruitment into estuaries en route southwards would be similar in species composition. The disparity between Kosi Bay and the more southerly systems of St Lucia, Richards Bay and Durban, where a greater variety of recruiting species was noted in all three systems, argues for a southern source of postlarvae and consequently a Thukela Bank – St Lucia/Richards Bay axis with a high degree of separation and independence of this southern stock from that in Mozambique.

14.5 Population dynamics The dynamics of penaeid prawn populations worldwide have long been a subject of study for research scientists and, obviously, of interest to the industry because of the fluctuations in annual population size and the consequent variations in annual catches which characterize these fastgrowing, highly fecund organisms. The catch is based on the generation of the year, or the 0+ year class in fisheries terms. The biomass that contributes to the catch is therefore generated each year and there is none of the averaging effect seen in longerlived species, where the total biomass of the

population is based on a succession of generations or year classes. The life strategy of fast growth and high fecundity and the variability of the habitats occupied, particularly as juveniles, because of their proximity and vulnerability to catchment processes can be seen as factors contributing to variations in annual generation size. Death rates in these fecund species are typically highest in the early stages of development. Variations in these mortality or survival rates at these early stages are therefore the factors most likely to affect subsequent population sizes.

277

278

Anthony T. Forbes and Nicolette T. Forbes

It is common cause amongst local trawler operators that a drought adversely affects the catches, although they have no readily available explanation. This effect is not unique to South Africa; catches of the banana prawn, Penaeus merguiensis, in the Gulf of Carpentaria in northern Australia have been strongly correlated with rainfall (Staples et al., 1981). Initial correlation analysis (Demetriades et al., 2000) supported the relation between summer rain and offshore catches the following winter. Subsequent more detailed consideration of the possible interactions between P. indicus and catchment runoff included the possible negative effect of very low salinities in the nursery areas, which would make them inaccessible. Runoff would also cause introduction of terrestrially derived nutrients into the generally oligotrophic inshore waters of KwaZulu-Natal, which would contribute to phytoplankton production and thereby potentially enhance survival of the planktonic larval stages. Timing of the latter events was crucial and the model produced indicated that runoff in late winter–early spring, when postlarvae were still recruiting to the estuaries, was a critical period. Favourable conditions for larval development in combination with favourable nursery ground conditions for the juveniles would then be reflected in greater offshore catches in the following winter. There is not necessarily a clear correlation between estuarine abundance and subsequent offshore catches (Forbes and Benfield, 1986), because of the

variability in the size at which juvenile P. indicus leave the estuary. Reductions in salinity in the nursery grounds would simply act as a trigger for earlier emigration. At this stage, the relationships described above are still largely qualitative and will initially require, at a minimum, data on nutrient inputs to the Thukela Bank and some measure of phytoplankton production on the Thukela Bank to test the hypothesis. Support for the acceptance of variations in postlarval abundance and estuarine recruitment was provided by sampling done between 1982 and 1992 at Kosi Bay, St Lucia, Richards Bay and Durban Bay (Forbes et al., 1994). Unfortunately, because of the nature of the work and the associated logistics, it was impossible to sample all areas simultaneously and the possible significance of catchment runoff was not appreciated when the programme was begun. The available data are therefore patchy and cover St Lucia over the periods 1982–1984 and 1988, Kosi Bay 1988 and early 1989, Richards Bay and Durban Bay mid 1991 to mid 1992. The most consistent species was P. japonicus, but from the point of view of the fishery and the dominance of P. indicus, recruitment of this latter species into St Lucia was up to two orders of magnitude higher in 1988 than at any time over the period 1982–1984. The years 1982– 1984 were drought years, while 1988 was still showing the effects of Cyclone Domoina four years earlier and coincided with record catches on the Thukela Bank.

14.6 The estuarine bait fisheries 14.6.1 History and status Although the distribution of penaeid prawns on the east coast of South Africa extends well beyond the centre of abundance in KwaZulu-Natal estuaries and on the Thukela Bank (Figure 14.2), and while they are well known to anglers in Eastern Cape rivers several hundred kilometres to the south, commercial

harvesting has historically been limited to the more tropical St Lucia estuarine system and to Richards Bay, both of which are in relatively close proximity to the adult habitat on the Thukela Bank. This situation is also in keeping with the relative sizes of the many South African estuaries. The St Lucia estuarine system, with an area of c. 300 to 350 km2,

Penaeid prawns

depending on water levels, contributes c. 75% of the total estuarine habitat in the country (Begg, 1978). According to Champion (1976), ‘the St Lucia estuarine system represents the largest estuarine penaeid prawn reservoir in South Africa’. This now has to be seen as indicative of the period. Since then, salinity fluctuations in the lake, ranging from fresh to hypersaline, and the extreme drought conditions during summer 2003–2004, when the mouth closed and the lake virtually dried up (January 2004) for the first time in recorded history, resulted in the loss of this nursery area. This loss of nursery function will occur both during fresh and hypersaline periods. Both the St Lucia and Richards Bay fisheries referred to above were operated by the Natal Parks Board (NPB). The prawns of St Lucia and the fishery were first referred to in terms of species composition, distribution, habitat preference and relative abundance in the pioneering estuarine surveys carried out by Day et al. (1954). According to Champion (1976), the NPB-run commercial bait fishery in St Lucia began in 1952, although Joubert and Davies (1966) refer to private ‘small scale’ commercial exploitation ‘before World War II’. The reasons for the undertaking of these operations by the NPB are uncertain, but presumably represented an attempt to control the harvesting of prawns and to resolve the dilemma of what species are or should be afforded protection in areas such as these two systems, where the ‘right’ to catch fish has long been taken for granted, despite the protection afforded to other organisms such as mammals and birds. The St Lucia estuarine system falls within one of the oldest proclaimed nature reserves in the country. On this basis, it could be argued that the operation of the bait fisheries by the NPB and the exclusion of recreational, and specifically other commercial operators, conferred a level of protection on the stocks as effort could be controlled. There is, however, no record of the manner in which the level of effort was set and, in the absence of any indication to the contrary, it is a reasonable assumption that it was based simply on demand for

FIGURE 14.3 The standard prawn trawling rig of aluminium dinghy, single outboard motor, gate net and three-man crew used in the St Lucia prawn fishery (Graphics: Marine & Estuarine Research, 2011).

bait prawns and the ability of the trawl operation to meet this demand. The prawn catching technique involved aluminium dinghies powered by single 25 to 30 hp outboard motors, with a three-man crew (Figure 14.3). Three such rigs were operated at St Lucia. The nets used were referred to as ‘gate nets’ and consisted of rectangular aluminium frames 4.9 m wide by 1.1 m high, mounted on skids which prevented them from sinking into the generally muddy areas trawled. The nets were approximately 4 m deep with 25 mm stretch mesh. Photographs of boats and nets taken in the mid 1960s (Joubert and Davies, 1966) and personal observations in the mid 1990s indicated that no change in technique occurred over this period. The nets were towed in shallow water – observations indicated that the nets rarely totally disappeared below the surface – where the wash from the motor, which the driver frequently swung from side to side, appeared to be a major factor driving the prawns into the nets. There were no tickler chains and the skids would have kept the lower portion of the net frame off the bottom. At St Lucia in the summer prawn season, the boats would leave before first light, about 04:00, typically returning between 11:00 and midday depending on catches. Exceptionally large or small catches would

279

280

Anthony T. Forbes and Nicolette T. Forbes

often result in an earlier return, either because the holding capacity would have been reached or because catches were small and further effort was considered pointless. Netting at St Lucia occurred in the Narrows, the 20 km partially tidal channel linking the sea and the lakes. Observations over several years in the 1970s and 1980s indicated that the boats never operated in the lake and the northern limit was in the area known as Brodie’s Crossing at the southern end of the lake (Figure 14.2). Netting in the Narrows was done as close to the fringing reed beds as the draught of the boat and the reach of the outboard motor permitted. In the Brodie’s Crossing area, the nets were hauled wherever there was enough water. A possible explanation for this contrast was provided by the work of Owen (1992), who showed that the invertebrate macrobenthos, which would have supported the prawns, was more abundant in the non-dredged fringes of the Narrows and in the undredged Brodie’s Crossing areas. Dredging of the Narrows, which included the construction of Potter’s Channel immediately south of Makakatana Bay (Figure 14.2), was undertaken as a management strategy in the 1960s to enhance water exchange between the sea and the lake. Care of the catch was rudimentary. Penaeus indicus, which dominated the catch, dies rapidly when removed from the water. On the boats the catch was typically held in heavy duty, open, black plastic tubs. At times these were partially filled with ice blocks which would have afforded some protection against the summer temperatures, but the more common practice was simply to cover the tub with a hessian sack. On return, the catch would be rinsed with ordinary tap water. In midsummer, tap water temperatures would not provide any chilling effect. The prawns would then be packed in cartons holding about 250 g, the cartons stacked in crates and the crates stored in large walk-in freezers. Even at a storage temperature of -20
C, by the time that freezing of the catch was complete, up to seven hours may have elapsed since the time of removal from the water. Such treatment was virtually guaranteed to

cause blackening of the gills and softening of the muscle tissue following subsequent thawing, representing a significant quality deterioration. This fact was not lost on the fishermen of St Lucia and queues would build up rapidly at the bait shop, following the return of the bait boats in the hope of obtaining fresh prawns. The earliest recorded prawn monitoring in the St Lucia estuarine system was carried out by Champion (1976), from about 1966 to 1971. He recorded catches in terms of numbers of cartons, the currency of the day, but also attempted to measure effort in terms of both days and hours fished. Unfortunately, he recorded catches on a calendar year basis which distorted the picture because the catch is seasonal over summer, but annual catch records combined catches from the end of one season with the start of the next. Reworking of the data derived from monthly catch records over the period 1968 to 1975 (Champion, 1976) confirmed the pattern of predominantly summer catches. Effort in fishing days did not necessarily track catches and supported his suggestion that the fishery catered significantly to local bait demand. Consequently, levels of effort, and to some extent catch peaks, were generated by social (i.e. bait demand) rather than biological considerations. Prime fishing periods, however (e.g. during the grunter, Pomadasys commersonnii, run in late winter/early spring), did not necessarily coincide with periods of prawn abundance. Subsequent to the above, Champion (1976) referred to monitoring of effort in boat hours from May 1970 to March 1973 and of records of catch composition on a monthly basis from January 1969 to December 1971, but further commented at the 1976 Charter’s Creek Workshop that ‘the monitoring of commercial prawn catches, as practiced (sic.) (by the NPB) at present, is probably motivated more by accounting than biological needs. Prawn catches are, for example, recorded in number of cartons, not mass. Catch compositions are not analysed and recording of the boat hour index of effort has been discontinued. This type of information is needed for even the most basic assessment of the fishery. Not

Penaeid prawns

only should appropriate data collection be resumed, but the quality of monitoring should also be improved.’ It can also be mentioned that, for bait purposes, anglers were more interested in obtaining an adequate number of bait size prawns, rather than a few large prawns in their cartons, and accordingly any large prawns, particularly the tiger prawn, P. monodon, would be kept aside during processing of the catch and typically not recorded. Demetriades (1990) used monthly samples collected during 1984–1987 and weekly samples from 1987 to 1988 to determine species composition in the catch. Catch masses were recorded either directly or estimated by conversion of carton numbers to catch masses, on the basis of average mass of prawns per carton, in order to investigate catch trends over the period 1973 to 1990. Effort was recorded as boat days. This period covered the wet years in the mid 1970s, the drought of the early 1980s and Cyclone Domoina in January 1984. ‘Annual’ catches were calculated using a seasonal year from June to May, which was taken as more representative of successive generations than data accumulated over a calendar year. Catch per unit effort (CPUE) was calculated as kilograms per boat day. Three species, namely in order of abundance P. indicus, Metapenaeus monoceros and P. monodon, contributed to the fishery during the above period (Table 14.2). The contribution by mass as opposed to numerical abundance decreased the significance of the relatively smaller M. monoceros and increased the

Table 14.2. Percentage species composition in the bait prawn fishery from 1 kg samples collected monthly from 1984 to 1986 and weekly during 1987–1988 (n ¼ 24 246) Species

Abundance

Mass

P. indicus

75

82

M. monoceros

19

9

P. monodon

3

8.6

Other species

0.7

0.4

significance of the relative larger P. monodon. While there was substantial variation between years, the most obvious feature was a downward trend in the fishing effort over the 17 years for which it was possible to obtain data (Figure 14.4a). The total seasonal catch varied between 10 and 30 metric tonnes with most values falling between 10 and 20, but tending towards a decline over the sampling period (Figure 14.4b). Catch rates were remarkably consistent, at a seasonal average of c. 25 kg boat1 day1, despite the 17 year period covering conditions ranging from virtually fresh water to full seawater and a major cyclonic flood in January 1984 (Figure 14.4c). What would have been the inevitable disappearance of the bait fishery, as a result of the loss of the estuarine function of the system due to the separation of the Mfolozi River mouth from the St Lucia estuarine system, the consequent closure of the St Lucia Mouth and the cutting off of both postlarval immigration and juvenile emigration routes, was pre-empted by an unanticipated management decision in the mid 1990s. Personal experience at St Lucia during the 1980s indicated that the management situation described by Champion (1976) at the Charter’s Creek Workshop did not change, the prawn fishery was low on the management priority list and responsibility for the bait fishery was often given to the newest recruit to the station, who seldom had any fisheries experience. It took several years before persuasion resulted in any relevant monitoring being resumed following the Champion era, but there was rarely any attempt by the management authorities to consider the catch data on a long-term basis. The result of this combination of events was to some extent predictable and reached a conclusion in this recorded decision: ‘Shrimping has been low for a number of years and staff decided at the meeting to stop the shrimping operation. It was felt that there is not enough information available as to what effect the shrimping operation has had on this resource’ (Anon., 1996). Even with the benefit of hindsight, a statement such as this, emanating from a provincial

281

282

Anthony T. Forbes and Nicolette T. Forbes

FIGURE 14.4 (a) Fishing effort (boat days), (b) seasonal catches (metric tonnes), and (c) catch rates (kg boat1 day1) in the St Lucia bait prawn fishery.

Penaeid prawns

FIGURE 14.4 (cont.)

conservation body which had its own research staff and which operated a commercial fishery in a proclaimed nature reserve for over 40 years, must be seen as a serious indictment. Other apparent, but unrecorded, considerations included financial losses arising from salaries, boat costs and freezer maintenance (R. H. Taylor, pers. comm.).

14.6.2 Bycatch The issue of the impacts of bottom trawling on benthic communities and the bycatch taken by trawl netting has been a controversial topic worldwide for many years. Bycatch was described by Fennessy (1994) as being a consequence of the indiscriminate nature of the typical trawl used in trawling, which resulted in the catching of demersal organisms associated with prawns or with the habitat occupied by the prawns. Bycatch would therefore be described as ‘any organism caught incidentally to the target species’. Some of this

bycatch may be retained so that bycatch would then include such species as well as discards, which are not used in any way and are discarded. The acceptability of high bycatch:prawn ratios is controversial, particularly when the bycatch involves large or threatened marine species, such as rays and turtles. This type of bycatch appears, however, to be extremely unusual in the South African situation. While bycatch has been a long-standing issue in offshore fisheries, the use of nets in estuarine prawn nursery grounds, which also act as nursery grounds for a variety of marine fish species (Whitfield, 1998), is potentially equally controversial. Although the bait prawn fishery in the St Lucia system has not operated since 1996, this fishery was an integral part of the recreational angling infrastructure of St Lucia town for over 40 years, and concern over the bycatch resulted in an investigation in the early 1990s which was reported on by Mann (1995a) and for the sake of completion is included here. Samples at St Lucia were

283

284

Anthony T. Forbes and Nicolette T. Forbes

collected intermittently between 1990 and 1992 and once or twice a month between August 1992 and September 1993. A total of 40 species from 21 families (Table 14.3) was recorded, dominated by the slimy, Leiognathus equula (23.5%), followed by the longspine glassy, Ambassis ambassis (9.8%), the Natal stumpnose, Rhabdosargus sarba (8%), the longjaw glassnose, Thryssa vitrirostris (7.8%), the Natal moony, Monodactylus argenteus (7.8%), the perch, Acanthopagrus berda (6.4%) and the longarm mullet, Valamugil cunnesius (6.3%). All other species contributed less than 5%. The representativeness of the data is, however, to some degree challengeable as the sampling was done during a drought period, which resulted in mouth closure between December 1992 and

October 1993. This would have prevented any recruitment of juvenile fish into the estuary during this period and it is therefore likely that there would have been fewer juvenile fish in the estuary and, accordingly, fewer fish in the bycatch. Mann (1995a) refers to investigations of the bycatch during 1969–1971, which supported his conclusion that the localization of the bait fishery in the Narrows, the availability of large areas of similar habitat elsewhere in the system, the rapid return of discarded fish to the water and the low ratio of fish to prawn in the catch argue that the bait fishery had little effect on fish populations in the lake. The lack of adequate data and the closure of the fishery in 1996 means that the real impact of the bait fishery will never be known.

Table 14.3. Percentage contribution by the different fish species recorded in the bait fishery bycatch at St Lucia between August 1992 and September 1993 (n ¼ 847) Family

Species

Common name

%

Ambassidae

Ambassis natalensis Ambassis productus

Longspine glassy Slender glassy

Belonidae

Strongylura leiura

Yellowfin needlefish