Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management [1 ed.] 9781607504474, 9781607500308

Citation preview

Catchment2Coast

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

a systems approach to coupled river-coastal ecosystem science and management

Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Deltares Select Series Volume 2 ISSN 1877-5608

Deltares is a Dutch independent research institute for water, soil and subsurface issues. It was formed in 2008 from a merger of Delft Hydraulics, GeoDelft, the Subsurface and Groundwater unit of TNO and parts of Rijkswaterstaat.

Cover illustration: Avicennia tree and roots along the Maputo Bay shoreline (Photo M. Marchand). Previously published in this series: Volume 1. F.J. Los, Eco-Hydrodynamic Modelling of Primary Production in Coastal Waters and Lakes Using BLOOM

Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Catchment2Coast a systems approach to coupled river-coastal ecosystem science and management

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

P. M. S. Monteiro and M. Marchand Editors

IOS Press

Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

© 2009 The author and IOS Press All rights reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without prior written permission from the publisher. ISBN 978-1-60750-030-8 Library of Congress Control Number: 2009905265 Publisher IOS Press BV Nieuwe Hemweg 6B 1013 BG Amsterdam Netherlands fax: +31 20 687 0019 e-mail: [email protected]

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Distributor in the UK and Ireland Gazelle Books Services Ltd. White Cross Mills Hightown Lancaster LA1 4XS United Kingdom fax: +44 1524 63232 e-mail: [email protected]

Distributor in the USA and Canada IOS Press, Inc. 4502 Rachael Manor Drive Fairfax, VA 22032 USA fax: +1 703 323 3668 e-mail: [email protected]

LEGAL NOTICE The publisher is not responsible for the use which might be made of the following information.

PRINTED IN THE NETHERLANDS

Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Note from the Editor-in-Chief

Catchment2Coast was an interdisciplinary research and modelling project that aimed to improve understanding of the linkages between coastal ecosystems and the adjacent river catchments. The project involved nine partner organisations from three European and three southern African countries, including Mozambique, where the project was conducted. It is with pleasure to see that this study is now published as a coherent piece of work which makes its results not only accessible to the scientific community but also to managers, policy makers and other interested parties. This fits well within the philosophy of the Deltares Select Series, to provide a medium where cutting edge scientific knowledge is combined with an understanding of political, administrative and economic processes. Catchment2Coast has tackled a problem which is at the interface of many different domains: between river and the sea, between bay and ocean, between water and soil, but also between ecology and economy. It used a variety of tools and methods, ranging from continuous hydrodynamic monitoring and biochemistry flux measurements to remote sensing and mathematical modelling techniques. But perhaps even more important was the integration of disciplines that took place during the project, which consisted of experts from different countries both from Europe and Southern Africa.

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

The knowledge gained though this research provides a sound basis for the setting of environmental flow requirements for the Maputo, Incomati and similar river systems in sub-Saharan Africa. And what is more, it may serve as an example and stimulus for continued cooperation between experts from different disciplinary and cultural backgrounds. Prof. dr. Huib de Vriend Scientific Director Deltares

v Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Acknowledgements The Catchment2Coast project was made possible by a research fund from the INCODEV programme of the European Union, contract number ICA-CT-2002-1009. We thank Cornelia Nauen of the former INCO Programme in EU DG XII, Brussels for her support throughout the project and in the opportunities to disseminate the project ideas in international meetings. Acknowledged written contributions: Adriano Macia, Antonio Hoguane, John H. Simpson, João Lencart, João Gomes Ferreira, Ana Nobre, Andrea Franco, Issufo Halo, Mags Moodley, Vivek Naiken, Sue Matthews (editing).

vi Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Foreword

The scientific community concerned with water is being driven by dominant paradigms that force us to think of water and its associated aquatic ecosystems in specific and often rigid ways. The dominant paradigm today, specifically in the field of water resource management, is what we call "Integrated Water Resource Management", or simply IWRM. This is underpinned by a veritable mantra of texts that seek to extol the alleged virtues of this approach. Central to this paradigm are certain key concepts, the most notable being the river basin as a unit of management, and the notion of subsidiarity that seeks to cascade management to the lowest appropriate level in society as envisaged by what we call the Dublin Principles.

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Yet practical experiences show us a different world. On the one hand complexity defies subsidiarity and the river basin is often inappropriate as a unit of management, specifically when that basin supports ecosystems and associated livelihood flows across international borders. On the other hand, levels of "optimization" needed to achieve genuine benefits to a wide range of stakeholders suggest that small is not necessarily better. In this regard the work on benefit-sharing currently being conducted by a team of which I am but one member (see Phillips, Daoudy, McCaffrey, Ojendal & Turton (2006) "Transboundary Water Cooperation as a Tool for Conflict Prevention and Broader Benefit-Sharing", Stockholm: Expert Group on Development Issues, Swedish Foreign Ministry) is showing us that rather than cascading management down to lower levels in society, we need to move in the opposite direction - upwards to higher levels of optimization from which a greater basket of benefits can be sourced and then shared. My own work on the notion of a hydro-political complex, a theoretical level of analysis and potentially management that exists above the river basin but below the regional economic grouping we call the Southern African Development Community (SADC), suggests the same general trend. For this reason it is refreshing to find work of the calibre and focus in this book. Here we have a team of scientists that move beyond the constraints imposed by the IWRM paradigm, where both fresh-water and marine aquatic ecosystems are linked, as are terrestrial and aquatic processes. For the first time we can now say with some degree of certainty, that marine processes are intimately linked to terrestrial processes via the freshwater flow of river systems. Furthermore we can now see, with a high level of confidence, that the management of such complex systems cannot realistically be cascaded down to lower levels of society in keeping with the principle of subsidiarity. In truth, cooperation is needed at many different levels, within a country, between countries, and between ministries and scientific disciplines. If we understand the transboundary waters problematique as being driven by the need to mitigate issues arising when natural flows of water (and the nutrients contained therein) cross artificial jurisdictional boundaries, then this study is a classic. The serious reader of this book will benefit from the refreshing new approach to a complex problem that would

vii Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

not have been researched if the authors were guided by the narrow concept we call IWRM. This book also stands out as a monument to the potential of endeavours across disciplines, across international political boundaries, and across institutions.

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Dr. Anthony Turton Executive Director International Water Resources Association (IWRA).

viii Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Executive summary

Introduction Catchment2Coast was an interdisciplinary research and modelling project involving nine partner organisations from three European and three southern African countries. It aimed to provide an ecosystem-scale understanding of the linkages between economically important tropical coastal resources and their associated river catchments (Monteiro & Matthews, 2003; Marchand, 2003). Maputo Bay, a large, shallow bay on the coast of Mozambique sustained by three transboundary river catchments, served as the study site.

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

The core hypothesis of the project was that the most important biophysical interactions between freshwater catchments and coastal domains occur at the sub-seasonal event scale (days). Without a full understanding of the mechanism of these event-scale interactions, their economic consequences cannot be adequately assessed. The project therefore used a number of numerical models (coastal, river basin and ground water) with the required dynamical capability to implement a system-scale approach to the functional dependence of coastal systems on river basin drivers. Resource economics models then allowed these results to be translated into impacts on urban and rural livelihoods. The project focused on a single coastal resource, the shrimp (Penaeus indicus), which was used as an indicator of the ecosystem productivity response to catchment forcing. Shrimp is Mozambique’s most economically important living resource, and catch per unit effort (CPUE) data provide the best proxy for long-term trends in coastal ecosystem productivity (although it is recognized that CPUE may not represent the true status of shrimp biomass). This approach was also generic enough to allow other recognized impacts of catchment-based human activities, such as mining effluents, pathogens, eutrophication, erosion and silting, to be addressed in a holistic way in the future. The biophysical component of the research was guided by a number of hypotheses, which were articulated by the leadership team at the first and second project steering committee meetings.

Scales of forcing hypothesis: 1) The overarching hypothesis across all process domains is that the key linkages between river fluxes and coastal responses occur at the event scale and not at the aggregated seasonal or annual scale.

ix Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Physical forcing hypotheses: 2) Density-driven circulation: the impact of the river inputs on shrimp production is through the stratification and circulation characteristics driven by density gradients that develop as a result of the mixing of river and sea water. This is the ROFI (region of freshwater influence) hypothesis. 3) Salinity hypothesis: the impact of local rain, river-derived freshwater and ground water is key to maintaining the required salinity ranges for the early life stages in the mangrove habitats. 4) Temperature hypothesis: temperature variability is an important factor in defining habitat suitability for the early life stages of the shrimp (shade).

Biogeochemical forcing hypothesis: 5) The impact of river inputs is through the input of suspended sediments, organic matter and nutrients into the mangrove biogeochemical remineralization – production system. Variability in this biogeochemical forcing regulates the space and time scales of energy input into the food web that supports the shrimp.

Findings

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Given that the physical forcing hypothesis was initially the preferred one, there was a strong focus on understanding the nature of stratification and density flows linked to freshwater inputs. Maputo Bay is a tidally energetic embayment with a highly variable seasonal freshwater input (up to 3000 m3·s-1 in extreme conditions). There is an unusually large spring to neap tide ratio (~6), so tidal currents and the associated circulation are modulated on a fortnightly cycle. Although tidal range rises and falls periodically and continuously changes the amount of energy available for bottom mixing, the water column remains vertically mixed throughout the dry season. During the wet season, freshwater input creates density-driven currents, which also appear to be modulated by the tides. However, the sudden episodes of intense discharge result in density-dependent stratification of the water column. An interesting inference from the model runs is that the density-driven flows evident at neap tides are largely suppressed by the increased vertical mixing at spring tides. The observations confirmed that stratification occurs at neap tides and there is some evidence of periodic stratification at spring tides. They also indicated that significant stratification is only induced by freshwater input, and even during the period of maximum surface heating, persistent thermal stratification does not develop. These variations in vertical and horizontal density gradients control the flushing patterns of the bay. During neap tides, low tidal energy input allows the development of stratification and density-driven currents, which tend to flush the bay efficiently. During spring tides, however, enhanced vertical mixing arrests the estuarine plume for

x Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

most of the tidal cycle and hence allows baroclinic circulation only near low-water slack. Although the physical forcing hypothesis was later rejected, these insights were key to understanding the mechanisms of freshwater retention, which are an essential part of the biogeochemical hypothesis. The physical forcing hypothesis was rejected mainly because an observational programme undertaken at the mangroves revealed that the early life stages of shrimp (post-larvae) arrived at the mangrove nursery areas prior to the onset of the wet season river flows and floods. The data showed that the recruitment of early life stages into the nursery areas was completed by December, and therefore largely independent of the physical oceanography resulting from freshwater inflows during the wet season.

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Following the rejection of the physical forcing hypothesis, the science focus shifted to the biogeochemical forcing hypothesis. The original formulation of this hypothesis was also rejected, because modelling and observations showed that the river was not the supplier of nitrogen to the coastal – estuarine system, and could not support the productivity needs in the mangrove nursery. In the absence of any oceanic or other land sources, the biogeochemical hypothesis was reformulated to account for an autochthonous source of nitrogen in the form of nitrogen fixation (N-fixation). Ravikumar et al. (2004), working in Indian mangroves, showed that N-fixation was driven by salinity dependence, which was sensitive to the spatial and temporal character of the 20 – 30psu range. The re-designed field work, focussing on sediment water production and remineralization fluxes, showed that the Avicennia mangroves, where most N-fixation occurs, were indeed the areas where the greatest salinity dependence on new production was found. Hindcasting ecosystem productivity (shrimp CPUE) for the period 1996 – 2004 showed that the modelled productivity was able to “predict” the trends in 8 of the 9 years. The study was therefore able to show that salinity-dependent mangrove production during the nursery phase of the shrimp life cycle was the critical factor governing interannual ecosystem production in Maputo Bay. The linkages between river flow, salinity, mangrove production and shrimp grow-out results in a two to three month lag between wet season river flows and shrimp recruitment into the fishery. This supports local traditional knowledge of a two month lag between good rains and high shrimp catches.

Advances The most important scientific contributions that this work made towards increasing understanding of tropical river – coastal ecosystem linkages were:

xi Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

1.

Identifying salinity variability - N-fixation coupling as the key linkage that governs the dependency of coastal ecosystem productivity on river flows in lownutrient systems

2.

Highlighting the underestimated role of N-fixation in driving critical new production that governs overall ecosystem productivity, rather than the much larger regenerated production fluxes generated by the mangrove forests themselves

3.

Clarifying the extent to which physics governs ecosystem productivity, as the retention character of the estuarine – coastal water body physics governs the magnitude and temporal scale of the freshwater pulse (flood) that sustains wet season new production in the nursery zones

4.

Showing that nitrogen losses from the mangroves (outwelling) were limited to the spring tidal periods in the wet season, when physical transport rates exceeded uptake rates by microphytobenthos in the mangroves.

5.

Demonstrating the importance of a system approach in both formulating the hypothesis and understanding the critical linkages, including the ecological role played by the freshwater and estuarine wetland systems.

Recommendations

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Based on the findings of Catchment2Coast, it is recommended that:

ƒ

Harvesting of mature mangrove trees (particularly Avicennia) is suspended immediately, and protected areas are established to protect ecosystem processes in the mangroves

ƒ

Operation of the four largest dams in the Incomati and Maputo basins are coordinated to ensure a minimum wet season (Jan – Apr) water flux of 500 Mm3 for the Incomati River and 250 Mm3 for the Maputo River

ƒ

Catchment management and water allocation plans in the Incomati and Maputo River basins are coordinated to increase resilience of the system to natural fluctuations in runoff

ƒ ƒ

Conservation areas in the Xinavane and Maputo wetlands are declared

ƒ

Annual estimates of prawn biomass are initiated, and the January experimental CPUE is used to forecast the seasonal average The main sources of uncertainty are investigated further: a. Maputo River flow data b. New production hypothesis: N-fixation – salinity relationship c. Mangrove food web and resource competition d. Salinity variability in the bay and mangrove domains.

xii Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Awareness should be raised about the dependence of coastal resources and ecosystem health on the integrity of adjacent river systems. It is not feasible to develop coastal conservation strategies without including the modifications of ecosystem services provided by river systems through changes in land use and water allocation.

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

ƒ

xiii Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

This page intentionally left blank

Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Contents Note from the Editor-in-Chief Foreword Executive Summary Chapter 1 The challenge..................................................................................1 1.1 Linking rivers and coasts ............................................................................. 1 1.2 The science base .......................................................................................... 2 1.3 Catchment2Coast: the project ...................................................................... 3 1.3.1 What is a system? .................................................................................. 5

Chapter 2 Setting the stage .............................................................................7 2.1 Maputo Bay and its river catchments........................................................... 7 2.1.1 Geography and economy of Mozambique............................................. 7 2.1.2 Maputo Bay ........................................................................................... 7 2.1.3 Incomati River ....................................................................................... 9 2.1.4 Maputo River....................................................................................... 10 2.1.5 Umbeluzi River ................................................................................... 11 2.2 The shrimp fishery of Maputo Bay ............................................................ 12 2.3 The life cycle of shrimps............................................................................ 14 2.4 Research hypotheses .................................................................................. 16 2.5 Research set-up .......................................................................................... 17

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Chapter 3 The research outputs ...................................................................22 3.1 Catchment runoff and river flows .............................................................. 22 3.1.1 Catchment modelling........................................................................... 22 3.1.2 Incomati model set-up ......................................................................... 23 3.1.3 Calibration and verification results...................................................... 25 3.2 Maputo Bay hydrodynamics ...................................................................... 28 3.2.1 Introduction ......................................................................................... 28 3.2.2 Hydrodynamic modelling .................................................................... 28 3.2.3 Observational and model results.......................................................... 30 3.2.4 Summary ............................................................................................. 34 3.3 Maputo Bay Biogeochemistry: Boundary Fluxes ...................................... 36 3.3.1 Observational approach ....................................................................... 36 3.3.2 Observations and conclusions.............................................................. 36 3.3.3 The biogeochemical hypotheses .......................................................... 38 3.4 The Mangrove Ecosystem.......................................................................... 39 3.4.1 Mangrove distribution and health........................................................ 39 3.4.2 Mangrove biogeochemistry ................................................................. 40 3.5 Maputo Bay Ecosystem Model .................................................................. 43 3.5.1 Modelling platform.............................................................................. 43 3.5.2 Model set-up........................................................................................ 43

xv Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

3.6 Shrimp Population Dynamics .................................................................... 46 3.6.1 Larval transport surveys ...................................................................... 46 3.6.2 Shrimp IBM model.............................................................................. 48 3.7 Resource Economics: Prawn Fisheries ...................................................... 52 3.7.1 Introduction and data ........................................................................... 52 3.7.2 Methodology........................................................................................ 54 3.7.3 Results ................................................................................................. 55 3.7.4 Analysing catchment and industry policy changes.............................. 56 3.7.5 Policy decision affecting effort exerted ............................................... 56 3.8 Hindcasting Ecosystem Production in Maputo Bay 1996 – 2004.............. 58 3.8.1 Maputo Bay System Model ................................................................. 58 3.8.2 Model Results...................................................................................... 59 3.8.3 Bay system production variability 1996 - 2004................................... 60

Chapter 4 Decision support and management implications.......................62 4.1 A Decision Support Tool for catchment to coast linkages......................... 62 4.1.1 Architecture of the Decision Support Tool.......................................... 62 4.2 Implications of C2C science for Integrated Water Resource Management 66 4.3 Recommendations to Strengthen Ecosystem Resilience in the coupled Incomati and Maputo – Maputo Bay systems ......................................................... 67

Chapter 5 Conclusions, lessons learned and legacy ....................................69 5.1 5.2

Conclusions................................................................................................ 69 The Legacy ................................................................................................ 70

References

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Annex A: Composition of the Catchment2Coast project team

xvi Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Chapter 1 The challenge 1.1

Linking rivers and coasts

One of the globally recognized obstacles for the successful implementation of Integrated Water Resource Management (IWRM) policies is the gap in understanding of the linkages between river basins and adjacent coastal ecosystem services. This is particularly relevant to those ecosystem services that play a vital socio-economic role in fisheries and tourism. In the case of transboundary river - coastal systems, competing use of ecosystem services to support agriculture and coastal economic benefits are rarely understood as part of integrated systems, where benefits in one domain often accrue at a cost to others. The scientific component of the Catchment2Coast project aimed to provide a basis for the negative costs of development strategies to be internalized at the feasibility stage, rather than externalized and passed on as a deferred cost to another country.

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

It is recognized that there are many levels at which the natural and human estuarine – coastal habitats can be impacted by the physical and chemical characteristics of river and ground water flow. These may include: x Fisheries x Coral reefs x Mangrove forests x Sea grass beds x Beaches x Harbours. Each of these habitats support economic activities, which are both critical to people who benefit from them at the coast and sensitive to the impacts of human activities in the river catchments. Rivers not only carry freshwater to the coast, but also sediments, nutrients and pollutants. Sediments play a dominant role in delta formation on a geological time frame, whereas excessive loads of nutrients and pollutants are known to significantly influence the water quality and ecology of estuaries and coastal seas. Marine ecosystems are harmed by careless land practices hundreds or even thousands of kilometres upstream. For example, deforestation has in some countries led to a significant increase in sediment load in coastal waters, endangering the health of coral reefs. In addition, over half of the large river systems in the world are affected by dams, which have led to dramatic effects in delta and marine areas. For instance, the Aswan High Dam has resulted in the complete loss of the sardine fishery offshore of the Nile Delta and to serious erosion problems on the delta shores. Dams in the Indus and Zambezi rivers have reduced freshwater flows that previously sustained the mangrove forests in their deltas. Dam-impacted catchments experience higher irrigation pressure and about 25 times more economic activity per unit of water than do unaffected catchments.

1 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Figure 1 – Maguga Dam in the Komati river (photo courtesy of Vela VKE) Increasingly, coastal managers are becoming aware that these problems cannot be solved within their management area alone, given that the coastal area is an essential component of the river basin. However, although it has long been recognized that catchment-based activities impact on coastal ecosystems downstream, their full consequences for socio-economically important ‘ecosystem services’ and coastal resources are poorly appreciated. It is not clear whether this is due to gaps in understanding of the linkages between catchment and coast, or to limitations of the historical data. These gaps pose particular obstacles in the implementation of integrated ecosystem development policies, especially those that deal with transboundary consequences of management actions. Transboundary concerns apply to both national and ecosystem boundaries, including the traditional ones between river catchments and their adjacent coastal ecosystems. The research community has an important task in this respect. It should provide the knowledge to substantiate the river-coast relationships, and develop tools to predict the impacts of proposed development plans, as well as climate change. It should provide information that clarifies the trade-offs between upstream economic development and downstream/coastal benefits of freshwater flows.

1.2

The science base

The science of the dependence of coastal systems on river fluxes has progressed over 30 years from an initial geochemical budgetary focus (RIOS 1981) to the characterization of water quality problems mainly linked to the consequences of

2 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

eutrophication and erosion (UNESCO 1988, Smetacek et al. 1991, Vollenwieder 1992). More recently it has become increasingly linked to the integration of the needs of sustainable growth in both river basins and coastal domains (LOICZ 1995, UNEP 1999). Different approaches have evolved to provide the scientific basis to address these resource management needs. While the widely used budgetary modelling approach (LOICZ 1996, LOICZ 1997) has been shown to be a useful methodology to quantify changes in seasonal and interannual water quality (e.g: nutrients), it mostly integrates time scales longer than those that characterize the variability of life stages of economically and ecologically important living resources (Buzzelli et al. 1999). These often respond to event or sub-event scale forcing of advection or production input fluxes, which requires a more explicit treatment of dynamical considerations (Monteiro & Largier 1999). In an estuary, salt and fresh water come together. The extent and nature of the salt intrusion largely depends on the flow phenomena, and influences the sedimentation process and the estuarine ecology. Because salt water has a higher density than fresh water, the saline marine water tends to develop a layer under the fresh water, resulting in a two-layered, stratified system. Energy input is needed to mix the salt water with the fresh water. In an estuarine environment, this energy is provided by wind and tidal movements. Considerable progress has been made recently in understanding the interaction between buoyancy input and mechanical forcing by wind and tidal energy inputs, which characterises the ROFI (region of freshwater influence) regime. For example, the ROFI produced by the Rhine discharge into the North Sea has been extensively studied and is now well documented (Simpson 1997), and the results are being assimilated into management strategies. Similar advances have occurred in understanding the interaction between groundwater aquifers and coastal dependencies (Acworth et al. 2000).

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

1.3

Catchment2Coast: the project

The Catchment2Coast project was an international and multidisciplinary ecosystemscale research project. It aimed to understand the linkages between human activity in river catchments, such as farming, irrigation and water abstraction and its impact on coastal resources. Although it has long been recognized that catchment-based activities impact on riverine and coastal ecosystems, the linkages and their full consequences are still poorly understood. Catchment2Coast has advanced this understanding, particularly in tropical systems. The Maputo Bay ecosystem and shrimp industry, which provides important livelihoods to thousands of fishers and their families, served as a pilot study site. The project was co-funded by the Fifth Framework Programme (EU INCO-DEV) of the European Union and the nine scientific institutions involved (see Annex A).

3 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Figure 2 – Location of the study area: Incomati catchment and Maputo Bay (Mozambique

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

map courtesy of University of Texas Libraries).

The research project took an interdisciplinary systems approach to understand the linkages between the main domains of river catchments and their associated coastal environments. It used linked dynamical numerical platforms as tools to translate the impacts of human development needs in the rivers basins into cost-based consequences for resources in rural and urban economies in the coastal domain. At the core of the research was the hypothesis that human activity in the Incomati catchment is impacting on the shrimp industry in Maputo Bay. Prior to June 2000, when the Mozal aluminium smelter near Maputo began production, shrimps yielded the highest export earnings for Mozambique. In 1996, shrimps represented approximately 35% of all export earnings; the total annual value of the shrimp industry for Maputo Bay for the period 1992-1994 was estimated to be approximately US$3.5 million from a catch of 3000 tons, from which 3000 artisanal and semi-industrial fishers derived a direct livelihood. The tools and skills developed

4 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

through Catchment2Coast will help elucidate the true costs of upstream developments on the delivery of goods and services from coastal ecosystems. Studies addressing the impact of activities in the catchment typically focus on seasonal or mean monthly inputs from rivers into estuarine environments. One of the limitations of this type of budgetary or steady state approach is that it provides little or no understanding of the mechanism or scale of event that has the most sensitive impact on the particular habitat or resource production at the coast. The key question is: Do the observed outcomes in resource production in the coastal domain depend on the integrated seasonal runoff signal, or does a single event of a certain magnitude govern the seasonal response? Either scenario has important implications for the way in which human activities in the catchment are managed, and their costs internalized. The costs that need to be internalized are those that result from the impact of water quality and flow control on economically important living resources in the coastal domain. Catchment2Coast was a demonstration project that provided a working example of how critical it is to have a thorough understanding of the biophysical environment in order to manage the interaction and socio-economic needs of the river catchments and coastal zones. It was not aimed at providing all the answers for the management of Maputo Bay, but at showing a new and more effective way in which that could be done. The shrimp Penaeus indicus was used as an indicator of the ecosystem response to catchment forcing. This was a necessary simplification because the focus of the project was on the system linkages rather than coastal foodweb complexity. The outcomes should, however, form a basis to deal with food web complexity in the future. The approach should also be generic enough to deal with other recognized impacts such as mining effluents, pathogens, eutrophication, erosion and silting.

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

The allocation of water in a catchment is a political process subject to competing needs and powers, particularly where there are national and river – ocean boundary considerations. The ecosystem modelling approach developed in this project aimed to use the scientific findings to facilitate the negotiation process through a shift in understanding of what constitutes system boundaries.

1.3.1

What is a system?

Given the focus of this work it is necessary to consider what we define as a system. The starting point is that the minimum size is defined by a scale in which the internal fluxes are larger than the boundary fluxes. Hence the variability within the system is not governed by external forcing but rather by internal one which can be modified. This is critical if the science is to be extended to become part of a policy or management plan. It explains why the system boundaries are not at the estuarine head but incorporate the complete river basin associated with Maputo Bay and adjacent shelf. This ecosystem scale definition raises its own difficulties in the sense that it transcends the historical political and administrative boundaries which makes it difficult to transform from a scientific discovery to a contribution to sustainable development. To achieve this requires a shared view on the scale and benefits of system as whole. The more difficult system scale to define is the human dimension because although many of the relevant consumptive activities are within the ecosystem defined boundaries,

5 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

others such as remote markets and political power are external. For the purposes of this work we have remained with the bio-physical flux based definition of the system scale.

6 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Chapter 2 Setting the stage 2.1 2.1.1

Maputo Bay and its river catchments Geography and economy of Mozambique

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Mozambique is situated on the eastern coast of southern Africa and has a coastline of about 2,700 km (see map Figure 2). The continental shelf is alternatively narrow and very wide, extending up to 90 nautical miles offshore at Beira. The primary productivity is relatively high due to marine currents and numerous rivers introducing nutrient-rich sediments. The capital, Maputo, is situated in the south of the country, and has a population now exceeding 15 million people. The ocean plays an important role in the country’s social, cultural and economic life. There are three large ports: Maputo, Beira and Nacala, which provide the main gateway for landlocked neighbouring countries (Hoguane 1998). In 1990, Mozambique was estimated to be the world’s poorest nation; since then, the country has been in transition toward a more market-oriented economy and the prospect of raising its standard of living. Construction of the Mozal aluminium smelter began in 1999, and its output has since doubled the country’s exports, providing in excess of US$800 million in foreign exchange earnings. Titanium and natural-gas deposits are being explored by foreign investors, and significant coal deposits and hydropower potential also exists. Many citizens are migrant labourers working in South African mines. However, most of Mozambique’s workforce is still engaged in traditional subsistence cultivation, the principal cash crops being cashews, sugarcane, cotton, copra and tea. Cattle and goats are raised, but their numbers are kept low by the tsetse fly. There are small forestry and fishing industries, the latter primarily targeting shrimp. The principal imports are foodstuffs, farm equipment, crude petroleum and petroleum products, and machinery; the chief exports are aluminium, shrimp and the agricultural commodities mentioned above.

2.1.2

Maputo Bay

Maputo Bay is located in the southern part of Mozambique at latitude 26qS. The rainy season is from December to April and accounts for most of the mean annual rainfall of about 1100 mm. The climate is subtropical and the mean diurnal air temperature varies between 18qC in winter and 27qC in summer. Winds are mainly from the south-east (trade winds). Mean monthly wind speed as observed at the Maputo meteorological station varies from 2 m·s-1 during winter to 4 m·s-1 during summer. Winds within the bay are generally stronger compared to those observed on the mainland and weaker compared to the open sea (Hoguane 1998). Maputo Bay is about 40 km long and 30 km wide, on average. It is open to the sea from the northern side and bounded at the eastern side by the Inhaca island and the

7 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Machangulo peninsula. The bay is shallow, with a water depth generally less than 10 m, except in the channels (Figure 3). There are numerous sand banks and some deep channels (over 15 m), generally oriented north-south. The tide is semi-diurnal, with an appreciable diurnal inequality, reaching about 3 m during spring tides. There is a small tidal phase lag between the western and eastern sides of the bay, with high water about half an hour later at the Maputo harbour than the eastern side of the bay and the entrance (Hoguane 1998).

Figure 3 – Maputo Bay bathymetry. According to previous records of water temperature and salinity, the bay may be divided into two parts: the eastern side, with oceanic and less variable salinity throughout the year, and the western side, with lower, more variable salinity influenced by river discharge. Water column structure in the bay may be considered vertically homogenous, except in the estuaries (western side) and in the deep channels. The water in the bay is warmer during summer and cooler during winter than in the open ocean. The diurnal mean water temperature ranges from 17qC in winter to 37qC in summer. This high temperature range is due to the shallow depth of the bay (Hoguane 1998).

8 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

There are three major rivers entering Maputo Bay: the Incomati, Maputo and Umbezuli Rivers, which together discharge about 8 000 million m3 annually into the bay. Each has a catchment that extends into Swaziland and South Africa. Together the river basins encompass an area of more than 82 000 km2.

2.1.3

Incomati River

The Incomati catchment (see Figure 4) has a total surface area of 46 800 km2, of which 33% lies in Mozambique, 61% in South Africa and 6% in Swaziland. Main tributaries are the Komati, Crocodile and Sabie Rivers (see Table 1), which are impounded by a number of storage reservoirs (Table 2). The Corumane Dam, built in the 1980s for both hydropower generation and irrigation, has a storage capacity of 850 million m3 which is more than the mean annual run-off of the Sabie river. It has modified the lowflow regime, as the flows discharged by the dam are in general higher than those that would occur under natural conditions. Table 1 – Sub-catchment characteristics of the Incomati River Catchment

1. Komati 2. Crocodile 3. Sabie 4. Massintoto 5. Uanetze 6. Mazimechopes 7. Incomati Total

mean annual runoff (Mm3·a-1) 1 430 1 225 750 20 15 20 130 3 590

storage dams

Nooitgedacht / Vygeboom / Maguga / Driekoppies / Sand River Dam Kwena dam Corumane / Da Gama Dam

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Source: Tripartite Interim Agreement on Incomati and Maputo watercourses

The catchment area is topographically divided by the Great Escarpment into a western plateau and a sub-tropical Lowveld in the east. Rainfall varies between 400 mm and over 1200 mm per year in the mountains. The communities of the Incomati basin comprise a total population of some 2.3 million people, which is predicted to increase to 3 million in the coming years. Whereas in many river basins the population density is low in the headwaters of a river, and higher in the lowlands, in the Incomati Basin it is just the other way around. The population density is highest in the Komati sub-basin in Swaziland and South Africa (58 and 65 persons per km2, respectively), whereas in the lower part of the catchment in Mozambique it is as low as 17 persons per km2 in places. In colonial times the upstream area of the Incomati was reserved as an agricultural area for White Afrikaners and their labour force. In contrast, the lower Incomati in Mozambique has historically been sparsely populated by Black cattle farmers. In 1955 the Portuguese colonial government stimulated large-scale agriculture along the river by constructing

9 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

infrastructure for eight irrigation schemes, two of which served sugar plantations. Sugar is the main agro-industry in the Incomati Basin and the Incomati supplies water to 120,000 ha irrigated sugarcane plantations (Heyink Leestemaker, 2000).

Figure 4 – Map of the Incomati catchment. Table 2 – Main storage reservoirs in the Incomati catchment

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Name of dam Nooitgedacht Sand River Vijgeboom Maguga Driekoppies Da Gama Corumana Injaka Kwena

2.1.4

storage capacity in Mm3 81 50 84 332 251 14 850 120 167

river Komati Komati Komati Komati Lomati Sabie Sabie Sabie Crocodile

year of inauguration 1962 1966 1971 2002 1998 1979 1988 2001 1982

Maputo River

The Maputo catchment covers an area of 29 970 km2 and is made up of a number of sub-catchments (Table 3). Its mean annual runoff is of similar magnitude to that of the Incomati and can therefore be expected to have a significant impact on Maputo Bay. The river is shared with South Africa and Swaziland.

10 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Figure 5 – Typical landscape of the Crocodile River catchment. Table 3 – Sub-catchments of the Maputo River

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Catchment 1. Lusushwana 2. Mpuluzi 3. Usuthu 4. Ngwempisi 5. Mkhondvo 6. Ngwavuma 7. Pongola 8. Maputo Total

mean annual runoff (Mm3·a-1) 420 260 610 500 570 180 1 160 100 3,800

storage dams

Pongolapoort dam

Source: Tripartite Interim Agreement on Incomati and Maputo watercourses

2.1.5

Umbeluzi River

This river is the smallest of the rivers entering Maputo Bay, having a mean annual runoff in the order of 600 Mm3. The catchment area is approximately 5 400 km2, and is shared with Swaziland and South Africa. The river is an essential source of water for the capital, Maputo. It is impounded by a dam, Pequenos Libombos, built in the 1980s. It has been reported that the construction of the M’njoli dam in Swaziland has almost halved the flow of water of the Umbezuli River to Mozambique (Salman 2002).

11 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Figure 6 – The Sabie River.

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

2.2

The shrimp fishery of Maputo Bay

Shrimps are the most important resource for commercial fisheries in Mozambique, accounting for a total national catch of about 10 000 tons. The main fishing grounds are located on the Sofala Bank, on the continental shelf. The fishery is shared between several companies, all Mozambican, and dominated by two companies established as joint ventures with Japanese and Spanish partners. The “semi-industrial” fishery is still at a relatively low level of exploitation, believed to comprise about 10% of the fishery. The closed season has recently been extended from one to three months (December to March). Prawns are also targeted by artisanal fishers, and the fishery could probably be developed in the northern area (Angoche) if the necessary infrastructure for processing and/or transportation was developed onshore. Markets for these high-value products are largely controlled by the Japanese and Spanish companies involved in the fishery. The South African market is no longer of primary importance as the products fetch better prices in Europe and Japan. Maputo Bay shrimp fisheries are worth some three million US$ annually, although there are strong interannual fluctuations in catches, usually attributed to environmental factors. They provide livelihoods to more than 3 000 fishers and their families. The target species harvested in the Maputo Bay shrimp fishery are penaeid shrimp species: • Penaeus indicus (white prawn) • Metapenaeus monoceros (brown or ginger prawn)

12 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

• • • •

Penaeus japonicus (flower or banana prawn) Penaeus monodon (tiger prawn) Penaeus semisulcatus (tiger prawn) Penaeus latisulcatus. The bulk of the catch is made up by P. indicus and M. monoceros (Dengo & Govender 1998). The catch has oscillated between 300 and 750 tons annually. Peak harvests were recorded in 1972 (more than 800 t) and 1982 (about 500 t) whereas at other years annual harvests were less than 100 t (see Figure 7). The causes of this oscillation are not well understood and form the major research question of Catchment2Coast. annual shrimp catch (semi-industrial) 900 800 700

tons

600 500 400 300 200 100 2002

2000

1998

1996

1994

1992

1990

1988

1986

1984

1982

1980

1978

1976

1974

1972

1970

1968

1966

0

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

year

Figure 7 – Annual total catches from Maputo Bay semi-industrial fleet (source: IIP and Dengo & Govender 1998). For the period 1965 to 1997 catch per unit effort (CPUE) in the Maputo Bay semiindustrial shrimp fishery was variable from year to year. High CPUEs were recorded in 1966, 1972, 1981 and 1984 (Figure 8). However, in the 1990s CPUE was much lower than historical levels. These lower levels are also reflected in the CPUE series obtained from data provided by the fishing companies. The highest CPUE was recorded in 1972 and equalled 121 kg/fishing day. In 1997 the CPUE had dropped to 46 kg/fishing day.

13 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

c.p.u.e 140

Kg / fishing days

120 100 80 60 40 20 0 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98

Year DNP

IIP

Figure 8 – Trends of catch per unit effort from Maputo Bay semi-industrial fleet (1965 - 1997) (source: Dengo & Govender 1998). There is also a distinct seasonality in shrimp catches. Figure 9 gives the monthly catches for the main shrimp species, P. idicus and M. monoceros, for the years 19941997, as derived from the biological sampling programme. High catches for both species were taken during the months of April and May, with P. indicus having a higher catch than M. monoceros. However, in 1996 the catches of M. monoceros exceeded that of P. indicus during the months of March and April. Generally, catches of M. monoceros exceeded those of P. indicus during the later parts of the year.

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

2.3

The life cycle of shrimps

Shrimps are euryhaline species, meaning that they can tolerate a wide range of salinity. The juveniles are typically found in estuaries, lagoons or coastal areas, whereas the adults move further offshore. The life history of the shrimp may be divided into six phases: • planktonic/benthic embryo phase (inshore/offshore) • planktonic larval phase (inshore/offshore) • juvenile (estuarine) • adolescent (estuarine) • sub-adult (coastal) • adult (inshore/offshore).

14 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

30000

25000

20000

Tons

15000

10000

5000

0

-5000 J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D

1995

1994 P. indicus Others

1996

1997

M. monoceros Total

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Figure 9 – Trends of monthly catch for the main species of prawns in 1994-1997, Maputo Bay (source: Dengo & Govender 1998). The main elements of the life cycle of the shrimp in Maputo Bay are shown in Figure 10, which indicates that spawning occurs in the offshore coastal domain outside Maputo Bay. The eggs and newly hatched larvae are transported by an as yet undefined advective pathway back into the bay – estuarine domain, where they rely on the production fluxes from mangrove forests for the larval development phase. Juvenile shrimps recruit into the fishery, based mainly in the mud bottom bay-scale habitat, from where the mature adults return to the open shelf domain for the next spawning cycle. The schematic flow chart shows the main pathways by which river inputs can impact on the prawn cycle. The impact can occur in the estuarine-based mangroves and/or directly into the bay. In both cases, the main vectors are suspended loads of cohesive (< 60Pm) sediments, detrital organic carbon (POC) and nitrogen (PON), and dissolved nutrients (NO3-, NH4+, PO43-, SiO2). The two main species in Maputo Bay have distinct habitat preferences. P. indicus prefers fringe mangroves, whereas M. monoceros is significantly more abundant in the adjacent sand flat (Rönnbäck et al. 2002).

15 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Nauplius Mysis Zoea

eggs

spawning: coastal – ocean domain

coast-bay estuarine coupling (advection – stratification)

larval development: estuarine domain mangrove habitat

postlarvae recruitment and grow out: Bay domain

river boundary forcing event seasonal water quality and flow

fisheries

adult

Figure 10 – Schematic representation of the shrimp life cycle in Maputo Bay (source: Da Silva & Monteiro 2001).

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

2.4

Research hypotheses

There are many pathways through which river flow could influence shrimp productivity, some directly, but mostly through a chain of relations (see Figure 11). Water quality, salinity and hydraulic characteristics can have a direct impact on the life cycle, but also influence the productivity and health of the marine ecosystems, such as mangroves and seagrass beds, that provide habitat and feeding grounds for shrimp. The strong annual fluctuation in shrimp catch (see Figure 7) suggests a signal that operates on a relatively short time frame, rather than structural changes in habitat that normally take several years to become apparent. This gives rise to the first leading hypothesis for the biophysical part of the research. Scales of forcing hypothesis: 1. The overarching hypothesis across all process domains is that the key linkages between river fluxes and coastal responses occur at the event scale and not at the aggregated seasonal or annual scale. Hypotheses have also been formulated that emanate from three aspects of shrimp population dynamics: migration, trophic status and physiological stress or trigger during different stages of the life cycle.

Physical forcing hypotheses:

16 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

2.

3.

4.

Density-driven circulation: the impact of river inputs on shrimp production is through the stratification and circulation characteristics driven by density gradients that develop as a result of the mixing of river and sea water. This is the ROFI (region of freshwater influence) hypothesis. Salinity hypothesis: the impact of local rain, river-derived freshwater and ground water is key to maintaining the required salinity ranges for the early life stages in the mangrove habitats. Temperature hypothesis: temperature variability is an important factor in defining habitat suitability for the early life stages of the shrimp (shade).

Biogeochemical forcing hypothesis: 5. The impact of river inputs is through the input of suspended sediments, organic matter and nutrients into the mangrove biogeochemical remineralization – production system. Variability in this biogeochemical forcing regulates the space and time scales of energy input into the food web that supports the shrimp.

2.5

Research set-up

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

The most important biophysical interactions between freshwater catchments and coastal domains seem to occur at the sub-seasonal event scale. Unless these scales are resolved the quantification of the economic consequences of those interactions will not be adequate or robust. Catchment2Coast integrated a number of numerical modules (coastal, river basin and ground water) with the required dynamical capability in order to implement a system approach to the functional dependence of coastal systems on river basin and ground water forcing. Each of the models and approaches was developed under a separate work package (WP 1 to WP7). The integration of models was executed as Work Package 8 and the dissemination and legacy activities of the project comprised work package 9. An overview of the project structure is given in Figure 12. One of the important challenges in modelling the response of natural aquatic systems to anthropogenic forcing is to balance the degree of functional complexity with the cost implications of operational reality, user needs and numerical processing limitations. A coupled modelling approach was adopted as the most flexible methodology to: x explicitly address the dynamical considerations, which are thought to be a key to improved predictability. x allow the functional complexity to be adjusted through sensitivity analyses. x allow the cost-effectiveness of monitoring programmes to be optimised by identifying the critical time and space scales through sensitivity analyses. x allow the underlying assumptions to be clearer and more easily questioned than is the case with empirically driven programmes. x strengthen the value of biophysical understanding by allowing the outcomes “currency” to be in a form that is more easily assimilated into development planning in rural and urban areas: resource economics.

17 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Figure 11 – Pathways of factors influencing the shrimp life cycle. Explanation of the relations: 1. Climate (change): exerts an influence on river flow through precipitation and evaporation rates. 2. Catchment landuse: influences through runoff the water volumes and nutrients that enter the river. 3. Economy – land use: complex interactions (long term). 4. Economy – fisheries: market prices influence fisheries effort and vice versa (short term). 5. River flow / water quality: provides an input to the bay of suspended sediments and nutrients. 6. River flow / water quantities: influences the salinity of the bay waters as well as its temperature (estuary/bay/coastal waters scale). 7. Groundwater flow: influences the salinity of mangrove soils and adjacent waters (local effect). 8. Bay-ocean circulation: influences the salinity gradients in the bay.

18 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.

Transboundary catchment Urban-Rural Development Pressures and Needs

Model integration and scenario simulations WP8

River hydrology & WQ model WP 1

Groundwater Hydrology & WQ model WP 2

Coastal – estuarine hydrodynamic model WP 3

Biochemical watersediment quality model WP 4

Mangrove ecosystem health & dynamics WP 5

Shrimp ecosystem dynamics WP 6

Renewable Resource Economics WP 7

Coastal aquatic Rural-Urban economies Sustainability needs

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

22. 23. 24.

Rainfall from tropical storms: can have a possible significant effect on local salinity. Salinity: influences the water column processes and the growth of seagrass and mangroves. Shrimp migration pattern: influenced by bay-ocean circulation patterns. Salinity: influences the physiology of the shrimp (could also have an effect on the transformation from one larval stage to another). Temperature: influences the physiology of the shrimp. Nutrient input: influences the water column processes, growth of seagrass and mangroves. Sediment input: influences the growth of seagrass and mangroves and negatively affects light penetration in the water column by increasing turbidity. Mangrove: is a source of organic matter (detritus) to the water and soil compartments. Mangrove: provides a habitat and food for juvenile shrimp (esp. P. indicus). Seagrass: influences the water column and sediment layer in a variety of ways (source of detritus, oxygen, sedimentation of suspended matter, etc.). Seagrass: provides a habitat and food for juvenile shrimp (not relevant for P. indicus?). Sediment layer – water column interaction: exchange of matter. Water column: primary and secondary production provides food for juvenile and adult shrimp. Competitors: compete for food with juvenile and adult shrimp. Predators: factor of (unknown) mortality rate of shrimp (all stages). Fisheries: factor of mortality rate of adult shrimp.

Dissemination and long term support

Figure 12 – Structure of the project.

19 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

A number of field work campaigns (Table 4) were executed in the Maputo Bay area to provide essential data on the different environmental parameters on which the models are calibrated. The field work was designed in such a way that it would yield a comprehensive picture of all factors that could possibly be relevant for shrimp production. Many of the field campaigns were done simultaneously to allow conclusions to be drawn with respect to interacting variables. The collected data were used to calibrate or validate the models and / or to falsify the hypotheses. Table 4 – Data collection activities of the Catchment2Coast project Field campaign

Period

parameters measured

method of sampling

Maputo Bay hydrodynamics surveys

September 2003, December 2003, March 2004

Conductivity, density, fluorescence, backscatter, oxygen saturation, irradiance, salinity, temperature

CTD* and surface / bottom water samples at coast – bay boundary and estuary – bay boundary

Continuous hydrodynamic monitoring

May-September 2003 January – April 2004

current speed and direction; temperature and salinity in 10 min. intervals

2 moorings with Aanderaa RCM7 current meters

Geophysics field survey

November and December 2003

resistivity

transect with Lung imaging system

Hydrogeochemistry

2003

EC, pH, macro-ions, nutrients, isotopes

samples at boreholes and groundwater wells

18O/deuterium

wood samples

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Mangroves water source Water quality sampling of the Incomati

October 2004 – March 2005

NO2, NO3, NH4, Ortho-P

water samples at Magude

Mangrove biogeochemistry (sediment – water fluxes)

December 2003 – March 2004 and March 2005

oxygen uptake, nutrient (NH4, PO4) fluxes, sediment size, RedOx, POC, C:N ratio, pore water (NH4, PO4, Si), macrofauna activity

benthic incubations, sediment pore water extraction, crab hole density

Mangrove vegetation surveys

January – June 2004

vegetation zonation and human disturbance

quadrate sampling at 7 locations in mangrove

20 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Table 4 (Cont.) Field campaign

Period

parameters measured

method of sampling

Sampling of planktonic larvae

March – May 2004

planktonic larvae (number of individuals / taxa)

plankton nets at surface and bottom of estuary

Sampling of post-larvae and juvenile shrimp

February 2003 – February 2004

abundance and shrimp size

Sampling of artisanal shrimp fishery

2003 and 2004

shrimp catch

beam trawling net in mangrove habitats visits to main fishery centres and markets

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

* CTD stands for Conductivity-Temperature-Depth recorder

21 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Chapter 3 The research outputs 3.1

Catchment runoff and river flows

Vivek Naiken, Marilyn Govender, Mags Moodley, Mike Savage

3.1.1

Catchment modelling

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Hydrological models can assist in projecting the impact of land and water management strategies. However, given the complex mosaic of different land uses in most catchments, a spatially explicit model is required. The chosen models therefore had to be integrated with recent advances in geographic information systems (GIS) and other spatial analysis software. Although many hydrological models are widely available, few have been easily integrated into the GIS environment. One such model is the Soil Water Assessment Tool (SWAT), developed at the Blackland Research Centre in Texas for the United States Department of Agriculture (USDA) Agricultural Research Service. The SWAT model is a daily time-step model and some of its key strengths lie in the ability to predict the relative impacts of changes in management practices, climate and vegetation on water quantity and quality. Despite the fact that freshwater input is dominated by the two main rivers, the Maputo and Incomati, the process studies focused on the Incomati River and estuary only. This was a necessity linked to capitalizing on the historical effort that has gone into the Incomati relative to the Maputo system. This problem was recognized and an approach was adopted to overcome it in collaboration with a parallel research effort on the Maputo River. The project used the Agricultural Catchments Research Unit model (ACRU) (Schulze et al. 1995) for the Maputo River, and input variables were determined based on available and sufficient data and information. ACRU is a physical-conceptual model in that “it conceives of a system in which important processes and couplings are idealized, and physical to the degree that physical process are represented explicitly” (Schulze 1995). Despite several advances made in this respect, the results were too coarse for use at the scale that was required by the hydrodynamics modelling. A brief outline of the concepts and general structure of the SWAT model is included here. The model is a continuous daily time-step model developed to simulate the longterm impacts of land and water management (e.g. reservoir sedimentation over several years) or agricultural practices (e.g. crop rotation, planting and harvesting dates and irrigation). It is physically based, uses readily available inputs and is computationally efficient to operate on large catchments within a reasonable time. SWAT allows a number of different physical processes to be simulated in a catchment, which may be grouped into the following major divisions: x hydrology, weather and soil temperature x sedimentation

22 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

x x x

crop growth nutrients and pesticides agricultural management.

The model may simulate a catchment in lumped or distributed mode, by automatically delineating the catchment either into sub-catchments or hundreds of grid cells based on a Digital Elevation Model (DEM). The use of sub-catchments in a simulation is particularly beneficial to differentiate the impact of various land uses and soils on the hydrology of a catchment. The development of the model as an extension to ArcView has increased the flexibility of the model, with the special features of ArcView also being available to SWAT model users. Other basic data requirements include: x spatial coverage of land cover and soil types x daily precipitation x daily maximum and minimum air temperature.

3.1.2

Incomati model set-up

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Spatial data, including a Digital Elevation Model (DEM), and land use and soil maps, were sourced from various South African, Mozambique and Swaziland databases (see Table 5). All spatial data sets extracted from Mozambique were re-projected to a more commonly used projection in South Africa (i.e. Transverse Mercator, Clark 1880) for comparative purposes during model configuration. The land cover and soil spatial coverages were classified according to the Mozambique databases (Figure 13). Land cover has been classified into nine classes ranging from deciduous and evergreen forests to croplands and savanna-type vegetation. Large areas of the catchment are covered with deciduous broadleaf forest. The nine land use classes were set up in the model to utilize SWATs land cover classes listed in the land cover database, but modified to represent South African conditions. On the other hand, soils have been classified into four very broad soil classes. Large gaps in the model inputs still exist, when characterizing the soils of the Incomati catchment. The five major drainage basins contributing to streamflow entering Mozambique viz. Crocodile, Sand, Sabie, Nwanedzi, and the Komati were used as a guide for subcatchment delineation. In SWAT, each sub-catchment may also be distributed into one or more unique land use/soil combinations (hydrological response units, HRUs). HRUs for the Incomati were defined according to dominant land use and soil present in each sub-catchment. Available soils data from South African and Swaziland data sets allowed the catchments to be configured with two soil layers, which were added to the SWAT soil database.

23 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Table 5 – Spatial data sourced from South Africa, Mozambique and Swaziland Country

Type of spatial data DEM Land cover Soils River network

Spatial resolution 400 m 1 : 250 000 1 : 250 000 1 : 250 000

Spatial projection Transverse Mercator Clark 1880

Data source CSIR

Swaziland

Soils

1 : 250 000

Transverse Mercator Clark 1880

University of Swaziland

Mozambique (Incomati Basin)

DEM Land cover Soils River network

1000 m

Lambert Equal-Area Azimuthal

ARASUL

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

South Africa

Figure 13 – Land cover classification for the Incomati catchment. The temporal data used in the model is listed in Table 6. Rainfall and air temperature stations were selected using the Southern Africa Quaternary catchment database programmes written by the School of Bioresources Engineering and Environmental Hydrology (BEEH, University of KwaZulu-Natal). Long-term daily rainfall and air temperature data sets from 1951 to 2000 were extracted from daily rainfall and air temperature databases. Data quality was checked and interpolated using real data from surrounding stations or relevant rainfall or air temperature patching techniques. Ultimately, 93 rainfall stations and 77 air temperature stations were selected. Available

24 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

stream flow records for the Incomati catchment have been requested from the relevant institutions in South Africa, Mozambique and Swaziland. Monthly stream flow records for two gauging stations in Mozambique - Ressano Gracia and Magude - have been obtained and used for calibration (record lengths 1953 to 2002 and 1983 to 2000, respectively). Abstraction data have been difficult to obtain. Abstraction data for Maguga Dam and the Komati River were collected. This is, however, a non-continuous record as measurements have been taken at discrete time intervals. In addition, a list of farmers that draw water from the Komati river basin for irrigation has been compiled. Data for dam operating and inter-basin transfer rules was not sourced. Table 6 – Temporal data sourced from South Africa, Mozambique and Swaziland Country

Type of temporal data Rainfall Air temperature Wind speed Land use attributes Soil attributes Water quantity Water quality

Data availability 1950 – 2000 1950 – 1995 Good Good Good 1983 – 2000 1980 - 1999

Data source

Swaziland

Rainfall Air temperature Wind speed Land use attributes Soil attributes Water quantity Water quality

1950 – 2000 1950 – 1995 Good Good Good 1983 – 1999 1980 – 1999

Rainfall database (BEEH) Air temperature database (BEEH) South African Weather Services CSIR CSIR DWAF Water quality (CSIR)

Mozambique (Incomati Basin)

Rainfall Air temperature Wind speed Land use attributes Soil attributes Water quantity Water quality

1950 – 2000 Poor Poor Poor Poor Poor Poor

Rainfall database (BEEH) IIP IIP ARASUL ARASUL IIP IIP

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

South Africa

3.1.3

Rainfall database (BEEH) Air temperature database (BEEH) South African Weather Services CSIR CSIR DWAF Water quality (CSIR)

Calibration and verification results

After parameterizing the model with the necessary data a calibration run for water quantity was undertaken using the two gauging stations Ressano Gracia and Magude. The data set was split up into four periods (see Table 7). The final verification period (January 2003 to January 2004) was introduced to correspond with the common

25 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

simulation period adopted by all research partners. This baseline data set was used as a reference point for future model simulation or scenarios. Table 7 – Breakdown of time periods for model simulation Phase

Time Period

Warm-up phase

October 1983 – September 1985

Calibration

October 1985 – September 1990

Verification (Initial)

October 1990 – September 2000

Verification (Final)

January 2003 – January 2004

Figure 14 presents the calibration curve for the entire Incomati catchment. Stream flow, which is strongly seasonal, is generally over-simulated for extreme events but under-simulated during times of low flows. The over-simulation is probably associated with floods that occurred in the region between January and February 1988 and January and February 1989. Such flooding could have overtopped weirs, causing damage to gauging stations, and thus leading to errors in the observed record. Nevertheless, a high r2 of 0.84 was obtained. Base flow has been under-simulated due mainly to poor records of abstraction, especially when this relates to subsistence agriculture. An additional reason for the under-simulation is that the SWAT model requires parameters that are much more sensitive than the data currently available.

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

The simulation of stream flow during the verification period yielded some interesting results (Figure 15). It was encouraging to note that the seasonal trend in stream flow is adequately demonstrated in both the observed and simulated records. During the verification period two large floods occurred in the region during April 1996 and April 2000. These events have generally been under-simulated by the model but, as mentioned earlier, this may be related to inaccurate records of actual stream flow because of overtopping of the weir or access to gauging stations. Nevertheless the general trends in stream flow between the observed and simulated stream flow is quite consistent, which lends confidence to the model output. A hindcast run was undertaken to check and test the model’s behaviour for the complete study period from January 1996 to December 2004. The models output for the Incomati catchment fairs very well when compared to the observed data. The model has also responded well to rainfall events with the lag or recession flow after storm events closely following the same path as the observed stream flow. The nineyear river flow time series was a key input into the ecosystem model hindcasting of the production variability between 1996 and 2004. The river flows formed time-varying boundary conditions for the Maputo Bay hydrodynamic model.

26 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

200

50.00

150

100.00

100

150.00

50

200.00

0

250.00

Rainfall (mm)

0.00

O

ct -8 6 D ec -8 6 Fe b87 Ap r-8 7 Ju n87 Au g87 O ct -8 7 D ec -8 7 Fe b88 Ap r-8 8 Ju n88 Au g88 O ct -8 8 D ec -8 8 Fe b89 Ap r-8 9 Ju n89 Au g89 O ct -8 9 D ec -8 9 Fe b90 Ap r-9 0 Ju n90 Au g90 O ct -9 0

Streamflow (m3/s)

250

Date Rainfall (mm)

Observed

Simulated

Figure 14 – Comparison between simulated and observed streamflow (m3·s-1) for the Incomati catchment at the Magude weir, October 1986-October 1990 (calibration phase).

1800

0.0

1600

50.0

1400

150.0

800

200.0

600 250.0 400 300.0

200 0 ct -9

1 Ap r-9 2 O ct -9 2 Ap r-9 3 O ct -9 3 Ap r-9 4 O ct -9 4 Ap r-9 5 O ct -9 5 Ap r-9 6 O ct -9 6 Ap r-9 7 O ct -9 7 Ap r-9 8 O ct -9 8 Ap r-9 9 O ct -9 9 Ap r-0 0 O ct -0 0 Ap r-0 1 O ct -0 1 Ap r-0 2 O ct -0 2 Ap r-0 3 O ct03

350.0

Date Rainfall (mm)

Observed

Simulated

Figure 15 – Comparison between simulated and observed streamflow (m3·s-1) for the Incomati catchment at the Magude weir, October 1991-October 2003 (verification phase).

27 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Rainfall (mm)

Streamflow (m3/s)

1000

O

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

100.0 1200

3.2

Maputo Bay hydrodynamics

John Simpson, João Lencart, Pedro Monteiro, Stephen Luger, Jenny HarcourtBaldwin, Antonio Hoguane

3.2.1

Introduction

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Regions of freshwater influence (ROFIs) in lower latitudes are characterized by a pronounced seasonal cycle in river discharge, which can induce a radical change in circulation patterns. Maputo Bay in Mozambique is one such system in which a strongly varying discharge interacts with vigorous tidal and wind mixing, changing the water density structure between a tidally driven oceanic bay in the dry season and a ROFI system in the wet. This large shallow embayment receives estuarine inputs at three major points, with discharges largely concentrated in three months of the year (December to March). In extreme years, the flow can be of the order of 2000 m3·s-1 during this period, while during the rest of the year, freshwater discharge is nearly absent. Although description of system dynamics of similar bays is absent from the literature, the review carried out in the first year of the project highlighted the mechanisms that control the density structure and produce marked changes in circulation. The main factor that controls the density gradient and the attendant estuarine circulation is freshwater availability. Wind and tidal stirring contribute to the erosion of vertical stratification, while the estuarine circulation and a positive surface heat input promote stratification. The physical oceanography programme was designed to investigate the response of Maputo Bay’s circulation, flushing and water column structure to the main forcing factors, namely tides, freshwater input, wind stress and surface heat exchange. Observational work involved the deployment of moored current meters and tide gauges together with monthly conductivity, temperature and depth (CTD) surveys of Maputo Bay and the Incomati estuary over a 15 month period. In parallel with the observational campaigns, a 3 dimensional (3D) hydrodynamic model, calibrated by comparison with the current meter observations, was used to explore the response of the bay to a wide range of forcing scenarios. Following is a summary of the main outcomes of the combined measurement and modelling studies.

3.2.2

Hydrodynamic modelling

The hydrodynamic processes were modelled using the Delft3D-FLOW modelling package. One of the key issues to be addressed in setting up the hydrodynamic model was how to resolve the range of scales across all the linked domains (mangrove creeks ~1 – 10m) to the coastal shelf system ~100 km. Initially it was thought that two-way nesting may be the most effective way of overcoming these extremes while still maintaining the integrity of the entire lower ecosystem. This was eventually decided not to be the most effective way of linking up the system domains due to likely problems with time steps in the small scale extreme controlling the overall system. The lack of data on mangrove topography was also a problem. The domain was finally set up with a variable resolution curvilinear continuous grid from the upper tidal

28 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

reaches in the rivers to the shelf break (Figure 16). This approach excludes the mangrove “catchment units”.

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Model calibration was undertaken initially for tidal water level variability. Boundary forcing with a basin-scale model (le Provost) resulted in good recovery of the phase and amplitude within the bay itself (Figure 17).

Figure 16 – The hydrodynamic model grid (3D with 8 layers) implemented for the Maputo Bay ecosystem, showing the very high resolution in the estuarine domains and coarse resolution at the open boundaries (colours show depth in meter).

29 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Figure 17 – Tidal water levels modelled and measured for Maputo Bay (x-axis shows hours).

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

3.2.3

Observational and model results

Maputo Bay is a tidally energetic embayment, which is forced by a combination of tides and highly variable seasonal freshwater input (up to 3000 m3·s-1 in extreme conditions). The tides are semidiurnal, with a range of ~0.5 m at neap tides, increasing to ~3 m at spring tides. This large springs to neaps ratio (~6) in the tidal range means that the tidal currents and associated tidal stirring are strongly modulated on a fortnightly cycle, with dissipation increasing by 2-3 orders of magnitude to peak spring tide values of ~ 1 W·m-3. This high level of tidal stirring, supplemented at times by wind stress, maintains a mixed water column throughout the bay during the dry season. The strong tidal currents are also responsible for a significant, tidally- rectified circulation which, in the absence of freshwater input, appears to dominate the circulation. During the wet season, freshwater input produces additional density-driven currents, which also appear to be modulated by the tides. These forcing factors are plotted together with the density structure represented by the potential energy anomaly ĭ in Figure 18. Although the tidal range rises and falls periodically and continuously changes the amount of energy available for bottom mixing, the water column remains

30 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

vertically mixed throughout the dry season. However, during the wet season, when river inputs rise to ~103 m3·s-1 in a few days followed by several episodes of intense discharge later in the austral summer, the freshwater input causes a reduction in salinity, along with an associated increase in buoyancy input to the bay. This leads to the development of stratification (as depicted by greater values in the ĭ plot) in response to the residual currents forced by density differences. The influence of the freshwater input is also evident in comparing the survey results from day 390 (Figure 19a) and day 420 (Figure 19b). Figure 19a shows a reasonably uniform density distribution, although surface heat input was close to maximum values. Figure 19b shows a horizontal gradient when the bay was under the influence of freshwater inputs. The horizontal gradient now is of the order of 10-4 kg·m-3·m-1 but the spring tide is still capable of mixing the water column and contributing to the sharpening of the frontal structures. It can therefore be inferred that freshwater plays the central role in controlling horizontal density gradients in Maputo Bay.

36 35 34 Tidal Range (m)

100

200

300

400

500

0

100

200

300

400

500

0

100

200

300

400

500

0

100

200 300 400 Time (days since 01/01/2003)

500

2 1

500 0 -3

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

0

3

1000

Phi (W.m )

River Flow (m3/s)

33

4 2 0

Figure 18 – Seasonal evolution of forcing agents and consequent change in salinity and potential energy anomaly in section A. From top to bottom panels: bay averaged salinity; synthetic tidal range from analysis of Clube Naval data; Incomati River flow from Magude River gauge; section mean potential energy anomaly and standard deviation from a central section of the bay.

31 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

0

-5

-5

-10

-10

Depth (m)

0

-15 0

-15 5

10

15

20

25

30

Sigma t (kgm-3)

35 -20

25

.5

.5

.5

.5

.5

.5

24

24

23 23

22 22

21 21

20 20

19 19

0

5

10

15

20

25

30

35

Distance from Maputo harbour (km)

Figure 19 a & b – Density structure at section A (Maputo harbour to Inhaca Island). Left: survey at day 390; right: survey at day 420. Detailed observations over a tidal cycle in the wet season revealed the development of some semi-diurnal periodic stratification induced by tidal straining. This short-period variation in stratification occurs even at spring tides when the level of stirring was strong enough to prevent the development of a stable situation. Generally, buoyancy inputs by surface heating alone were insufficient to induce stratification, and the observations indicate that significant persistent stratification is only induced at times of high levels of freshwater input. Even then, stability is rapidly eroded by tidal stirring, except during neap tides.

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Observations of the density field at monthly intervals did not allow direct investigation of the changes in flushing over the fortnightly cycle, but it was possible to use a tidal prism model to estimate the wet season mean flushing time to be ~10 days with a retention factor of ~0.8 (the retention factor is the fraction of water leaving on the ebb tide that re-enters the bay on the flood tide). For wet season conditions with low tidal energy input, the Delft3D model indicates a flushing time of ~20 days, which is significantly different from the tidal prism model estimates. For mean wet season conditions, the Delft3D model indicates even longer flushing times (~40 days). An interesting inference from the observations, confirmed by model runs, is that the density driven flows due to freshwater input, which are evident at neap tides, are largely suppressed by the increased vertical mixing at spring tides, as in the model experiments of Linden and Simpson (1981). This means that, under wet season conditions, the flushing time of the bay will increase with increasing tidal range. This is in marked contrast to the behaviour expected in the dry season, when flushing times should decrease with increasing tidal range as residual circulation and dispersion are both enhanced by larger tidal currents. The opposing effects of tidal current amplitude under conditions of high and low river input are illustrated in a simple analytical model of flushing, which allows for both shear dispersion and density-driven circulation modified by tidal stirring. The result for the full range of tidal currents and freshwater inputs is shown in Figure 20.

32 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Maputo Flushing times 4

60

3.5

3

50

log(Qf(m 3/s)

2.5

40

2 30 1.5 20 1

10

0.5

0

0.2

0.4

0.6

0.8

1

1.2

U(m/s)

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Figure 20 – Flushing time in days for an embayment forced by tidal currents (m·s-1) and freshwater input (log of river discharge in m3·s-1). Flushing is due to a combination of shear diffusion and density-driven circulation. This simple model indicates that for discharge above 10 m3·s-1, flushing time is less than 30 days as tidal stirring increases from neap to spring tides. For low discharge and low tidal flow, flushing times would increase to more than 30 days, but since low tidal flow is never sustained for longer than the neaps period of a few days, the model indicates maximum flushing times of ~10-20 days, which is in fair agreement with the observations. Under high discharge conditions (Qf ~1000 m3·s-1), flushing times are short (n( s, t )K ( s, t )@  P >( s )n( s, t )@ ws

(2)

where: t = time (d) n = number of individuals of weight s Ș = scope for growth (g d-1) ȝ = mortality rate (d-1).

49 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Three weight classes were considered: Class 1 corresponds to juveniles that live in the nursery areas, Class 2 to subadults and Class 3 the adult stage. Subadult and adult shrimp live in deeper waters of the bay. The number of individuals in each weight class depends on the individual scope for growth (calculated in the individual growth model) and on a natural mortality rate. The appropriate number of weight classes was tested as well as the spatial delimitation of each development stage in the model boxes. The biomass of the different classes is calculated at each time-step by multiplying the number of individuals they contain by the average weight of the class. Mortality represents the loss of individuals and comprises a natural mortality and fishing mortality. The natural mortality was considered to be a function of shrimp size and it decreased from the first to the last class. In addition to the natural mortality, a fishing mortality was also applied to classes 2 and 3, as described below. A man object simulates the human component and takes into account the fishing practice and the cohorts exploited. A class for Maputo Bay was developed which includes both artisanal and semi-industrial catches. For the semi-industrial fleet, monthly catch data from the Mozambican Fisheries Institute (IIP) were used, while for the artisanal fishery a constant annual catch was assumed. This class allows the simulation of different management strategies regarding catch rates and periods of fishing. The density dynamics of the three weight classes simulated is shown in Figure 32.

Juvenile

Subadult

Adult

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

-2

Shrimp density (ind m )

0.30 0.25 0.20 0.15 0.10 0.05 0.00 0

60

120

180

240

300

Time (Days) Figure 32 – Variation of shrimp density over the simulation period.

50 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

360

Juvenile density increases during the recruitment period and decreases through mortality and shrimp growth to the next stage of the life cycle, resulting in a density decrease from the first to the last stage. It is possible to follow the periods of maximum density for juveniles, subadults and adults, which are achieved in March, April and June respectively. The higher availability of shrimp occurred during the first half of the year, corresponding to the period when higher catches are recorded (Paula e Silva, pers. comm.). Figure 33 shows shrimp biomass dynamics per unit area. These results follow the same pattern obtained for the shrimp density, showing maximum values during the rainy season and minimal values during the southern hemisphere winter. As expected the adult class is dominant in terms of biomass with a maximum value of about 2.5 gFW·m-2 during the peak.

-2

Shrimp biomass (gFW m )

Juvenile

Subadult

Adult

3.00 2.50 2.00 1.50 1.00 0.50 0.00 0

60

120

180

240

300

360

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Time (Days) Figure 33– Variation of biomass per unit area over the simulation period.

51 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

3.7

Resource Economics: Prawn Fisheries

Martin de Wit, Rowan Le Roux

3.7.1

Introduction and data

The ability to effectively estimate the economic impacts of policy changes on the livelihoods of the Mozambique semi-industrial and artisanal shrimping industry in Maputo Bay requires the estimation of industry-specific production functions that predict catch rates based on the available shrimp stock (biomass) and on fishing effort. Biomass (S) is estimated using a biophysical model (Shrimp Model), which resolves the problems associated with using a proxy for biomass (such as catch per unit effort or seasonality) or assuming a constant stock. Catch rates in turn feed into the industry profit function as follows: PROFIT = Total Revenue – Total Cost

(3)

where: Total Revenue = Catch rate * Market price per unit catch and Total Cost = total expenses involved in harvesting this catch.

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

The model can thus be used to estimate the effects of policy changes on industry profit by affecting shrimp biomass or fishing effort (through the predicted impact on catch rates), or expenses. Time series data were collected for both the semi-industrial and artisanal fisheries (Instituto de Investigaçao Pesqueira (IIP) 1996-2003). However, data on the artisanal fishery proved to be unreliable and erratic. Analysis was therefore restricted to the semi-industrial fishery. Data for this fishery was collected for the period 1996 to 2003 on (1) the monthly total Shrimp Catch (C) in kilograms and (2) the total monthly Fishing Effort (D) in days at sea per month. The monthly catch and effort data for 1996 to 2003 is presented graphically in Figure 33. The observed effort data were highly volatile and did not appear to exhibit any clear trend. This observation is critical when it comes to forecasting effort into the future, as the simple extrapolation of the average effort in any given period would not be satisfactory. From the above data the monthly average Catch per Unit Effort (CPUE) for the fleet was calculated for the 1996 to 2003 period (see Figure 34). The March/April period is usually associated with a peak in CPUE, corresponding to the opening of the shrimping season (January and February are closed months). A smaller peak is often experienced around the September/October period, and is potentially associated with an increase in shrimp density due to spawning aggregations.

52 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mar Apr May Jun Jul Aug Sep Oct Nov Dec

kg/day

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Effort (Days) 700.00

400.00 25,000

300.00 20,000

200.00 15,000

10,000

100.00 5,000

-

1996

1996

1997

1997

1998

1998

1999

Months 2000

1999

2000

2001

Effort (days)

2001

2002

2002

Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated, 2003

Catch (Kg)

Figure 33 – Aggregated Catch and Fishing Effort (1996 – 2003).

Catch Per Unit Effort

100.00

90.00

80.00

70.00

60.00

50.00

40.00

30.00

20.00

Months

2003

Figure 34 – Catch per Unit Effort (CPUE) (1996 – 2003).

53

Catch (kg)

Aggregated Catch and Effort (Monthly) 45,000

600.00 40,000

500.00 35,000

30,000

CPUE values for 2003 are significantly lower than in previous periods due to an increase in the number of vessels in this year (25 boats versus an average of 20 over the previous seven years) and hence in total effort, without a corresponding increase in catch.

3.7.2

Methodology

As in previous studies, this study is based on a production function that relates catch rates (C) to fishing effort (E) and biomass (S). Due to data limitations with respect to vessel types and characteristics, fishing effort (E) is estimated in terms of the number of days spent at sea by each vessel operating out of the Maputo harbour in each month. Other variables that might be expected to affect the amount of effort expended, such as the market price of shrimps and the cost of inputs such as fuel, were excluded due to insufficient data. A further variable that may be expected to affect the semi-industrial catch is the artisanal catch. A higher artisanal catch will mean that less of the available stock can be caught by semi-industrial harvesters. However, data regarding the artisanal catch was too unreliable for this variable to be included in the analysis. The production function thus took the following general form: Catcht = ƒ (Effortt; Biomasst)

(4)

Where: Catcht = total catch (kg) in period t Effortt = total fishing days (days) in period t Biomasst = total stock or biomass (kg) available during period t, where

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Biomass = ƒ (biophysical model)

(5)

Utilising data provided by the Fisheries Institute (IIP), the semi-industrial shrimping fleet production function was calculated. The semi-industrial production function was used to estimate the economic consequences of various policy decisions which effect both biomass (such as policies affecting the supply of nutrients, mangroves and fresh water, which are all essential determinants of shrimp biomass), and effort (such as changes with regard to the closed season and number of boats). Based on the impacts of these policy changes on biomass or fishing effort, the production function enables prediction of the resulting effects on catch rates, and thus on revenue generated by the fishery. In line with economic theory, and with literature suggesting that a Cobb-Douglas production function is an appropriate functional form for a fishery production function (e.g. Hannesson 1983), it was assumed that the production function for the Maputo Bay semi-industrial shrimp industry would take a Cobb-Douglas form, as in Equation (6), which is reproduced below:

54 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

C = AEȕ1Sȕ2

(6)

Where: C = Catch E = Effort S = Biomass A = constant This form allows for a variable catch coefficient, q, which depends on both ȕ1 and ȕ2, the estimated coefficients (Hannesson 1983). A double log functional form was used; that is, the regressions were performed on the natural logs of both the dependent and independent variables. The regression results are presented and discussed in the following section.

3.7.3

Results

An ordinary least squares (OLS) regression of the natural log of catch rates on the natural logs of fishing effort (E) and biomass (S) yielded the following results: SUMMARY OUTPUT Regression Statistics Multiple R R Square Adjusted R Square Standard Error Observations

0.68 0.46 0.44 0.37 80.00

ANOVA df Regression Residual Total

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Intercept Effort Biomass

SS 2.00 77.00 79.00

8.81 10.47 19.28

Coefficients 1.60 0.95 0.34

Standard Error 1.02 0.15 0.08

MS 4.41 0.14

F 32.41

t Stat 1.56 6.16 4.31

P-value 0.12 0.00 0.00

Significance F 0.00

Lower 95% Upper 95% (0.44) 3.63 0.64 1.25 0.18 0.50

From the above results the following production function was observed for the semiindustrial shrimping industry: Ln(Catch) = c + Dln(effort) + Eln(Biomass)

(7)

Where: D = 0.95 E = 0.34 and c = 1.60. The resulting Cobb-Douglas equation is as follows: Catch = e1.60 Effort0.95Biomass0.34

(8)

55 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

The forecasting power of the model (R2 = 46%) can be demonstrated by comparing actual (observed) catch rates with catch rates predicted by the equation. Figure 35 illustrates this predictive power for the period analyzed. 45 000

40 000

35 000

Kilograms

30 000

25 000

20 000

15 000

10 000

-

Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mar Apr May Jun Jul Aug Sep Oct Nov Dec Mar Apr May Jun Jul Aug Sep Oct Nov Dec

5 000

1996

1997

1998

1999

2000

2001

2002

2003

Time Predicted Catch

Actual Observed Catch

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Figure 35 – Actual versus predicted catch. Two concerns around the data and the resulting ‘fit’ of the model arise at this stage. Firstly the model is unable to predict the high peak in catch for the 1999 period; this is as a result of an error in the original hindcast data provided by the biomass model. Secondly, as the model is only able to use number of days spent at sea per month as its effort proxy, additional factors affecting effort such as tidal fluctuations, vessel capacity, and fuel prices, to name a few, are excluded from the model. If the model were able to include this additional data the predictability of the model would improve.

3.7.4

Analysing catchment and industry policy changes

While being cognisant of the two concerns highlighted above, the bioeconomic production function estimated above was used to model the effects of a change in policy directly affecting the amount of effort exerted by the semi-industrial shrimping industry, as well as changes in the catchment ultimately affecting the availability of shrimp biomass. The proposed policy decision affecting effort was the extension of the closed (no-fishing) period from the current two months to three months.

3.7.5

Policy decision affecting effort exerted

The policy decision that required analysis was whether the additional month should be added to the beginning or the end of the closed period, i.e. December or March.

56 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

The baseline forecast effort scenario was calculated by utilizing the existing effort distribution and drawing a random effort variable from between the two observed effort data points 7 and 8 years previously. That is, the forecast effort variables for 2004 were generated by drawing a random number between the corresponding 1996 and 1997 observed effort data points. Similarly, the 2005 forecast effort variables were generated using the 1997 and 1998 observed data points. The eight-year lag period was chosen as it allows for the historic distribution of effort captured in the existing data to be fully extrapolated into the future. Shortening the lag period reduces the available range in which the effort variable could fall. That is if the lag period was reduced to the final two years of observed data all the forecast effort variables would be restricted to between 443 days and 575 days for March, for example, as opposed to between 298 days and 575 days when the longest possible lag time is used. The result of the above methodological application was a baseline effort scenario that maintained the historic total effort average over the 25 year forecast period of 4 100 days while still maintaining a relatively high level of volatility, with the total effort ranging from a low of 3 400 days to a high of 4 800 days. For both closed period scenarios, the total average effort exerted over each fishing cycle was held constant at an average 4 100 fishing days over the 25 year forecast period, thus the fishing days removed as a result of the increased closed period were re-apportioned to the remaining periods. Holding the total effort constant is consistent with the assumption that in order to maintain the current profit levels experienced by the fishing fleet, more effort will have to be exerted in the remaining periods.

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

The average effort exerted during the March period was 378 shrimping days, while the average effort exerted during December was 353 shrimping days. However, the critical concern between these two periods is the CPUE obtained for the period. March, on average, showed the highest CPUE return for the entire shrimping cycle, with CPUE measured at an average of 79 kg/effort. Alternately, December observed the lowest average CPUE for the periods under review, with CPUE averaging only 35.5 kg/effort. This implies that if March is included in the closed period, while maintaining the effort restriction to 4 100 days, the resulting effort redistribution causes an anticipated reduction in the total catch over the period of approximately 2 tonnes (this number is estimated using a baseline predicted shrimp biomass that is calculated using the same theory as that used to estimate effort). This occurs regardless of the fact that the average effort exerted per remaining period increases, as the availability of the actual biomass and hence the resulting high CPUE in the March period is a restricting factor. Conversely, if December is included in the closed period, the resulting effort redistribution causes an average increase in the total catch of approximately 2.5 tonnes over the period (this number is estimated using a baseline predicted shrimp biomass that is calculated using the same theory as that used to estimate effort). This is explained as a result of an increase in the effort exerted over the high CPUE periods in the beginning of the year.

57 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

3.8 2004

Hindcasting Ecosystem Production in Maputo Bay 1996 –

Pedro Monteiro, Ana Franco, João Lencart, Jenny Harcourt Baldwin, Marcel Marchand

3.8.1

Maputo Bay System Model

An investigation on whether the salinity – production relationship could be tested against an independent data set was conducted. The stringent test to the model was set to verify whether modelled ecosystem production could hindcast the variability in prawn CPUE over a period of nine years. The salinity variability at the mangrove areas was generated by forcing the model with nine years of observed and modelled runoff from the Incomati and Maputo rivers respectively. The investigation was undertaken using the system model framework depicted in Figure 36. This shows how the estuarine – coastal hydrodynamics are forced by the river flows from the SWAT basin model and potentially the ground water model too, although this link was not implemented or tested. The hydrodynamic model generates the salinity time series at the mangrove domains, which are used to drive the mangrove new production fluxes using the Ravikumar N-fixation – salinity model.

High resolution: max 1 year scale River Basin: SWAT

Mangrove Production Model

Aggregated Scales: 20 year scale GIS: spatial upscaling mangrove production

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Flows, WQ

Estuarine–Coast hydrodynamics Delft3D

Groundwater

Salinity Variability

EcoWin: shrimp IBM population growth and fishing

Total Economic Value Model

FE Flow

Figure 36 – The system model framework used for the Catchment2Coast project.

58 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

The production – salinity relationship in Figure 37 was used to convert the salinity variability to a production flux, which was then scaled up using the mangrove ecosystem GIS to define the spatial characteristics of non-degraded mangrove, half of which was deemed to be mature Avicennia marina. Mangrove New Production - Salinty Relationship Maputo Bay (from: Kumar, 2004) 1.00E+03

Production Rate (mgCm-2d-1)

9.00E+02 8.00E+02 7.00E+02 6.00E+02 Lower Rate

5.00E+02

Upper Rate

4.00E+02 3.00E+02 2.00E+02 1.00E+02 0.00E+00 0

5

10

15

20

25

30

35

Salinity

Figure 37 – Mangrove new production - salinity relationship.

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

3.8.2 Model Results The results reveal that 1996 and 2000 were the years with the highest productivity rates, linked to higher river flows and the persistence of bay – mangrove low salinity conditions (Figure 38). The predicted production in 1999 was a significant underestimate of what was expected given the flow from the Incomati River. It can be seen that the increase in production in the wettest years is driven by the contribution from the Machangulo mangroves and to a lesser extent the Costa do Sol mangroves. In the driest years these domains make no contribution to the total bay ecosystem production. When the new production rates are scaled to the respective mangrove areas it can be seen that the Costa do Sol mangroves make a minimal contribution to the total daily production (Figure 38). This minor role is due to the combined effects of relatively high salinity and small area. The typical predicted carbon fluxes available to the nursery systems are 2 – 4 tons per day (but as low as 1 -2 tons C per day in dry years) to support the growth rates through the post larval – juvenile stages. Most of this is concentrated in the Incomati and Maputo river estuarine mangroves. In wet years the Machangulo system in the south-east of the bay significantly increases the production fluxes.

59 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Maputo Bay Ecosystem Production Monthly Average Production per m2 (High Coefficients) 2500

Machangulo Maputo River Xefina

Production (mgC m -2d-1)

2000

Costa do Sol

1500

1000

500

Jan04 Feb04 Mar04 Apr04

Jan03 Feb03 Mar03 Apr03

Jan02 Feb02 Mar02 Apr02

Jan01 Feb01 Mar01 Apr01

Jan00 Feb00 Mar00 Apr00

Jan99 Feb99 Mar99 Apr99

Jan98 Feb98 Mar98 Apr98

Jan97 Feb97 Mar97 Apr97

Jan96 Feb96 Mar96 Apr96

0

Wet Season Months, 1996 - 2004

Figure 38 – Plots of the modelled variability of new production fluxes (mg·C·m-2·d-1) in each of the main mangrove domains for the four months of the rainy season (Jan – Apr) over a nine-year period (1996 – 2004), revealing that the highest rates are in the river estuaries.

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

3.8.3

Bay system production variability 1996 - 2004

The mangrove system production data aggregated to a wet season average (kg·C·d-1) for the nine-year simulation shows that there can be a 4-fold (2 – 8 tons C·d-1) variability in the energy supply to the nursery system ( Figure 39). The years showing the largest contrast are 2003 (driest) and 2000 (wettest, coinciding with significant flooding). The remaining years vary around a narrowed range of ~3 – 5 tons C·d-1. The CPUE data for the same period (Figure 39) appear to reflect a sensitive interannual variability trend that is close to that described above in the wet season mangrove production. The trends are strong in most instances, with the exception of 1999. A closer inspection of the river flow data that was used to force the bay hydrodynamic model showed that the flow of the Maputo River was in that year well below what was expected. The flows of the Maputo River, and hence the salinity variability in the southern mangroves, are the main sources of uncertainty in the ecosystem production model. Essentially, years with elevated river inputs drive higher new production rates in the nursery domains, which result in higher recruitment of juvenile prawns to the adult

60 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

fishery in the bay. This summarises the functional or process linkage between the river and coastal ecosystem production variability. The corresponding variability of the wet season and bay-averaged salinity at all four mangrove domains shows an almost inverse relationship. This plot suggests that a process link may not be entirely necessary and that a simple salinity – CPUE correlation may work just as well as a predictor. However, such a correlation would assume that salinity is the only forcing factor driving shrimp variability, which is incorrect, as the area of mature Avicennia is also important. If this area changes, a particular correlation based on a different initial condition will fail. Furthermore, the salinity correlation works with a monthly averaged salinity value, which is not a practical indicator of necessary ecosystem property. NPrdcn CPUE

10.0E+3

100

9.0E+3

90

8.0E+3

80

7.0E+3

70

6.0E+3

60

5.0E+3

50

4.0E+3

40

3.0E+3

30

2.0E+3

20

1.0E+3

10

000.0E+0

CPUE (Kg/Boat/Day)

New Production Rates (Kgd-1)

Maputo Bay: Hindcast Run 1996 - 2004 Modelled New Production and Observed Prawn CPUE

0 1996

1997

1998

1999

2000

2001

2002

2003

2004

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Date

Figure 39 – Plot showing the comparative variability over nine years between observed shrimp fishery CPUE and time – space averaged modelled mangrove ecosystem production during the high flow season (Jan – Apr). The modelled trends are remarkably close to the observed ones except in 1999 (orange bar) when the model failed to predict the required high production due to an underestimate of the modelled flow of the Maputo River. 2004 (blue bar) was the year in which we projected the new production flux value before we knew the result of the CPUE survey. The model correctly predicted the increased CPUE for 2004.

61 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Chapter 4 Decision support and management implications 4.1

A Decision Support Tool for catchment to coast linkages

Chris Sprengers, Marcel Marchand Based on the increased understanding of the system behaviour and interrelations, a computational framework has been prepared with which various kinds of future developments in the Incomati river catchment and Maputo Bay environment can be explored. The framework consists of a user-friendly graphical interface and makes use of a library of model runs and on-demand calculations. The user can define different socio-economic and climate change scenarios as well as a selection of measures. The Catchment2Coast Decision Support Tool is especially useful for river and coastal management analysts, but can also be used for training of students at BSc and MSc level.

4.1.1

Architecture of the Decision Support Tool

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

The Decision Support Tool (DST) can be regarded as an innovative kind of knowledge carrier (Figure 41). Its interactive feature is appropriate for the highly integrated approach of the river and coastal relationship, including the linkage between the natural science disciplines with the economic domain. The design of the C2C DST is based on three major steps that have to be taken to streamline and support the process of decision making (Figure 40): 1. Compose a case (Case Composer Component) 2. Make calculations (River Catchment, Maputo Bay Flow & Quality, Maputo Bay Ecology and Resource Economics) 3. Analyze the results (Maps, graphs and tables). These three steps form one of the main components of the working environment of the C2C-DST: the Workflow component. Compose

1

Compute

2

Analyze

3

Figure 40 – The three steps of the Decision Support Tool.

62 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Figure 41 – Example screen of the Catchment2Coast Decision Support Tool. Step 1: compose a case The design of the framework is based on the concept of cases, scenarios, measures and strategies. A case is a combination of a selected scenario and a selected set of measures (a strategy). A scenario is a combination of: x autonomous developments on subjects in the study area under consideration, such as land use, population growth etc. x autonomous developments on the boundary conditions for the study area under consideration, such as change in rainfall, sea level rise, river discharges, etc. A measure consists of proposed interventions on subjects in the study area under consideration, such as land use (area), education (society), infrastructure, etc. A strategy is a combination of measures for the study area under consideration in order to enforce a certain development, such as towards nature, towards industrial development, etc. A list of scenario options for the Maputo Bay application is given in Table 9, whereas Table 10 gives the pre-defined measures.

63 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Table 9 – List of scenario options in the C2C-DST ID 1 1.1

Name Time horizon Selection of time series / event

Description

2 2.1

Climate change Hydraulic conditions

2.2

Local rainfall in estuary

3 3.1

Rural & Urban development Population growth 2 SWAT calculations

3.2

Groundwater abstraction

3.3

Water pollution from Maputo

4 4.1

Economic & environmental conditions Discount rate 3 options (%/year) Inflation rate 3 options (%/year) Taxation rate 3 options (%/year) License fee 3 options (%/year) Shrimp price 3 market price options for shrimps ($/kg)

Remarks

either a single event (one year) or a time horizon can be chosen (several years)

6 flood events, wind / temperature for time series Input of percentage / year

Selection of one of the options

Selection of one of the options overrides selection at option 2.1

2 scenario options GWmodel Input of percentage / year

Step 2: Compute impacts on water system and other components In order to compute impacts on the water system and other components, different types of tools are used. Given that some of the components require huge simulation times, two approaches have been adopted (Table 11). The dynamic models with simulation times greater than about 5 minutes (DELFT3D-FLOW), or models which require complex user-interaction (SWAT), are not integrated into the framework. Instead a selection of simulation results of completed model runs are used. The models (E2K, SHRIMP, ECONOMIX) with small simulation times (< 5 minutes) are embedded in the framework.

64 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Table 10 – List of measures included in the C2C-DST ID 5 5.1 5.2

Name Description River basin management Dam operation 2 options: No measures, Low flow optimalization Land use 2 SWAT calculations: forestry and sugar cane

6 6.1

Coastal management Mangrove management

6.2

Fisheries

7 7.1

Economic conditions Number of boats

7.2

Closed period

Remarks Default SWAT basecase is used Selection of one of the options overrides selections at options 2.1 and 3.1

3 options (indicative): No change, Current trend (-5 % / year), Increase area (+5 % / year) 3 options (indicative): No change, Decrease of fishing pressure (-10%) , Increase fishing capacity (+10%) 3 options for selecting nr. of boats (18, 20 ,22) Select dec-jan-feb, janfeb, jan-feb-mar for closing fishing period

Table 11 – Approach to incorporating different models in the C2C-DST Models SWAT, D3D-FLOW

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

E2K, SHRIMP, ECONOMIX

Approach Selection predefined ‘off-line’ simulation results from ‘upstream’ model Setting of input data for the ‘on-line’ model

The impact models of the framework cover the impacts on fishery and shrimp catches, which are computed depending on the calculated salinity levels. Overall socioeconomic impacts are computed using decision rules, based on scenarios with respect to developments in population growth, increase of land use for agriculture, industry and nature, reduction of loads etc. The number and kind of other impacts can also be adjusted or extended.

Step 3: Analyze and present results. The analysis and presentation phase is carried out in the Framework using generic presentation components such as a graphical viewer, a table viewer and a map viewer. All view options enable the user to export data for reporting and further analysis. The computed impacts of all cases are stored in a MsAccess database. The tablepresentation tool reads the data and presents the results well-ordered in a table of effects. The tables can be exported for use in MsExcel to prepare reports for decision makers.

65 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

4.2 Implications of C2C science for Integrated Water Resource Management The main contribution of this study to IWRM is in providing an understanding of the linkages between river and coastal systems, specifically in southern Mozambique but possibly also for the entire tropical eastern seaboard of Africa. Coastal ecosystems are an integral part of their associated river basins, and ecosystem services provided by the coastal domain are shaped in space and time by the dynamics of the river system. Perhaps of particular interest is the need for improved coordination of management actions between river and coastal resource managers due to the potential negative feedback that arises from the coupled nature of the systems.

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

This research programme helped elucidate the nature of the coupling between river and coastal systems by resolving: x the link between the seasonal dynamics of river flow and the response of coastal ecosystem productivity x the link between ecosystem export production that supports the biodiversity and fisheries (shrimp P. indicus) and the health of mature, non-degraded mangrove systems x the extent to which ecosystem services that support biodiversity in the coastal domain depend on the integrated nature of freshwater and estuarine mangrove wetlands. The salinity dependence of the nitrogen fixation in mangrove systems provides the required linkage that helps understand how river resource management decisions related to water flow impact on coastal economic resources and biodiversity. Mangrove primary production needs to provide for both the nursery function of the mangrove systems themselves, as well as an excess supply through outwelling that maintains the production needs of nearshore systems such as rocky and coral reefs. The implication that this has for IWRM is that river flows need to have a minimum threshold such that the salinities in the mangrove systems are kept below 20psu for the four months (Jan – Apr) when the nursery and coastal production needs peak. The retention nature of the coastal domains, Maputo Bay in this case, helps to reduce the flow requirements when the salinity threshold is achieved. However, salinity control alone is not enough to ensure a responsive linkage. Nitrogen fixation appears to be mainly associated with Avicennia tree root systems. The trees compete with the ecosystem for the nitrogen flux associated with the fixation pathway. For this reason mature trees and forests provide a greater flux of nutrients to support export production because they have lower growth needs. The management implications for IWRM are that the exploitation of mature mangrove trees reduces the nitrogen flux available to support the nursery function, and the only way to compensate for it is to increase the low salinity period beyond the 4-month length. This can only be achieved by increasing the period of elevated river flow. In this way a feedback exists between the management of coastal resources (mangrove forests) and river water allocation strategies. Regenerating mangrove forests does not alleviate the

66 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

problem in less than a decade because growing trees have a high nitrogen requirement themselves. This calls for the integration of coastal resource management into IWRM of the coupled river systems. Nitrogen fixation is an adaptation by the coastal domain to the nutrient retention function of the freshwater wetlands in the floodplain. The latter act as a buffer to prevent eutrophication problems in coastal systems. Land use pressures mainly driven by agri-industry threaten the integrity of floodplain wetlands and can alter the nutrient balance that sustains the ecosystem services. Increased nutrients and sediments fluxes that would be expected to be linked to a reduction of the wetlands would alter the existing functional biodiversity in the mangroves (eutrophication) and coral reefs (light reduction due to turbidity). Both are key to two economic mainstays of non industrialised economies: fisheries and tourism. These scientific conclusions underline the need for careful management and conservation of the mangrove areas along Maputo Bay and elsewhere in tropical systems where coastal productivity depends on river inputs. Furthermore, it provides a sound basis for the setting of environmental flow requirements for the Maputo, Incomati and Umbeluzi Rivers. Evidently, a seasonal high river flow with the right timing is crucial for a large and healthy shrimp stock in the bay. Upstream land and water developments can have significant impact on these stocks and therefore need to be subject to an impact assessment. The Catchment2Coast Decision Support Tool (see previous section) can be used for this assessment or adapted for other comparable systems.

4.3 Recommendations to Strengthen Ecosystem Resilience in the coupled Incomati and Maputo – Maputo Bay systems

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

The following recommendations are specific to the Maputo Bay integrated system, but the principles are applicable to most comparable systems along the eastern African tropical coast.

ƒ

Harvesting of mature mangrove trees (particularly Avicennia) should be suspended immediately and carefully monitored.

ƒ

Reserves for process diversity in the mangrove ecosystems should be declared. This means that remaining pristine systems should be protected.

ƒ

An initial minimum wet season (Jan – Apr) water flux of 500 Mm3 for the Incomati River and 250 Mm3 for the Maputo River should be implemented to sustain coastal ecosystem production and associated fisheries. These fluxes could be refined with better observational programmes.

ƒ

Coordinate the operation of the four largest dams in the Incomati (> 80% capacity) and Maputo basins to achieve the magnitude and timing of these flow requirements.

67 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

The management and water allocation plans in the Incomati and Maputo River basins should be better coordinated in order to increase resilience of the system to natural fluctuations in runoff.

ƒ

The conservation of the Xinavane and Maputo wetlands should become a very high priority in the IWRM plans.

ƒ

Annual prawn and fish biomass estimates should be initiated as a basis to test the effectiveness of strengthened IWRM decisions.

ƒ

The data suggest that a January experimental CPUE can be used to forecast the seasonal average fisheries yield.

ƒ

Future plans should focus on the main sources of uncertainty highlighted by this study: - Maputo River: good flow data is essential; - New production hypothesis: the N-Fixation – Salinity relationship set out in this study needs to be validated with independent methods and extended to test it on other systems; - The relation between N-Fixation and the species and age composition of the mangrove forest needs to be studied in order to provide guidance on tree harvesting; - The mangrove food web and fisheries resource competition for the productivity needs to be better understood using ecosystem models and improved data; - The salinity response in the bay and mangrove domains to changes in total basin flow needs to be better understood in order to optimize water flow requirements.

ƒ

Communication to all stakeholders is essential in order to shorten the period taken for the scientific outputs to be taken up as part of the political nature of water resource allocation.

ƒ

The shortage of science graduates and technical expertise of managers remains as possibly the most serious obstacle towards the development of a responsive IWRM strategy in transboundary water systems. Therefore education, training and capacity building should be given maximum attention in any future management-, investment- and international cooperation programme.

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

ƒ

68 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Chapter 5 Conclusions, lessons learned and legacy 5.1

Conclusions

After three years of data collection and analysis, modelling and interpretation a large body of new knowledge has been acquired. This knowledge has given valuable insight into the dynamical characteristics that explain the functioning of bay hydrodynamics in relation to river inflow, in the variability of coastal ecosystem production, seasonal and interannual shrimp species yields and the valuable role of mangrove ecosystem services. This study took a different approach to observing and understanding mangrove production. Rather than measure total production, most of which is recycled, we focussed on new production sources and fluxes, which are essential to balance the export fluxes when juvenile life stages recruit to the estuarine – bay systems.

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

The overarching hypothesis across all process domains in the Catchment2Coast Project was that the key linkages between river fluxes and coastal responses occur at the event scale and not at the aggregated seasonal or annual scale. This hypothesis was not supported by the study’s final results because coastal ecosystem production is the result of a cumulative seasonal scale river discharge that maintains the salinity in the mangroves below the threshold required for nitrogen fixation to occur. Although the river inputs occur through flood flow events on a time scale of days, the retention character of the estuarine – bay system buffered the response of lowered salinity driven by these events. The importance of the magnitude and timing of the event scale lies in its interaction with the tidal cycle and the impact on stratification and flushing rates. Thus seasonal scale persistence of lowered salinity at the mangroves, rather than large scale flood events, modulates the interannual ecosystem production. The second hypothesis was the hydrodynamic control hypothesis, where the impact of river inputs on shrimp production is through the stratification and circulation characteristics driven by density gradients that develop as a result of the mixing of river and sea water. This was also called the ROFI (region of freshwater influence) hypothesis. The hypothesis was rejected on the basis of its initial formulation because there was no evidence that recruitment of early life stages depended on freshwater inflows. Hydrodynamic control of coastal production emerged as a key factor through the retention character of the estuarine – bay circulation, which shifted the dependence of the system from regular flood events to less frequent freshwater inputs in sustaining low salinities in late summer.

69 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

The third hypothesis, that coastal production depended on river-derived nutrients, was also rejected because very little dissolved nitrogen was supplied by river floods. Rather, it was the nitrogen fixation in the mangrove systems that were hypothesized as the most likely source of new nitrogen supporting export fluxes from the nursery systems. The importance of healthy mature mangrove forest for the shrimp productivity was supported by several observations. For instance, field sampling of juvenile shrimp showed that shrimps are considerably more abundant in non-degraded mangroves than in degraded mangroves and estuarine channel. Also a strong dependency was found of the biogeochemical fluxes in the mangrove area on changes in salinity driven by either the river flow or local rainfall and groundwater flows. The impact of the wet season is particularly pronounced in the case of Avicennia zones, for which shrimp have a marked preference. Bearing the importance for the coastal productivity in mind, the loss of mature mangrove forest in the Maputo Bay area is of serious concern. A survey of the condition of the mangroves of Maputo Bay was performed using remote sensing. A comparison between the years 1991 and 2003 revealed that 16.8% (204.3 ha) of the semi-intact mangrove area was degraded during this period, giving a rate of deforestation of 17 ha per year.

5.2

The Legacy

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

The Catchment2Coast project has a regional as well as a yet to be exploited continental-scale legacy. The principal regional legacy is that it forms a consolidated research knowledge base on the importance of downstream water resources from the cross-border rivers between South Africa, Swaziland and Mozambique. Thus the prime end-users are the members of the Tripartite Agreement, national government agencies in the respective countries, regional water managers and commercial resource stakeholders. However, the possibility that this work may be used in the even larger coupled river – coastal systems along the tropical eastern seaboard of Africa remains an aspect of potentially unlocked value. A wide range of regional and international agencies and networks have been informed and supplied with the Catchment2Coast project results and its products. Over 16 students from Mozambique, South Africa, Portugal and the Netherlands were able to partially or completely fulfil their MSc/Licentiate/PhD thesis under the Catchment2Coast project. Furthermore, two stakeholder workshops were held: the first at the Royal Swazi Convention Centre in Swaziland on 22 and 23 April 2004 and the second on 16 and 17 November 2005 in Maputo, Mozambique. A Stakeholder Liaison Group was established, to ensure the continued communication between the stakeholder community in the region and the Catchment2Coast project. This Liaison Group consisted of Mr Dumisane Mndzebele of Swaziland, Ms Graca Massicame of Mozambique and Mr Niel van Wyk of South Africa.

70 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

This research project has already led to other initiatives, some of which implement ideas developed in C2C, while others further advance the thinking and technology. These include: x

x

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

x

Ecology and Economics of Ecosystems (ECO3) – a three-year programme developed by CSIR of South Africa to engage the complexity and uncertainties of linked or coupled human – ecological systems SPEAR: INCO funded-project - developed by IMAR of Portugal to extend some of the ideas about coupling ecosystem and economics models with feedback capacity, focussing on sustainable aquaculture in China SMILE – a project developed and led by IMAR of Portugal with the participation of partner 4, using some of the experience and ideas developed in C2C to set up an ecosystem model for the Loughs of Northern Ireland to support the sustainable use of coastal resources and biodiversity needs.

71 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

References Acworth, R.I., Soriano, R. and I.L. Turner 2000. The vertical head distribution in a coastal sand-dune aquifer. In Groundwater: Past Achievement and Future Challenges. Sililo et al. (Eds) Balkema Press. Buzzelli, C.P., Wetzel, R.L. and M.B. Meyers 1999. A linked physical and biological framework to assess biogeochemical dynamics in a shallow estuarine ecosystem. Estuarine, Coastal and Shelf Science 49: 829-851. Da Silva, R. and P. M. S. Monteiro 2001. The dynamic interaction of the coastal domain of Maputo Bay with its main river / groundwater catchments and its implications for the sustainability of the artisanal and semi-industrial prawn fishery. Unpublished draft proposal. Dengo, A. and A. Govender 1998. The shrimp resource of Maputo Bay semi-industrial and artisanal fisheries. IIP, Mozambique and Oceanographic Research Institute, South Africa. Hannesson, R. 1983. Optimal harvesting of ecologically interdependent fish species. J. Environmental Econ. Manag. 10: 329–345. Heyink Leestemaker, J. 2000. The domino effect, a downstream perspective in water management in Southern Africa. In Water for peace in the Middle East and Southern Africa, Green Cross International. Geneva. Hoguane, A.M. 1998. Information on Eastern African Sea Level, Mozambique. Intergovernmental Oceanographic Commission of UNESCO. LOICZ 1995. Coastal zone resources assessment guidelines. Loicz Reports & Studies No. 4. LOICZ 1996. LOICZ biogeochemical modelling guidelines. Loicz Reports & Studies No. 5.

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

LOICZ 1997. Towards integrated modelling and analysis in coastal zones: principles and practices. Loicz Reports & Studies No. 11. Marchand, M. 2003. Problem analysis. Internal document Catchment2Coast. October 2003. Monteiro, P.M.S. and S. Mathews 2003. Catchment2Coast: Making the link between coastal resource variability and river input. South African Journal of Science, 99: 299-301. Monteiro, P.M.S. and J.L. Largier 1999. Thermal stratification in Saldanha Bay, South Africa and subtidal, density driven exchange with the coastal waters of the Benguela upwelling system. Estuarine, Coastal and Shelf Science. Nilsson, C., Reidy, C.A., Dynesius, M. and C. Revenga 2005. Fragmentation and flow regulation of the world’s large river systems. Science 308: 405-408. www.sciencemag.org Ravikumar, S., K. Kathiseran, S.Thadedus Maria Ignatiammal, M. Babu Selvam, S. Shanthy 2004. Nitrogen-fixing azotobacters from mangrove habitat and their

72 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

utility as marine biofertilizers. Journal of Experimental Marine Biology and Ecology, 312 (1): 5-17. Rabalais, N.N., R.E. Turner, D. Justic, Q. Dortch, W.J. Wiseman, and B.K. Gupta 1996. Nutrient changes in the Mississippi River and System Responses on the Adjacent Continental Shelf. Estuaries 19 (28): 386-407. Rönnbäck, P., A. Macia, G. Almqvist, L. Schultz and M. Troell 2002. Do penaeid shrimps have a preference for mangrove habitats? Distribution pattern analysis on Inhaca island, Mozambique. Estuarine, Coastal and Shelf Science 55 (3): 427-436. Salman, S.M.A. 2002. The Abuja Ministerial Declaration on Water – A milestone or just another statement? Water International 27 (3) [International Water Resources Association] Schulze,R.E., 1995. Hydrolgy and Agrohydrology. A text to accompany the ACRU 3.00 Agrohydrological modelling system. SA Water Research Commission. Report No. TT69/95. Simpson J.H. 1997. Physical processes in the ROFI regime. Journal of Marine Systems 12: 3-15. Smetacek, V., U. Bathmann, E-M. Nöthig and R. Scharek 1991. Coastal eutrophication: causes and consequences. In Ocean Margin Processes in Global Change, Mantoura, R.F.C., J-M. Martin and R. Wollast (Eds): 251-279. Syvitski, J.P.M., C.J. Vörösmarty, A.J. Kettner and P. Green 2005. Impact of humans on the flux of terrestrial sediment to the global coastal ocean. Science 308: 376380. www.sciencemag.org Tripartite Interim Agreement on Incomati and Maputo watercourses http://www.dwaf.gov.za/documents.asp

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Unesco 1988. Eutrophication in the Mediterranean Sea: receiving capacity and monitoring of long-term effects. Proc. Scientific Workshop, Bologna, Italy, 2-6 March 1987 Van Katwijk, M. M., N. F. Meier, R. van Loon, E. M. van Howe, W. B. J. T. Giesen and G. van der Velde 1993. Sabaki River sediment loading and coral stress: correlation between sediments and condition of the Malindi- Watamu reefs in Kenya (Indian Ocean). Marine Biology 117: 675-683. Vollenwieder, R.A. 1992. Coastal marine eutrophication: principles and control. In Marine Coastal Eutrophication. Vollenwieder, R. A., R. Marchetti, and R. Viviani (Eds). Elsevier, Amsterdam: 1-20.

73 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Annex A: Composition of the Catchment2Coast project team The project team of Catchment2Coast comprised modelers, field experimenters and researchers from 9 research institutes and universities from southern Africa and Europe. Valuable input was provided from end-users during several workshops. Scientific coordinator was Dr Pedro S. Monteiro of CSIR and the project management was in the hands of Marcel Marchand of Delft Hydraulics. The project was organized in 9 workpackages, each of them led by a scientist. Below the full list of participants in the project is presented.

Partner organizations of Catchment2Coast: CSIR Division of Water, Environment and Forestry technology (CSIR) (South Africa) Deltares / Delft Hydraulics (The Netherlands) Institute of Marine Research – Centro de Modelação Ecologica (IMAR) (Portugal) Instituto de Investigação das Pescas (IIP) (Mozambique) SPACRU, University of Natal (NATAL) (South Africa) University Eduardo Mondlane (UEM) (Mozambique) University of Cape town (UCT) (South Africa) University of Swaziland (UNISWA) (Swaziland) University of Wales, Bangor – School of Ocean Sciences (UWB) (United Kingdom)

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

List of main scientists Dr Pedro S. Monteiro Dr Colin Everson Dr Martin de Wit Dr. Stephen Luger Mr Rowan le Roux Ms Christine Colvin Mr Camaren Peter Mr Marcel Marchand Mr Chris Sprengers Mr Johannes Smits Dr Rui de Paula e Silva Prof JoƗo Gomes Ferreira Ms Ana Nobre Ms Andrea Franco Dr Mags Moodley Mr Vivek Naiken Mrs Marilyn Govender Prof Michael Savage Prof Geoff Brundrit Dr Jenny-Lisa Harcourt-Baldwin Mr Emlyn Balarin

CSIR CSIR (CSIR) (CSIR) CSIR CSIR CSIR Deltares Deltares Deltares IIP IMAR IMAR IMAR NATAL NATAL NATAL NATAL UCT UCT UCT

[email protected] [email protected]

[email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected]

74 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Prof Antonio Hoguane Dr Adriano Macia Mr Elonio Muiuane Prof Jonathan Matondo

[email protected] [email protected] [email protected] [email protected] [email protected]

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

Prof John Simpson Mr Joao Lencart

UEM UEM UEM UNISW A UWB UWB

75 Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,

Copyright © 2009. IOS Press, Incorporated. All rights reserved.

This page intentionally left blank

Catchment2Coast: a Systems Approach to Coupled River-Coastal Ecosystem Science and Management, IOS Press, Incorporated,