Physics and modelling of wind erosion

Strong wind-erosion events, such as severe duststorms, may threaten human lives and cause substantial economic damage.The physics of wind erosion is complex, as it involves atmospheric,

Physics and Modelling of Wind Erosion

ATMOSPHERIC AND OCEANOGRAPHIC SCIENCES LIBRARY

VOLUME 3

Editors

Lawrence A Mysak, Department of Atmospheric and Oceanographic Sciences,

McGill University, Montreal, Canada

Kevin Hamilton, International Pacific Research Center, University of Hawaii,

Honolulu, HI, U.S.A.

Editorial Advisory Board

L Bengtsson Max-Planck-Institut für Meteorologie, Hamburg, Germany

A Berger Université Catholique, Louvain, Belgium

J.R Garratt CSIRO, Aspendale, Victoria, Australia

G Geernaert DMU-FOLU, Roskilde, Denmark

J Hansen MIT, Cambridge, MA, U.S.A

M Hantel Universität Wien, Austria

H Kelder KNMI (Royal Netherlands Meteorological Institute),

De Bilt, The NetherlandsT.N Krishnamurti The Florida State University, Tallahassee, FL, U.S.A

P Lemke Alfred-Wegener-Institute for Polar and Marine Research,

Bremerhaven, Germany

P Malanotte-Rizzoli MIT, Cambridge, MA, U.S.A

D Randall Colorado State University, Fort Collins, CO, U.S.A.J.-L Redelsperger METEO-FRANCE, Centre National de Recherches

Météorologiques, Toulouse, France

A Robock Rutgers University, New Brunswick, NJ, U.S.A

S.H Schneider Stanford University, CA, U.S.A

G.E Swaters University of Alberta, Edmonton, Canada

J.C Wyngaard Pennsylvania State University, University Park, PA, U.S.A

Physics and Modelling

Dr Yaping Shao

ISBN 978-1-4020-88 - eISBN 9781402088

-Library of Congress Control Number: 2008932207

All Rights Reserved

c

2008 Springer Science + Business Media B.V.

No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose

of being entered and executed on a computer system, for exclusive use by the purchaser of the work Printed on acid-free paper

Preface 1

Wind erosion occurs in many arid, semiarid and agricultural areas of theworld It is an environmental process influenced by geological and climaticvariations as well as human activities In general, wind erosion leads to landdegradation in agricultural areas and has a negative impact on air quality.Dust emission generated by wind erosion is the largest source of aerosols whichdirectly or indirectly influence the atmospheric radiation balance and henceglobal climatic variations Strong wind-erosion events, such as severe duststorms, may threaten human lives and cause substantial economic damage.The physics of wind erosion is complex, as it involves atmospheric, soiland land-surface processes The research on wind erosion is multidisciplinary,covering meteorology, fluid dynamics, soil physics, colloidal science, surfacesoil hydrology, ecology, etc Several excellent books have already been written

about the topic, for instance, by Bagnold (1941, The Physics of Blown Sand

and Desert Dunes), Greeley and Iversen (1985, Wind as a Geological cess on Earth, Mars, Venus and Titan), Pye (1987, Aeolian Dust and Dust Deposits), Pye and Tsoar (1990, Aeolian Sand and Sand Dunes) However,

Pro-considerable progress has been made in wind-erosion research in recent yearsand there is a need to systematically document this progress in a new book.There are three other reasons which motivated me to write this book Firstly,

in most existing books, there is a general lack of rigor in the description ofwind-erosion dynamics; secondly, the emphasis of the existing books appears

to be placed primarily on sand-particle motion, while topics related to themodelling of dust entrainment, transport and deposition have not been pre-sented in great detail and thirdly, the results presented in the existing booksappear to be mainly experimental and lacking in documentation of the com-putational modelling effort involved

My intention is to provide a summary of the existing knowledge of winderosion and recent progress in that research field The basic contents of thebook include the physics of particle entrainment, transport and depositionand the environmental processes that control wind erosion It is intended totreat the physics of wind erosion as rigorously as possible, from the viewpoint

v

vi Preface 1

of fluid dynamics and soil physics A considerable proportion of the book

is devoted to the computational modelling of wind erosion I hope that thisbook can be used as a reference point for both wind-erosion researchers andpostgraduate students My basic consideration is that wind erosion can only

be understood from a multidisciplinary viewpoint and the computationalmodelling of wind erosion should focus on the development of integrated sim-ulation systems Such a system should tightly couple dynamic models, such

as atmospheric prediction models and wind-erosion schemes, with real datathat characterises soil and surface conditions In the introductory chapter ofthe book, this basic concept is reiterated, while in Chapter 9 examples of theadvocated modelling approach are given Chapter 2 provides a summary ofwind-erosion climatology in the world and selected regions Chapters 3 and

4 are devoted to the description of atmospheric modelling and land-surfacemodelling, as these are the prerequisite for the modelling of wind erosion.Chapter 5 is a description of the basic aspects of wind-erosion theory, whileChapters 6, 7 and 8 are dedicated to the entrainment, transport and depo-sition of sand and dust particles In Chapter 9, the integrated wind-erosionmodelling system and the data requirement are described The concludingremarks are given in Chapter 12

November 1999

Many colleagues have helped with the preparation of this edition Inparticular, I wish to thank Drs Masao Mikami, Irina Sokolik, Karl-HeinzWyrwoll, Qingcun Zeng, Gongbing Peng, Chaohua Dong, Zhaohui Lin,Masaru Chiba, Naoko Seino, Taichu Y Tanaka, Masahide Ishizuka, EunjooJung and Youngsin Chun for their support I also wish to thank Ms DagmarJansen for her careful proofreading of the manuscript and Ms Martina Klosefor helping with the manuscript preparation using LaTeX.

March 2008

vii

About 10 years ago, Dr M R Raupach introduced me to the research ofwind erosion I have ever since maintained a strong interest in this field.During these years, I came to know many colleagues, including Professor

L M Leslie, Dr J F Leys, Dr G H McTainsh, Mr P A Findlater,Professor W G Nickling, Dr D A Gillette, Professor H Nagashima,

Dr B Marticorena, Dr G Bergametti and Dr I Tegen among manyothers, who helped me to develop a understanding of the topics presented

in this book I am grateful to them for the valuable discussions and ments during the years and to many of them for providing me with theirresearch results for inclusion in this book In the wind-erosion research com-munity, there prevails truly a collaborative spirit The development of theintegrated wind-erosion modelling system described in Chapter 9 has been

argu-a teargu-am effort, argu-and I argu-acknowledge explicitly the significargu-ant contributions tothe project made by my colleagues and friends, especially, Dr H Lu, Dr

P Irannejad, Dr R K Munro, Dr C Werner and Mr R Morison The sistance of Dr P Irannejad and Mr H X Zhuang in preparing the graphsand the manuscript has been very helpful The painstaking final corrections

as-by Dr R A Byron-Scott have resulted in improvements to a text whichhas been written uncomfortably in my second language Several chapters ofthe book were drafted during my stay at the Institute for Geophysics andMeteorology, University of Cologne, in 1999 when I was an Alexander vonHumboldt Research Fellow My stay in Germany has been a happy one, and Ithank Professor Dr M Kerschgens and the Humboldt Foundation for makingthat possible My thanks also go to Dr M de Jong from Kluwer AcademicPublishers for her enthusiastic and patient approach toward publishing thisbook Finally, I would like to take this opportunity to express my gratitude

to Professor P Schwerdtfeger, Dr J M Hacker and Dr T H Chen for theircontinuous encouragements throughout my scientific career

ix

Preface 1 v

Preface 2 vii

Acknowledgements ix

1 Wind Erosion and Wind-Erosion Research 1

1.1 Wind-Erosion Phenomenon 1

1.2 Wind-Erosion Research 7

1.3 Integrated Wind-Erosion Modelling 9

2 Wind-Erosion Climatology 13

2.1 Climatic Background for Wind Erosion 13

2.2 Geographic Background for Wind Erosion 18

2.3 Atmospheric Systems 20

2.3.1 Monsoon Winds 21

2.3.2 Cyclones and Frontal Systems 23

2.3.3 Squall Lines 24

2.4 Global Wind-Erosion Patterns 26

2.5 Major Wind-Erosion Regions 29

2.5.1 Dust Weather Records and Satellite Remote Sensing 29

2.5.2 North Africa 30

2.5.3 The Middle East 34

2.5.4 Central Asia 36

2.5.5 Southwest Asia 37

2.5.6 Northeast Asia 39

2.5.7 The United States 44

2.5.8 Australia 45

xi

xii Contents

Modelling 49

3.1 Atmospheric Boundary Layer 49

3.2 Governing Equations for Atmospheric Boundary-Layer Flows 52 3.3 Reynolds Averaging and Turbulent Flux 56

3.4 Equations for Mean Flows 59

3.5 Equations for Turbulent Fluxes and Variances 60

3.5.1 Turbulent Dust Flux and Dust Concentration Variance 60

3.5.2 Turbulent Kinetic Energy 61

3.5.3 Features of Different Atmospheric Boundary Layers 63

3.6 Surface Layer 67

3.6.1 Flux-Gradient Relationship 67

3.6.2 Friction Velocity 68

3.6.3 Logarithmic Wind Profile and Roughness Length 71

3.6.4 Stability Measures 72

3.7 Similarity Theories 74

3.7.1 Monin–Obukhov Similarity Theory 75

3.7.2 Mixed–Layer Similarity Theory 78

3.8 Turbulent Flow Models 79

3.9 Meso-scale, Regional and Global Atmospheric Models 85

4 Land-Surface Modelling 91

4.1 General Aspects 92

4.2 Surface Energy Balance 94

4.3 Soil Moisture 96

4.4 Soil Temperature 101

4.5 Calculation of Surface Fluxes 101

4.6 Land-Surface Parameters 106

4.7 Examples of Land-Surface Simulation 110

4.8 Treatment of Heterogeneous Surfaces 112

5 Basic Aspects of Wind Erosion 117

5.1 Soil-Particle Characteristics 117

5.2 Forces on an Airborne Particle 124

5.3 Particle Terminal Velocity 129

5.4 Modes of Particle Motion 131

5.5 Threshold Friction Velocity for Sand Particles 134

5.5.1 The Bagnold Scheme 135

5.5.2 The Greeley-Iversen Scheme 138

5.5.3 The Shao–Lu Scheme and the McKenna Neuman Scheme 139

5.6 Threshold Friction Velocity for Dust Particles 142

5.6.1 Relative Importance of Forces 142

5.6.2 Stochastic Nature of Threshold Friction Velocity 145

Contents xiii

6 The Dynamics and Modelling of Saltation 149

6.1 Equations of Particle Motion 150

6.2 Uniform Saltation 152

6.3 Non-Uniform Saltation 154

6.4 Streamwise Saltation Flux 156

6.5 The Bagnold-Owen Saltation Equation 157

6.5.1 The Bagnold Model 157

6.5.2 The Owen Model 158

6.6 Other Saltation Equations 163

6.7 The Owen Effect 165

6.7.1 The Formulation of Owen 165

6.7.2 The Formulation of Raupach 166

6.7.3 Other Formulations 171

6.7.4 Profile of Saltation Flux 171

6.8 Independent Saltation 172

6.9 Supply-Limited Saltation 175

6.10 Evolution of Streamwise Sand Transport with Distance 176

6.11 Splash Entrainment 178

6.11.1 Wind-Tunnel Observations 178

6.11.2 Numerical Simulations 179

6.12 Numerical Modelling of Saltation 186

6.12.1 Simple Flow Model 186

6.12.2 Large-Eddy Simulation Model 187

6.12.3 Particle Motion 188

6.12.4 Aerodynamic Entrainment 190

6.12.5 Splash Scheme 191

6.13 Understanding of Saltation from Numerical Simulations 194

6.13.1 Importance of Splash Entrainment 194

6.13.2 Particle-Momentum Flux, Saltation Flux and Roughness Length 197

6.14 Saltation in Turbulence 199

6.14.1 Intermittency of Saltation 201

6.14.2 Aeolian Streamers 205

6.14.3 Dynamical Similarity of Saltation 206

7 Dust Emission 211

7.1 Dust Flux and Friction Velocity 211

7.2 Mechanisms for Dust Emission 216

7.3 Aerodynamic Dust Entrainment 222

7.4 Energy-Based Dust-Emission Scheme 222

7.5 Volume-Removal-Based Dust-Emission Scheme 226

7.5.1 Motion of Ploughing Particle and Volume Removal 230

7.5.2 Vertical Dust Flux 232

xiv Contents

7.6 Comparison of Dust Schemes 233

7.7 Spectral Dust-Emission Scheme 235

7.8 Discussions on Dust Schemes 245

8 Dust Transport and Deposition 247

8.1 Evidence of Dust Transport and Deposition 247

8.2 Lagrangian Dust-Transport Model 252

8.3 Eulerian Dust-Transport Model 255

8.4 Vertical Dust Transport by Diffusion 261

8.5 Vertical Dust Transport by Convection 273

8.5.1 Convective Adjustment 273

8.5.2 Cumulus Parameterisation 275

8.6 Dry Deposition 277

8.6.1 Two-Layer Dry-Deposition Model: Smooth Surface 278

8.6.2 Two-Layer Dry-Deposition Model: Vegetation 282

8.6.3 Single-Layer Dry-Deposition Model 286

8.7 Wet Deposition 288

8.7.1 The Theory of Slinn 289

8.7.2 Scavenging Rate 295

8.7.3 Scavenging Ratio 299

9 Integrated Wind-Erosion Modelling 303

9.1 System Structure 304

9.2 Wind-Erosion Parameterisation Scheme 307

9.3 Threshold Friction Velocity for Natural Surfaces 308

9.3.1 Drag Partition: Approach I 310

9.3.2 Drag Partition: Approach II 316

9.3.3 Relationship of λ and z0 317

9.3.4 Double Drag Partition 321

9.3.5 Soil Moisture 323

9.3.6 Chemical Binding and Crust 327

9.4 Sand Drift and Dust Emission of Soils with Multiple Particle Sizes 328

9.5 Climatic Constraints on Dust Emission 333

9.5.1 Erodibility Derived from Synoptic Data 333

9.5.2 Erodibility Derived from Satellite Data 336

9.5.3 Wind-Erosion Hot Spots 336

9.6 Land-Surface Parameters 336

9.6.1 Soil Particle-Size Distribution 337

9.6.2 Soil-Binding Strength 342

9.6.3 Frontal-Area Index 344

9.6.4 Soil Moisture 347

9.7 Manipulation of GIS Data 347

Contents xv

9.8 Examples of Integrated Wind-Erosion Modelling 350

9.8.1 Wind-Erosion Time Series 350

9.8.2 Wind-Erosion Predictions on Global, Regional and Local Scales 351

9.9 Data Assimilation 359

10 Sand Dunes, Dynamics and Modelling 361

10.1 Classification of Sand Dunes 363

10.2 Migration Speed of Transverse Dunes 370

10.3 Basic Features of Flow over a Sand Dune 373

10.4 Sand Transport 377

10.5 Computational Fluid Dynamic Simulation 381

10.5.1 Flow-Model Implementation: Non-hydrostatic Model 382

10.5.2 Flow-Model Implementation: Large-Eddy Model 384

10.5.3 Computation of Erosion and Deposition Rates 385

10.6 Discrete Lattice Simulation 388

11 Techniques for Wind-Erosion Measurements 391

11.1 Wind-Tunnel Measurements 391

11.2 Sand Measurements 393

11.2.1 Passive Samplers 393

11.2.2 Active Samplers 396

11.2.3 Impact Sensors: Sensit, Saltiphone and Safire 397

11.2.4 Sand Particle Counter 399

11.3 Dust Measurements 400

11.3.1 High- and Low-Volume Air Samplers 400

11.3.2 Optical Particle Counter 402

11.4 Deposition Collectors 403

11.5 Field Measurements 404

11.6 Particle-Size Analysis 407

11.6.1 Dry Sieving 407

11.6.2 Settling Tube and Elutriator 408

11.6.3 Electro-Sensing Methods 410

11.6.4 Laser Granulometry 412

11.7 Abrasion Emitter 412

12 Concluding Remarks 415

12.1 Current Research Topics 415

12.2 Dust Cycle in the Earth System 419

References 421

Index 447

pro-In the first instance, wind erosion is a geological and climatic phenomenonwhich takes place over long periods of time in deserts and arid regions Most ofthe time, wind-erosion events proceed unnoticed but sometimes they are mostspectacular Figure 1.2 shows the satellite image of a massive dust storm overthe Atlantic on 26 February 2000 During this event, dust from the SaharaDesert was lifted to up to 5,000 m above ground and blown off the African con-tinent by an easterly wind Dust storms of this magnitude have been observedelsewhere in the world, for instance in the Middle East, China and Australia(Figs 1.3, 2.4) Between 15 and 19 April 1998, severe dust storms developedover the Gobi Desert in Mongolia and China In the following days, the dust-storm front moved across China and, by April 20, the elongated dusty beltcovered a 2,000-km stretch of the east coast of China The dust clouds weremoving across the Pacific on 23 and 24 April and arrived in North America

by 27 April (Husar et al 2001)

Wind erosion is the main mechanism for the formation and evolution ofsand seas in the world and the long-range transport of sediments from conti-nent to ocean Recent studies suggest that the global dust emission amounts to3,000 Mt yr−1(estimates vary between 1,000 and 10,000 Mt yr−1), and a con-siderable proportion of this dust is deposited in the ocean (Duce et al 1991)

Y Shao, Physics and Modelling of Wind Erosion, 1c

Springer Science+Business Media B.V 2008

2 1 Wind Erosion and Wind-Erosion Research

Wet deposition

Saltation

Impact on radiation (Optical thickness, backscatter)

Roughness elements Trapped particles

Soil texture

& surface crust

Dust emission Turbulent diffusion Convection

Fig 1.1 An illustration of the three phases of wind erosion: entrainment, transport

and deposition Atmospheric conditions, soil properties, land-surface characteristicsand land-use practice control the erosion process (Modified from Lu, 1999)

Large quantities of minerals and organic matter are carried with the dustparticles and redistributed around the world The Loess Plateau in Chinahas a soil depth ranging from 30 to 120 m and its formation is believed to

be largely due to the deposition of wind-transported particles from the GobiDesert over many millions of years On geological time scales, wind erosioncontributes greatly to the global mineral and nutrient circulation and to theevolution of surface topography

Particles suspended in the atmosphere, the aerosols, play an importantrole in the climate system, as they influence the atmospheric radiation bal-ance directly, through scattering and absorbing various radiation components,and indirectly, through modifying the optical properties and lifetime of clouds.Aerosols generated by wind erosion (mineral aerosol) are the most importantaerosols in the atmosphere The global dust emission of 3,000 Mt yr−1 is com-parable with the global sea-salt emission, which is estimated to be around3,300 Mt yr−1 (Penner et al 2001) Both estimates for the global dust andsea-salt emissions have large uncertainties, probably a factor of two The radia-tive forcing of tropospheric aerosols on the atmosphere is currently an activeresearch topic For climate studies, the key research topics related to wind ero-sion are the global dust cycle, namely, the emission, transport and deposition

of dust, and the atmospheric processes which involve mineral aerosols, such

as radiation, cloud formation and precipitation

1.1 Wind-Erosion Phenomenon 3

Fig 1.2 Satellite image of a dust storm over the Atlantic Dust from the Sahara was

blown off the African continent by an easterly wind on 26 February 2000 (NOAA,acknowledgement)

Wind erosion also causes air-quality hazards in populated areas adjacent

to major dust sources In Beijing, for example, the measured near-surface dustconcentration during severe dust storms has been reported to be as high as5–10 mg m−3 Near dust sources, dust concentration can exceed 20 mg m−3(Yabuki et al 2002) The northeast Asian dust storm that occurred between

18 and 24 March 2002 caused severe disruptions of social activities in thenorthern part of China and Korea (e.g closure of airports and schools) Dur-ing the 21–23 October 2002 Australian dust storm, the PM10 concentrationsmeasured in some coastal cities of Australia (e.g Brisbane) were close to

1 mg m−3 (Chan et al 2005) Many contaminants that pose risks to humanhealth and the environment are found or associated with dust, including metal,pesticides, dioxins and radionuclide Thus, to quantify dust sources and to es-timate airborne dust concentrations are also important to air-quality studies

In the recent history, human activities have created profound disturbances

to the natural environment Excessive clearance of native vegetation, overgrazing and inadequate agricultural practices have resulted in increased fre-quency and intensity of wind erosion in some parts of the world Tegenand Fung (1995) estimated that 20–50% of the global dust load is derivedfrom human-disturbed soils This estimate has been recently repudiated byProspero et al (2002), but there is evidence that over human-disturbed sur-faces, the rate of wind erosion can be many times that over undisturbed

4 1 Wind Erosion and Wind-Erosion Research

Fig 1.3 Image of dust storms in the Sahara captured by the space shuttle (NASA,

acknowledgement)

natural surfaces During the 1930s, for example, decreased precipitation pled with intensive agricultural activities lead to severe wind erosion in theGreat Plains of the United States, which became known as the dust bowl

cou-of the USA In the Sahel, drought conditions combined with overpopulationalso resulted in a considerable increase of wind-erosion events In China, cul-tivation on the Loess Plateau may have contributed much to the severe duststorms in northeast Asia In Australia, some of the recent severe dust stormshave originated in the agricultural areas, where the native vegetation has beencleared over the past 200 years Figure 1.4 shows a dust storm over the Mur-ray River near Mildura (Australia), a farming area claimed from the forests

of native Mallee trees

Wind erosion in agricultural areas leads to land degradation During anerosion event, fine soil particles rich in nutrient and organic matters, are car-ried away by wind over large distances and this results in the loss of soil nutri-ents According to Raupach et al (1994), the February 1982 Melbourne duststorm generated a loss of 2 million tonnes of topsoil, including 3,400 t of ni-trogen and 10 t of phosphorus The May 1994 dust storms in Australia caused

a soil loss between 10 to 20 million tonnes The preferential removal of fineparticles by wind erosion leaves coarser and less fertile material behind Conse-quently, eroded soils become less productive and have a smaller water-holdingcapacity For land-care purposes, the major tasks of wind-erosion research are

1.1 Wind-Erosion Phenomenon 5

The origin of the dust were the nearby farming areas (J F Leys, acknowledgement)

to quantify the risks of wind erosion on different temporal and spatial scales,

to identify the responsible factors and to develop wind-erosion preventionmeasures

Wind erosion involves complex physics that is not yet fully understood.Its study requires the knowledge of a wide range of disciplines including at-mospheric sciences (climatology, synoptics, remote sensing, cloud physics andatmospheric boundary layers), fluid dynamics, soil physics, surface hydrol-ogy, colloidal sciences, and ecology as well as agricultural sciences and landmanagement

Almost all physical processes related to wind erosion are particle-size pendent We often use the term ‘sand’ to describe particles in the size range

de-between 60 and 2,000 µm and the term ‘dust’ to describe particles smaller than 60 µm More precise definitions are given in Chapter 5.

Wind erosion is the consequence of two types of forces at work: theaerodynamic forces that tend to remove particles from the surface, andthe forces, such as gravity and inter-particle cohesion that resist the removal

6 1 Wind Erosion and Wind-Erosion Research

The former can be quantified by the friction velocity, u ∗, a measure of wind

shear at the surface, and the latter by the threshold friction velocity, u ∗t,which defines the minimum friction velocity required for wind erosion to oc-

cur While u ∗ is related to atmospheric flow conditions and surface

aerody-namic properties, u ∗t is related to a range of surface properties, such as soiltexture, soil moisture and vegetation cover For dry and bare sandy soil sur-

faces, u ∗t is small, and therefore it is not surprising that wind erosion occurs

mainly under such conditions The balance between u ∗ and u ∗t is governed

by, and is sensitive to, a number of environmental factors, namely, (1) weather(wind, temperature, rainfall, etc.); (2) soil type (soil texture, hydraulic proper-ties, etc.); (3) soil state (wetness, compactness, aggregation, etc.); (4) surfacemicroscopic conditions (aerodynamic roughness length, vegetation coverage,etc.) and (5) surface macroscopic conditions (landforms, windbreaks, etc.) As

a consequence, wind erosion is strongly variable in space and intermittent intime The sporadic nature of wind-erosion events makes the modelling andprediction of wind erosion, even in a qualitative sense, a formidable task.For soil particles to become airborne, lift forces associated with wind shearnear the surface or caused by particle impacts must overcome the gravitationaland cohesive forces acting upon them We call the entrainment of particlesaerodynamic entrainment if it is dominated by aerodynamic forces, and re-fer to it as splash or bombardment entrainment if it is mainly caused bythe impact of other moving particles In either situation, the forces involved

in the entrainment process vary strongly from case to case, depending on arange of factors, but in particular on particle size Consequently, the domi-nant mechanism for particle entrainment also depends on particle size Forsand particles, the entrainment is essentially aerodynamic, while for dust par-ticles the entrainment is primarily due to the impact of saltating sand grains,

a phenomenon known as saltation bombardment (Gillette et al 1982; Shao

et al 1993b)

The motion of airborne particles in the atmosphere has two modes, known

as saltation and suspension Saltation refers to the small hopping motion ofsand-sized particles in the direction of the wind, while suspension refers tothe floating motion of dust-sized particles in the atmosphere

Through saltation, soil particles are transported in the direction of windduring an erosion event Saltation is the mechanism for the evolution of sandseas on regional scales and the development of sand dunes and fence-line drifts

on local scales It is an interesting dynamic problem which involves the teractions between the fluid phase, the particulate phase and the surface Asparticles saltate through the atmospheric surface layer with a strong windshear, they absorb momentum from the airflow and generate a momentumtransport in the vertical direction During the impact on the surface, thesaltating particles may splash more particles into the atmosphere The de-position of saltating particles is of great significance to the evolution of theland surface Particles in saltation may be deposited as wind speed reducesdue to changes in atmospheric conditions or changes in surface roughness

in-1.2 Wind-Erosion Research 7

(e.g shrubs) or topography (e.g hills) Saltation is also associated with alarge degree of randomness originating from the lift-off velocities, lift-off an-gles and turbulent fluctuations in the atmosphere It is of particular interest inwind-erosion studies to estimate the intensity of horizontal sand drift related

to saltation and to statistically describe the stochastic features of this mode

of particle motion

In contrast to sand particles, dust particles, once airborne, can remainsuspended in the atmosphere for some time and be dispersed beyond the at-mospheric surface layer by turbulence and transported over large distances.This process leads to net soil losses from areas of wind erosion The dispersion

of dust particles is a difficult fluid-dynamic problem, known as heavy-particlediffusion Because dust particles have a density more than 2,000 times largerthan the air density, dispersion of dust particles differs from that of neutrally-buoyant fluid parcels In addition, unless the atmospheric patterns and tur-bulence properties are adequately pre-specified, it is not possible to predictthe transport of dust particles in the atmosphere with a reasonable accuracy.Dust particles suspended in the atmosphere are eventually delivered back

to the surface through dry and wet deposition Dry deposition is the transfer ofairborne dust particles to the surface through turbulent and molecular diffu-sion and gravitational settling, while wet deposition is the transfer of airbornedust particles through precipitation Both processes are of similar importance

In the first instance, dry deposition is a fluid-dynamic problem dealing withthe diffusion of particles through a thin layer immediately adjacent to thedeposition surface It also involves many physical processes that are difficult

to quantify, such as the static electrical charge The difficulty in studying drydeposition also lies in the lack of detailed knowledge of the flow structure inthe very thin layer immediately adjacent to the deposition surface On theother hand, wet deposition involves the process of rain droplets capturingdust particles suspended in air The study of wet deposition requires detailedunderstanding of raindrop size distribution, particle concentration and thecapture mechanisms

8 1 Wind Erosion and Wind-Erosion Research

Field Measurements

Field measurements of wind erosion have been carried out on different scales inmany parts of the world Sand drift has been measured using saltation traps,and the impacts of land-surface parameters on wind erosion have been stud-ied Measurements of wind and dust concentration profiles have been madeusing anemometers and dust samplers mounted on towers In addition, net-works of air samplers and deposition traps have been setup in recent years formeasuring dust concentration and deposition over large areas Such networksare providing valuable data for studying dust movement in the entrainmentphase (tower measurements), the transport phase (tower measurements andair samplers) and at the deposition phase (deposition traps)

Wind-Erosion Assessment

Assessment of wind erosion on continental scales has been performed by sidering wind erosivity and wind erodibility Wind erosivity describes thepotential of wind to generate erosion, while erodibility describes the potential

con-of the surface to be eroded Chepil and Woodruff (1963) proposed to use awind-erosion index and developed a model for calculating such indices withthe data of wind speed, precipitation and evaporation McTainsh et al (1990)applied the model of Chepil and Woodruff to determining wind-erosion in-dices for Australia With the development of Geographic Information Systems(GIS), more attention has been paid to soil and land-surface factors Studies

of wind-erosion climate based on dust-storm records have been carried out by,for instance, Middleton (1984), Littman (1991), Goudie and Middleton (2001),Qian et al (2002) and Kurosaki and Mikami (2005) among many others

Satellite Remote Sensing

Satellite remote sensing is advantageous in dust-storm monitoring Sensors

on board of satellites detect the radiances of various surfaces of the Earththrough different spectral channels These channels are set in correspondence

to the atmospheric radiation windows and water vapour absorption bands.Various satellite-sensed signals are combined (1) to identify and monitor duststorms in real time (Carlson, 1979; Ackerman, 1989); (2) to derive land-surfaceand atmospheric parameters for dust modelling; (3) to retrieve dust quanti-ties, such as dust load, optical thickness, particle size, etc (Ackerman, 1997;Zhang et al 2006); and (4) to derive long-term dust climatology For example,Prospero et al (2002) have used the NIMBUS 7 Total Ozone Mapping Spec-trometer (TOMS) aerosol index over a 13-year period (1980–1992) to examinethe distribution of dust sources on the globe

Empirical Wind-Erosion Modelling

Empirical wind-erosion models have been under development for some time.The most widely used is the Wind-Erosion Equation (WEQ) (Woodruff and

1.3 Integrated Wind-Erosion Modelling 9

Siddoway, 1965), an empirical model in which the driving parameters are scriptors of soil type, vegetation, roughness, climate and field length Theoriginal WEQ used annual averages of these descriptors to estimate annualaverage soil loss For estimates over shorter periods, the WEQ was modified

de-by Bondy et al (1980) and Cole et al (1983) More recent revisions haveled to the Revised Wind-Erosion Equation (RWEQ) which includes input pa-rameters such as planting date, tillage method and amount of residue fromthe previous crop; a weather generator is then used to predict future ero-sion (Comis and Gerrietts, 1994) The empirical nature of the WEQ limitsits transferability from the central Great Plains of the USA, for which it wasoriginally developed, to other areas of the world Also, the complex interac-tions between the variables controlling wind erosion are not fully accountedfor in the WEQ For this reason, a more process-oriented model called theWind-Erosion Prediction System (WEPS) has been developed The WEPSincludes submodels for weather generation, crop growth, decomposition, soil,hydrology, tillage and erosion (Hagen, 1991)

Large-Scale Field Experiments

Several large-scale field experiments have been recently carried out, dedicated

to wind erosion, dust storms and dust cycle The Aeolian Dust Experiment onClimate Impact (ADEC, Mikami et al 2006) and the Asian Pacific RegionalAerosol Characterization Experiment (ACE-Asia, Huebert et al 2003; Ari-moto et al 2006) are two examples Networks have also been constructed toobtain dust observations over large areas For example, the Aerosol RoboticNetwork (AERONET) is a federation of ground-based remote sensing aerosolnetworks AERONET assesses aerosol optical properties and validates satelliteretrievals of these properties The data include globally distributed observa-tions of spectral aerosol optical depths and precipitable water The networkhas been operating since 1993 and has been carrying out routine measure-ments at around 150 stations distributed all over the world

1.3 Integrated Wind-Erosion Modelling

The approach advocated in this book is integrated wind-erosion modelling

An integrated wind-erosion modelling system enables the simulation and diction of all aspects of wind erosion, from particle entrainment, transport todeposition The aim of such a system is to provide quantitative assessment andprediction of wind erosion on scales from local to global To this end, the in-tegrated system needs to be constructed with six basic components: an atmo-spheric model, a wind-erosion model, a land-surface scheme, a dust-transportscheme, a data-assimilation scheme and a geographic-information data base.The atmospheric model provides the data required to drive the wind-

pre-erosion scheme, such as friction velocity, u ∗, wind field for dust advection,

10 1 Wind Erosion and Wind-Erosion Research

turbulence intensity for dust diffusion and deposition, and precipitation forwet deposition In addition, the atmospheric model provides the data, such

as radiation, required by the land-surface scheme for modelling the mental variables, such as soil moisture and vegetation cover, which stronglyinfluence wind erosion Most atmospheric models are coupled with radiationschemes to deal with the impact of mineral aerosols on radiation transfer inthe atmosphere

environ-The wind-erosion model enables the quantification of the entrainment,transport and deposition of soil particles of all sizes For instance, as far asparticle entrainment is concerned, it enables the prediction of (1) the thresholdfriction velocity for wind erosion, (2) the rate of sand transport and (3) therate of dust emission

Data assimilation is a technique which combines model and data to achieve

an optimal simulation or prediction of a problem of concern This techniquehas been very successfully applied to atmospheric and oceanic predictions.Because of the lack of dust measurement data, very little has been done so far

in applying data assimilation to dust modelling However, dust measurementsare becoming increasingly available For example, satellites can now providecontinuous dust monitoring over large areas and the developments of inversemethods are producing quantitative estimates of dust load in an atmosphericcolumn Networks of lidar are being established, which provide dust-profileestimates at a number of locations Further, stations equipped with dust sam-plers and radiometers are being set up We expect that in the near future,data assimilation will be an important component of integrated wind-erosionmodelling

Reliable land-surface data is important to wind-erosion modelling Theyare required, because the properties of the land surface control the erosionthreshold friction velocity, the capacity of the soil to release dust and thepartitioning of wind-shear stress acting on non-erodible roughness elementsand the erodible surface Three categories of parameters can be distinguished.The first consists of parameters related to soil properties, e.g., soil particle-size distribution and soil-binding strength The second consists of aerodynamicparameters related to surface roughness and drag partitioning The third cat-egory consists of parameters which specify the soil thermal and hydraulicproperties For the purpose of modelling wind erosion on regional to conti-nental scales, these soil and land-surface parameters can be stored as layers

in a geographic information system (GIS)

The first attempt of developing an integrated wind-erosion modelling tem was probably made by Gillette and Hanson (1989), who used extensiveatmospheric and land-surface data to determine the spatial and temporalvariations of dust production in the United States Gillette and Hanson didnot use an atmospheric model and did not consider dust transport and depo-sition Earlier versions of dust models, more from the atmospheric perspec-tive, have been developed by Westphal et al (1988) and Tegen and Fung(1994, 1995) In these early dust models, rather crude wind-erosion schemes

sys-1.3 Integrated Wind-Erosion Modelling 11

and land-surface data were used Marticorena and Bergametti (1995), Shao

et al (1996) and Marticorena et al (1997) developed physics-based erosion models and applied them to improve the simulations of wind erosion.Shao and Leslie (1997) and Lu (1999) developed an almost fully integratedwind-erosion modelling system which couples a physics-based wind-erosionscheme with an atmospheric model, a land-surface scheme and a geographic-information database They have implemented the system for the prediction ofdust storms in Australia Since the late 1990s, a number of dust storm modelsfor global, regional and local dust problems have been developed Examples ofglobal dust models include those of Zender et al (2003), Ginoux et al (2004)and Tanaka and Chiba (2006) Examples of regional dust models include thestudies of Nickovic et al (2001), Liu et al (2001), Shao et al (2003) and Uno

wind-et al (2005) Seino wind-et al (2005) simulated dust storms in the Tarim Basinusing a meso-scale dust model

Integrated modelling is a new approach to studying wind erosion It takesthe advantage of the recent rapid expansion in computing power, develop-ments in atmospheric and land-surface modelling, and the increasing avail-ability of land-surface and remote-sensing data This approach is a majorstep forward in the quantitative prediction of wind erosion, the comprehen-sive analysis of wind-erosion processes and the identification of the naturaland human factors that affect wind erosion However, integrated systems arecomplex As will become evident in this book, nearly all wind erosion processesare sensitive to parameters which cannot be derived with great certainty Forexample, threshold friction velocity is sensitive to soil moisture and vege-tation cover As a consequence, it is difficult to predict wind erosion withgreat accuracy Nevertheless, recent studies have demonstrated that integratedwind-erosion modelling systems can produce results (sand-drift intensity, dustemission, dust concentration, etc.) which are comparable in magnitude withobserved data, and the uncertainties embedded in the modelling systems arecomparable with the uncertainties of observations

Wind-Erosion Climatology

In this chapter, we describe the climatology of wind erosion We are ested in the spatial patterns and temporal variations of wind erosion and thegeographic, climatic and synoptic conditions which determine them

inter-The understanding of wind-erosion climatology is largely based on thevarious measurements obtained through the techniques of remote sensing,comparison of aerosol samples with soil samples, monitoring air mass trajec-tories, analysis of dust weather records and numerical modelling Among thesetechniques, the analysis of dust weather records and the analysis of remotesensing data are the most popular (Kurosaki and Mikami, 2005; Prospero

et al 2002) Although weather records are relatively scarce for desert areasand the weather-station network is not dense enough in non-desert areas toprovide an accurate picture, the basic climatic features of wind erosion on theglobe are now quite well known Earth-observing satellites are now providing

a huge amount of data for studying large-scale dust activities By examiningthe aerosol optical thickness (Legrand et al 1994; Moulin et al 1998) and theabsorbing aerosol index (Herman et al 1997), which can be retrieved fromsatellite signals, much about large-scale dust activities has been learned

2.1 Climatic Background for Wind Erosion

The global pattern of wind erosion is closely related to the general circulation

of the atmosphere The distributions of solar radiation and albedo over theglobe determine that there is a surplus of available energy (net radiation)

in the region of low latitudes and a deficit in the region of high latitudes.This distribution of available energy leads to the general circulation of theatmosphere

As the Earth rotates around its axis, an apparent force acts continuously

on a moving air stream if it is studied in a coordinate system which follows theEarth’s rotation This apparent force, known as the Coriolis force, is−2Ω × v

for a unit mass of fluid moving with speed v Ω is the angular velocity vector

Y Shao, Physics and Modelling of Wind Erosion, 13c

Springer Science+Business Media B.V 2008

14 2 Wind-Erosion Climatology

of the Earth’s rotation The Coriolis force is proportional to the magnitude of

v but acts in the direction perpendicular to v (right hand side in the

north-ern hemisphere, and left hand side in the southnorth-ern hemisphere) In the freeatmosphere, the pressure-gradient force and the Coriolis force dominate thebehaviour of the flow The wind at the balance between the two forces is known

as the geostrophic wind In a local coordinate system (with the origin

posi-tioned at a specified location on the earth surface), (x, y, z), with x pointing

eastward, y northward and z upward, the geostrophic wind is vg= (u g , v g , 0),

where u g and v g are given by

at-The basic features of the atmospheric general circulation are as illustrated

in Fig 2.1 In the meridional direction, the circulation is characterised by threecirculation cells, the Hadley, the Ferrel and the polar cells Near the equator,warm air rises and flows to the poles in the upper atmosphere Under theinfluence of the Coriolis force, the poleward-moving air obtains the westerlymomentum and forms westerly flows at around 30◦N and 30◦S At these lati-tudes, air converges at high levels and subsides, leading to the developments ofthe sub-tropical highs The higher pressure in the sub-tropical region leads toflows in the lower atmosphere toward the equator, which complete the Hadleycell Again, due to the Coriolis force, the airflows moving toward the equa-tor acquire an easterly component and are known as the trade wind Thesub-tropical high also generates poleward flows in the lower atmosphere Inthe polar regions, the situation is the opposite, where strong surface coolingcauses air to sink and to flow towards the equator in the lower atmosphere.The air moving towards the equator converges at about 50◦N and 50◦S withthe air originating from the sub-tropical high, forming the polar-front zone.The three meridional circulation cells give rise to three surface windregimes in each hemisphere: the trade winds of low latitudes, the midlati-tude westerly and the polar easterlies This general circulation pattern hasmajor implications to the distributions of wind, precipitation, temperatureand hence wind erosion The trade winds from the two hemispheres form theinter-tropical convergence zone (ITCZ), which can be clearly identified overthe ocean and less so over the land, where it is modified and suppressed byother atmospheric systems The ITCZ may be broken at several locations by

2.1 Climatic Background for Wind Erosion 15

Fig 2.1 An illustration of the atmospheric general circulation, showing the

merid-ional circulation in the vertical cross-section of the atmosphere and the pattern of

winds over the Earth surface PF denotes the polar front and J p and J sthe locations

of the high-level (about 200 hPa) jet streams (Redrawn from Defant and Defant,1958)

monsoon flows, but can be traced around the globe Air in the ITCZ is mostlyunstable and convective cells develop frequently, producing heavy rainfall Inthe northern-hemisphere summer, the ITCZ is situated around 10◦N, broken

by the monsoon over the Indian Ocean In the northern-hemisphere winter, theITCZ is situated around 5◦N, stretching from the East Pacific, crossing LatinAmerica, the Atlantic Ocean and North Africa The ITCZ is broken at the eastcoast of Africa, but reappears at about 10◦–15◦S over the Indian Ocean andthe West Pacific The meridional migration of the ITCZ has a significant im-pact on wind-erosion activities For example, wind erosion in the Sahara andSahel region is enhanced as the ITCZ advances to the north, causing convec-tive thunderstorms, but becomes significantly weaker in late summer In thepolar-frontal zone, baroclinic instability allows the developments of cyclonesassociated with frontal systems which cause systematic rainfall in large areas.The high-pressure system in the sub-tropical latitudes is the most permanentfeature of the global atmospheric circulation The Pacific and Azores anti-cyclones are large areas of subsiding air The dominance of the sub-tropicalhighs near latitudes 20◦–30◦N and S results in deficiency in rainfall, becausethe subsiding air increases the stability of the atmosphere and prevents theformation of clouds Some of the most arid regions of the world are found in

16 2 Wind-Erosion Climatology

such places, including the Sahara Desert, the Kalahari and Namib Deserts,the Middle East, the Thar Desert, the coastal deserts of northern Chile, Peru,southern California and large areas of western and central Australia

The idealised global circulation pattern as depicted in Fig 2.1 is cantly modified by the irregular distribution of continents, oceans and moun-tain ranges which have different thermal and dynamic properties as well asdifferent seasonal variations The continents have a smaller heat capacity thanthe ocean and hence undergo stronger annual temperature variations Thethermal contrasts between the continents and the ocean lead to the devel-opments of monsoons which profoundly affect the distributions of wind andprecipitation

signifi-Monsoons constitute a major component of the global circulation and play

a major role in the sub-tropical and tropical regions of Asia and Africa West,east and middle Africa, north Indian Ocean, India, South China Sea, thesouthern parts and north-eastern parts of China, Japan, west Pacific Indonesiaand the northern part of Australia are areas under the influence of monsoons.The Asian monsoon system is the most prominent in the world and consists

of the north-east winter monsoon and the south-west summer monsoon soons greatly affect the patterns of precipitation in many parts of the world

Mon-In Mon-India, for instance, there is little precipitation during the winter monsoonbetween November and March, while rainfall is plentiful during the summermonsoon between June and October In China, the movement of the precipi-tation zone from south to north between spring and autumn is closely related

to the propagation of the southeast Asian summer monsoon In April andMay, the rain zone is normally situated in south China and, in early summer,

in the Changjiang valley before moving further north in late summer Thesummer monsoon that produces rainfall, cannot penetrate to the inland areas

of China This results in a diminished precipitation and increased wind sion in these areas In addition, the Tibetan Plateau and the high mountains

ero-in western Chero-ina block the moisture from the Indian Ocean ero-in the southwest

As a consequence, north-western and northern China, parts of north-easternChina and Mongolia, suffer deficiencies in rainfall The Taklimakan Desert sit-uated between the Tian and Kunlun mountain ranges receives no more than

100 mm of rainfall per year, and the Gobi Desert situated along the border

of China and Mongolia receives no more than 200 mm of rainfall per year.The Taklimakan and the Gobi Deserts are the largest deserts in the temperateclimate zone

In general, wind erosion is more active in dry years Littman (1991) ied dust-storm frequencies in Asia and found that the occurrence of duststorms not only shows seasonal fluctuations, but also inter-annual variationsbetween 3.6 and 5.5 years In Australia, severe dust storms mostly occur dur-ing drought years, such as 1982 and 1983, 1991, 1994 and 2002 During thesummer of 1982–1983, the eastern regions of Australia were transformed into

stud-a nestud-ar desert Inlstud-and of the Dividing Rstud-ange, there were vstud-ast stud-arestud-as of fstud-ailedcropland and bared grazing land from which topsoil began to erode Dust

2.1 Climatic Background for Wind Erosion 17

storms started in the spring and were widespread during the summer, withparticularly severe dust storms occurring in January and February of 1983(Garratt, 1997)

The underlying mechanism for extensive drought in Australia is the ElNi˜no which represents the sea surface temperature variations in the equatorialPacific During the El Ni˜no years, warm waters occur in the eastern Pacificalong the tropical coast of South America The El Ni˜no phenomenon is coupledwith the Southern Oscillation in the tropical atmosphere, which represents thevariations of the Walker circulation El Ni˜no and Southern Oscillation (ENSO)are two closely related processes, the former taking place in the Pacific Oceanand the latter in the tropical atmosphere

The Walker circulation (Fig 2.2) is a longitudinal circulation in latitude atmosphere, arising from the variations of sea surface temperatureacross the ocean Air rises at longitudes of relative heating and sinks at otherlongitudes of relative cooling Under the normal state of the ocean and atmo-sphere across the Pacific Ocean, in the tropical latitudes off the west coast ofSouth America, the south or south-east trade winds generate a surface cur-rent of cold water from the south, and this is deflected towards the west bythe Coriolis force The surface water drifting away is replaced by up-welling

low-of cold deep-ocean water Consequently, there develops a strong gradient insea surface temperature from this area across the Pacific to the warm wa-ters of the Indonesian archipelago This temperature contrast drives an east-west circulation cell in the atmosphere – the Walker circulation Over theIndonesian region, the relatively warm and humid air rises to produce cloudsand rainfall Over the eastern Pacific, the relatively cool air sinks and thus,the rainfall is scanty by tropical standards This is accompanied by easterlywinds across the Pacific at the surface and westerlies in the upper atmo-sphere

On occasions, the easterlies of the tropical Pacific weaken and therebyreduce the wind stress, resulting in warm water flowing eastward across thePacific to displace the cold water off the South American coast With theappearance of warm surface waters in the central to eastern Pacific, the dif-ference in temperature across the Pacific is reduced The Walker circulationbecomes weak or breaks down altogether, and a more complicated circulationcell structure develops in its place As a result, cloud formation and rainfall

is reduced in the Australian/Indonesian region but increased in the central toeastern Pacific There is also evidence that these warm waters in the centralPacific can induce a series of atmospheric waves propagating into the North-ern Hemisphere For some El Ni˜no events these atmospheric disturbances can

be the cause of severe weather events over the entire globe Areas tently affected by the El Ni˜no include Australia, Southeast Asia, large areas

consis-of China and the United States as well as certain areas in South America andAfrica

18 2 Wind-Erosion Climatology

Fig 2.2 An illustration of the atmospheric circulations over the tropical Pacific.

During El Ni˜no years, warm surface-ocean waters appear in the tropical East Pacific,accompanied by a weak Walker circulation El Ni˜no years are associated with lowerrainfall and greater wind-erosion risks in eastern Australia (top) During La Ninayears, warm surface-ocean waters appear in the tropical West Pacific, accompanied

by a strong Walker circulation La Nina years are associated with higher rainfall andsmaller wind-erosion risks in eastern Australia (bottom)

2.2 Geographic Background for Wind Erosion

Wind erosion can only happen in areas where there is supply of sand anddust However, the formations of sand and dust sources are determined by,apart from aeolian transport, weathering and fluvial (including glaciofluvial)processes Prospero et al (2002) pointed out that almost all major present-daydust sources are located in arid topographic depressions where fluvial action

is evident

2.2 Geographic Background for Wind Erosion 19

Fig 2.3 Implications of fluvial actions to the formation of present-day dust sources

Fluvial actions are of paramount importance to the formation of thepresent-day dust sources, such as the Bod´el´e Depression (North Africa), theTarim Basin (China) and the Lake Eyre Basin (Australia) These sources arelocated in arid regions centred over topographic lows or on lands adjacent totopographic highs Fluvial processes are efficient in producing fine particles byseparating them from the soil matrix and carrying them to deposition basins

or alluvial plains (Fig 2.3) Fluvial action is evident in the dust source regions

by the presence of ephemeral rivers and streams, alluvial fans, playas and saltlakes Alluvial fans form at the base of mountains where water erosion sup-plies the sediment The upper part of the alluvial fan is characterized by coarsesediment, and the lower part by fine sediment The development of alluvialfans over geologic time may extend to large areas to form alluvial basins, or

in areas of gentle slopes alluvial plains Playa, also known as alkali flat, is

a flat-bottomed dry lakebed which consists of fine-grained sediments infusedwith alkali salts A consistent association of dust sources with playas has beenfound (Gill, 1996) While playas may be not dust sources themselves, strongdust emission may occur in the alluvial fans that ring the basins in which theplayas are found (Reheis et al 1995)

Most prudent-day dust source regions have deep and extensive alluvialdeposits resulting from a relatively recent pluvial history During the pluvialphases, these regions were flooded and thick layers of sediment were depositedand are now exposed to wind erosion Many of the dust sources were floodedduring the Pleistocene (roughly 2 million to 10,000 years ago, e.g., the LakeChad Basin)

20 2 Wind-Erosion Climatology

Most of the present-day sand and dust sources are confined to regionswith annual rainfall below 250 mm While the global pattern of aridity isdetermined by the general circulation of the atmosphere, the interferences oflandforms on the atmospheric flow field can significantly modify the climate

in a region The most arid places in the world are basins situated in thewake of high mountains For example, the Great Basin (USA) in the wake ofthe Sierra Nevada receives less than 120 mm annual rainfall, and the TarimBasin (China) surrounded by the Tian and Kunlun mountains and the PamirPlateau receives less than 100 mm annual rainfall As air approaches themountain ranges, it is either diverted to flow around them or forced to rise.The adiabatic cooling associated with the upward motion promotes conden-sation and precipitation and thereby depletes the moisture in the airflow.Further down wind, air descends over the mountains The adiabatic heatingassociated with the downward motion suppresses condensation and precipita-tion and thereby generates hot and dry airflows, known as foehn In connection

to wind erosion, this process has three important consequences First, an aridshadow develops in the mountain wake Second, precipitation increases onmountain slopes which leads to fluvial activity and sediment transport fromthe mountain slopes to the adjacent depressions Third, complex flow patternsdevelop in regions around the mountain ranges, which affect dust transport,e.g., dust may be transported along preferred paths or allowed to accumulate

in certain areas Preferred routes of dust transport have been identified onsatellite images of dust events For example, dust is often seen to elongatefrom south-eastern Iran into the Indus Delta along the southern flanks ofthe Makran mountains (Prospero et al 2002) Dust from the Gobi Desert ispreferentially transported south-eastwards along the north-western boundary

of the Tibetan Plateau and dust from the Thar Desert eastwards along thesouthern boundary of the Tibetan Plateau (Shao and Dong, 2006) Dust orig-inating from the northeast Tarim Basin is often advected to the southwest ofthe basin causing high dust concentrations over there (Shao and Wang, 2003)

2.3 Atmospheric Systems

Severe dust storms are mostly generated by vigorous weather systems, such asmonsoon winds, cyclones and fronts, squall lines and thunderstorms In thesesystems, the necessary meteorological conditions for the formation of a sub-stantial dust storm are present, including (i) strong near-surface wind to liftsand and dust particles from the surface; (ii) strong turbulence or convection

to disperse particles to high levels; and (iii) strong winds to transport particlesover large distances horizontally Despite these common features, the atmo-spheric systems which generate dust storms in different parts of the world can

be quite different

2.3 Atmospheric Systems 21

Fig 2.4 The structure of the atmospheric circulation over North Africa in August.

A: Location of the monsoon trough and the low-level easterly jet over the Sahara Tothe north of the monsoon trough at about 20◦N blows the north-easterly Harmattan(hot and dry) and to the south the south-westerly monsoon wind (warm and moist).B: Vertical structure of the atmosphere along the dashed line in A Strong windsassociated with the convections in the monsoon trough may generate dust storms(Redrawn from Barry and Chorley, 2003)

2.3.1 Monsoon Winds

In North Africa, Southeast Asia and the Middle East, dust storms are ten generated by monsoon winds or monsoon-related convective disturbances.The Harmattan (low-level hot and dry north-easterly) of North Africa and theShamal (low-level hot and dry northerly) of the Middle East are well-knownexamples The patterns of wind and precipitation in North Africa are very

of-22 2 Wind-Erosion Climatology

Fig 2.5 Satellite image of the dust storm originating in the Gobi Desert of

north-west China on 14 April 1998 The image, from 16 April 1998, shows the dust cloudsbehind the cold front and near the centre of the storm (SeaWiFS image by NormanKuring, NASA GSFC, acknowledgment)

much determined by the regular and continuous migration of the monsoontrough between the annual extreme locations of about 2◦N in January and

25◦N in August South of the trough is a shallow south-westerly airflow of c.a.1,000 m deep, and overriding the south-westerly is an easterly of around 3,000

m deep, known as the midlevel easterly jet North of the monsoon trough blowsthe Harmattan from the sub-tropic high pressure center (Fig 2.4) Along themonsoon trough, shallow easterly waves develop, accompanied by surface cy-clones The cyclones generate strong near-surface winds which entrain largeamounts of dust into the atmosphere (Westphal et al 1988) Isolated stormsand a broad zone of disturbances also occur just below the midlevel easterlyjet Another mechanism for the outbreaks of the Saharan dust are the dryconvections which mix momentum from the midlevel easterly jet down to thesurface where dust is mobilized (Prospero and Carlson, 1981) In the Harmat-tan affected regions, wide spread dust haze occurs in light wind conditions,known as the Harmattan haze

The Shamal blows almost daily in the Middle East between June andSeptember It may generate dust storms across Iraq, Kuwait and the Arabian

2.3 Atmospheric Systems 23

Peninsula The synoptic situation responsible for the Shamal is a zone

of convergence between the subtropical ridge extending into the northernArabian Peninsula and Iraq from the Mediterranean Sea and the Mon-soon Trough across southern Iran and the Southern Arabian Peninsula Thenortherly flow associated with this synoptic situation is further accelerateddue to the channelling effect exerted by the Iranian and the Arabian Plateaus

As the Shamal blows over the Tigris and Euphrates alluvial plains, a dous amount of dust may be lifted into the atmosphere to create dust plumeswhich stretch southward across Kuwait and the northern Persian Gulf TheShamal dust storms move into an area like a dust wall of 1,000–2,500 m talland 100–200 km thick In the dust wall, the visibility drops to near zero

tremen-2.3.2 Cyclones and Frontal Systems

Frontal systems associated with cyclones and depressions also generate scale dust storms For example, in spring in Northeast Asia, up to 15 Mon-golian cyclones (midlatitude cyclones initiating from baroclinic disturbancesover the Mongolian Plateau) travel from northwest to southeast along the EastAsian trough over the east coast of the Eurasian continent Mongolian cyclonesare accompanied by rigorous cold fronts and post-frontal north-westerly windsreaching up to 20 m s−1 The cyclones move over the Gobi Desert and gener-ate dust storms in Northeast Asia Figure 2.5 shows a satellite image of the

large-16 April 1998 dust storm originating in the Gobi Desert During this event,dust was transported to as far as the west coast of the United States.Intense wind-erosion events in Australia are mostly associated with South-ern Ocean cold fronts propagating from southwest to northeast (Garratt, 1984;Garratt et al 1989) The 8 February 1983 Melbourne dust storm, for instance,was generated by a dry cold front which represented a sharp demarcation be-tween a hot northwest flow ahead of it and a west-south-west cooler flowbehind it (Fig 2.6) The structure of the front was typical of the summer coolchanges in southern Australia, which can be described as propagating gravitycurrents (horizontal flows of a fluid that is denser than its surroundings) Thecold front consists of an advancing wedge of cooler air marked in this case by

a wall of dust, which undercuts the warm air mass ahead of the front Thedepth of the gravity current at the head was about 300 m The propagationspeed of the gravity current exceeded 20 m s−1

Figure 2.7 depicts the typical weather pattern that generates dust storms

in Australia A dust episode in Australia lasts two to three days Initially,

a low is positioned over the Southern Oceans and a high to the west of Australia A cold front extends from the low pressure centre tothe Great Australia Bight Strong post-frontal southerly airflows are di-rected onto the southern coast of the continent, and a strong hot dry north-westerly wind blows over the south-eastern parts of Australia Then, the lowmoves rapidly eastward, accompanied by the north-eastward propagation ofthe cold front Strong south-westerly wind (12–15 ms−1) prevails behind the

Fig 2.6 (a) Transect through a cool change along the axis normal to the front,

showing the height of the advancing wedge of cold air and the wind vectors at threesites: ahead of the front (site A), immediately behind the front (site B) and about 30

km behind the front (site C); (b), (c), (d) and (e): respectively, surface wind speed,wind direction, surface pressure and air temperature along the transect

cold front, causing wide-spread dust emission The dust particles are carried

by the south-westerly to converge to the frontal area, resulting in high dustconcentration near the front They are then transported north-westwards andsouth-eastwards by the diverging flows along the front A portion of the dust

is eventually deposited to the Indian Ocean and the Pacific Ocean

2.3.3 Squall Lines

As thunderstorms develop in an unstable atmosphere, massive warm air mayascend and release latent heat during the formation of clouds These storms

2.3 Atmospheric Systems 25

Fig 2.7 Weather patterns associated with the 8–11 February 1996 dust event in

Australia Near surface wind (vectors, ms−1) and air temperature (contours, K) areshown

are sometimes accompanied by intense horizontal vortices with violent surfacewinds, known as squall lines During summer (May–October) over somedeserts in the southern Sahara, the south-western United States and otherarid areas of the world, squall lines accompanied by vigorous dust storms arefrequently observed As illustrated in Fig 2.8, in front of the storm, buoyantwarm air rises beyond its condensation level and produces heavy rain Dry airreaching the storm at high levels from the rear descends to the ground as it

is cooled by the evaporation of the precipitation droplets As the descendingcurrent hits the ground, it is deflected forward and moves out of the cloud inlarge lobes, forming a high wall of dust which rises up to 1,000–2,500 m Ob-servations show that the flow speed in these lobes can be as large as 50 m s−1and a flow speed around 20 m s−1 is quite common (Lawson, 1971)

26 2 Wind-Erosion Climatology

0

Freezing Level Motion of Storm

Cold Air Downdraught Cold Air Downdraught

Outflow Leading Edge Dust Wall

Front

Tropopause

258 273

243 228 225

Fig 2.8 Formation of a dust storm in a thunderstorm The arrows indicate the

direction of the air currents; the hachured region represents falling or suspendedprecipitation; the stippled region depicts a dust storm caused by the down draught

of cold air (From Pye and Tsoar, 1990)

The development of thunderstorms and squall lines is closely related tothe stability of the atmosphere Over 20 severe dust storms related to squalllines occur each summer in the Khartoum region of the Sudan when themonsoon trough lies over the area Strong surface convection is possible withlight, relatively moist westerlies overlain aloft by moist easterlies, allowingthe formation of cumulonimbus Similar synoptic situations prevail in the aridsouth-western United States, where over 10 dust storms occur each year nearPhoenix, Arizona They are initiated by moist tropical air from the Pacificoverlain by an upper flow of moist air from the Gulf of Mexico Dust storms ofthis kind are usually short-lived local events, lasting only a few hours, and aresometimes followed by heavy rain which settles the dust and stops erosion

2.4 Global Wind-Erosion Patterns

Wind erosion occurs both on natural and disturbed soil surfaces Naturalaeolian surfaces are found mainly in desert areas, while disturbed soil surfacesare found mainly in the semi-arid areas that are under the influence of humanactivities, such as clearing of native vegetation and cultivation The aridity

of the semi-arid regions, for instance the Saharan/Sahelian boundaries, isstrongly influenced by climate change on decadal time scales

2.4 Global Wind-Erosion Patterns 27

Fig 2.9 Global distribution of leaf-area index for March 2004 and the locations of

the world’s major deserts, including: (1) Great Basin, (2) Sonoran, (3) Chihuahua,(4) Peruvian, (5) Atacama, (6) Monte, (7) Patagonia, (8) Sahara, (9) Somali-Chabli,(10) Namib, (11) Kalahari, (12) Karroo, (13) Arabian, (14) Rub al Khali, (15)Turkestan, (16) Iranian, (17) Thar, (18) Taklimakan, (19) Gobi, (20) Great Sandy,(21) Simpson, (22) Gibson, (23) Great Victoria and (24) Sturt

The locations of the major deserts are shown in Fig 2.9 Most of theworld’s dry lands are concentrated on the major continents in the subtropicalregion between 20◦and 30◦in both hemispheres, except those in East Asia inhigher latitudes around 40◦N Overall, the dry lands occupy about one third ofthe Earth’s land surface, with about 4% of them being classified as extremelyarid, 15% as arid and 14.6% as semi-arid (Cooke, 1973) The vegetation cover

in the world is closely related to the availability of water for plant growth As

an example, Fig 2.9 shows the distribution of leaf-area index over the globefor March 2004 Leaf-area index is defined as the total one-sided leaf areaover a unit ground-surface area and is an indicator of the vegetation cover.Deserts, arid and semi-arid regions correspond well with regions of small leaf-area index

Wind erosion on disturbed soil surfaces can be many times stronger than

on natural soil surfaces This is because freshly exposed soils usually containmore erodible particles than natural soils, in which fine material is likely tohave been already depleted Also, disturbed soil surfaces generally have a lowerwind-erosion threshold friction velocity, since they are disrupted mechanically

by agricultural activities

Tegen and Fung (1995) carried out an analysis of the potential areas ofwind erosion over the world They introduced six surface classes with 1◦ × 1 ◦

resolution, which are

• Desert and sparsely vegetated soils (NS)

• Old cultivated eroded soils (OA)