Geomorphology of Desert Dunes [2 ed.] 1108420885, 9781108420884

Sand dunes are a globally important depositional landform and sedimentary system. Their origins and dynamics are importa

221 46 45MB

English Pages 360 [361] Year 2023

Report DMCA / Copyright

DOWNLOAD PDF FILE

Recommend Papers

Geomorphology of Desert Dunes [2 ed.]
 1108420885, 9781108420884

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

G EO MO RP H O LO G Y OF D E S E RT D U N E S

Sand dunes are a globally important depositional landform and sedimentary system. Their origins and dynamics are important in understanding how deserts have evolved in response to climate change and changes in sand supply and mobility, and how they will continue to evolve in the future. This book provides a state-of-the-art review of the characteristics of desert dunes and their sediments, and explores their dynamics on timescales from days to millennia as they respond to changes in wind speed and direction, precipitation and sand supply. This extensively revised edition reflects the advances in our understanding of desert dunes, their dynamics and history; and covers recent developments including the luminescence dating revolution, ground penetrating radar and advances in numerical modeling. Also covering dunes on Mars and Titan, this authoritative reference is a musthave for researchers and graduate students working on desert dunes and aeolian geomorphology.

nicholas lancaster is a leading expert on desert sand dunes, and Emeritus Research Professor from the Desert Research Institute, USA. He has worked on desert dunes in Africa (Namib, Kalahari, northern and western Sahara), Arabia, Antarctica, and the western United States (Mojave and Sonoran Deserts). His research focuses on dune dynamics and morphology, the application of remote sensing, ground penetaring radar and optical dating, and the impacts of climate change on desert regions. He has won multiple awards including the Farouk El-Baz Award from the Quaternary Geology and Geomorphology Division of the Geological Society of America (2001), the NSHE Regent's Researcher Award (2007), and the Liu Tungshen Medal from the International Quaternary Association (INQUA, 2019).

Published online by Cambridge University Press

‘Studying desert dunes continues to be essential to our understanding of the geomorphology and climate of Earth and other worlds in the solar system. The update to this anchoring text seamlessly merges the significant advancements in aeolian science over the past few decades with core concepts from decades prior. This book will serve as the go-to source for any scientist needing a reference for wind-blown sand dunes and as the textbook for training the next generation of aeolian scientists.’ Professor Ryan Ewing, Texas A&M University ‘Nick Lancaster’s fifty-plus years of field research on desert dunes and his keen insights on the ‘big-picture’ of dune formation and change make him the best qualified person to write the definitive book on the subject. This thoroughly revised and updated second edition is a must-read for anyone seeking to understand desert sand dunes.’ Jeff Lee, Department of Economics and Geography, Texas Tech University

Published online by Cambridge University Press

G E O M O R P H O L O G Y O F D E S E RT DUNES NICHOLAS LANCASTER Desert Research Institute

Published online by Cambridge University Press

Shaftesbury Road, Cambridge CB2 8EA, United Kingdom One Liberty Plaza, 20th Floor, New York, NY 10006, USA 477 Williamstown Road, Port Melbourne, VIC 3207, Australia 314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre, New Delhi – 110025, India 103 Penang Road, #05–06/07, Visioncrest Commercial, Singapore 238467 Cambridge University Press is part of Cambridge University Press & Assessment, a department of the University of Cambridge. We share the University’s mission to contribute to society through the pursuit of education, learning and research at the highest international levels of excellence. www.cambridge.org Information on this title: www.cambridge.org/9781108420884 DOI: 10.1017/9781108355568 © Nicholas Lancaster 2023 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press & Assessment. First published 2023 Printed in the United Kingdom by TJ Books Limited, Padstow Cornwall A catalogue record for this publication is available from the British Library. ISBN 978-1-108-42088-4 Hardback Cambridge University Press & Assessment has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

Published online by Cambridge University Press

Contents

Preface to the Second Edition Acknowledgments Part I Introduction and Fundamental Concepts 1

Desert Dune Systems 1.1 Dunes as Complex Systems 1.2 Dune Systems 1.3 The Concept of Sediment State 1.4 The Quaternary Legacy 1.5 Development of Modern Dune Studies

page xi xiii 1 3 3 5 7 10 10

Part II Dune Morphology and Sediments

13

2

15 15 15 18 21 21 30 36 41 41 44 44 46 47

Dune Morphology 2.1 Introduction 2.1.1 Aeolian Bedform Hierarchies

2.2 Dune Classifications 2.3 Major Dune Morphological Types 2.3.1 2.3.2 2.3.3 2.3.4 2.3.5 2.3.6 2.3.7 2.3.8 2.3.9

Crescentic Dunes Linear Dunes Star Dunes Dome Dunes Parabolic Dunes Nebkhas Sand Sheets and Zibars Topographically Controlled Dunes Source-Bordering Dunes

v

Published online by Cambridge University Press

vi

3

Contents

Dune Sediments 3.1 Mineral Composition 3.2 Grain Size and Sorting 3.2.1 Conceptual Models for Grain Size and Sorting 3.2.2 Grain Size and Sorting Characteristics of Different Dune Types

3.3 Grain Shape 3.4 Sand Color 3.4.1 Spatial Variations in Dune Color

3.5 Dune Sedimentary Structures 3.5.1 Sedimentary Structures of Major Dune Types

Part III Dune Processes and Dynamics 4

Sand Transport by Wind 4.1 Introduction 4.2 The Surface Wind 4.3 The Sand Transport System 4.3.1 4.3.2 4.3.3 4.3.4 4.3.5 4.3.6

Entrainment of Sand by the Wind Saltation Trajectories Modification of the Wind Profile by Saltating Grains Mass Flux of Sand Controls on Mass Flux Spatial and Temporal Changes in Sand Transport Rates

4.4 Wind Ripples 5

Airflow and Sand Transport on Dunes 5.1 Introduction 5.2 Airflow over Dunes 5.2.1 5.2.2 5.2.3 5.2.4

6

The Stoss or Windward Slope The Lee Side Lee Airflow on Linear Dunes Airflow over Star Dunes

Dune Dynamics 6.1 Erosion and Deposition Patterns on Dunes 6.1.1 Stoss (Windward) Slopes 6.1.2 Lee Face Deposition

6.2 Erosion and Deposition Patterns on Different Dune Types 6.2.1 Crescentic Dunes 6.2.2 Linear Dunes 6.2.3 Star Dunes

Published online by Cambridge University Press

49 49 51 52 54 61 63 65 66 69 85 87 87 88 90 91 92 94 96 98 101 101 107 107 109 110 113 116 121 122 122 122 124 128 128 129 134

Contents

6.3 Annual to Decadal Timescales of Dune Change 6.3.1 Rates and Temporal Patterns of Dune Migration and Extension

6.4 Long-Term Dune Dynamics 6.4.1 Numerical Modeling of Dune Dynamics 6.4.2 Dune Initiation 6.4.3 Dune Growth

Part IV Boundary Conditions for Dune Formation and Development 7

Controls of Dune Morphology 7.1 Introduction 7.2 Factors That Determine Dune Morphology 7.2.1 7.2.2 7.2.3 7.2.4

8

9

10

Wind Regime Sand Supply Influence of Vegetation Sediment Characteristics

vii 137 138 142 143 145 148 151 153 153 153 154 157 158 161

Controls of Dune Orientation 8.1 Gross Bedform-Normal Rule 8.2 Bed Instability and Fingering Modes of Dune Type and Orientation 8.3 Superimposed Dunes and Complex Patterns 8.4 Influence of Sediment Characteristics

163 163

Controls of Dune Size and Spacing 9.1 Dune Height and Spacing Relationships 9.2 Grain Size Effects 9.3 Airflow Effects 9.4 Controls of Dune Size 9.5 Dune Spacing as a Product of Self-Organization and Pattern Coarsening 9.6 Influence of Sediment Source and Dune Field Geometry

175 175 176 178 179

Response of Dune Systems to Changing Boundary Conditions 10.1 Dune Activity 10.2 Empirical Studies of Responses to Changing Boundary Conditions

186 186

10.2.1 Late Holocene and Historical Record of Dune Activity 10.2.2 Decadal Scale Changes

10.3 Climatic Indices of Dune Activity

Published online by Cambridge University Press

165 169 174

180 182

188 188 191 195

viii

Contents

10.4 Eco-geomorphic and Landscape-Scale Approaches 10.5 Reorientation of Dunes in Response to Changing Wind Regime

196 200

Part V Sand Seas and Dune Fields

203

11

11.4.1 Sources of Sand 11.4.2 Sand Thickness and Volume

205 205 205 212 212 216 218 220 221 221

12

Dune Patterns in Sand Seas and Dune Fields 12.1 Patterns of Dune Morphologic Types 12.2 Analysis of Dune Patterns 12.3 Controls of Dune Patterns 12.4 Dune Pattern Development

226 226 231 235 236

13

Formation of Sand Seas and Dune Fields 13.1 Conceptual Framework 13.2 Models for the Development of Sand Seas 13.3 Timescales for Sand Sea and Dune Field Development

240 240 241 247

Characteristics and Distribution of Sand Seas and Dune Fields 11.1 Introduction 11.2 Global and Regional Distribution of Sand Seas and Dune Fields 11.3 Boundary Conditions for Sand Seas and Dune Fields 11.3.1 Relations to Topography 11.3.2 Wind Regimes 11.3.3 Relations to Tectonics

11.4 Sediments of Sand Seas

Part VI Dune Systems in Time and Space

251

14

Ancient Dune Systems: The Rock Record 14.1 The Stratigraphic Record of Aeolian Deposits 14.2 Examples of Ancient Aeolian Sandstones 14.3 Reconstruction of Ancient Dunes and Their Wind Regimes 14.4 Preservation Potential of Modern Sand Seas and Dune Fields

253 253 257 259 260

15

Quaternary Paleodune Systems 15.1 Occurrence of Paleodune Systems 15.2 Chronology of Paleodune Systems 15.3 Interpretation of Luminescence Datasets 15.4 Paleoclimatic Inferences from Paleodune Systems

263 263 266 272 272

Published online by Cambridge University Press

Contents

16

Planetary Dune Systems 16.1 Mars

ix

16.2 Venus 16.3 Titan 16.4 Pluto

281 281 284 285 285 286 287

Part VII Conclusions

289

17

291 291 293

16.1.1 Active Dunes and Sand Fluxes 16.1.2 Martian Dune History

Review and Prospects 17.1 The Current State of Desert Dune Knowledge 17.2 Future Prospects

References Index Color plates can be found between pages 106 and 107.

Published online by Cambridge University Press

296 331

Published online by Cambridge University Press

Preface to the Second Edition

The original version of Geomorphology of Desert Dunes was published in 1995. The content reflected concepts and examples existing in the early 1990s. At this time, it was possible for an individual to comprehend the scope of the field and to be familiar with the key concepts and studies. This is no longer the case, as the number of articles published in the field of aeolian science has increased exponentially in the last three decades and the diversity and complexity of approaches has expanded considerably. The first edition was published prior to many important developments in the field, which include, but are not limited to, (1) the luminescence dating revolution, enabling a chronology of periods of aeolian accumulation and dune formation in many areas, as well as assessment of rates of dune accumulation and movement; (2) the advent of ground penetrating radar, facilitating imaging of the sedimentary structures of dunes; (3) advances in technology, facilitating more-detailed measurements of winds and sand transport on dunes, and their relationships to the formation and dynamics of desert dunes; (4) studies of the mineralogy and provenance of dune sand using geochemical, isotopic, and remote sensing approaches; and (5) numerical modeling of dune processes and dynamics. The insights gained by these new approaches have enabled greater understanding of the dynamics of desert dune systems on different timescales, but also have posed many new questions to be addressed by field studies and numerical modeling. In many cases, modeling of processes and forms has resulted in new insights but has also greatly exceeded the ability to test the models in the field. The geographical range and diversity of investigations has expanded dramatically, with important contributions from new areas, especially from Chinese scientists. Numerical modeling has engaged a new community of investigators from backgrounds in physics and numerical methods, in addition to traditional fields of geomorphology.

xi

https://doi.org/10.1017/9781108355568.001 Published online by Cambridge University Press

xii

Preface

This new edition has been extensively revised to reflect the advances in the understanding of desert dune processes, dynamics, and history that have taken place over the past three decades, building on the work discussed in the original edition. As before, the content and approach of this book reflect my experience in dune fields and sand seas in different desert regions. Organization of This Book This book is organized into seven main parts, each focusing on a specific aspect of the geomorphology of desert dunes. Part II discusses the characteristics of dunes and their morphology and sediments; Part III provides a review of sand transport processes as an introduction to a discussion of dune dynamics; Part IV focuses on the boundary conditions that determine dune morphology and the response of dune systems to changes in boundary conditions; Part V considers aeolian sand bodies (dune fields and sand seas) as the depositional sink for the dune system; Part VI discusses how dune systems have developed in time (including ancient aeolian sandstones) and space (planetary dune systems); Part VII is a review of the field and offers some ideas about the future of desert dune studies.

https://doi.org/10.1017/9781108355568.001 Published online by Cambridge University Press

Acknowledgments

It has been my privilege to be able to devote my career to the study of desert dunes, to be able to travel to and live in different parts of the world, and to truly experience the beauty and magic of dune landscapes. Many people have provided scientific collaboration, support, and friendship along the way. They include Mary Seely, the late Ron Greeley, Gary Kocurek, Bill Nickling, Jack Gillies, Charlie Bristow, Dan Muhs, Ashok Singhvi, Paul Hesse, Xiaoping Yang, the late Ken Glennie, and many others. It has been a privilege to know you and to learn from your knowledge and insights. In production of this book, I want to acknowledge Sarah Lambert, my editor, and Cambridge University Press for their support and patience; and my wife, Maria Tinangon, for putting up with me while I devoted time to writing when I could have been doing other things.

xiii

Published online by Cambridge University Press

Published online by Cambridge University Press

Part I Introduction and Fundamental Concepts

Published online by Cambridge University Press

Published online by Cambridge University Press

1 Desert Dune Systems

The majority of desert dunes (>95%) occur in accumulations known as sand seas or ergs (Wilson, 1973) that comprise areas of dunes of varying morphological types and sizes, as well as areas of sand sheets. Smaller dune areas are known informally as dune fields. Sand seas and dune fields represent a globally important depositional landform and sedimentary system and cover an area of 7  106 km2 with a volume estimate of 4:88  104 km3 (Lancaster, 2022). The greatest concentrations of desert sand seas and dune fields occur at midlatitudes (3550° N), especially in the arid regions of central Asia and China, and in low-latitude desert areas of Africa, Arabia, and Australia (1530° N and S) (Fig. 1.1). 1.1 Dunes as Complex Systems Desert sand dunes form part of self-organized complex systems of aeolian bedforms, which comprise a hierarchy of (1) wind ripples, (2) individual simple dunes or superimposed dunes on mega dunes, (3) megadunes (also called draa or compound and complex dunes) – characterized by superimposition of simple or elemental dunes on larger forms, and (4) dune fields and sand seas (Fig. 1.2). The selforganized patterns of dunes develop over time and in space as the nonlinear response of sand surfaces to the wind regime (especially its directional variability) and the supply of sand (Kocurek and Ewing, 2005; Werner, 1995). As the spatial scale increases, the number of variables decreases, and smaller-scale processes (e.g. the saltation of sand grains) are subordinated to and decoupled from emergent, larger-scale behavior, such as interactions between dunes. This reduction of variables with increasing scale facilitates understanding of key aspects of dune-field patterns, such as dune spacing and dune reorientation with changes in wind regime (Kocurek and Ewing, 2005) as well as the assessment of responses to external changes on differing spatial and temporal scales. It also provides a framework in which to transfer knowledge of dune systems on Earth to other planetary bodies, such as Mars and Saturn’s moon Titan (Ewing et al., 2015b). 3

https://doi.org/10.1017/9781108355568.002 Published online by Cambridge University Press

4

Introduction and Fundamental Concepts

Figure 1.1 Global map of sand seas and dune fields.

Dune fields

Spatial Scale (m)

10,000

Sand Seas

Mega dunes 100 Dunes 1

Ripples 0.01

0.01

0.1

10

1000

Time scale (years)

Figure 1.2 Temporal and spatial scales of dune systems.

https://doi.org/10.1017/9781108355568.002 Published online by Cambridge University Press

100,000

Desert Dune Systems

5

1.2 Dune Systems Sand seas and dune fields are dynamic sedimentary bodies that form part of local to regional-scale sand transport systems in which sand is moved by the wind from source zones to depositional sinks via transport pathways (Fig. 1.3). Dune formation is therefore dependent on a supply of sand-sized sediment, wind energy to entrain and transport that material, and topographic and/or climatic conditions that promote deposition of that sediment. Sources and sinks for sediment are linked by a cascade of energy and materials that can be viewed in terms of sediment inputs and outputs, transfers, and storages that determine the sediment state of the system. Following Kocurek and Havholm (1993) dune systems (Fig. 1.4) can be divided into (1) dry systems, in which the water table and its capillary fringe are sufficiently far below the depositional surface that they have no effect on dune migration, sediment transport, and deposition; (2) wet systems, where the water table and its capillary fringe is at or near the depositional surface. Dune dynamics and sediment transport and deposition are controlled both by the wind regime and by the moisture content of the substrate; and (3) stabilized systems, in which vegetation or some other factor periodically or continually stabilizes the substrate while the system remains active overall. Good examples of dry

O

ourc e

s

Devil’s Playground Transport pathway

Kelso Dunes Sediment sink

Seco nda ry

alluv

ial s

Mojave River Sink Primary Source Area

10 km

Figure 1.3 Example of a sand transport system – Kelso Dunes, California. Sand is transported west to east from the Mojave River fan-delta to the dunes. Alluvial fans adjacent to the dunes provide secondary sand sources. See Muhs and colleagues (2016).

https://doi.org/10.1017/9781108355568.002 Published online by Cambridge University Press

6

Introduction and Fundamental Concepts

Dry system Namib Sand Sea Algodones Dunes

Accumulation space

Accumulation

Wet system White Sands, Liwa

Water table Accumulation

Accumulation space defined by water table

Stabilizing system Kalahari, Negev-Sinai

Accumulation

Accumulation space defined by vegetation

Figure 1.4 Characteristics of aeolian systems in relation to water tables and stabilization by vegetation (after Kocurek, 1998).

aeolian systems are the majority of the Namib Sand Sea and the Algodones Dunefield (California). Wet aeolian systems include the coastal parts of the Namib Sand Sea, White Sands dune field (New Mexico, USA), as well as the Liwa and adjacent areas of the Rub’ al Khali in the United Arab Emirates and adjacent areas of Saudi Arabia. Stabilized systems include the Thar Desert sand

https://doi.org/10.1017/9781108355568.002 Published online by Cambridge University Press

Desert Dune Systems

7

sea in India, much of the southwestern Kalahari and Australia, and parts of the Negev–Sinai sand sea (Egypt and Israel). These designations can change over time, as water tables rise and fall and vegetation expands and contracts in relation to climate and sea level change; they can also vary spatially – parts of a sand sea may behave as a wet or stabilized system, while other areas are dry. Thus, many eastern parts of the Namib Sand Sea represent a stabilized dune system; coastal areas of the same sand sea near Walvis Bay are a wet system; the majority is a dry system. 1.3 The Concept of Sediment State The concept of sediment state as articulated by Kocurek and Lancaster (1999) provides a framework in which to assess the state of dune systems and their response to changes in external forcing (e.g. climate change) as well as the internal dynamics of the system. In this model, development of dune systems is governed by the supply of sediment of a size suitable for transport by the wind, the existence of wind energy to erode and transport this material (erosivity or mobility), and the susceptibility of the surface to entrainment of material by wind (erodibility or sediment availability). The interactions between these variables in space and time determine the overall state of the aeolian system, which may be supply limited, availability limited, or mobility limited (Kocurek and Lancaster, 1999). In this conceptual model, sediment supply is defined as the emplacement of sediment of a size suitable for wind transport that serves as a source of material for the system. Sediment availability is the susceptibility of a sediment surface to entrainment of material by wind, as determined by factors such as vegetation cover and soil moisture content. Sediment mobility (or potential transport rate) is determined by the magnitude and frequency of winds capable of transporting sediment. The actual sand transport rate (Qa) is determined by sediment supply and availability and may be less than the potential rate (Qp) in proportion to the ratio Qa/Qp, The relations between sediment supply, availability, and mobility define the state of the sand transport system and its variation through time (Fig. 1.5), including its response to external forcing factors such as climate change. The overall state of the system may be described as follows: (1) Supply limited, in which Qa≪Qp and the system is starved of sediment, as in many central and northern sand Saharan sand seas, some of which are apparently eroding in present conditions (Mainguet, 1984). Variations in sediment supply (e.g. as a result of sea level changes, and changes in river sediment loads) appear to be most important where dunes are close to their source, as in

https://doi.org/10.1017/9781108355568.002 Published online by Cambridge University Press

8

Introduction and Fundamental Concepts Availability Limited

Transport Limited

Transport and Availability Limited Contemporaneous Input

CIAL

CITL

CITAL

LIAL

LITL

LITAL

Lagged Input

SAL

STL

STAL

Stored Sediment

Figure 1.5 Matrix of possible states of a sand transport system. Based on Kocurek and Lancaster (1999).

the Mojave desert (Lancaster and Tchakerian, 2003), the Takla Makan, and in some parts of Arabia (Garzanti et al., 2013; Rittner et al., 2016). (2) Availability limited, in which Qa < Qp. In many areas, for example, southwestern Kalahari, Great Plains of the USA, northeastern China, and Australia, sediment availability is determined by vegetation cover. Changes in sediment availability as a result of variations in rainfall and vegetation cover have played a dominant role in episodic development of dunes in the southwestern Kalahari and Australia (Telfer and Hesse, 2013; Telfer and Thomas, 2007), in addition to determining periods of dune activity in many semiarid regions, such as the High Plains of North America (Halfen and Johnson, 2013) and northeastern China (Lu et al., 2005; Yang et al., 2011). (3) Transport limited, in which the actual rate of sand transport (Qa) equals the potential rate (Qp) and the system is limited only by the capacity of the wind to move sediment from source zones, for example, the Namib Sand Sea (Lancaster, 1989b). The variations over time in sediment supply, availability, and mobility (transport capacity) can be used to evaluate the long-term dynamics of a dune system as a function of its sediment state (Fig. 1.6). Sediment supply can provide contemporaneous input to the system (CI) or be stored (S), if it is limited by availability (e.g. vegetation, high water table) and/or the transport capacity of the wind (e.g. low wind energy). Stored sediment can released later to provide lagged input (LI), if availability and/or transport capacity increase. In time, the supply of sediment may be exhausted, in which case the system becomes supply-limited. If availability remains high, then Qp≫Qa and the system will become net erosional.

https://doi.org/10.1017/9781108355568.002 Published online by Cambridge University Press

https://doi.org/10.1017/9781108355568.002 Published online by Cambridge University Press

Stored sediment (Transport limited)

Sediment supply

Availability

Mobility

Sediment input

Availability Limited

Transport Limited

Availability Limited

Time

Sediment input

Stored sediment (Availability limited) Stored sediment (Transport and availability limited)

(b)

Stored sediment (Availability limited)

Lagged input (Availability limited)

Lagged input (Transport limited)

Lagged input (Availability limited)

Sediment input

Time

(c)

Stored sediment (Availability limited)

Contemporaneous and lagged input (Availability limited)

Contemporaneous and lagged input (Transport limited)

Contemporaneous input (Availability limited)

Figure 1.6 Variation of sediment state through time shown through trends in sediment supply, availability, and mobility: (a) aeolian construction contemporaneous with the generation of the sediment supply; (b) aeolian construction separated in time from an earlier period when the sediment supply was generated; (c) aeolian construction sourced by both a contemporaneous sediment supply and erosion of previously stored sediment. Redrawn from Kocurek and Lancaster (1999).

Time

(a)

10

Introduction and Fundamental Concepts

1.4 The Quaternary Legacy Sand seas and many dune fields contain large volumes of sand, implying that their accumulation has taken place over many thousands or tens of thousands of years, during which Quaternary climatic and sea level changes have had a significant impact on sediment supply, availability, and mobility. Analysis of patterns of dune morphology and morphometry, as well as stratigraphic and dating studies of dune sediments in many areas indicates the importance of the legacy of Quaternary periods of dune construction, stability, and reworking (Thomas and Bailey, 2017; Thomas and Bailey, 2019). The Quaternary legacy varies in importance from apparently minor in young, rapidly developing dune fields, for example, Skeleton Coast dune fields of Namibia; to dominant, as in the vegetation-stabilized dune fields of the Kalahari, the Great Plains of North America, and Australia. In addition, there are many examples of dune systems that are currently inactive and stabilized by vegetation and biocrusts. Such systems are common on the margins of active desert dune areas and in semiarid regions. They are indicative of past environments in which dune formation was promoted by changes in sediment supply, availability, and mobility. These paleo-dune systems have been demonstrated to be a product of Quaternary climatic changes . 1.5 Development of Modern Dune Studies Modern dune studies aim to understand dune morphology and development in the context of changing boundary conditions of sediment supply, availability, and mobility. The paradigms that guide dune studies today are the product of decades of observations and experiments, aided by developments in technology and crossfertilization from other areas of geomorphology and geology, physics, and ecology (Livingstone et al., 2007; Thomas and Wiggs, 2008). Desert dunes have been the object of study for many years (see summary in Goudie [1999]). Early scientific descriptions of sand seas resulted from exploration of desert regions and assessment of their natural resources (Bagnold, 1951; CapotRey, 1947; Hack, 1941; Madigan, 1936). Widespread availability of aerial photographs in the 1950s and 1960s provided a broader perspective (Holm, 1960; Smith, 1965; Twidale, 1972) including recognition of relict sand seas in desert margin areas (Grove, 1958; Grove and Warren, 1968) and provided data for some significant conceptual advances by Ian Wilson (Wilson, 1971, 1972, 1973). Satellite images first became available in 1972 and have revolutionized studies of sand seas, first providing a synoptic global view (McKee, 1979), in which it was recognized that similar dune types occur in widely separated sand seas, enabling correlation of dune types with wind regimes (Fryberger and Dean, 1979; Wasson

https://doi.org/10.1017/9781108355568.002 Published online by Cambridge University Press

Desert Dune Systems

11

and Hyde, 1983a) and mapping of dune types (Breed et al., 1979). The spectral information in a variety of remote sensing platforms has enabled analysis of the distribution of different sand populations and tracing of sand sources, using visible near infrared (VNIR) (Blount et al., 1990; White et al., 2001; White et al., 2007) and thermal infrared (TIR) wavelength regions (Ramsey et al., 1999; Scheidt et al., 2011). In addition, digital elevation models derived from satellite and aircraft systems are providing valuable data sets for analysis of patterns of dune morphology and distribution (Al-Masrahy and Mountney, 2013; Bishop, 2013; Hugenholtz et al., 2012). Remote sensing analyses of dune sand composition are complemented by increasingly precise mineralogical, geochemical, and isotopic analyses that provide insights into sand sources and transport pathways (e.g. Muhs, 2004; Muhs, 2017; Muhs and Budahn, 2019; Muhs et al., 2017). The advent of luminescence dating in the 1980s (Singhvi et al., 1982) and its wider application in the 1990s (Stokes, 1992) have revolutionized knowledge of the chronology of periods of dune construction and reworking in many sand seas (Lancaster et al., 2016), although many issues of interpretation of the ages remain (e.g. Thomas and Bailey, 2017). A major development in dune studies has been the development of numerical modeling of dunes, facilitated by vastly increased computing power. The formation and evolution of elementary dunes as well as entire dune fields and dune interactions are now simulated with high spatiotemporal resolution with reducedcomplexity models such as cellular automata (e.g. Bishop et al., 2002; Narteau et al., 2009; Nishimori et al., 1998; Werner, 1995) as well as with physically based coupled airflow-sand transport models (e.g. Andreotti et al., 2002b; Hersen et al., 2004; Parteli et al., 2014; Schwämmle and Herrmann, 2004). This has led to significant conceptual advances in our understanding of how dunes and dune patterns develop, including their responses to different wind regimes (Courrech du Pont et al., 2014; Eastwood et al., 2011; Gao et al., 2015b; Reffet et al., 2010) and to climate change via changes in vegetation cover (Nield and Baas, 2008a).

https://doi.org/10.1017/9781108355568.002 Published online by Cambridge University Press

https://doi.org/10.1017/9781108355568.002 Published online by Cambridge University Press

Part II Dune Morphology and Sediments

Published online by Cambridge University Press

Published online by Cambridge University Press

2 Dune Morphology

2.1 Introduction Dunes are created by interactions between granular material (sand) and shearing flow (the atmospheric boundary layer). The resulting landforms are bedforms that are dynamically similar to those developed in subaqueous shearing flows (e.g. rivers, tidal currents). Desert dunes occur in a variety of morphologic types, each of which displays a range of sizes (height, width, and spacing). Aerial photographs and satellite images of sand seas show that most dune patterns are quite regular, and that very similar dune morphological types occur in widely separated sand seas. For example, vegetated linear dunes in Australia and the southwestern Kalahari are almost identical in morphology, as are compound crescentic dunes in Namibia, Arabia, and North America. This indicates that (1) the local response of sand surfaces to airflow is governed by generally applicable physical principles and (2) that there are general controls of dune size and spacing, as will be discussed in Part IV. 2.1.1 Aeolian Bedform Hierarchies In all sand seas and dune fields there is a hierarchical system of aeolian bedforms superimposed on one another (Lancaster, 1988a; Wilson, 1972). Similar bedform hierarchies also occur in subaqueous environments (Allen, 1968; Ashley, 1990; Jackson II, 1975). Wind ripples cover at least 80% of sand surfaces in all dune areas, whereas in many sand seas large dunes (compound or complex dunes, megadunes, draas) are characterized by the development of smaller dunes superimposed on their stoss and/or lee slopes. Three orders of aeolian bedforms can therefore be identified (Fig. 2.1), although only the first two occur in all sand seas: (1) wind ripples (spacing 0.1–1 m), (2) individual simple or elemental dunes or superimposed

15

https://doi.org/10.1017/9781108355568.003 Published online by Cambridge University Press

16

Dune Morphology and Sediments Superimposed Dunes

Primary Megadunes

40

(a)

20

0 40

(b)

% of Observations

20

0 40

(c)

20

0 20

(d)

0 0

20

40

60 80 100

200

400 600

1,000

2,000

4,000

Crest-to-Crest Spacing (m)

Figure 2.1 The hierarchy of superimposed dunes and megadunes (compound and complex dunes) in the Namib and Gran Desierto sand seas: (a) Namib compound crescentic dunes; (b) Namib linear dunes; (c) Namib Star Dunes; (d) Gran Desierto compound crescentic dunes. After Lancaster 1988a.

dunes on compound and complex dunes (spacing 50–500 m), and (3) compound and complex dunes, mega dunes, and draa (spacing > 500 m). A range of sizes of dunes occurs in all desert sand seas, yet satellite images and air photographs of desert sand seas show that most dune patterns are quite regular. Regular, ordered spacing of dunes of a limited range of sizes in a given area gives rise to the relationships between dune height, width, and spacing that have been documented by many workers for widely separated sand seas (Fig. 2.2), using field surveys (Lancaster, 1989b), remote sensing images (Thomas, 1988; Wasson and

https://doi.org/10.1017/9781108355568.003 Published online by Cambridge University Press

Dune Morphology

(a)

17

(b)

Crescentic Dunes

102

Dune Height (m)

Barchan Transverse Linear Star

101

Namib Sand Sea Skeleton Coast Gran Desierto

100 101

102

103

102

101 104 102

103

Dune Spacing (m)

Linear Dunes

Dune Height (m)

1,000 Namib Sand Sea (complex) Namib Sand Sea (compound) Southwest Kalahari Simpson Strzelecki Great Sandy Desert

100

10

1 100

1,000

10,000

Star Dunes 10

3

102

101 Namib Sand Sea Gran Desierto

100 100

1,000

10,000

Dune Spacing (m)

Figure 2.2 Relationships between dune height and spacing (wavelength). (a) field and map data; (b) global sample from digital elevation data (from Gunn et al., 2021, Table S2).

https://doi.org/10.1017/9781108355568.003 Published online by Cambridge University Press

104

18

Dune Morphology and Sediments

Hyde, 1983b), and, more recently, digital elevation datasets (Bullard et al., 2011; Gunn et al., 2021; Hugenholtz et al., 2012). This organization of dunes into regular patterns is a classic example of self-organization in a geomorphic system (Hallet, 1990; Kocurek et al., 2009; Werner, 1995; Werner and Kocurek, 1997). 2.2 Dune Classifications Many different classifications of dune types have been proposed (see Mainguet 1983 and 1984 for a list of references to different schemes). They fall into two groups: (1) those that imply some relationship of dune type to formative winds or sediment supply (morphodynamic classifications), and (2) those based on the external morphology of the dunes (morphological classifications). There are many morphodynamic classifications in which dunes are classified by their form and relation to formative winds, especially their alignment relative to the dominant or resultant (vector sum) sand transport direction (e.g. Hunter et al., 1983). Thus, dunes may be classified as transverse, longitudinal, or oblique (Fig. 2.3), yet studies of dune dynamics (see Chapter 6) show that different parts of the same dune may be simultaneously transverse, oblique, or longitudinal to the primary wind direction. Recently, Hu and colleagues (2021) have proposed a classification of dunes based on the concept of fingering and bed instability modes of dune development. Other workers (e.g. Mainguet, 1983; Mainguet, 1984) have attempted to order dunes by including aspects of their mobility and relation to sediment budgets and thus to distinguish between erosional types

–15°



15°

Longitudinal dunes

Oblique dunes

Oblique dunes

–75°

75°

Transverse dunes –90°

Transverse dunes 90°

Figure 2.3 Morphodynamic classification of dunes based on the relationship between dune trend and wind direction. After Hunter and colleagues (1983).

https://doi.org/10.1017/9781108355568.003 Published online by Cambridge University Press

Dune Morphology

19

Aeolian Dunes

Anchored

Free

Transverse

Linear

Star

Vegetated

Topographic

Transverse Barchan Dome Reversing

Linear Seif

Star Network

Parabolic Blowout Nebkha

Climbing Falling Lee Sand ramp

Figure 2.4 Livingstone and Warren classification of dunes (after Livingstone and Warren, 1996).

(parabolic dunes, sand ridges) and purely depositional forms (barchanoid dunes, transverse chains, linear dunes, and star dunes). In a similar approach, Livingstone and Warren (1996) proposed that dunes could be classified as “free” dunes, in which the form is determined solely by the interaction between sand and wind; and “anchored” dunes, where topography or vegetation determine dune morphology (Fig. 2.4). The morphological classification of McKee and his coworkers (McKee, 1979) groups dunes on the basis of their shape and number of slip faces into five major types: crescentic, linear, reversing, star, and parabolic (Fig. 2.5). In turn, three varieties of each dune type can occur: simple, compound, and complex. Simple or elemental dunes are the basic form of each dune type. Compound dunes are characterized by superimposition or juxtaposition of dunes of the same morphological type (e.g. superimposition of smaller crescentic dunes on the stoss side of large crescentic dunes). Complex dunes occur where dunes of two different types are superimposed or merged (e.g. crescentic dunes on the flanks of larger linear or star dunes, or linear dunes with star dune peaks). Compound or complex dunes are a common feature of many modern sand seas and comprise 46.6% of the dunes in sand seas examined by Fryberger and Goudie (1981). They are equivalent to the draa of Wilson (1972), and this term has been applied to all large dunes with superimposed bedforms (Kocurek, 1991). In addition to the major dune types just identified, many sand seas and dune fields contain areas of gently undulating to flat sand surfaces (sand sheets), and areas of low rolling dunes without slip faces (zibar). Interactions between vegetation and sand accumulations lead to the formation of shadow dunes (also called shrub-coppice dunes or nebkha). Where sand accumulates adjacent

https://doi.org/10.1017/9781108355568.003 Published online by Cambridge University Press

20

Dune Morphology and Sediments

(a)

(b)

(c)

(d)

(e)

(f)

Figure 2.5 Major dune types (after McKee, 1979a): (a) barchan; (b) crescentic ridges; (c) linear; (d) star; (e) reversing; (f) parabolic.

to topographic obstacles, echo dunes or climbing dunes occur on their windward side, and lee dunes or falling dunes on the lee side. In addition, topographically controlled accumulations comprising aeolian sand and alluvial and slope deposits or sand ramps may form adjacent to mountain fronts (Lancaster and Tchakerian, 1996). Most dune types can be accommodated in an expanded morphological classification of dunes (Fig. 2.6), based on sand supply and wind regime complexity, as modulated by vegetation and topography.

https://doi.org/10.1017/9781108355568.003 Published online by Cambridge University Press

Dune Morphology

21 Draas, Megadunes

Dunes

Wind regime complexity

Relative sediment supply CRESCENTIC Barchans

Compound Linear dunes

Complex Linear dunes

LINEAR Simple - Straight - Sinuous REVERSING STAR

SAND SHEETS

vegetation anchored

NEBKHAS

topographic influence

Compound crescentic dunes

Dome dunes

coarse sand

source bordering

Crescentic ridges

ZIBARS

BLOW OUT DUNES

PARABOLIC DUNES

LUNETTES

ECHO DUNES

FALLING DUNES

CLIMBING DUNES

SAND RAMPS

Figure 2.6 A morphological classification of desert dunes in relation to sand supply and wind regime complexity.

2.3 Major Dune Morphological Types 2.3.1 Crescentic Dunes Crescentic dunes form in unidirectional sand-moving wind regimes that exhibit a narrow range of directions with one dominant directional sector. Dune trends are approximately transverse to the vector sum or resultant direction of sand movement (Fig. 2.7). Barchans Barchans are migrating crescent-shaped dunes with two horns that face in the downwind direction, and a concave lee face (Fig. 2.8). Most barchans range in

https://doi.org/10.1017/9781108355568.003 Published online by Cambridge University Press

22

Dune Morphology and Sediments (a)

January

April

July

October

(b)

Annual D.P. = 111 v.u.

Spring (Mar.–Jun.) D.P. = 52 v.u.

Summer (Jul.–Sept.) D.P. = 17 v.u.

Winter (Oct.–Feb.) D.P. = 42 v.u.

20 v.u.

Figure 2.7 Wind regimes of crescentic dunes: (a) simple and compound crescentic dunes, coastal Namib Sand Sea; (b) compound crescentic dunes, Algodones Dunes, California (data from Havholm and Kocurek 1988, figure 3). Note that the units of measurement of wind speed are different in each case, so the vector magnitudes vary.

https://doi.org/10.1017/9781108355568.003 Published online by Cambridge University Press

Dune Morphology

23

height between 3 and 10 m. Very large barchans (megabarchans) with superimposed crescentic dunes have been described from a few localities in the western Sahara, Peru, and Arabia (Elbelrhiti et al., 2005; Goudie, 2020; Simons, 1956). Wind regimes in areas of barchans are characterized by the dominance of one directional sector, although seasonal changes in wind directions involving shortlived, high-intensity storms do occur (e.g. in the Namib, western Sahara) and may contribute to instability of the dunes (Elbelrhiti et al., 2005). Barchans are widely distributed globally and occur in 11 major provinces: Egypt and the Sudan, Namib, Peru, Somalia, Western Sahara, Iraq/Iran border, Baluchistan and Seistan, Eastern Arabia, Central Sahara, Mauritania, and the Hexi Corridor and neighboring parts of China and Mongolia (Goudie, 2020). They are rare in the deserts of North America, the Kalahari, and Australia. Barchans typically occur in two main domains: (1) the margins of sand seas and dune fields and (2) sand transport corridors linking sand source zones with depositional areas (Goudie, 2020). Most areas of barchans are characterized by a nonerodible substrate and an irregular spacing of dunes. Despite their wide distribution and the attention that has been given to this dune form, they comprise a very small proportion (3.0 for

https://doi.org/10.1017/9781108355568.003 Published online by Cambridge University Press

Dune Morphology

43

(a)

500 m (b)

Figure 2.21 Parabolic dunes: (a) White Sands, New Mexico; (b) Winnemucca dune field, Nevada. Image from Google Earth, earth.google.com/web/. For the color version, refer to the plate section.

elongate types (Pye and Tsoar, 1990). The degree of elongation is a function of the interactions between sand supply, vegetation deposition tolerance, and migration rate (Yan and Baas, 2017). Transitions from crescentic to parabolic dunes have been ascribed to migration of crescentic dunes into areas of greater phreatic (groundwater-dependent) vegetation cover (Reitz et al., 2010) and reduced sand transport rates. In semiarid areas of the Canadian Prairies, barchans developed during a period of climate change to cooler and drier conditions and increased sand transport reverted to parabolic dunes as climates warmed and became more mesic in the twentieth century (Wolfe and Hugenholtz, 2009). Wind regimes in areas of parabolic dunes are similar to those in areas of crescentic dunes – narrow to wide unimodal, with a range of wind energy and sand transport potential.

https://doi.org/10.1017/9781108355568.003 Published online by Cambridge University Press

44

Dune Morphology and Sediments

Figure 2.22 Nebkhas – coastal Qatar. For the color version, refer to the plate section.

2.3.6 Nebkhas Nebkhas (also known as coppice dunes) develop where sand is trapped by vegetation clumps (Goudie, 2022; Thomas and Tsoar, 1990). They are widely distributed in semiarid areas and also occur in hyperarid areas like the coastal Namib, Qatar (Fig. 2.22), and Kuwait where sand movement into areas of phreatophyte vegetation occurs (Khalaf et al., 1995; Lancaster, 1989b). Nebkhas in the Namib range up to 3.5 m high. In Mali, they are 0.35 to 0.72 m in height (Nickling, 1993) and have developed in less than 30 years following land use change. Nebkhas are widespread in areas of the Jornada Experimental Range and sand sheets in southern New Mexico where they have a relief of 1–2 m, anchored by mesquite shrubs (Gillies et al., 2014; Hall and Goble, 2015). 2.3.7 Sand Sheets and Zibars Low-relief sand deposits or sand sheets (Fig. 2.23) are common in many sand seas and occupy from as little as 5% of the area of the Namib Sand Sea to as much as 70% of the area of Gran Desierto (Lancaster, 1995). Fryberger and Goudie (1981) estimated that 38% of aeolian deposits are of this type.

https://doi.org/10.1017/9781108355568.003 Published online by Cambridge University Press

Dune Morphology

45

Figure 2.23 Sand sheet in the northwestern Gran Desierto sand sea.

Sand sheets develop in conditions unfavorable to dune formation (Kocurek and Nielson, 1986). These may include a high water table, periodic flooding, surface cementation, coarse-grained sands, and presence of a vegetation cover that acts to limit sand availability for dune formation. Sand sheets commonly occur on the upwind margins of sand seas and dune fields in conditions of net negative or neutral sediment budgets (erosion or bypassing). Some of the most extensive sand sheets known are in the eastern Sahara, where they cover more than 100,000 km2. Sand sheets in this area have a relief of less than 1 m and a total thickness of a few cm to as much as 10 m (Breed et al., 1987b; Maxwell and Haynes, 2001). The sand sheets consist of sandy plains with a granule-to-pebble lag deposit that forms the surface layer, together with areas of very low amplitude (10–30 cm), longwavelength (130–1,200 m) bedforms that form a giant chevron pattern on Landsat images (Maxwell and Haynes, 1989). In this area, sand sheets develop because of an abundance of coarse particles that inhibit all but localized dune development. Sand sheets cover an area of 1,000–1,500 km2 in the northwestern parts of the Gran Desierto (Fig. 2.23). They form a sparsely vegetated, flat to gently undulating surface with a maximum local relief of 1–5 m. The sand sheets are composite features consisting of successive generations of aeolian accumulations separated by stabilization surfaces (Lancaster, 1993). Sand sheets in the northwestern Gran

https://doi.org/10.1017/9781108355568.003 Published online by Cambridge University Press

46

Dune Morphology and Sediments

Desierto appear to have developed in conditions of sediment bypassing and a restricted sand supply together with a sparse vegetation cover that is insufficient to prevent sand transport taking place but sufficient to cause divergence and convergence of airflow around individual plants in the manner suggested by Fryberger and colleagues (1979), giving rise to localized deposition by wind ripples and shadow dunes. Many sand sheets as well as interdune areas between linear and star dunes are organized into low, rolling dunes without slip faces known as zibar (Holm, 1960) with a spacing of 50–400 m and a maximum relief of 10 m. Typically, zibars are composed of coarse sand and occur on the upwind margins of sand seas. Zibars in the Skeleton Coast dune field are 1–2 m high with a crest-to-crest spacing of 100 m (Lancaster, 1982a) and occur in association with areas of megaripples. Those on the western margins of the Algodones dune field have a crest-to-crest spacing of 60 m and rarely exceed 2 m in height (Nielson and Kocurek, 1986). Somewhat similar coarse-grained bedforms occur in the Selima Sand Sheet of Egypt but have a much longer wavelength (130–1,200 m) and an amplitude of 0.1–10 m (Breed et al., 1987a; Maxwell and Haynes, 1989). 2.3.8 Topographically Controlled Dunes Dunes that owe their existence and form to interactions of sand-transporting winds with topographic obstacles are common in many high-relief desert regions, yet are poorly documented. At the smallest scale are shadow dunes that accumulate in the lee of large boulders or breaks in slope (Hesp, 1981). By contrast, large climbing and falling dunes in the Mojave Desert have a total relief of as much as 200 m and comprise large volumes of sand. Many of these features are composites of aeolian, fluvial, and colluvial deposits and are called sand ramps (Fig. 2.24a). They are widespread in the Mojave Desert and on the eastern margin of the Namib Sand Sea (Lancaster and Tchakerian, 1996; Rowell et al., 2018). Very large climbing dunes also occur in the Atacama Desert of Peru (Haney and Grolier, 1991; Howard, 1985). Echo dunes form on the windward side of escarpments, where the average slope is greater than 55° and the dune is separated from the scarp by a sand-free area that results from a fixed eddy between the cliff and the dune (Tsoar, 1983b). When average slope angles are less than 55°, the fixed eddy is small or nonexistent, and the result is a climbing dune. Falling dunes occur in the lee of hills that are asymmetric in form, with lee slopes steeper than those to windward. In many areas of the Mojave Desert, climbing and falling dunes occur on opposite sides of the same mountain range (Fig. 2.24b).

https://doi.org/10.1017/9781108355568.003 Published online by Cambridge University Press

Dune Morphology

47

(a)

(b)

Figure 2.24 Topographically controlled dunes: (a) sand ramp east of Dale Lake, California; (b) climbing and falling dunes, Dumont, California.

2.3.9 Source-Bordering Dunes Dunes immediately adjacent to their sediment source occur widely where fluvial and lacustrine environments penetrate dune fields (Bullard and McTainsh, 2003), as in the Strzelecki Desert of Australia (Cohen et al., 2010), where they form transverse to dominant winds. Linear dunes often extend downwind from these sources (Wasson, 1983b). Lunettes (Fig. 2.25) are dunes with a subparabolic shape that form on the downwind margins of playas and dry lake beds, closely following the margins of the source area. Lunette dunes are widely distributed in Australia ((Bowler, 1983),

https://doi.org/10.1017/9781108355568.003 Published online by Cambridge University Press

48

Dune Morphology and Sediments (a)

5 km (b)

500 m

Figure 2.25 Lunette dunes: (a) Lake Mungo, New South Wales, Australia; (b) lunette dunes and linear dunes in the southwestern Kalahari. Note how linear dunes extend from the lunette. Images from Google Earth, earth.google.com/web/. For the color version, refer to the plate section.

the Kalahari (Goudie and Thomas, 1985; Lancaster, 1978), and parts of the High Plains of the USA (Holliday, 1997). They can be formed from a range of sediments from sands to clays and mixtures of these end members. Many lunettes are composite forms, resulting from multiple episodes of accumulation (Telfer and Thomas, 2006). In some places, linear dunes also extend downwind from sandy lunettes in Australia (Twidale 1980) and the southwestern Kalahari (Lancaster, 1988c).

https://doi.org/10.1017/9781108355568.003 Published online by Cambridge University Press

3 Dune Sediments

3.1 Mineral Composition The majority of desert dune sands are composed of quartz and feldspar derived from plutonic rocks (mainly granites), sandstone, and some metamorphic rocks. Dunes composed of volcaniclastic materials occur in some areas (Edgett and Lancaster, 1993); for example, at the Great Sand Dunes, Colorado (51.7% volcanic rock fragments, 27.8 % quartz). Inland dunes of carbonate-rich sand have been documented from Oman and the United Arab Emirates (Besler, 1982; Gardner, 1988), and gypsum dunes occur at White Sands (Jerolmack et al., 2011; McKee, 1966); in the Lake Eyre Basin (Chen et al., 1991); and adjacent to the chotts of Tunisia (Swezey et al., 1999). Studies in a wide variety of locations indicate that quartz dominates the composition of many dune fields and sand seas, and these sands are classified as mineralogically mature (Muhs, 2004). Mineralogical maturity reflects some combination of sand source, transport history, and postdepositional alteration and is expressed by the ratio between the quartz and feldspar content of the sand (Fig. 3.1). A high degree of mineralogical maturity may reflect derivation from quartz-rich rocks, such as sandstones on the Colorado Plateau for the Moenkopi dunes in northeast Arizona and, via the Colorado River, for the Algodones and Parker dune fields; Saharan and Arabian sand seas are partly composed of sand derived from Paleozoic sandstones (Garzanti et al., 2013; Muhs, 2004; Pastore et al., 2021). Long-distance transport of sand by aeolian and fluvial processes has also been proposed to explain the mineral maturity of dune fields in Australia and Namibia (Garzanti et al., 2012; Muhs et al., 2018), whereby less-resistant feldspar grains are abraded to silt size during transport and lost by suspension. The high degree of mineral maturity for Australian dune fields may also reflect long-distance transport of sand, but some studies suggest more local sand sources (e.g. Pell et al., 2000; Pell et al., 2001; Pell et al., 1997). Relatively low mineral maturity in the Takla Makan sand sea and Mojave Desert dune fields reflects proximity to sand sources in adjacent alluvial fan systems (Muhs, 2017; Rittner et al., 2016). 49

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

50

Dune Morphology and Sediments

(a)

(b)

Figure 3.1 Mineral maturity of sands expressed by quartz/feldspar ratio in (a) global sample of dune fields and sand seas; (b) North American dune fields. From Muhs (2017).

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

Dune Sediments

51

Many dune fields in the Mojave and Sonoran deserts of the SW USA are composed of 40–50% feldspathic sand, reflecting short-distance transport pathways from local sand sources (Muhs, 2004; Muhs et al., 2003). For example, dunes adjacent to Owens Lake, California, are derived from sand transported by fluvial systems from the adjacent Sierra Nevada Mountains (Lancaster et al., 2015). At Kelso dunes, geochemical and mineralogical studies show that while much of the dune field is composed of quartz-rich sand derived from the Mojave River that drains from the Transverse Ranges of southern California, local sources in the mountains on the east of the dune field provide a significant input to the eastern part of the dune field (Muhs et al., 2017). Similarly complex patterns of dune sand composition occur in the Gran Desierto sand sea, reflecting multiple sand sources in local alluvial fans (feldspar-rich), coastal areas (carbonate-rich), and the Colorado River (quartz-rich) (Beveridge et al., 2006; Scheidt et al., 2011). 3.2 Grain Size and Sorting There are many particle size analyses of dune sands. They show that most dunes are composed of fine to medium (mean grain size 1.60–2.65 phi, 160–330 µm), very well to moderately sorted sands (phi standard deviations 0.26–0.55), although there is a wide range between and within individual dunes and sand seas (Ahlbrandt, 1979). Comparative data (Table 3.1) for average values of mean grain size and sorting for different dune types in various sand seas show that very fine (2.51–2.77 phi; 150–180 µm) dune sands occur in the Grand Erg Oriental, Thar, Takla Makan, Simpson Desert, and Gran Desierto sand seas. Those from the Namib, the UAE, and northwestern Sahara are intermediate in size (2.05–2.12 phi; 200–240 µm). Relatively coarse sands occur in the Skeleton Coast dune field, in western Mauritania, and at Kelso Dunes, the Algodones Dunes, and on the northern margins of the Gran Desierto sand sea. Following Jerolmack and Brzinski (2010), the median or dominant grain size may indicate the magnitude of the “formative wind,” which is the wind speed that transports the majority of sand in saltation, equivalent to approximately 1.5–2.0 times the threshold wind speed for saltation. Sand grain size and sorting are therefore not just a descriptor of dune sediments, but also an important index of the process environment, with coarse sands reflecting environments subject to more intense transport events.

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

52

Dune Morphology and Sediments Table 3.1 Grain size characteristics of selected dune crest sands Location

Mean (phi units)

Mean (µm)

Sorting (phi standard deviation)

Crescentic dunes Gran Desierto Algodones White Sands Skeleton Coast Namib Sand Sea Salton Sea Tunisia Kelso Takla Makan Liwa

2.43 2.46 1.61 2.02 2.20 2.27 3.25 1.80 3.08 2.71

186 182 328 247 218 207 105 287 118 153

0.41 0.42 0.59 0.51 0.55 0.46 0.53 0.49 0.37 0.53

Linear Dunes SW Kalahari Simpson Desert Namib Sand Sea Saudi Arabia Mauritania Takla Makan UAE

2.16 2.53 2.44 2.67 2.20 2.87 2.63

224 173 184 157 218 137 161

0.49 0.43 0.37 0.32 0.60 0.52 0.44

Star Dunes Gran Desierto Namib Sand Sea Great Sand Dunes Kelso Takla Makan

2.43 2.29 2.09 2.26 2.81

186 204 235 209 143

0.41 0.29 0.26 0.30 0.58

3.2.1 Conceptual Models for Grain Size and Sorting Aeolian sands can be viewed as a mixture of varying proportions of saltation, reptation and creep, and suspension populations (Eastwood et al., 2012; Folk, 1970; Visher, 1969). The proportions vary with wind shear stress (Fig. 3.2). Generally, wind ripple deposits tend to be relatively coarser and less well sorted because they contain grains transported by both creep/reptation and saltation. Grainfall deposits are the finest and best sorted and are mainly composed of saltating grains. Grainflow deposits are somewhat coarser than grainfall and contain both saltating and reptating grains that have been reworked by avalanching (Eastwood et al.,

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

Dune Sediments

53

March Event 90

u = 0.35 m/sec *

80 70 Percent

60 50 40 30 20 10 0 Creep

Saltation Wind Ripple

Incipient suspension Grainfall

Suspension

Grainflow

April event 90

u = 0.45 m/sec *

80 70 Percent

60 50 40 30 20 10 0 Creep

Saltation Wind Ripple

Incipient suspension Grainfall

Suspension

Grainflow

Figure 3.2 Relative proportions of creep/reptation, saltation, modified suspension, and suspension populations measured in different depositional environments on crescentic ridges at White Sands dune field. Data from Eastwood and colleagues (2012).

2012). Variations in the spatial distribution of these modes of aeolian deposition over dunes, together with changes in the effectiveness of saltation and reptation/ creep processes on sloping surfaces give rise to patterns of grain size and sorting parameters over dunes that reflect the action of sand transport and depositional processes on dunes.

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

54

Dune Morphology and Sediments

Grain Size and Sorting Parameters Considerable attention has been paid to the development and interpretation of statistical parameters of the size and sorting of sand-sized sediments. The graphical parameters devised by Folk (1966) are widely used, as are various moment statistics (Friedman, 1962). Both graphical and moment measures characterize sediment size with the modal or mean grain size and sorting using the standard deviation of grain sizes. The asymmetry of the distribution is given by its skewness: positive skewness indicates a “tail” of fine grains, whereas negatively skewed sands have a coarse “tail”. The phi scale is commonly used to designate the size and sorting of sand. phi ¼ log 2d where d is the grain diameter in millimeters. Recently developed grain size and sorting parameters include use of the loghyperbolic distribution (Barndorff-Nielsen et al., 1982; Barndorff-Nielsen and Christiansen, 1985; Hartmann and Christiansen, 1988; Vincent, 1986). These measures are much better than the Folk and Ward approach at characterizing skewed grain size distributions, but the interpretation of the parameters in terms of sedimentary processes and their general applicability (especially to bimodal sands) is debated (Vincent, 1988).

3.2.2 Grain Size and Sorting Characteristics of Different Dune Types Crescentic Dunes Changes in grain size and sorting parameters over barchans and crescentic ridges show a consistent pattern (Fig. 3.3) that is characterized by a decrease in modal and mean grain size from the base of the stoss slope toward the crest, with the finest sands occurring on the middle of the slip face (Barndorff-Nielsen et al., 1982; Lancaster, 1989b; Watson, 1986); sorting also improves in the same direction, although Watson (1986) found that dune crest sands were poorly sorted by comparison to other positions on the dune. In the Namib, skewness becomes more negative (fewer coarse grains) across the dune. A tendency for coarse grains to be concentrated at the base of the dune, especially on barchans, has been widely noted (e.g. Finkel, 1959; Hastenrath, 1967; Lancaster, 1982a; Watson, 1986). Unlike the case of linear and star dunes, distinct subenvironments with characteristic grain size and sorting patterns are not discernible. Linear Dunes Many investigators have noted changes in grain size and sorting over linear dunes. Bagnold (1941) observed that, in the Libyan desert, crest sands of linear dunes were finer than those from the base or plinth. Similar conclusions were arrived at by McKee

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

Dune Sediments

crest

Cumulative %

slip face

55

mid stoss slope

base stoss slope

99.5 95 50 5 0.5 0

2

4

0

2

4

0

2

4

0

2

4

2

4

0

2

4

0

2

4

0

2

4

%

30 0

0

slip face

stoss slope crest

mid

base

2.4 Mean 2.2

2.0 1.0 Standard deviation

0.8 0.6 0.4

+0.2 Skewness 0 –0.1

0.6 Kurtosis

0.5 0.4

Figure 3.3 Patterns of particle size distribution and grain size and sorting parameters over crescentic dunes in the Namib Sand Sea.

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

56

Dune Morphology and Sediments

and Tibbitts (1964) and Glennie (1970) for simple linear dunes in southwestern Libya and Oman; and by Alimen and colleagues (1958), Lancaster (1981a), and Livingstone (1987) for complex linear dunes in Algeria and the Namib. However, in the Simpson and southwestern Kalahari dune fields, crests of simple linear dunes are coarser than their flanks (Crocker, 1946; Folk, 1970; Lancaster, 1986; Lancaster, 1987). Further, some workers (e.g. Wang et al., 2003; Warren, 1972) have reported no differences in grain size and sorting over linear dunes they investigated. Grain size and sorting patterns on simple linear dunes in the Simpson Desert studied by Folk (1970) show that the crest sand is coarser, but better sorted, compared to flanks and interdune areas. Similar patterns were observed in the southwestern Kalahari by Lancaster (1986) (Fig. 3.4). Folk explained grain size and sorting patterns on Simpson Desert linear dunes in relation to the source material. He suggested that the wind selected material in the 180-µm size range from the source sediment and concentrated it into dunes. If, as in the Simpson Desert, the source was a fine-grained alluvium, then it would tend to become finer, more poorly sorted, and bimodal over time as the sand was removed, leaving dune crests coarser, but better sorted than the source sediment. In the case of a coarse source material, the converse would be the result. Dune flanks are intermediate in composition, as they trap weakly saltating coarse sand and receive some fine sand by avalanching from slip faces. (b)

(a) 60

60

40

40

% retained by each sieve

Crests 20

20

0 40

0 40

20

20

0 40

0 40

20

20

Flanks

Interdunes

0

0 1

3 2 Phi

4

1

3

2

4

Phi

Figure 3.4 Patterns of particle size distribution and grain size and sorting parameters over simple linear dunes in the southwestern Kalahari (after Lancaster 1986): (a) fine interdune; (b) coarse interdune.

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

Dune Sediments

slip face

plinth

upper slope

crest

plinth

Mid

99.5 95

Cumulative %

57

0

2

4

50 5 0.5 0

2

Base

4 0

2

0

2

4

0

2

4

0

2

4

0

2

4

0

2

4

phi

4

Mid

%

50 0

0 0

2

2

4

4

Base 0

plinth

2

0

2

4

phi

4

slip face

crest

upper slope

plinth

2.4

Mean

2.2 2.0

Standard deviation

0.8 0.4 0.4

Skewness

0.2 0 0.6

Kurtosis

0.5 0.4

Figure 3.5 Patterns of particle size distribution and grain size and sorting parameters over complex linear dunes in the Namib sand sea.

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

58

Dune Morphology and Sediments

Complex linear dunes of the Namib Sand Sea are composed of two distinct groups of sand (Fig. 3.5) (Lancaster 1981a). Sand from the crests, slip faces, and upper west slopes of the linear dunes is characteristically fine, well to very well sorted and near-symmetrical. By comparison, that from the plinth area is coarser, moderately sorted, and frequently strongly positively or fine skewed. Interdune areas are coarser still and moderately to poorly sorted. Detailed sampling by Livingstone (1987) indicates that the two groups of sand identified by Lancaster (1981a) are end members of a continuum of grain size and sorting changes over the dune. The differences can be explained in terms of the pattern of sand movement on the dunes. In this model, the population of coarse grains moved by creep and reptation are progressively left behind and concentrated in the interdune areas and on dune plinths as sand moves up dune slopes under the influence of winds from directions oblique to the dune trend (Lancaster 1981a; Livingstone 1987). Star Dunes As with other large dunes, the surface sand of star dunes is characterized by significant spatial variations in grain size and sorting parameters. Sand from the crests of star dunes is generally very well sorted, and often much better sorted than other dune types in a sand sea. Star dunes in the Namib Sand Sea are composed of fine to very fine sands. Mean grain size decreases from interdune areas and plinths to the crest of the dune and sorting improves in the same direction (Fig. 3.6). Bivariate plots of mean values of grain size and sorting parameters show that the crests, upper slopes, and slip faces of star dunes are consistently finer and better sorted than adjacent plinths and interdune areas. Star dunes in the Gran Desierto are similarly composed of very well sorted fine to very fine sands. Sand from dune crests is typically very well sorted and near-symmetrical. Dune plinths are slightly coarser and less well sorted. Sorting processes on star dunes are similar to those on large linear dunes, with loss of coarse grains upslope leading to their concentration in plinth areas. Comparisons between Dune Types Specific associations between dune types and grain size and sorting parameters are difficult to establish. Many of the reported relationships are probably the product of progressive sorting and fining of sands downwind from source zones. The clearest relationship to emerge is that zibar and sand sheets are composed of coarse poorly sorted sands (e.g. Kocurek and Nielson, 1986; Warren, 1972). Where multiple dune types have been sampled, the results are varied. In the Algerian Sahara, Bellair (1953) found that barchans and crescentic dunes were composed of well-sorted, unimodal sands, but complex linear and star dunes were composed of bi- or trimodal sands. However, Alimen and colleagues (1958) and

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

Dune Sediments

interdune area

(a)

slip face

59

upper slopes

crest

plinths

Cumulative %

99.5 95 50 5 0.5 0

2

4

0

2

4

0

2

4

0

2

4

0

2

4

4

0

2

4

0

2

4

phi

%

50

0 0

2

4

0

2

4

0

2

phi

plinth

(b)

slip face

crest

upper slopes

plinth

interdune area

2.6 2.4

Mean

2.2 2.0

0.9

Standard deviation

0.7 0.5 0.3

0.2

Skewness 0

Kurtosis

0.55 0.50

Figure 3.6 Patterns of particle size distribution and grain size and sorting parameters over star dunes in the Namib Sand Sea.

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

60

Dune Morphology and Sediments

Capot-Rey and Gremion (1964) could find no consistent relationships. At White Sands, New Mexico, McKee (1966) observed a progressive decrease in grain size and improvement in sorting from dome dunes, through barchans and crescentic dunes to parabolic dunes. Subsequent work has confirmed these observations. Sand (in this case composed of gypsum) becomes steadily finer and better sorted, largely as a result of the loss of coarse grains, mainly via abrasion and mostly within the first 5 km of transport (Jerolmack et al., 2011). In the Takla Makan sand sea, Wang and colleagues (2003) found that there was an increase in mean grain size from compound crescentic and compound dome dunes to compound/complex linear dunes to star dune, which are composed of coarser sand than other dune types. They argued that this might reflect changes in sand sources, wind regimes (higher to lower energy), and dune age. In the Namib Sand Sea (Fig. 3.7), there is an improvement in sorting, and an overall decrease in mean grain size from zibars and sand sheets, through barchans and crescentic dunes to linear and star dunes, which tend to be best sorted, although not always the finest. These patterns result in part from the nature of sediment transport on different dune types. On crescentic dunes, sand movement is essentially unidirectional, and saltating sands are buried on avalanche faces to be recycled later as these deposits are exposed by advance of the dune. Crestal areas of linear dunes are reversed seasonally and undergo more frequent resorting. Star dune crests are best sorted, because they undergo constant reworking by multidirectional winds (Lancaster, 1989b). Sand sheets and zibar 1.0

Crescentic

Standard deviation

*

Star Compound linear Complexlinear

0.5 * * **

*

0 1.5

2.0

2.5

3.0

Mean grain size (phi)

Figure 3.7 Bivariate plots of grain size and sorting parameters showing comparisons between dune types in the Namib Sand Sea.

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

Dune Sediments

61

However, the patterns observed in the Namib Sand Sea are not well represented in some other sand seas and dune fields where episodic input of sand from different sources has occurred. In the Gran Desierto, there are multiple generations of dunes, some of which are texturally and compositionally distinct (Beveridge et al., 2006; Lancaster, 1992; Scheidt et al., 2011). Similar patterns of grain size and sorting (Fig. 3.8) can be observed at Kelso Dunes, where there are significant differences between dunes of each distinct morphologic unit, which represent multiple episodes of sediment input to the dune field (Muhs et al., 2017). In addition, sands from active dune areas are consistently finer and better sorted than those from vegetationstabilized dunes (Paisley et al., 1991) (Fig. 3.9), a pattern that has also been noted from dunes in the Great Plains of the USA (e.g. Bolles et al., 2017). 3.3 Grain Shape Although early workers (e.g. Cailleux, 1952; Shotton, 1937) suggested that aeolian sands were rounded or well-rounded in shape, more recent investigations (Folk, 1978; Garzanti et al., 2015; Goudie et al., 1987; Goudie and Watson, 1981) indicate that, in aeolian sands, true roundness in the dominant 125–250 µm size group is rare and most grains are subangular to subrounded in shape (Fig. 3.10). Rounding of grains occurs as a result of abrasion by grain-to-grain contacts and is most effective in aeolian, compared to fluvial and shallow marine environments; most of the rounding takes place close to the beginning of aeolian transport and declines exponentially downwind (Garzanti et al., 2015). Goudie and Watson (1981) noted that grains from different sand seas cluster around distinctive grain roundness characteristics that reflect sand source characteristics and transport pathways (Fig. 3.11). Thus, sand from Tunisian dunes is more rounded than that from the Namib and Kalahari, which in turn is more rounded than Thar and some North American dune sands, much of which have been derived directly from Mesozoic and later source rocks. The surfaces of desert sand grains may have distinctive characteristics when examined at high magnifications using a scanning electron microscope (Krinsley and Trusty, 1985). Some of these features are (1) rounding of edges on all grain shapes, (2) “upturned plates” resulting from breakage of quartz along cleavage planes in the crystal lattice, (3) elongate depressions resulting from conchoidal fractures during collisions between grains, (4) smooth surfaces resulting from solution and reprecipitation of silica, and (5) arcuate, circular, or polygonal fractures resulting from collisions and/or weathering. The proportions of these features change through time in accumulating aeolian deposits so that older sand grains exhibit a greater proportion of features that can be attributed to weathering (Tchakerian, 1991).

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

62

Dune Morphology and Sediments Group 1

60.0 50.0 40.0

% 30.0 20.0 10.0

00 Pa n

50

00

4.

φ

3.

50

3.

00

2.

50

2.

00

1.

50

1.

0.

0.

00

0.0

Group 2

60.0 50.0 40.0

% 30.0 20.0 10.0

50 3. 00 3. 50 4. 00 Pa n

00

2.

50

2.

00

1.

50

1.

0.

0.

00

0.0

φ Group 3

60.0 50.0 40.0

% 30.0 20.0 10.0

0. 00 0. 50 1. 00 1. 50 2. 00 2. 50 3. 00 3. 50 4. 00 Pa n

0.0

φ Standard Deviation (φ)

1.0

Group 1

0.9

Group 2

0.8 0.7 0.6 0.5 0.4

Group 3

0.3 6 2.

4 2.

2 2.

0 2.

8 1.

1.

6

0.2

Mean (φ)

Figure 3.8 Particle size distributions and bivariate plots of grain size and sorting parameters showing differences between dune generations at Kelso Dunes.

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

Dune Sediments

63

Active Dune Crest 70.0 60.0 50.0 %

40.0 30.0 20.0 10.0 0.0 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 Pan φ Inactive Dune Crest 70.0 60.0 50.0

%

40.0 30.0 20.0 10.0 0.0 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 Pan φ

Figure 3.9 Comparison of particle size distributions between active and inactive dunes at Kelso Dunes.

3.4 Sand Color The colors of dune sands vary widely from one sand sea to another, and even within the same sand sea. Some of the darkest and most reddened sand occurs in the Kalahari and Simpson–Strzelecki deserts where sand colors are red to yellowish red (Munsell Notation 2.5YR 5/8 to 7.5YR 5/8) (Folk, 1976a; Wasson et al., 1983; Lancaster, 1986). Relatively pale colors (10YR 6/2 to 7/4) are associated with dunes in some coastal deserts (Lancaster, 1982a). Many North American sand seas and dune fields are composed of relatively pale-colored sand (10YR 6/2) (e.g. Blount and Lancaster, 1990). Redder hues occur in Colorado Plateau dune fields,

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

64

Dune Morphology and Sediments

(a)

(b) Figure 3.10 Scanning electron microscope (SEM) images of dunes from Namib Sand Sea: (a) subangular-angular grains from eastern sand sea; (b) subrounded grains from central sand sea.

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

Dune Sediments

65

Subrounded

4.0

Bahrain

+++ + + + ++ + + + ++ + + + + + + + Kalahari ++ + + + + ++ + + + + + ++ + + + + ++ + ++ + Namib + + + ++ + ++ + + +

y = 140 + 0.52x r 2 = 0.51

+

3.0

Thar +

Subangular

+ + ++

+ + +

+

+ + ++ + +

Tunisia

+

U.S.A.

+

+

2.0 2.0

2.5 Subangular

3.0

3.5 Subrounded

4.0

4.5 Rounded

Figure 3.11 Variation in grain shape between sands in different sand seas (after Goudie and Watson 1981).

reflecting the source of sand in Mesozoic sandstones (Stokes and Breed, 1993). Sand from the northwestern Sahara (Alimen et al., 1958), Libya (Walker, 1979), the central Namib Sand Sea (Walden and White, 1997), and Arabia (Besler, 1982) is intermediate in color with reddish yellow (7.5 YR 5/6 to 5/8) to yellowish red (5 YR 5/8) colors common. Studies of individual sand grains under the microscope show that increased reddening of the grains is achieved by a greater extent and thickness of iron oxides deposited in pits and other surface irregularities on clear or frosted quartz grains. Reddening or rubification is most pronounced on smaller (125–250 µm), more angular grains, as was observed by Folk (1976a,b) from Simpson desert dunes. In some areas, red colors are the result of the amber color of many quartz grains (e.g. Wasson et al., 1983). 3.4.1 Spatial Variations in Dune Color It has been widely reported that dunes become redder with distance in the direction of transport and as the residence time sand in them is longer (e.g. Alimen et al., 1958; El Baz and Maxwell, 1982; Folk, 1976a; Logan, 1960; Walker, 1979; Wopfner and Twidale, 1967), implying that sand color is a proxy for dune age.

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

66

Dune Morphology and Sediments 103° W NM

TX

5 YR

7.5 YR

10 YR

Aeolian sand

25

0

10 YR

km 34° N

Figure 3.12 Changes in sand color downwind in the Muleshoe dune field, expressed by dominant Munsell hues in late Holocene aeolian sands. Redrawn from Muhs and Holliday (2001).

However, Gardner and Pye (1981) suggest that color is not necessarily a function of dune age and that time is not always an independent variable. In the Negev, there is no relationship between dune age and degree of reddening of sands (Roskin et al., 2012). As observed by Folk (1976a), reddening is a product of moisture availability, temperature, and time. Also important are the intensity of sand movement and abrasion, grain mineralogy, presence of weatherable minerals, and the aeolian dust input. Further, different sources for sands, especially the existence of pre-reddened sediments, may be important. In the case of the Muleshoe Dunes of New Mexico and Texas, the degree of reddening decreases downwind (Fig. 3.12) as a result of abrasion of red clay coatings on sand grains derived from the underlying reddened Blackwater Draw Formation (Muhs and Holliday, 2001; White and Bullard, 2009). In the Namib Sand Sea, sands become redder toward the eastern margin of the sand sea (Fig. 3.13). This pattern was attributed to a greater degree of dune stability toward the east (Lancaster, 1989b). However, satellite remote sensing studies, in combination with geochemical and mineral magnetic analyses, indicate that the color variation is a product of the mixing of sand from two different sources: palecolored sand derived from southern and coastal sources and red sand derived from the preexisting Tsondab Sandstone, which outcrops in the eastern parts of the sand sea (White et al., 2007). 3.5 Dune Sedimentary Structures Sedimentary structures preserved in dunes provide a rich source of information on the ways in which the dune has accumulated and the processes involved. In particular, they can be used to provide information on the wind speed and direction and the sand-moving event duration that created the dune sediments (Eastwood et al., 2012).

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

Dune Sediments

67

15° E

Walvis Bay

23° S

10 YR 75 YR

5 YR

Gobabeb

reddishyellow

Tsondab Vlei

10 YR 6/4

.

5 YR 5/8 – 75 YR 5/8

b

R

a ch

yellowishbrown

au

Ts

7.5 YR 5/4–6/8

yellowishred 25° S

brown 10 YR 7/4 very pale brown

Lüderitz

0

km

50

Figure 3.13 Spatial variation in dune sand colors in the Namib Sand Sea.

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

68

Dune Morphology and Sediments

Grain fall

Grain flow

Wind ripple

Figure 3.14 Lee face of dune showing three main types of aeolian deposits.

There are three primary modes of deposition on dunes (Hunter, 1977): (1) migration of wind ripples, (2) fallout from temporary suspension of previously saltating grains in the flow separation zone in the lee of the crest, and (3) avalanching on the lee slope of the dune (Fig. 3.14). These processes form three main types of aeolian sedimentary structures, which constitute the primary units of aeolian deposition (Hunter, 1977; Kocurek and Dott, 1981): (1) climbing translatent strata (wind ripple laminae), (2) grainfall laminae, and (3) grainflow cross-strata. Climbing translatent strata are formed by the migration of wind ripples under conditions of net deposition that give rise to bedform climbing (Rubin and Hunter, 1982). In conditions where wind shear stress decreases downwind, the transport capacity of the wind declines and excess sediment is deposited or transferred to the bed, producing wind ripple laminae. Wind ripple deposits are very widespread in most dune areas and occur on the stoss slopes of many dunes, as well as on the plinths or aprons of larger linear and star dunes (Eastwood et al., 2012). Wind ripple laminae are also prominent in the deposits of most sand sheets and interdune areas (Ahlbrandt and Fryberger, 1981; Kocurek and Nielson, 1986) and on the steeply sloping lee faces of dunes affected by strong secondary flows (Eastwood et al., 2012; Tsoar, 1982; Walker, 1999) . Grainfall laminae are formed when grains that saltate over the brink line of the dune come to rest on the lee face (Nickling et al., 2002). Preservation of grainfall deposits only occurs if no subsequent oversteepening and avalanching take place. They appear to be more common on small dunes or in deposits laid down in very strong winds so that grainfall occurs on the lower parts of the lee face (Hunter 1977).

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

Dune Sediments

69

Grainflow cross-strata are formed by avalanching of grainfall deposits on the lee face of dunes, which are oversteepened beyond the angle of repose of dry sand (28– 34°) (Sutton et al., 2013). They consist of a series of overlapping tongues 3–4 cm thick with coarse grains concentrated on their upper surfaces and at the toe of the tongue. Most grainflow strata thin out toward the top of the dune and have a tangential lower contact with the base of the dune. Slump structures and blocks may be preserved in damp sand avalanches (Cornwall et al., 2018; McKee et al., 1971). The primary stratification types identified previously may be combined in a variety of ways, depending on the incidence angle between the formative wind and the brinkline, with secondary influences of wind speed and dune morphology. Sets of wind-ripple strata are formed at wind incidence angles of 25–40°, grainfall/ grainflow foresets cover wind-ripple bottomsets at incidence angles of 40–70°, and grainflow/grainfall foresets at 70–90°. Erosional reactivation surfaces form at incidence angles up to 15°, and bypass surfaces up to 25° incidence angle (Eastwood et al., 2012). The relationships between grain size, transport mode, shear velocity, and grain-settling velocity can be inverted to estimate formative wind speeds, thus providing a methodology to infer the paleo-wind environment of the dune sediments. Primary dune sedimentary structures are separated by bounding surfaces of different types (Fig. 3.15) (Brookfield, 1977; Kocurek, 1984). Reactivation surfaces are very common and occur within sets of laminae and represent periods of nondeposition or erosion resulting from short-term or seasonal changes in wind strength and/or direction. Superposition surfaces bound sets of strata and form by the migration of superimposed dunes across or along the flanks of a larger form (e.g. a megadune). Generally, the complexity of sedimentary structures increases with dune size and as the importance of secondary flows and superimposed bedforms increases. Interdune surfaces may divide the accumulations of laterally migrating dunes, or in some environments may represent episodes of deflation to the water table (Loope, 1984; Stokes, 1968). In addition, regional-scale or super surfaces (Kocurek and Havholm, 1993) form as a result of hiatuses in sand sea and dune field accumulation due to depletion of sediment supply and/or climatic changes. 3.5.1 Sedimentary Structures of Major Dune Types Information on dune sedimentary structures has traditionally been gained through trenching of dunes and sand sheets. This expensive and destructive process has largely been superseded by the use of ground-penetrating radar (GPR) to characterize sedimentary structures in two or three dimensions. Aeolian sediments are

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

70

Dune Morphology and Sediments Interdune Deposits

(a)

1

Set

3

(b)

Interdune Deposits

1

Interdune Deposits

1

2

3 Cosets 2 3

Interdune Deposits

1

(c) Depositional Surfaces )

) Angle of climb

Angle of climb

Figure 3.15 Model for the formation of bounding surfaces in aeolian deposits (after Kocurek, 1988): (a) migration of simple dunes and interdune areas; (b) migration of compound/complex dunes and interdune areas. Different orders of bounding surfaces indicated by 1, 2, 3.

particularly amenable to study using GPR because they have a high resistivity, which allows good penetration of the signal and contain sedimentary structures at a scale that can be resolved by GPR. Ground-penetrating radar exploits changes in permittivity within the dune sediments to image dune sedimentary structures (Jol and Bristow, 2003) and has been used to image the two- and three-dimensional structure and shallow stratigraphy of aeolian sands in a variety of environments.

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

Dune Sediments

71

Crescentic Dunes Studies by Ahlbrandt (1974), McKee (1982), McKee (1966), and Hunter (1977) have shown that the sediments of barchans and crescentic dunes are dominated by grainfall and grainflow cross-strata deposited on the lee face and preserved as the dune migrates downwind. Detailed studies of small barchanoid ridges at Padre Island by Hunter (1977) show the spatial distribution of grainfall, grainflow, and wind ripple laminae (Fig. 3.16). Larger dunes are dominated by grainflow laminae, whereas small dunes include more grainfall and wind ripple laminae. Wind ripple laminae form a thin set of strata that parallel the slope of the stoss slope and dune crest. Barchans in northern Arizona, at White Sands, the Killpecker dune field, and the western Sahara are composed of sets of cross-strata with dips between 31 and 34° in the direction of dune migration. Crescentic dunes at Killpecker and White Sands (Fig. 3.17) are

t ne a f du eo dg te on Fr

Climbing translatent stratification

Grainfall lamination

tim e

Sandflow cross-stratification

of n tio na p la

Boundaries between sets

A' 0

2

6m

4

A

A

A'

Figure 3.16 Spatial distribution of sedimentary structures in a small crescentic ridge (after Hunter 1977).

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

72

Dune Morphology and Sediments A (i) SW

NE

35 30 25 20 15 10

0

10

20

30

40

50

60

70

80

90

100

110

120

130

140 150

M ain t N ren 19 ch ° E (A

)

5 0 Feet

S S ide 78 tre ° E nc

(ii) NW 35 30 25 20 15 10 5 0 Feet

h(

SE

B)

27

0

10

20

30

40

50

60

70

80

90

100

110

120

130

140 150

Figure 3.17 Structures in barchan dunes: White Sands (modified from McKee, 1966): (i) parallel to wind direction, (ii) perpendicular to wind direction. Reactivation surfaces in red (color online).

dominated by large sets of leeward-dipping avalanche face laminae at angles of 30 to 34°, with a tangential contact to the lower surface. Wind ripple laminae are restricted in all cases to a thin veneer over the stoss slope of these dunes. The sets of cross-strata are separated by horizontal to steeply-dipping reactivation surfaces, interpreted on GRR profiles of a barchan in the western Sahara by Bristow (2019) as the product of seasonal wind reversals (Fig. 3.18). Structures in large compound crescentic dunes are complex (Fig. 3.19). Multiple bounding surfaces separating sets of cross-strata were identified in GPR and trenching of the upper parts of large compound crescentic dunes in the Liwa area of the UAE (Bristow et al., 1996). Trough cross-stratification also occurs, as a result of migration of superimposed bedforms across and along the main dune. Linear Dunes Linear dunes are composed of varying proportions of wind ripple and avalanche face laminae. Bagnold (1941) published a hypothetical section of a linear dune, in

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

Dune Sediments

73

Figure 3.18 Barchan GPR profile parallel to formative wind in western Sahara. From Bristow (2019). Bounding surfaces in red (color online). 1 J 2

I

3 4 5

Primary wind direction

H G F

6 7

E

8

D C

9 10

B A

0 12 8 11 6 11 4 11 2 11 0 11 ters me

11 8 11 6 11 4 11 2 11 me 0 ter s

0 10 20

20

s ter me

me ter s 0

30

N

Figure 3.19 Isometric reconstruction from GPR profiles of a compound crescentic dune showing sets of cross-stratification bounded by gently dipping, troughshaped, second-order bounding surfaces and less extensive third-order bounding surfaces. From Bristow and colleagues (1996).

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

74

Dune Morphology and Sediments

which laminae were divided into steeply dipping avalanche deposits on the crest and central areas of the dunes, and low-angle accretion laminae on the dune plinths (Fig. 3.20a). McKee and Tibbitts (1964) confirmed this basic pattern for a 15-m-high simple linear dune in the Libyan desert (Fig. 3.20b). They found that the upper parts of the dunes were composed of cross-strata with dips of 26–34°, with low-angle strata (4–14°) in lower areas. The high-angle strata were interpreted as avalanche face deposits formed in a diurnally bidirectional wind regime, with winds blowing at around 45° to the dune. Tsoar (1982) recognized two groups of laminae on a simple linear dune in the Sinai (Fig. 3.20c). The first group were deposited by grainfall and grainflow in a 1–2-m-wide area parallel to the crest and dipped at 33° perpendicular to the crest line, whereas the second, with dips of 20–25° oblique to the crest line, were deposited by accretion of wind ripple deposits on preexisting slopes as a result of local changes in flow velocity. These deposits formed the bulk of the dune sediments. Avalanche and wind ripple deposits accumulate on each side of the dune according to the primary wind direction in each season. Bristow and colleagues (2000) used ground-penetrating radar to study the internal structure of a sinuous simple linear dune (Fig. 3.21) in a seasonally bidirectional wind regime. They showed that the dune is composed of sets of unimodal cross-strata resulting from the lateral migration of the dune sinuosities, with deposition occurring on the outside of each curve. As the dune grows, the structure is composed of stacked sets of cross-strata with the eastward-dipping sets truncated by a westward-dipping set deposited as a sinuosity moves up the dune. Large linear dunes are additionally characterized by trough cross-strata deposited by migrating superimposed transverse dunes. GPR studies of a large (70-m-high) linear megadune in the northern Namib Sand Sea (Bristow et al., 2007) (Fig. 3.22) show that the dune is mainly comprised of large sets of cross-strata dipping toward the east, indicating net eastward lateral migration. On the western dune flank, there is a strong unconformity (superposition surface) between cross-strata within the dune that dip toward the east, which are truncated by those with an apparent dip toward the west. This is the result of superimposed dunes migrating along the dune flank and eroding older eastward-dipping sets of cross-stratification. On the eastern flank of the dune, sets of cross-stratification and their bounding surfaces dip roughly parallel to the dune surface, across which northerly migration of superimposed dunes occurs. Near the dune crest, small sets of cross-strata separated by reactivation surfaces document seasonal reversals of winds, similar to those identified by McKee (1982) in trench studies in the same area. Exposures of large, linear dunes created by sand quarrying in the UAE further demonstrate the

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

Dune Sediments

75

(a)

(b) scale for boxes 0

2m

20

10

0

m

0

20

40

60

80

S

N

(c) adv

anc

ed

33°

20–25°

irec

tion

20–25°

33 ° 20–25°

Figure 3.20 Sedimentary structures of linear dunes: (a) conceptual model of linear dune structure (modified from Bagnold, 1941); (b) modified from McKee and Tibitts (1964); (c) after Tsoar (1982).

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

76

Dune Morphology and Sediments

Figure 3.21 GPR profiles across a sinuous linear dune. On profile 1, structures are below the resolution of the GPR. Profiles 2 and 3 show planar sets of crossstratification with dominant dips toward the east. Profiles 4 and 5 show bidirectional dips formed by northward migration of the sinuous crestline. Profile 6 shows dominant dips to the east flanked by sets of trough cross-stratification from superimposed dunes. Modified from Bristow and colleagues (2000).

complexity of linear megadune structures, with multiple sets of high- and lowangle cross-strata separated by reactivation and/or superposition surfaces (Leighton et al., 2013b). Vegetated simple linear dunes are widespread in the Kalahari region of southern Africa and Australia. Those studied by Breed and Breed (1979) consisted mostly of medium-scale thin cross beds with dips commonly less than 20° occurring in tabular and wedge-shaped sets separated by near-horizontal bounding surfaces (Fig. 3.23a). Similar structures (Fig. 3.23b) have been observed in vegetated linear dunes in the Negev and the southwestern Kalahari and were interpreted to represent episodic extension of the dunes (Roskin et al., 2011b), as well as periods of extensive reworking of more-recent dune sediments (Telfer and Thomas, 2007), to expose older sands in the dune core. Buried soils are also present in some Australian vegetated linear dunes (Fitzsimmons et al., 2007; Hesse 2011).

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

0

5710 0 +/- 250 5730 +/- 360 57

WD 50

OSL ages

(c)

WD 50

Interpretation

(b)

GPR profile

(a)

50

5240 0 +/+/ 16 160

5520 5 520 0 +/- 270

WD 100

WD 100

100

52 +/+/ 4

WD 150

WD 150

150

2310 +/- 80

2410 +/- 110

2050 050 +/+ 80

34 +/+ 7

WD 200

WD 200

200

13 +/- 2 24 +/+ 3

10 +/- 3

WD 250

WD 250

250

300

Line position (m)

50 +/- 3

WE 200

WE 200

350

1790 +/- 70

900 +/- 30

400 +/- 20

140 +/- 10

WE 150

WE 150

400

450 +/- 20

340 +/- 20

WE 100

WE 100

450

290 +/- 20 350 +/- 20

WE 50

WE 50

550

5730–5240 years

2410–140 years

Less than 52 years

Key to dune age

500

Time (ns)

0 50 100 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850

Time (ns)

0 50 100 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850

Time (ns)

0 50 100 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850

Figure 3.22 GPR profile across a Namib complex linear dune showing (a) dipping reflections interpreted as sets of crossstratification and bounding surfaces within the dune. On the western flank of the dune, sets of cross-stratification formed by superimposed dunes truncate older east-dipping sets of cross-stratification. (b) Interpretation of dune evolution based upon the ground-penetrating radar profile and optically stimulated luminescence (OSL) ages. (c) OSL ages with locations of boreholes and OSL samples. From Bristow and colleagues (2007).

-5 Elevation (m)

0

5

10

15

20

25

30

35

40

45

50

55

60

-5 Elevation (m)

0

5

10

15

20

25

30

35

40

45

50

55

60

-5 Elevation (m)

0

5

10

15

20

25

30

35

40

45

50

55

60

78

Dune Morphology and Sediments scale for boxes

(a)

0

60 cms

30

0

m 0

100

200

300

E

(b)

400

500

600 W

(i)

(ii)

(iii)

Figure 3.23 Sedimentary structures in vegetated linear dunes: (a) Moenkopi Plateau, Arizona. From Breed and Breed (1979); (b) Negev Desert: (i) GPR profile, (ii) interpretation of GPR with OSL ages, (iii) detailed interpretation of crest area. From Roskin and colleagues (2014).

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

Dune Sediments

79

Star Dunes The sedimentary structures developed in star dune deposits are complex, as demonstrated by GPR profiles (Vriend et al., 2012). Investigations by trenching sample only a very small proportion of the total accumulation (Fig. 3.24a). McKee (1966; 1982) trenched star dunes in Saudi Arabia and the Namib Sand Sea and observed cross-cutting

(a) 0.5

m

0

W

E

S

N

(b)

Detail of crest and north flank

Figure 3.24 Star dune structures: (a) crest area of Namib star dune (drawn from photograph in McKee 1982); (b) GPR profile of star dune at Dumont dunes, California. From supplementary data in Vriend and colleagues (2012), sedimentary structure of large sand dunes: examples from Dumont and Eureka dunes, California. Geophysical Journal International 190, 981–992.

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

80

Dune Morphology and Sediments

sets of high-angle cross-strata near dune crests formed by winds from multiple directions, including seasonal wind reversals. Similar overlapping sets of avalanche face laminae were observed in reversing dunes at Kelso and Great Sand Dunes (Schenk et al., 1993; Sharp, 1966) and star dunes in China (Wang et al., 2005). Complex structures were also intersected on star dune arms in the examples investigated. McKee also observed that, in basal areas of star dunes in the Namib, most avalanche face laminae dip steeply to the northeast with low-angle strata dipping westward. This pattern was interpreted to mean that the star dune accumulated largely by winds from south to west and east to southeast directions. Similar asymmetry in the deposits of reversing dunes has been noted from Great Sand Dunes (Andrews, 1981), indicating that deposits developed by secondary winds are largely reworked by those from the primary wind direction. Ground-penetrating radar studies of dunes at Great Sand Dunes also indicate a complex series of trough-shaped bounding surfaces formed by scour in a reversing wind regime and/or by secondary lee-side airflow (Schenk et al. 1993). GPR transects at Dumont Dunes (Vriend et al., 2012) show that the southern flank of the dune is dominated by northward-dipping cross-strata, while the northern flanks are comprised of a complex series of grainflow and grainfall deposits, interspersed with multiple reactivation surfaces (Fig. 3.24b). Nielson and Kocurek (1987) and Wang and colleagues (2005) documented the widespread occurrence of low-angle wind ripple deposits on star dunes, where dune plinths form most of the depositional surfaces. Because of their low topographic position, such deposits have a higher potential for preservation in the rock record than the avalanche face deposits, which are confined to a small part of the crestal areas of the dune. Nielson and Kocurek (1987) argued that star dunes therefore may not leave a characteristic suite of deposits, which could explain why star dunes have rarely been recognized in the rock record (Clemmensen, 1987). Parabolic Dunes Parabolic dune structures described by McKee (1966) from White Sands and Ahlbrandt (1974) from Wyoming (Fig. 3.25) exhibit sets of cross-strata that dip downwind at angles of 22–34° (McKee, 1966). On the stoss slopes and dune arms, deposits are typically wind ripple laminae in 1–2-m-thick sets. Set boundaries tend to be convex downwind, reflecting undercutting of the dune nose by cross winds according to McKee, or low rates of deposition parallel to existing dune slopes. Bioturbation of dune structures also is common, reflecting the importance of vegetation in the development of these dunes. Zibars and Sand Sheets Sediments of sand sheets and zibars are dominated by wind ripple laminae, reflecting the dominant surface process in these environments. Deposits of zibars

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

Dune Sediments (a)

Approximate top of dune

81 Approximate top of dune

Intersection with side trench Top of cut

Concealed

4

SW

Concealed

2

4.5

3

NE

0

9

20

13

10

(b) 5

(c) Subsets

4

4

3

3

32

20

2

Concealed 2

34

1

0 feet 0

1

2

3

0 feet 0

5

4

1

2

22

17

31

1

16

3

5

4

6

(d) North wall of main trench

15 10

22

SE

NW

5 25

20

0 feet 0

5

10

15

20

Area A (including B & C)

40

45

50

55

60

65

70

75

80

Area D Bounding surface h nc tre W de 55∞ N

∞W 55 ch n tre

de

N

Si

∞E h 50 enc N tr n ai M

35

30

Si

Area Illustrated

25

Area Illustrated

∞E h 50 enc N tr n ai M

Selected cross stratum

Apparent dip, in degrees

Figure 3.25 Parabolic dune structures (after McKee, 1966).

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

22

82

Dune Morphology and Sediments

at the Algodones Dunes documented by Nielson (Nielson and Kocurek, 1986) consist entirely of lee-face wind ripple laminae deposited as the zibar migrates downwind. The wind ripple deposits consist of fine- and medium-sand laminae, with occasional coarse-sand laminae resulting from high-wind events, that alternate as both packages up to 10 cm in thickness and on a lamina-by-lamina basis (Fig. 3.26). Laminae dip at less than 15°, and packages of laminae are bounded by low-angle truncation surfaces resulting from changes in wind direction. (a)

Zibar

Interzibar Packages

(b)

Interzibar

Zibar

Interzibar Zibar

Figure 3.26 Sedimentary sequences in zibar, based on observations at the Algodones Dunes (after Nielson and Kocurek 1986): (a) amalgamated interzibar deposits that lack distinctive sedimentary structures or textures; (b) zibar/interzibar deposits. Note zibar wind ripple laminae with dips of less than 15°. Interzibar laminae dip at 5° or less.

https://doi.org/10.1017/9781108355568.004 Published online by Cambridge University Press

Dune Sediments

83

Figure 3.27 Sedimentary structures of a sand sheet area – note alternating coarse and fine laminae.

Interzibar deposits are thin ( 6 m/sec). These are the avalanches that contribute most to dune migration, suggesting that stronger winds

https://doi.org/10.1017/9781108355568.007 Published online by Cambridge University Press

126

Dune Processes and Dynamics

Figure 6.3 Observed and modeled patterns of lee-side deposition by grainfall settling flux decay with distance from brink. (a) Comparison between laboratory wind tunnel studies of Sutton and colleagues (2013) and Cupp and colleagues (2005) (K prefix); (b) field data collected by Hunter (1985). (c) Field data (points) and modeled curves (lines) of Anderson (1988) and McDonald and Anderson (1995). (d) Field data of Nickling and colleagues (2002) (black lines) compared to observations from Sutton and colleagues (2013).

and higher rates of sand movement contribute to dune migration in a nonlinear fashion as a result of higher values of brink saltation driven by the higher wind speeds, larger avalanche failures caused by the downslope shift in peak grain fall

https://doi.org/10.1017/9781108355568.007 Published online by Cambridge University Press

Dune Dynamics

δD (mm h–1)

100

127

u 0:4 (Fig. 8.5d). Both barchans and fingering mode linear dunes occur in areas of

130

Transverse Longitudinal Oblique

120 110

αl (°)

100 90 80 70 60 50 N=1 N=2 N = 1.5 N = 5

90

Finger dunes Barchans & finger dunes Barchans

80 70

αF (°)

60 50 40 30 20 10 0 0

20

40

60

80

100 120 140 160 180 θ (°)

Figure 8.4 Dune orientation with respect to the angle θ between the two winds for different transport ratios N and two conditions of sand availability given by σF and σI (from Gao and Narteau, 2015).

https://doi.org/10.1017/9781108355568.009 Published online by Cambridge University Press

168

Boundary Conditions for Dune Formation and Development

Figure 8.5 Patterns of dune morphology around the Tibesti Massif in relation to wind regime: (a) map of σF =σ I -values derived from wind data; (b) areas of barchans; (c) linear dunes breaking up into barchans; and (d) linear dunes elongating in the direction of the resultant sand flux (image credits: Google Earth) (from Gao and Narteau, 2015).

https://doi.org/10.1017/9781108355568.009 Published online by Cambridge University Press

Controls of Dune Orientation

169

intermediate values of σF =σI (Fig. 8.5). Comparison of dune morphology and orientations from image data with sand flux potential calculated from reanalysis wind data sets (Courrech du Pont et al., 2014; Hu et al., 2021) shows a good correlation between the orientations predicted by the model and the observed trends (Fig. 8.6). 8.3 Superimposed Dunes and Complex Patterns Large dunes (megadunes, draa) are almost always compound and complex forms that are characterized by the development of smaller dunes superimposed upon them. Typically, the superimposed dunes are of crescentic form (Fig. 8.7). Analyses of dune patterns indicates that two or more statistically distinct populations of dunes occur in areas of compound and complex dunes (Ewing et al., 2006). In the Algodones dune field (Derickson et al., 2008), these comprise the primary crescentic dunes and the smaller crescentic dunes superimposed on them (Fig. 8.8a). In the central Namib Sand Sea, there are three populations of dunes, comprised of the main S–N linear dunes, crescentic dunes on their flanks, and small oblique linear dunes extending across interdune areas (Fig. 8.8b). For dune areas in Arabia and the Namib, Glennie (1970) and Besler (1980) suggested that compound and complex dunes are a product of Quaternary climatic changes, in which the large dunes were formed during periods of strong winds during Pleistocene glacial periods. Weaker modern winds could only form the small dunes that are superimposed on their flanks. Support from such a hypothesis comes from complex linear dunes in the Akchar and adjacent sand seas of Mauritania. These dunes are composite forms, with a core formed in the last glacial period, and superimposed crescentic dunes that are late Holocene in age (Kocurek et al., 1991; Lancaster et al., 2002). The widespread existence of compound and complex dunes in clearly active modern sand seas (e.g. Algodones compound crescentic dunes; Kumtagh raked linear dunes) suggests, however, that many superimposed bedforms are the product of contemporaneous aeolian environments. Although the sample size is small, data on the width and spacing of simple, compound, and complex crescentic and linear dunes indicate that their mean size is statistically significantly different (Lancaster, 1988b). This suggests that a minimum dune size must be reached before superimposed dunes can develop (Elbelrhiti et al., 2005). In the Namib Sand Sea, simple crescentic dunes have a spacing of less than 500 m, whereas the spacing of compound crescentic dunes is more than 500 m (Fig. 8.9). The recognition of two modes for dune orientation provides a framework for understanding complex dune patterns and superimposition of dunes of the same or

https://doi.org/10.1017/9781108355568.009 Published online by Cambridge University Press

Figure 8.6 Main dune types in the western Sahara Desert: (a) crescentic dunes, (b) star dunes, (c) asymmetrical barchan dunes, (d) linear dunes (bed instability mode), (e) oblique linear dunes, (f) linear dunes (fingering mode), (g) barchan chains, and (h) asymmetric linear dunes (bed instability mode). Insets in (a–h) show the local sand flux roses. The abbreviation RDD represents the predicted resultant drift direction. Red double-arrows show the predicted orientation of the bed instability mode. Black arrows show the predicted orientation of the fingering mode. From Hu and colleagues (2021). https://doi.org/10.1017/9781108355568.009 Published online by Cambridge University Press

Controls of Dune Orientation

171

Figure 8.7 Example of superimposed dunes on linear dunes, Northern Rub al Khali, UAE. Scale bar is 1 km.

different types. Although in some cases dunes on different trends may result from different periods of their formation (Lancaster et al., 2002), in many cases, it appears that the superimposed dunes coexist with larger linear forms and are an integral part of their sedimentary structures (e.g. in the Namib Sand Sea [Bristow et al., 2007]) and their dynamics e.g. raked linear dunes in the Kumtagh Sand Sea (Lü et al., 2017) and contribute to the growth of the larger dunes (Andreotti et al., 2009). Superimposed dunes are frequently of crescentic form and result from the bed instability mode, as the slopes of large linear, crescentic, and star dunes provide a supply of erodible sand (Courrech du Pont et al., 2014). In the case of the Kumtagh, crescentic ridges develop on the lee side of the main crest line and migrate parallel to the dune trend at a mean rate of 5 m/yr. The main linear dunes occur in supply-limited conditions and extend parallel to the resultant sand flux in the fingering mode, whereas the superimposed dunes form in the bed instability mode because of the unrestricted supply of sand from the flanks of main linear dunes (Lü et al., 2017). Similarly, in areas of complex linear dunes in the Namib Sand Sea, the three populations of dunes identified by Lancaster (1989b) and Ewing

https://doi.org/10.1017/9781108355568.009 Published online by Cambridge University Press

172

Boundary Conditions for Dune Formation and Development

(a) Main compound crescentic dunes

Superimposed crescentic dunes

% observations

(b)

40

20

0 50

100

500

1000

2000

Spacing (m)

Figure 8.8 (a) Spacing of superimposed and main crescentic dunes, Algodones Dunes. Plot after Derickson and colleagues (2008). Scale bar is 1 km. (b) Spacing of main S–N linear dunes and superimposed crescentic dunes, central Namib Sand Sea (after Lancaster 1988b). Images from Google Earth. Scale bar is 1 km.

and colleagues (2006) represent both fingering and bed instability modes in a multidirectional wind regime, with the crescentic dunes that migrate along the flanks of the main linear dunes developing in a bed instability mode and the primary and interdune corridor-crossing linear dunes forming in a fingering mode (Fig. 8.8b).

https://doi.org/10.1017/9781108355568.009 Published online by Cambridge University Press

Controls of Dune Orientation

173

% observations

40

20

0

50

100 500 1,000 2,000 Spacing (m)

Figure 8.9 Comparison of the spacing of superimposed and main dunes for Namib compound crescentic dunes. Image – Google Earth.

Standard deviation (φ)

1.00

S–N linear dunes East flank barchanoid dunes Corridor crossing linear dunes

0.80

0.60

0.40

0.20

0.00 1.50

2.00

2.50

3.00

Mean grain size (φ)

Figure 8.10 Grain size and sorting characteristics of elements of a complex dune pattern, central Namib Sand Sea.

https://doi.org/10.1017/9781108355568.009 Published online by Cambridge University Press

174

Boundary Conditions for Dune Formation and Development

8.4 Influence of Sediment Characteristics Wilson (1972) suggested that different elements of a complex dune pattern are composed of sands with different grain sizes. Grain size, by controlling the threshold velocity for sand movement, may determine the effective wind regime and therefore strongly influence dune alignments. However, there is little evidence to support such a hypothesis. Where grain size and sorting data from different elements of a complex dune pattern exist (Fig. 8.10), the differences are not significant in terms of threshold velocities and therefore effective sand-transporting wind regimes (Lancaster, 1989b).

https://doi.org/10.1017/9781108355568.009 Published online by Cambridge University Press

9 Controls of Dune Size and Spacing

9.1 Dune Height and Spacing Relationships Most sand seas show clear patterns of dune size and spacing, and the correlations that exist between dune height, width, and spacing suggest a high degree of selforganization. However, the nature of the controls on dune size and spacing are not well understood. The general form of the relations between dune height and spacing can be expressed by a power function (Lancaster, 1988a) DH ¼ c DSn where DH is dune height, DS is dune spacing, c is a constant, and the exponent n is a measure of the rate of change of the dependent variable relative to the rate of change of the independent variable (the slope of the regression line). Values of the exponent n range between 0.52 and 1.72 from one sand sea to another as well as from one dune type to another in the same sand sea (Fig. 9.1). An exponent of unity indicates that dune height increases at the same rate relative to dune spacing. Thus, a given amount of sand can be formed into a few widely spaced dunes or many smaller closely spaced dunes. Such a model was advocated for Simpson Desert linear dunes by Twidale (1972). Where the exponent is greater than unity (e.g. complex linear dunes in the Namib Sand Sea and star dunes in the Gran Desierto), dune height increases more rapidly than dune spacing, indicating a tendency for vertical growth of the dunes. Exponents less than unity indicate that dune height increases less rapidly than dune spacing. In these examples (e.g. compound linear dunes in the Namib Sand Sea, linear dunes in the Great Sandy Desert, crescentic dunes in the Gran Desierto) dune size may be limited by the availability of sand. Relatively small dunes for their spacing have also been noted from areas undergoing net sand loss (Mainguet and Chemin, 1983), and from areas of damp interdunes at Padre Island (Kocurek et al., 1992), reflecting a restricted supply of sand for dune building. The relationships between dune height and 175

https://doi.org/10.1017/9781108355568.010 Published online by Cambridge University Press

176

Boundary Conditions for Dune Formation and Development 500 1

2

200 100

Height (m)

50 3

4 5

9 10

6 8

7

5

1 10

100

1000 2000 3000

Spacing (m)

Figure 9.1 Relationships between dune height and spacing. 1 = Namib Sand Sea star dunes; 2 = Namib Sand Sea complex linear dunes; 3 = Namib Sand Sea compound linear dunes; 4 = Namib Sand Sea simple and compound crescentic dunes; 5 = Simpson Desert simple linear dunes; 6 = Great Sandy Desert simple linear dunes; 7 = Skeleton Coast dune field crescentic dunes; 8 = Gran Desierto simple crescentic dunes; 9 = Southwestern Kalahari simple linear dunes. Data for Australian sand seas from Wasson and Hyde (1983b).

spacing therefore appear to reflect both the availability of sand for dune construction and wind regime characteristics, which determine whether dunes will tend to accrete vertically (star dunes and many complex dune varieties), migrate (simple crescentic dunes), or extend (many simple and compound linear dunes). 9.2 Grain Size Effects Wilson (1972) found a clear relation between dune and compound and complex dune (megadune, draa) spacing and the grain size of the coarse 20th percentile of dune crest sands (P20) in three northern Saharan sand seas (Fig. 9.2). Similarly, Lancaster (1988b) documented a correlation between crescentic dune spacing and grain size in the Skeleton Coast, Namib, and Gran Desierto sand seas (Fig. 9.3). However, the spacing of barchan dunes in the western Sahara (Elbelrhiti, 2012; Ould Ahmedou et al., 2007) and linear dunes in the Namib, Australia, and Kalahari do not show any relation to grain size (Thomas, 1988; Wasson and Hyde, 1983b).

https://doi.org/10.1017/9781108355568.010 Published online by Cambridge University Press

Controls of Dune Size and Spacing

177

Ripples 2.0 P20 (mm)

Dunes 1.0 Draa .50 0.250 0.125 .01

.04

.16

.64

2.56

10

40

160

640

2560

Wavelength (m)

Figure 9.2 Relations between dune and draa (megadune) spacing and P20 for Saharan sand seas (redrawn from Wilson, 1972).

Figure 9.3 Relations between dune spacing and P20 for crescentic dunes in selected areas.

Various explanations have been put forward to explain the grain-size–dunespacing relationships in areas of crescentic dunes. All involve the effect of grain size on threshold shear velocity and/or aerodynamic roughness length. Wilson (1972) argued that the spacing of dunes and draas was related to the size of secondary flow elements, which were in turn related to the threshold wind-shear velocity required to transport the sand from which they were composed. Differences in sand transport rates between the base and crest of the stoss slope of crescentic dunes are affected by particle size and are accentuated in dunes composed of coarse sands with a higher threshold velocity for transport. This may produce dunes with a low aspect ratio (flatter profiles) and wider crest-to-crest spacing (Lancaster, 1985a; Tsoar, 1985).

https://doi.org/10.1017/9781108355568.010 Published online by Cambridge University Press

178

Boundary Conditions for Dune Formation and Development

Figure 9.4 Plots of P20 and grain diameter (D) versus bedform spacing for dunes. Solid curves are model predictions assuming no positive correlation between grain size and excess shear velocity. Data from Wilson (1972) and Lancaster (1988c). In addition to grain size and shear velocity, predicted dune spacing is a sensitive function of maximum ripple slope, δ. From Pelletier (2009).

Pelletier (2009) suggested, on the basis of a numerical model of ripple and dune development, that crescentic dune height and spacing scale with the aerodynamic roughness of a rippled sand bed, which is in turn determined by grain size and excess shear velocity (u  ut , where ut is a function of grain diameter). In addition to grain size and shear velocity, predicted dune spacing is a sensitive function of maximum ripple slope. The model predictions scale quite well with observed dune-spacing/grain-size relationships (Fig. 9.4). As variant on the Pelltier (2009) approach, Howard and colleagues (1978) suggested that possible controls of dune size include upwind roughness, with dune size increasing with the size of fixed roughness elements such as rocks or vegetation. 9.3 Airflow Effects Many investigators have suggested that the spacing of dunes is related to the scale of secondary flow circulations or to the dimensions of the zone of flow disturbance downwind of dunes (e.g. Baddock et al., 2007; Folk, 1976b; Wilson, 1972). However, it is unclear if this is the product of airflow dynamics over dune landscapes or a fundamental control of dune morphology. Others (e.g. Yalin, 1972;

https://doi.org/10.1017/9781108355568.010 Published online by Cambridge University Press

Controls of Dune Size and Spacing

179

Helical flow in boundary layers

Ground flow lines

Linear dunes

Figure 9.5 Helical roll vortex model for linear dune spacing (redrawn from Cooke et al., 1993).

Cooke et al., 1993) suggest that the scale of natural atmospheric turbulence is a major influence on dune size and spacing. The parallelism and regular spacing of linear dune systems has been noted by many workers and has given rise to hypotheses that their spacing is the product of organized vortex flow in the planetary boundary layer (e.g. Bagnold, 1953; Glennie, 1970) (Fig. 9.5). Hanna (1969) compared data on the spacing of linear dunes with the expected dimensions of helical roll vortices in subtropical desert areas, concluding that their probable wavelength was 2–6 km (two to three times the convective layer thickness of 1–2 km). However, the observed scale of helical roll vortices (Kelley, 1984) is in many cases much larger than average dune spacing, suggesting that linear dunes are not the product of such atmospheric motions. Further, although identified from meso-scale atmospheric boundary layer studies, evidence for the existence of helical roll vortices in linear dune landscapes is slight (Tseo, 1990) and, as discussed previously, process studies and numerical models indicate that formative winds are typically oblique to the dune trends. 9.4 Controls of Dune Size Observations and numerical models indicate that dune size tends to increase logarithmically, with dune growth rates decreasing as dunes become larger (Eastwood et al., 2011). This suggests some overall limit to dune size. Andreotti

https://doi.org/10.1017/9781108355568.010 Published online by Cambridge University Press

180

Boundary Conditions for Dune Formation and Development

and colleagues (2009) have suggested that dune height is limited by the average height of the mixing layer in the planetary boundary layer (Fig. 9.6a). They argue that a resonance between dune topography and the capping layer creates gravity waves that bound and limit dune growth. This elegant hypothesis has been challenged by Gunn and colleagues (2022) who could find no evidence for a link between boundary layer height and dune size (Fig. 9.6b). Their numerical modeling of dune growth suggests that dunes may grow indefinitely in principle; but growth slows with increasing size (Fig. 9.7) and may ultimately be limited by sand supply. The global data set of Gunn and colleagues (2022) indicates that dune height is inversely related to sand transport directionality so that dunes in areas with low RDP/DP ratios (e.g. star dunes) are relatively taller than those elsewhere. Lower aspect ratios with increasing dune spacing are associated with more unidirectional (high RDP/DP ratio) sand-flux regimes, whereas large aspect ratios correspond to low directionality. These observations suggest that highly variable wind regimes create higher dunes, while more unidirectional winds create lower dunes. Empirical evidence from the Namib Sand Sea and other areas supports this model. Thus, star dunes in the Namib occur in low-energy, complex wind regimes, compared to linear dunes, which occur in higher-energy, bimodal wind regimes. 9.5 Dune Spacing as a Product of Self-Organization and Pattern Coarsening A characteristic feature of the self-organizing nature of dune systems is the emergence of a quasi-equilibrium spacing caused by the interaction of dunes of different sizes and rates of migration. These interactions (Fig. 9.8) may range from constructive to regenerative (Kocurek et al., 2009). Constructive interactions involve merging, lateral linking, cannibalization, and transfer of sediment between dunes and create fewer, larger dunes. Such pattern coarsening leads to a more ordered and stable pattern with fewer interactions between dunes. Regenerative interactions including splitting, calving, and defect creation result in regression to a prior, less organized state. The processes of dune interactions are best documented from areas of barchans (Elbelrhiti et al., 2008) and crescentic dune systems (e.g. White Sands, Ewing and Kocurek, 2010a), where they lead to a quasi-regular pattern of crescentic ridges, but are less studied and poorly understood in the case of linear dune systems (Telfer et al., 2017). The regular spacing of linear dunes has been investigated in numerical simulations using a cellular automaton approach (Gadal et al., 2020a). They found that with dunes elongating in the fingering mode, extension does not produce a pattern with a specific wavelength. Regularly spaced extending linear dunes therefore appear to be influenced by periodicities in the boundary conditions. These may include vegetation, preexisting dunes, or topography.

https://doi.org/10.1017/9781108355568.010 Published online by Cambridge University Press

Controls of Dune Size and Spacing (a) 4

181

Star dune Longitudinal dune Transverse dune

λ (km)

3

2

1

0 0

1

3

2

4

δΘ / γ (km)

(b) rHx = –0.041

Dune Wavelength x (km)

3

H=x

2

1

0

0

1

2

3

Mixed Layer Height H (km)

Figure 9.6 (a) Measured giant dune wavelength λ as a function of the characteristic mixing height, computed from the annual ground temperature variations δθ. The solid line is the identity: λ¼δθ=γ. From Andreotti and colleagues (2009); (b) measured average dune heights (bars – standard deviation) and boundary layer height derived from Calypso satellite observations. From Gunn and colleagues (2021).

https://doi.org/10.1017/9781108355568.010 Published online by Cambridge University Press

182

Boundary Conditions for Dune Formation and Development

Figure 9.7 Numerical experiments of dune growth. (a) Dune height time-series for ReSCAL experiments. Line colors correspond to experiments shown in (b), a snapshot of the experiment at t = 162 years shown to define the horizontally periodic domain; W = H = 522 m. (b) Planform snapshots of each experiment at the final timestep t = 1,624 years. The number of flux directions FN is shown, as are the flux vectors for each experiment. (c) Wavelength x against height z for each experiment coarsening over time; bounding power laws from the observed data. (d) Dune height z against celerity (i.e. migration speed). From Gunn et al. (2021, 2022).

9.6 Influence of Sediment Source and Dune Field Geometry Ewing and Kocurek (2010b) have shown that the geometry of the source of sediment for an area of dunes plays an important role in determining dune morphology. Point and line sources provide a location for the development of dunes downwind in which small dunes develop close to the source and merge and grow as they migrate downwind, as in barchan corridors in Mauritania and western Sahara, and crescentic dunes at White Sands. Similar merging and pattern coarsening occurs in linear dune landscapes downwind of fluvial or lacustrine sources in

https://doi.org/10.1017/9781108355568.010 Published online by Cambridge University Press

Controls of Dune Size and Spacing

183

Merging

Lateral linking

Defect migration

Repulsion

Termination creation

Figure 9.8 Varieties of dune interactions as part of self-organization of dune fields resulting in changes in dune size and spacing.

Australia and the Kalahari (Fig. 9.9). Where the sediment source is planar, as in situations where vegetated dunes and sand sheets are reactivated, dunes emerge

https://doi.org/10.1017/9781108355568.010 Published online by Cambridge University Press

184

Boundary Conditions for Dune Formation and Development

1 km

Figure 9.9 Self-organization of linear dunes downwind of a sediment source in a lunette dune, southwestern Kalahari. Google Earth image.

Figure 9.10 Measured average crest spacing vs. dune-field area. The solid line shows the best fit to the data. The gray dots show the spacing to paleo–dune-field area relationship for the Algodones Dunefield, Gran Desierto, and Agneitir Sand Sea. Dashed arrows correspond to the modern areas of these dune fields (from Ewing and Kocurek, 2010).

https://doi.org/10.1017/9781108355568.010 Published online by Cambridge University Press

Controls of Dune Size and Spacing

185

throughout the area, at locations determined in part by disturbance (Barchyn and Hugenholtz, 2013a). Dune field and sand sea geometry also appear to set limits on dune size and spacing. Ewing and Kocurek (2010b) determined that dune spacing increases with dune field area for a wide range of environments, dune ages, and dune types (Fig. 9.10). Model results indicate that the area of a dune field limits the maximum spacing that can occur. As a result, dune size is strongly constrained by the geometry of the dune field for small dune areas, whereas dunes can increase in size and spacing where space is available, in larger dune fields and sand seas.

https://doi.org/10.1017/9781108355568.010 Published online by Cambridge University Press

10 Response of Dune Systems to Changing Boundary Conditions

10.1 Dune Activity Desert dunes occur in a range of states of activity: (1) active, in which sand transport is widespread and vegetation cover is sparse, leading to dunes that migrate, extend, or accrete vertically in response to seasonal and annual changes in current wind regimes; (2) partially active, characterized by spatially or temporally restricted sand transport (e.g. only the crest areas of dunes may experience sand transport), and interdune areas are largely stabilized by vegetation; (3) inactive, in which sand transport is absent or at a very low level. Dune morphology is distinct, but the perennial vegetation cover often exceeds 20%; (4) degraded, in which dune crests are indistinct, and the vegetation cover of grasses, shrubs, and trees is continuous (Lancaster and Hesse, 2016). In the latter areas, dunes are likely the products of past climatic and sediment supply conditions and may be out of equilibrium with modern wind regimes. It should be noted that these categories (Fig. 10.1) are not exclusive, nor are they constant in time and space: degrees of activity and the spatial expression thereof may change with seasons and variations in climate (e.g. drought cycles) (Bullard et al., 1997; Fisher and Hesse, 2019; Thomas and Leason, 2005). The state of activity is determined by the balance between the wind energy required to transport sand (erosivity) and the resistance to sand movement provided by vegetation cover and surface crusting (erodibility). In addition, vegetation may act to trap sand, thus enhancing deposition. There are three main approaches to analyzing the impact of the variables that determine dune activity: (1) empirical studies of the response of geomorphic systems to modern, short-term climatic changes (such as El Niño events and extended regional droughts) (Lancaster, 1997). Observational data enables threshold conditions to be identified and well-constrained process– response models (Fig. 10.2) to be developed from these scales of natural climate variability, which constitute natural experiments (Muhs and Holliday, 186

https://doi.org/10.1017/9781108355568.011 Published online by Cambridge University Press

Response of Dune Systems to Changing Boundary Conditions

187

Figure 10.1 Examples of the activity state of dunes, following Hesse and Lancaster (2016). Examples are from crescentic dunes.

1995). Such responses can also be validated against climate change over the historic or pre-instrumental record (e.g. Little Ice Age, Medieval Warm Period), and in some cases by reference to the late Holocene record

https://doi.org/10.1017/9781108355568.011 Published online by Cambridge University Press

188

Boundary Conditions for Dune Formation and Development Increased temperature

Increased gustiness

Decreased rainfall

Increased evapotranspiration

Increased variability of flood flows

Decreased vegetation cover

Formation of wide, braided sandy channels

Change to intermittent flow

Increased sand supply

Increased aeolian activity

Figure 10.2 Conceptual model for response of aeolian systems to climate change and sediment supply (modified from Muhs and Holliday, 1995).

(Forman et al., 2005); (2) development of climatic indices for the relationship between dune activity and climatic variables based on the principle that sand mobility is proportional to the wind speed above threshold and inversely proportional to effective precipitation, which determines soil moisture and vegetation cover; (3) conceptual and numerical models that incorporate both ecological and geomorphic processes and their complex, nonlinear interactions, including the effects of vegetation cover on sand-transport rates and the probability of erosion or deposition occurring in relation to vegetation cover and growth rate. 10.2 Empirical Studies of Responses to Changing Boundary Conditions 10.2.1 Late Holocene and Historical Record of Dune Activity Dune systems in semiarid, desert margin areas are sensitive to climate change and variability, especially to periods of extended drought and reduced precipitation. On the Great Plains of the USA and Canada, many dune fields are partially or completely stabilized by vegetation in current climatic conditions (Johnson et al., 2020; Muhs and Wolfe, 1999). Reports from early explorers and surveys of the Great Plains of the USA indicate, however, much more widespread dune activity, probably as a result of drought conditions, in the nineteenth century (Muhs and Holliday, 1995). In the late 1700s, a decade-long drought on the

https://doi.org/10.1017/9781108355568.011 Published online by Cambridge University Press

Response of Dune Systems to Changing Boundary Conditions

189

southern Prairies of Canada resulted in an estimated 10–20% increase in the area of active dunes and the formation of barchanoid dunes in the Great Sand Hills, Saskatchewan (Hugenholtz and Wolfe, 2005; Wolfe and Hugenholtz, 2009; Wolfe et al., 2001). These dunes are now largely stabilized in warmer and less arid conditions. Stabilization was accompanied by a change in dune morphology from barchanoid to parabolic over a period of 70 years between 1810 and 1880 (Wolfe and Hugenholtz, 2009). The sedimentary record of the last 2,000 years provides evidence for multiple periods of sand accumulation dated by luminescence methods in currently vegetation-stabilized dune fields. In areas such as the Great Plains of the USA and Canada, these are associated with periods of extreme and persistent drought conditions (megadroughts). The linkages between periods of dune activity and drought are supported by studies of dune-system response to historical and instrumental records of drought and rainfall deficits (Muhs and Holliday, 1995; Redsteer, 2020; Wolfe et al., 2001). The dune activity initiated by the late 1700s drought was one of a series of periods of dune activity that occurred in the past 2,000 years on the Great Plains (Forman et al., 2001; Miao et al., 2007). For example, multiple periods of dune activity (around 1,390, 670, 470, 240, 140, and 70 years ago) have been recognized in the area of the Nebraska Sand Hills during the past 1,500 years and have been correlated with evidence of drought episodes elsewhere in the region (Forman et al., 2005; Sridhar et al., 2006). The most significant of these periods of activity were coincident with a widespread and extended drought (megadrought) that occurred in the sixteenth century. Periods of sand mobility also appear to be related to a combination of climatic conditions (e.g. severe droughts) and disturbance (grazing, fire), indicating the complex response of this vegetation-stabilized dune field to regional and local stressors (Buckland et al., 2019). In addition, sand accumulations related to the 1930s drought can be recognized (Bolles et al., 2017). Elsewhere in this region, a major period of linear dune formation occurred 700–1,000 years ago, coincident with a period of widespread drought during the Medieval Warm Period (Mason et al., 2004; Sridhar et al., 2006). During this episode of dune formation, winds shifted direction so that the summertime southeasterly flow of humid air was replaced by dry southwesterly flow (Fig. 10.3). At Great Sand Dunes, Colorado, periods of aeolian deposition in the late Holocene can be correlated with decadal-scale drought episodes recognized in tree ring records (Forman et al., 2006), suggesting an overall control of dune activity by rainfall and vegetation cover in this area (Fig. 10.4).

https://doi.org/10.1017/9781108355568.011 Published online by Cambridge University Press

Northerly vector

Northerly vector

e dun ntic e c s Cre trend

Tre nd

of l

ine

ar d

une

s

Southwesterly vector

Southerly vector Modern winds and dune trend

Medieval warm period

Figure 10.3 Changes in wind regime on the Great Plains during formation of linear dunes in the Medieval Warm Period inferred from dune trends (redrawn from Srihdar et al., 2006).

40 36 32 1μ

28



-0.5 -1.5 -2.5 -3.5

-2 PDSI

4 3 2 1 600

700

800

900

1000 1100

1200

1300 1400

1500 1600 1700

1800

1900 2000

Years AD

Figure 10.4 Relationships between records of inferred precipitation and summer Palmer Drought Severity Index (PDSI) from tree ring data, and periods of OSLdated dune activity at Great Sand Dunes National Park, Colorado (after Forman et al. 2006). Vertical shaded bars indicate periods of possible correspondence between aeolian and tree ring records; arrows indicate periods of drought indicated by tree ring records, but for which there is no equivalent aeolian record.

https://doi.org/10.1017/9781108355568.011 Published online by Cambridge University Press

Number of OSL ages

1.5 0.5

Summer PDSI

24 3.5 2.5

400

Annual precipitation (cm)

44

Response of Dune Systems to Changing Boundary Conditions

191

10.2.2 Decadal Scale Changes Long meteorological records, coupled with repeated observations of the state of dune systems using aerial photographs and satellite images have provided valuable information on how dune systems respond to climate variability on decadal scales. In China, dune systems cover as much as 23% of the arid and semiarid areas of northern and western China. Significant variations (Fig. 10.5) in the areas covered by dunes that are classified as mobile (active), semi-anchored (partially stabilized), or anchored (stabilized by vegetation)

A

Mid 1970s

Mid 1980s

2000

Extent of desertification (%)

100 90 80 70 60 50 40 30 20 10 0 A

B

C

D

E

R

G

H

I

J

K

L

M

Dune Area 1999

B

2004

450,000 400,000

Area (km2)

350,000 300,000 250,000 200,000 150,000 100,000 50,000 0 Mobile dunes

Semi-anchored

Anchored

Figure 10.5 Changes in dune activity status in northeast Chinese dune fields (from data in Wang et al., 2009): (a) percent of active dunes (interpreted as “areas of desertification” in different dune areas); (b) total area of different dune activity states at different times.

https://doi.org/10.1017/9781108355568.011 Published online by Cambridge University Press

192

Boundary Conditions for Dune Formation and Development

have occurred since the 1950s over much of the Chinese arid zone (Wang et al., 2006; Wang et al., 2008a). The causes of these changes in dune activity status have been much debated, with some authors ascribing them to human activity, including overgrazing or land use change, and others favoring climatic variability as the main factor influencing changes in the areas of active or stabilized dunes (Wang et al., 2006). Wang et al. (2006) have analyzed climatic data from the 1950s in conjunction with data on areas of dunes of different status in northeastern China. They conclude that the period from the 1950s to the 1980s was characterized by higher wind speeds, compared to the period from the 1980s to the present. This resulted in sand drift potentials (Fryberger and Dean, 1979) that were two to five times greater in the 1950–1980 period compared to 1980–present. As a result, vegetation cover in many dune areas doubled so that some areas of active dunes became partially vegetated, while in other areas sand transport was restricted to dune crests. Many partially vegetated dunes became completely stabilized by vegetation. From 1999 to 2004, the area of active dunes and partially vegetated dunes decreased by 3.7% and 11.4% respectively, whereas the areas of stabilized (vegetated) dunes increased by 13.8%. In the Takla Makan desert, for example, dune mobility was higher in the 1960s and the mid-1980s compared to the 1990s (Wang et al., 2004). Mason et al. (2008) confirm the large decrease in wind energy in the Mu Us, Otindag, and Horqin deserts since the 1970s, but observe that this change did not result in widespread stabilization of dunes, except in the eastern Mu Us area. In fact, areas of active dunes remained constant or even slightly increased over the past 40 years. They conclude that wind energy may not be as important to changes in dune activity as has been suggested and that changes in human land use in this area may have a more significant role. At Great Sand Dunes (Colorado), Marîn et al. (2005) have documented migration rates of parabolic and barchan dunes over a period of 70 years using satellite images and aerial photographs (Fig. 10.6). They found that migration rates were up to six times higher in drought episodes compared to intervening wetter periods. More than 50% of the movement of parabolic dunes occurred during periods of drought that occurred between 1936 and 1941; 1953 and 1966; and 1998 and 1999. They identified a threshold condition with a Palmer Drought Severity Index (PDSI) value of less than –2 (equivalent to a reduction in summer and fall precipitation of 25%) for increased dune activity. In the southwestern Kalahari, the 1970s were characterized by higher rainfall and somewhat cooler temperatures and lower wind energy, whereas the 1980s were affected by severe regional drought, which was accompanied by generally higher

https://doi.org/10.1017/9781108355568.011 Published online by Cambridge University Press

Response of Dune Systems to Changing Boundary Conditions 35

193

4 PDSI

30 30

32

A

2 0

Average Rate m/yr

-2 -4

25

-6 2000

1995

1990

1985

1980

1975

1970

1965

1960

1955

1950

20

16

16

15 15

10

8

7

7

5

5 5

5

5

4

4

0

98–99

95–98

88–95

83–88

81–83

79–81

66–79

66–81

55–66

53–55

41–55 41–53

38–41

36–38

25

B

22

Average Rate m/yr

20

15

10 9

10

9

8

7 6 5

4

3 2

3

4 2

1

5

4 2

11

8

1

1

0 98–99 88–99 88–98 83–99 83–88 81–88 79–88 79–81 75–83 75–81 75–79 66–81 66–79 66–75 57–83 53–81 53–66 41–81 41–66 41–57 41–53 36–41

Figure 10.6 Variations in the rate of dune migration at Great Sand Dunes, Colorado, compared to the Palmer Drought Severity Index (PDSI) (after Marîn et al., 2005).

wind speeds (Bullard et al., 1997). The result was a significant variation in indices of dune surface activity, as modelled by the dune mobility index (Lancaster, 1988c) (Fig. 10.7). These changes are confirmed by comparison of the area of active dune sand on Landsat images from a drought period (1984) and a wetter period (1993) (Thomas and Leason, 2005). Using an empirically derived vegetation cover threshold for active sand movement of 14% (Wiggs et al., 1995), Thomas and Leason (2005) showed that the area affected by active sand movement varied between 16% (1984) and 6% (1993) for areas not affected by grazing. These results highlight the spatial dynamics of climate variability on dune systems and show that even if large areas have a low vegetation cover in dry periods, they can recover over a short period of time if human pressures are not excessive.

https://doi.org/10.1017/9781108355568.011 Published online by Cambridge University Press

194

Boundary Conditions for Dune Formation and Development Twee Rivieren

Keetmanshoop

Upington

Tsabong

Van Zylsrus

Kuruman

Gobabis

400

350

300

Fully active dunes

Mobility Index

250

200

150

Crests active

100 Partial activity 50 Inactive dunes 0 1955

1960

1965

1970

1975

1980

1985

1990

1995

Figure 10.7 Temporal changes in potential dune sand mobility in the southwestern Kalahari (data from Bullard et al., 1997).

Crest areas of linear dunes in the Simpson Desert of Australia vary in their activity in a complex manner related to periods of drought, disturbance by fire, and periods of increased rainfall (Fisher and Hesse, 2019). Activation of dune crests is promoted by fire and extended periods of drought that result in a reduction of vegetation cover. Increased crestal activity lags the initiating event, and four years or more of low vegetation cover (