Land Use Change, Vegetation Dynamics and Rainfall Spatio-Temporal Variability Over West Africa

LAND USE CHANGE, VEGETATION DYNAMICS AND RAINFALL SPATIO-TEMPORAL VARIABILITY OVER WEST AFRICA

BAMBA, Adama BSc (UAA, côte , MSc (UAA, côte (MET/11/7675)

A THESIS

SUBMITTED TO THE SCHOOL OF POSTGRADUATE STUDIES, THE FEDERAL UNIVERSITY OF TECHNOLOGY, AKURE (FUTA) IN PARTNERSHIP WITH THE WEST AFRICAN SCIENCE SERVICE CENTRE ON CLIMATE CHANGE AND ADAPTED LAND USE (WASCAL), IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE AWARD OF DOCTOR OF PHILOSOPHY DEGREE IN METEOROLOGY AND CLIMATE SCIENCE.

DEPARTMENT OF METEOROLOGY AND CLIMATE SCIENCE

FEDERAL UNIVERSITY OF TECHNOLOGY AKURE,

NIGERIA

MARCH, 2015

ABSTRACT

The decadal variability of rainfall and vegetation over West Africa have been studied over the last three decades, 1981-1990, 1991-2000 and 2001-2010 denoted as 80s, 90s and 00s respectively. The Climate Research Unit (CRU) monthly precipitation and temperature data, the Global Prediction Climatology Project (GPCP) monthly precipitation data, daily rainfall data from two stations in and Normalized Difference Vegetation Index (NDVI) from the National Oceanic and Atmosphere Administration (NOAA), all covering the period

1981-2010 have been used in this study. The aim of the study is to ascertain how changes in the land surface state affect the spatio-temporal distribution of rainfall over the West Africa region. The relationship between rainfall and vegetation indices over the region was determined . Also the decadal comparison between rainfall and

NDVI over the region was based on the significant t-test and the Pearson . The impact of land use change (deforestation) on West African monsoon, particularly the rainfall of June-July- August-September (JJAS) was simulated using the regional climate model

(RCM) RegCM version 4. The RegCM4 was coupled with the Biosphere-Atmosphere

Transfer Scheme (BATS) land surface state model and forced with ERAINT. The simulation covered three years from 2005 to 2007. But the study focused on JJAS of 2005 and 2006, two different years in terms of monsoon onset and sea surface temperature (SST) over the Gulf of

Guinea (GG). 2005 was characterized by early monsoon onset and cold SST while 2006 had late monsoon onset and high SST over GG. The model performance was evaluated by comparing the model output with GPCP and CRU observation datasets. Results show that significant return to wetter conditions is observed between the decade 80s and decade 90s over West Africa and it was maintained during decade 00s except over central part of Benin and the western side of Nigeria where a decrease in annual rainfall magnitude was observed.

During the same period, a re-greening of the Central Sahel and Sudano-Sahel regions was noted. From decade 90s to 00s, this re-greening belt was observed toward the South and the

Coastal areas, mainly over the Guinea Coast, Sudano-Guinea and Western Sahel regions. A northward movement of vegetation increase was also observed. A linear relationship was observed between rainfall and NDVI in the savannah region. Linear relationship between rainfall and NDVI was not observed in other regions. This may suggest that the re-growth of vegetation in the savannah region may be linked to the availability of the rainfall. The vegetation re-greening which was observed over the Sahel region in 90s following the recovery of rainfall from the drought of the 80s was not sustained in the decade 00s due to a slight reduction in rainfall. The RegCM4 simulation results indicate that, the model was able to reproduce the early and late monsoon onsets over the Sahel in 2005 and 2006 respectively, when no changes were made to the vegetation cover. After changes were made to the vegetation cover, the RegCM4 simulation results show that the position of structures like the

Intertropical Discontinuity (ITD), and the zones of ascending and descending motion over the

Sahara desert were not affected much; however the changes in vegetation were observed to have delayed the monsoon onset over Sahel. The time lag between the end of monsoon in

Guinea region and its onset in Sahel is more than one month (~45 days) in 2005 and around one month (30 days) in 2006. However, the time lag between the end of monsoon in Guinea region and its onset in Sahel is always less than a month in the simulation experiments without changes in vegetation cover. The results of the study have shown the impact of deforestation over the West African savannah zone on rainfall spatial and temporal variability and provided maps of rainfall and vegetation index variability that can guide decision making for policy makers to prevent further deforestation and soil degradation.

iii

RESUME

Cette thèse étudie l'influence du changement d'état de surface de la végétation sur la variabilité spatio-temporelle des précipitations en Afrique de l'Ouest. , la variabilité décen végétation sont révisées au cours des trois dernières décennies 1981-1990, 1991-2000 et 2001-2010 notées respectivement 80s, 90s et

00s. Le jet de données utilisé est constitué des données de précipitation du Climate Research

Unit (CRU), prévision mondial de climatologie projet (GPCP) et quelques données de station comme observation et

ilisation de normalisé de végétation (NDVI) du groupe Global Inventory

Monitoring and Modeling Studies (GIMMS). impact de la déforestation sur la variabilité des précipitations dans la région est étudié en utilisant le modèle climatique régional RegCM version 4 International Centre for Theoretical Physics Le principe est d'anticiper le phénomène de la déforestation en attribuant à la zone de transition située entre la région guinéenne et la région soudanaise (9 W-15 E et 6 N-10 N) des herbes courtes. Cette zone qui est constituée par la savane arborée et parsemée de hautes herbes. Le modèle a été évaluée à l'aide GPCP, CRU ensembles de données d'observation. La méthodologie basée sur l'utilisation du test de significativité statistique entre les différentes décennies et la corrélation de Pearson entre les précipitations et NDVI. En ce qui concerne, l'expérience de la déforestation, le modèle a été couplé avec le model Biosphere-Atmosphere Transfert (BATS) modèle d'état de surface et forcé avec ERAINT. La simulation

(2005-2006-2007), mais l'accent a été particulière mise sur la période JJAS de 2006 et 2007

correspondant à la saison des pluies dans la région sahélienne. Les principaux résultats sont indiqués comme suit: De 80 décennie à 90s de la décennie, un important retour

à des conditions plus humides est observée sur l'Afrique de l'Ouest et confirmé lors de la 00s

décennie, sauf sur le Bénin centrale et tout le côté ouest du Nigeria où il ya une diminution annuelle précipitations. Et au cours de la même période, un reverdissement des régions du

Sahel Central et soudano-sahéliens est noté. De la décennie 90 à 00, cette ceinture de reverdissement est observée vers le les régions côtières du Sud et, principalement sur les régions de la côte Guinéenne, soudano-guinéennes et du Sahel occidental. Le reverdissement de la végétation est observé dans le sens Sud-Nord. Néanmoins, dans la décennie 90s les changements positifs étaient en dessous de la latitude 10 N et pendant la décennie 00s il a atteint la latitude 12 N plus haut dans les régions comme une frontière Mali, la Guinée et le

Sénégal. Une relation linéaire qui se manifeste par une forte corrélation a été trouvée entre les précipitations principalement dans les régions de savane. La déforestation est un processus en cours sur la région Afrique de l'Ouest en dépit de son effet négatif sur l'environnement et le climat régional. Pour l'aspect simulation, le modèle a été capable de reproduire la mousson précoce et tardive sur le Sahel respectivement en 2005 et 2006 avant le changement apporte au couvert végétal. Après les changements de la couverture végétale, il est constaté que la position des différentes structures telles que le FIT, les zones de convection et des zones de subsidence ne sont pas affectés beaucoup; Cependant, les changements ont eu un impact sur la mousson au le Sahel. En effet, il semble retarder l'apparition de la mousson au le Sahel. La

entre la fin de la mousson dans la région guinéenne et son apparition au

Sahel est plus d'un mois en 2005 et environ un mois en 2006, alors que cette période est inferieure moins à mois avant la déforestation. Par conséquent, elle a réduit les précipitations de JJAS à respectivement 5% et 3% en 2005 et 2006. L'analyse décennale des observations des précipitations et de la végétation ainsi que la simulation ont montré que le reverdissement de la végétation dans la région diffère dans le temps zone à ; cependant, la déforestation induirait une baisse des précipitations dans la région.

ACKNOWLEDGEMENT

This Ph.D programme is fully supported by the German Ministry of education and Research

(BMBF) through the West African Science Service Centre on Climate Change and Adapted

Land Use (WASCAL). I am therefore grateful to WASCAL for granting me the financial support for the study and research visit to the International Centre for Theoretical Physics

(ICTP) in Italy and to the Laboratoire d'Etude des Transferts en Hydrologie et Environnement

(LTHE) in France and my participation in conference to present the results of the study.

I sincerely thank the executive Director and the staff of WASCAL Head office, Accra, Ghana and the Director, Prof. J.A. Omotosho and staff of WASCAL GRP-WACS, FUTA, Nigeria for their strong support and encouragement throughout the period of the study. I am also grateful to Prof. K.O. Ogunjobi, the Head, and the Staff of the Department of Meteorology and

Climate Science, FUTA, Nigeria, for their assistance and cooperation.

I am deeply thankful to my supervisor Prof. Arona Diedhiou at Institut de Recherche pour le

Développement (IRD) and LTHE and Co-Supervisor, Dr. Ahmed Balogun of FUTA, Nigeria, as well as my advisors Prof. Abdourhamane Konaré from Université Félix Houphouet Boigny

(UFHB), Pole Scientifique et (PSI) and Dr. Thierry Pellarin from the

Laboratoire d'Etude des Transferts en Hydrologie et Environnement (LTHE) Grenoble,

France for their assistance, guidance and immense contribution to make this thesis a reality.

I am grateful to Prof. Savané Issiaka and Prof. Kamagaté Bamory from Université Nangui

Abrogoua (UNA) former Université Abobo-Adjamé (UAA) for giving me the foundation tool for research and their constant counsel.

Many thanks to my external examiner, Prof. T.O. Odekunle from Obafemi Awolowo

University, Ile-Ife, Osun state, who help me to improve my thesis.

vii

My sincere gratitude to my colleagues from the WASCAL GRP WACS, FUTA, Nigeria for the good relationship we had during the training and the thesis writing.

I am also grateful to colleagues from the CNC Bingerville and from UNA in the circumstances Ouédraogo Moussa and Ouattara Ismael with whom I shared many experiences and also to Ismaila Diallo at ICTP for his assistance during my stay.

To my family, I am profoundly grateful to my uncles Koné Dramane and Koné Amara, my brothers Bamba Yaya and Bamba Amadou, my sisters Bamba Salimata and Bamba Nabintou etc. for their assistance, patience and supports.

To you all whose names are not officially mentioned in this document that we shared times, ideas, materials etc. together during all these years for the achievement of this goal in Cote

Morocco (CRASTE), Ghana (UCC), Nigeria (FUTA), Italy (ICTP) and France

(LTHE, Grenoble) please find through these simple words my sincere acknowledgements.

A thought for my deceased father Bamba Siriki who I did not know, may your soul stay in peace in paradise.

In God I believe, and I thank him for his omnipresence in my life

viii

DEDICATION

TABLE OF CONTENTS

ABSTRACT ...... ii

RESUME ...... iv

CERTIFICATION ...... vi

ACKNOWLEDGEMENT ...... vii

DEDICATION...... ix

TABLE OF CONTENTS ...... x

ACRONYMS ...... xiv

LIST OF FIGURES ...... xvii

LIST OF TABLES ...... xxii

Chapter 1 ...... 1

INTRODUCTION...... 1

1.1 Overview ...... 1 1.2 Statement of the Problem ...... 4 1.3 Justification ...... 7 1.4 Aim and Specific Objectives ...... 8 Chapter 2 ...... 9

LITERATURE REVIEWS ...... 9

2.1 Rainfall Spatio Temporal Variability over West Africa ...... 9 2.1.1 West African 1970s Drought Causes ...... 9 2.1.2 The Intertropical discontinuity and Intertropical convergence zone over West Africa ...... 11 2.1.3 Impact of climate change and the additional radiative forcing ...... 16 2.2 Vegetation Dynamics and Rainfall Variability over West Africa...... 17 2.3 The Relationship Rainfall and Vegetation ...... 19 2.4 Vegetation Indices over West Africa ...... 20 2.5 The West African Monsoon ...... 23 2.6 An Overview of Synoptic Scale Atmospheric Features over West Africa ...... 26 2.7 Climate Models ...... 28

2.7.1 Climate and general-circulation models ...... 28 2.7.2 Regional climate models: RegCM ...... 31 Chapter 3 ...... 34

RESEARCH METHOD ...... 34

3.1 Study Area ...... 34 3.1.1 Study Area Location ...... 34 3.1.2 The Vegetation over West Africa ...... 36 3.1.3 Regional climatology of West Africa ...... 39 3.2 Data Collection ...... 46 3.2.1 Rainfall from station data...... 46 3.2.2 Climate Research Unit data ...... 46 3.2.3 Global Precipitation Climatology Project ...... 47 3.2.4 Vegetation Indices Data form GIMMS ...... 48 3.2.5 Forcing Parameters ...... 49 3.3 Data Analysis ...... 51 3.3.1 Observation data processing ...... 51 3.3.1.1 Significance t-test of Differences ...... 51 3.3.1.2 Standardized Precipitation Index ...... 52 3.3.1.3 Correlation between rainfall and NDVI ...... 53 3.3.2 Model Setting and Simulation ...... 53 3.3.2.1 Model Description and Simulation ...... 53 3.3.2.2 Model Evaluation ...... 56 3.3.2.3 Land Surface Model ...... 56 3.3.2.4 Structure of BATS ...... 59 3.3.2.5 Experimentation: Change in Land Surface State ...... 60 3.3.2.6 Case study of 2005 and 2006 ...... 65 Chapter 4 ...... 67

RESULTS AND DISCUSSION ...... 67

4.1 Mean NDVI and Rainfall over Last Three Decades ...... 67 4.1.1 Vegetation and Spatial Distribution of Rainfall over West Africa ...... 67 4.1.2 NDVI Decadal Variability over West Africa during 1981-2010 ...... 69 4.1.2.1 Decadal Mean of NDVI ...... 69

4.1.2.2 Seasonal Variability of NDVI ...... 71 4.2 Spatio-Temporal Distribution of the Rainfall over Three Last Decades ...... 73 4.2.1 Changes compare to thirty years climatology ...... 73 4.2.2 Decade to decade changes ...... 75 4.2.3 Rainfall distribution ...... 77 4.2.4 Upward Trend of the Rainfall over the Region ...... 79 4.2.5 Changes over Seasons ...... 81 4.3 Spatio-Temporal Distribution of the Vegetation over Three Last Decades ...... 83 4.3.1 NDVI Decadal Anomaly Variability over West Africa ...... 84 4.3.2 NDVI Decadal Variability over West Africa ...... 86 4.3.3 Decadal Change on NDVI ...... 88 4.3.4 NDVI seasonal variability over West Africa ...... 90 4.3.5 Frequency distribution of the NDVI ...... 93 4.4 Relationship between Rainfall and NDVI over West Africa ...... 95 4.4.1 Intra Annual Variability of Rainfall and NDVI ...... 96 .. 97 4.4.3 Rainfall Intra-seasonal Variability ...... 101 4.4.4 NDVI Intra-seasonal variability ...... 103 4.4.5 Rainfall and NDVI Monthly Climatology ...... 106 4.4.6 Relationship between Rainfall and NDVI over West Africa ...... 109 4.5 Changes in Atmospheric Parameters...... 114 4.5.1 Upper, middle and lower levels tropospheric winds ...... 118 4.5.2 Tropical Easterly Jet ...... 118 4.5.3 African Easterly Jet ...... 120 4.5.4 Monsoon fluxes ...... 122 4.5.5 Zonal Wind, Convection and Wind Velocities ...... 125 4.5.6 Relative humidity and Convection...... 131 4.5.7 Change in Surface Temperature, Evapotranspiration Flux and Albedo ...... 135 4.5.8 Rainfall Seasonal Variability ...... 136 Chapter 5 ...... 147

CONCLUSION AND RECOMMANDATIONS ...... 147

5.1 Conclusion ...... 147 5.2 Recommendations ...... 149

xii

5.3 Limitations of the Study ...... 149 REFERENCES ...... 151

xiii

ACRONYMS

AEJ: African Easterly Jet

AEWs: African Easterly Waves

AGRHYMET: Centre Régional de Formation et d'Application en Agrométéorologie et Hydrologie Opérationnelle

AHVRR: Advanced Very High Resolution Radiometer

AMMA: Analyse Multidisciplinaire Mousson Africain

ACM2: Atmospheric convective Model Version2

BATS: Biosphere-Atmosphere Transfer Scheme

BMBF: German Ministry of education and Research

CIAT: International Center for Tropical Agriculture

CRU: Climate research Unit

DJF: December-January-February

ECMWF: European Centre for Medium-Range Weather Forecasts

ENSO: El Nino North South Oscillation

EOP: Enhanced Observing Period

FAO: Food and Agriculture Organization of the United Nation

FGGE: First GARP Global Experiment

GDP: Gross Domestic Product

GG: Gulf of Guinea

GIMMS: Global Inventory Monitoring and Modeling Studies

GPCP: Global Prediction Climatology Project

GTS: Global Telecommunications System

ICTP: International Centre for Theoretical Physics

IPCC: Intergovernmental Panel on Climate Change

IRD: Institut de Recherche pour le Développement

xiv

ITD: Inter-Tropical Discontinuity

ITF: Intertropical front

JJA: June-July-August

JJAS: June-July-August-September

LTHE: Laboratoire d'Etude des Transferts en Hydrologie et Environnement

MAM: March-April-May

MCS: Meso-scale Convective System

NCAR: National Center for Atmospheric Research

NCDC: National Climatic Data Center

NCEP: National Centers for Environmental Prediction

NDVI: Normalized Difference Vegetation Index

NOAA: National Oceanic and Atmosphere Administration

NORMER: Normal Mercator

RCMs: Regional Climate models

RegCM: Community Regional Climate Model of ICTP

SON: September-October-November

SPI: Standardized Precipitation Index

SSR: Sudano-Sahel Region

SST: Sea Surface Temperature

SUBEX: Subgrid Explicit moisture scheme

TEJ: Tropical Easterly Jet

TRMM: Tropical Rainfall Measuring Mission

UER: Upper East Region

WAM: West African Monsoon

WASCAL: West African Science Service Centre on Climate Change and Adapted Land Use

WMO: World Meteorology Organisation

WRF: Weather Research and Forecasting model

xvi

LIST OF FIGURES

Figure 1. 1: Flood of September 2009 in Burkina Faso [a]; and cultivated land in wooded 6

Figure 2. 2: Schematic depicts relative latitudinal positions of the ITCZ/ITD, TEJ, AEJ-N, AEJ-S and the WAJ during wet and dry years for 0 10 N (top) and 0 -12 S (bottom) (Refer online version for color images) from Williams and Kniveton, (2011). 13

Figure 2. 3: Schematic of the atmospheric circulation in the West African monsoon system during the boreal summer. Closed solid lines represent the isotachs of the African Easterly Jet (AEJ), which lies around 600 hPa. The red arrows show the thermally direct meridional monsoon circulation, and are typical of the time-mean winds in the peak monsoon season (Lebel et al., 2010). 25

Figure 3. 1: Map of Africa showing the Study area over West Africa with six selected sites over Sahel (Burkina Faso and Niger), Sudanian Savannah (Mali and Benin) and Guinea region (Ivory Coast and Ghana) for propose of the intra annual variability studies. 35

Figure 3. 2: Wooded savannah landscape in the Lamto reserve located in central part of Cote

Africa. 38

Figure 3. 3: Temperature and rainfall monthly mean in Bawku, Upper East Region of Ghana (1993 2011). 41

Figure 3. 4: Lamto Geophysical station where many climate and seismic parameters are measured; The station was created in 1962 by Maxime Lamotte and Jean-Luc Tournier. 43

Figure 3. 5: Temperature and rainfall monthly mean at Lamto station, Guinea region in Cote 2004). 43

Figure 3. 6: Intra seasonal Temperature mean at the Lamto station averaged over 1971-2000. 44

Figure 3. 7: Temperature Anomaly at the Lamto Station plotted with regard to climatology of 1971-2000. 45

xvii

Figure 3. 8: Model simulation domain with the topography (in contour) the zone of interest in red box. 54

Figure 3. 9: Schematic of individual physical process (Dickinson, 1993) 58

Figure 3. 10: Initial land cover obtained after RegCM4 domain simulation before changes in land cover (left) and after making changes in land cover (right); more details about the legend in Table 3.2. 62

Figure 3. 11: Diagram showing the differences on SST in Gulf of Guinea and monsoon onset between 2005 and 2006 over West Africa. 66

Figure 4. 1: Map of West Africa showing the rainfall climatology (1971-2000) based on CRU observation data (contour); NDVI climatology of yearly sum (shaded); the filled triangles represent the site where rainfall and NDVI have been selected for the intra variability study. 68

Figure 4. 2: Mean annual Rainfall (mm yr -1) shown in [a]; [b] and [c] for respectively decades 80s 90s and 00s and mean annual NDVI shown in [d]; [e]; [f] for respectively decades 80s; 90s and 00s. 70

Figure 4. 3: NDVI decadal mean showing changes in vegetation cover with progressive southward increase in decade 80s (a, b, c and d); decade 90s (e, f, g and h) and decade 00s (i, j, k and l). 72

Figure 4. 4: Spatial distribution of rainfall significant changes in decade [a] 80s; [b] decade 90s and [c] decade 00s compare to 30 years average (1981-2010) over West Africa at a level of 95%. 74

Figure 4. 5: Decadal changes in rainfall seasonal spatial distribution over West Africa at a level of 95%. Blue lines are areas with significant changes. 76

Figure 4. 6: Time latitudinal diagrams of rainfall seasonal mean (a, b and c) in decade 80s; decade 90s and decade 00s respectively and seasonal anomaly (d, e and f) for decade 80s; decade 90s and decade 00s respectively. 78

Figure 4. 7: SPI at six locations in Burkina Faso and Niger (Sahel); Mali and Benin (Sudan) and Ivory Coast and Ghana (Guinea Coast) over West Africa showing more wet condition mainly apart from decade 1990. The climatology is based on 1981-2000 rainfall mean. 80

Figure 4. 8: Decadal changes in rainfall seasonal spatial distribution between decade 80s-90s (a, b, c and d), decade 90s-00 (e, f, g and h) and decade 80s-00s (i, j, k and l) over West Africa at a level of 95%; Blue lines are areas with significant changes. 82

xviii

Figure 4. 9: Annual significance t-test computed between decades showing significant positive and negative changes in vegetation cover over West Africa in a, b and c for decades 80s; 90s and 00s respectively compare to 30 years average (1981-2010) at a level of 95%. 85

Figure 4. 10: Annual significance t-test computed between decades showing the positive and negative changes in vegetation cover over West Africa between decade 80s-90s [a]; decade 80-00s [b] and decade 00s-90s [c] at a level of 95%. 87

Figure 4. 11: Time latitudinal diagrams of seasonal NDVI shown in a, b and c for decade 80s, 90s and 00s respectively and NDVI seasonal anomaly - own in d, e and f for decades 80s, 90s and 00s. 89

Figure 4. 12: Positive and negative changes in vegetation cover over West Africa between seasons in decade 80s-90s (a, b, c and d); seasons in decade 80-00s (e, f, g and h) and seasons in decade 00s-90s (i, j, k and l) at a level of 95%. 91

Figure 4. 13: Annual frequencies and distribution of NDVI at Niger (left) and Burkina Faso (right) sites over Sahel region. 93

Figure 4. 14: Annual frequencies and distribution of NDVI at Mali (left) and Benin (right) sites over Sudanian region. 94

(right) sites over Guinea region. 95

Figure 4. 16: Decadal rainfall at the monthly timescale plotted for the six selected sites in the Sahel region ([a] Niger and [b] Burkina Faso), the Sudanian region ([c] Mali and [f] Benin) 102

Figure 4. 17: NDVI decadal mean averaged over months at six different points over Sahel region [a] Niger and [b] Burkina Faso, Sudanian regions [c] Mali and [f] Nigeria and Guinea 105

Figure 4. 18: NDVI and rainfall monthly mean averaged respectively over 1981-2012 and 1981-2006 at six different points over Sahel region [a] Niger and [b] Burkina Faso, Sudanian region [c] Mali and [f] Benin and Guine 107

Figure 4. 19: Spatial correlation between NDVI and rainfall over [a] decade 1980s, [b] decade 1990s and [c] decade 2000s significance areas at 95% confident level. 110

Figure 4. 20: Scatter plot showing correlation and linear equation between rainfall and NDVI over 1981-2010 at six different points over Sahel region [a] Niger and [b] Burkina Faso,

xix

112

Figure 4. 21: Rainfall monthly mean biases computed over June-July-August-September (JJAS) based on CRU a) and b) and GPCP c) and d). 115

Figure 4. 22: Temperature monthly mean biases computed over June-July-August-September (JJAS) based on CRU a) 2005 and b) 2006. 116

Figure 4. 23: JJAS mean Tropical Easterly Jet at 200 hPa before changes [a] and [c]; and after changes in vegetation cover [b] and [d] in 2005 and 2006. 119

Figure 4. 24: JJAS mean Africa Easterly Jet at 700 hPa before changes [a] and [c]; and after changes in vegetation cover [b] and [d] in 2005 and 2006. 121

Figure 4. 25: JJAS mean Monsoon fluxes at 850 hPa before changes [a] and [c]; and after changes in vegetation cover [b] and [d] in 2005 and 2006. 124

Figure 4. 26: Time latitudinal variability of JJAS zonal wind mean at 850 hPa averaged along vegetation cover (c and d) in 2005 and 2006. 126

Figure 4. 27: Vertical cross-section of the wind vel - of 2005 showing convergence and divergence zones a) before the change and b) after the change. 128

Figure 4. 28: Vertical cross-section of the wind velocity - of 2006 showing convergence and divergence zones a) before the change and b) after the change. 128

Figure 4. 29: Time latitudinal variability of JJAS wind velocity mean at 925 hPa averaged vegetation cover [c] and [d] in 2005 and 2006. 130

Figure 4. 30: Vertical cross-section of relative humidity in percentage (shaded) and the wind - 2005 [a] and in JJAS of 2006 [b]. 132

Figure 4. 31: Vertical cross-section of relative humidity in percentage (shaded) and the wind - 2005 [a] and in JJAS of 2006 [b]. 134

Figure 4. 32: Time latitudinal diagram of daily and monthly mean rainfall (mm day -1) - b) and d) with change in surface in JJAS 2005. 138

Figure 4. 33: Time latitudinal diagram of daily and monthly mean rainfall (mm day -1) - b) and d) with change in surface in JJAS 2006. 140

Figure 4. 34: JJAS rainfall biases between the experiment without changes in vegetation cover and the experiment with changes in vegetation cover in 2005 [a] and 2006 [b]. 141

Figure 4. 35: Rainfall averaged over sub Sudanian band before changes (CTL) and after changes (Sens) in vegetation cover in 2005 [a] and in 2006 [b]. 143

Figure 4. 36: Rainfall averaged over sub Sudanian band before changes and after changes in vegetation cover in 2005 and 2006. 145

xxi

LIST OF TABLES

Table 3. 1 Drought categories from SPI (McKee et al ., 1993) ...... 52

Table 3.2: Summary of the model configuration ...... 55

Table 3.3: Land cover/vegetation classes ...... 62

Table 3.4: BATS vegetation/land-cover (Dickinson, 1993) ...... 64

Table 4.1: Descriptive Statistics for the rainfall and NDVI time series ...... 96

Table 4.2: Monthly Cross-correlation between AHVRR NDVI and Rainfall at Lamto station (1981-2000); Correlation is significant at 0.05 level and (**) correlation is significant at 0.01 level. The indices p and n are respectively rainfall and NDVI...... 98

Table 4.3: Monthly Cross-correlation between AHVRR NDVI and Rainfall at Daloa (1981- 2000); Correlation is significant at 0.05 level and (**) correlation is significant at 0.01 level. The indices p and n are respectively rainfall and NDVI...... 100

Table 4.4: Brief description of observational datasets and model used to set the simulation of RegCM4.4 ...... 117

Table 4.5: Changes in Evapotranspiration, Temperature and Albedo due to vegetation cover change in JJAS of 2005 and 2006 over changed band...... 135

xxii

Chapter 1

INTRODUCTION

1.1 Overview

In many African regions, the land surface has faced some considerable amount of pressure over decades due to high rate of deforestation and exploitation. To compound matters, West

African countries have experienced some drought sequences between 70s and 80s. During these periods, the Sahel has experienced the most substantial and sustained decline in rainfall recorded anywhere in the world (Hulme and Kelly, 1997). These have induced some changes in vegetation cover over the region due to a close link between vegetation and rainfall. Many studies have shown the influence of precipitation on vegetation production (e.g. Herrmann et al., 2005; Che et al. , 2014), which in turn controls the spatial and temporal occurrence of grazing and favours nomadic lifestyle (Sivakumar, 2007). Furthermore, the rate of population increase over the world over the region is one of the highest in the world (Yuen and Kumssa,

2011). Also people live in poor conditions, so natural resources in general and particularly the natural vegetation cover are Deforestation has become one of the major phenomena impacting the climate over West Africa. However, desert conditions are induced by gradual and prolonged loss of vegetation cover over extensive land areas in a country, and across two or more countries. With time, desertification over West African regions has transformed extensive land areas into arid- and semi-arid zones and has therefore created a vast land area that has no sufficient quality life-supporting natural resource base and weather conditions. Furthermore, this permanent loss of vegetation cover leads to reduction in soil moisture that curtails biodiversity productivity and permit

drought conditions to persist (Charney, 1975; Hulme and Kelly, 1997). In this regard, the capacity of the original vegetation land cover to regenerate is severely impaired because of near-total absence of rainfall. However, ambient temperatures are typically very high, a factor that sustains continuous huge losses of soil moisture and the few available water bodies through direct evaporation. Unusually high wind speeds, low and high atmospheric pressure spots, etc. sustain these adverse climatic conditions.

Thus, poor land utilization practices, especially in subsistence farming and nomadic pastoral economies in the majority of the African countries have accelerated the loss of natural vegetation and exacerbated the problem of climate change.

The Guinea zone of West Africa is the area with intense economic activities which are mainly agriculture, wood trade, and mine exploitation, particularly in , where the economic growth is supposed to be based on agriculture. This region is drastically affected by the above activities. With regard to the importance of the land surface state in the water cycle, many researches focused on the relationship between rainfall and vegetation. Since the long drought period of the 70s and 80s, the relationship between land surface state and atmospheric parameters has been the focal point of important debates, and subject of many studies

(Nicholson et al., 1990; Diedhiou et al., 1999; Dickinson, 2003 and Laux et al., 2007).

According to Herrmann et al. (2005), in semiarid environments, the relationship between rainfall and vegetation has received a great deal of interest. Indeed, changes in land surface condition are suspected to be the precursor to changes of rainfall spatial and temporal distribution, precisely during drought years over Sahel region. During these periods, changes in land surface condition have been clearly linked to the deforestation in the Guinea region of

West Africa (Charney, 1975). Charney (1975) linked the Sahel 70s drought to the change in the surface state based on the relationship between rainfall and the land surface state: the human activities have decreased the forest cover which in turn has increased the surface

albedo to create a subsidence over the region and decreased the precipitation which finally have a negative impact on vegetation cover. This point of view is shared later by Sivakumar

(1992) and Herrmann et al. (2005) who linked this drought to anthropogenic land use changes.

As a result of some poorly implemented environmental policies adverse anthropogenic activities on land cover are still going on despite the warnings from scientists and. In contrast to previous findings recent studies have shown some recovery of rainfall and re-greening of vegetation over these regions (Rasmussen et al., 2001). The vegetation cover measured through the chlorophyllous activity of the plants indicates a vegetation recovery. Meanwhile, the regional rainfall fluctuations responded to the strong seasonal influence of the West

African Monsoon and associated deep convections processes. This response is related to the thermal gradient between the ocean and continental zone (Eltahir and Gong, 1996). This is in agreement with Omotosho (1990); Rodwell and Hoskins, (1996) and Mathon and Laurent

(2001).

In order to enhance the knowledge on West African climate and fill data gaps in Africa, capacity building is required on methods and skills. Areas needing immediate attention are remote sensing applications on land use change and climate systems. The remote sensing was firstly based on coarse resolution, and subsequently improved to higher resolution. Also, interest in regional climate modelling has steadily increased in the last two decades (Giorgi,

2006). As a result, a number of regional climate models (RCMs) have been developed, with a wide base of model users. One such RCM is the RegCM system, which has evolved from the first version developed in the late 1980s (RegCM1; Dickinson et al., 1989; Giorgi, 1990) to later versions in the early 1990s (RegCM2; Giorgi et al., 1993), late 1990s (RegCM2.5;

Giorgi and Mearns, 1999), and 2000s (RegCM3 Sylla et al., 2010; and RegCM4 Pal et al.,

2007; Giorgi et al., 2012 ). The RegCM model was the first limited area model developed for

long-term regional climate simulation: it has been used in numerous regional model intercomparison projects, and it has been applied by a large community for a wide range of regional climate studies, from process studies to paleo-climate and future climate projections

(Giorgi and Mearns, 1999 and Giorgi et al., 2006).

The need to understand the West African climate system led to the establishment of important research programmes over the region among them are the African Monsoon Multidisciplinary

Analysis (AMMA) (Janicot et al., 2008; Lebel et al., 2009 and Lebel et al., 2010) and, the

West African Science Service Center on Climate Change and Adapted Land Use (WASCAL)

(www.wascal.org; Bliefernicht et al., 2012). Our attention was drawn to the particular area in between Sahel and the Guinea zone where in addition to wood exportation trade and agriculture, the bush fire is also one the major cause of the loss of vegetation cover. Scientists and decision makers have a crucial role to play for in reducing the rate of deforestation and damage within the ecosystem by formulating policies that are guided by scientific evidence.

For the purposes of this study, attention was on parameters such as Normalized Difference

Vegetation Index (NDVI) and the gridded Climate Research Unit rainfall (CRU) data. The statistical relationship between the recent changes in the NDVI and rainfall was evaluated.

Then, the RegCM model is used for the modeling aspect of the work which concerns simulating rainfall and atmospheric features under two different experiments.

1.2 Statement of the Problem

According to IPCC (2013), Africa is one of the most vulnerable continents to the impact of climate change and climate variability. This situation is aggravated by the interaction of

coupled with the low adaptive capacity of the continent. Furthermore, changes in a variety of ecosystems are already being detected, particularly in African ecosystems, at a faster rate than anticipated (Boko et al., 2007).

Specifically, the West African region is going through some extreme events were mainly flood these last years. For instance some dramatic extreme events are observed respectively

1st September 2009 in Ouagadougou (Burkina Faso) (Fig. 1.1[a]), in Dakar (Senegal), 2010 and 2012 in Lagos and Abuja (Nigeria) and recently in June 2014 in Abidjan ( ) that led to wanton loss of lives and property. Land surface degradation is a continuous process going on at a disturbing rate through agricultural and other activities (Fig. 1.1[b])

The consequences of these are socio-economicchallenges that inconvenient the population.

Earlier studies have not established the relationship between land use change and rainfall spatio-temporal variability in the West African region.

This study will investigate the impact of deforestation over the West African savannah zone on rainfall spatial and temporal variability as well as the trend of the distribution of vegetation and rainfall between 1980 and 2010 following the long drought period of the

1970s in West Africa.

Figure 1. 1: Flood of September 2009 in Burkina Faso [a]; and cultivated land in wooded savannah in Niarala village of [b].

So many countries over West Africa have experienced the phenomenon with its associated distress. The risks of flooding and drought have been projected to increase in many areas, the frequency of heavy precipitation events (or proportion of total rainfall from heavy falls) will be very likely to increase over most areas during the 21st century, with consequences for the risk of rain-generated floods (IPCC, 2007; Bates et al., 2008; Quevauviller, 2011). At the same time, the proportion of land surface in extreme drought at any one time is projected to increase (likely), in addition to a tendency for drying in continental interiors during summer, especially in the sub-tropics, low and mid-latitudes. Thus, the change in the surface state deeply affects the rainfall regime through the modification of the onset and cessation of the agricultural process, and the seasonal forecasting. Hence, this study wishes to address the following questions:

(i) what is relationship between the rainfall and the land use over this particular region of

West Africa?

(ii) what are the spatial distribution of the recent trends of rainfall and the vegetation over

West Africa region?

(iii) what could be the model simulation of the impact of deforestation over savannah

zone of West Africa on the rainfall spatio-temporal variability over the region?

1.3 Justification

The land surface cover plays a key role in moderating the climate. However, the vegetation cover has been depleted these last few years over the region where strong variability is going on, characterized by the recurrence of the extreme rainfall and long drought of early 70s up to mid 80s. The anthropogenic effects on changes of the vegetation cover are also a continuous phenomenon. However, according to the FAO report, Africa has recorded the highest deforestation rate (0.7%) throughout the world (FAO, 2000). The deforestation will induce

the land degradation which is an obstacle for food security. Furthermore, the population over

West Africa is one the most vulnerable throughout the world due to the high rate of poverty; in fact the Gross Domestic Product (GDP) is one of the lowest through the world.

Understanding the relationship between rainfall and vegetation cover changes could allow

problems associated with deforestation on their day-to-day life. Improving knowledge on the relationship between rainfall and vegetation could be a sensitisation tool for vegetation and climate management.

1.4 Aim and Specific Objectives

The aim of this study is to ascertain how the changes in the land surface state can affect the spatio-temporal distribution of the rainfall over West Africa region.

The specific objectives are to:

i. assess the evolution of land cover change and rainfall spatio-temporal distribution over

the last three decades;

ii. determine the relationship between rainfall and vegetation indices over West African

region;

iii. simulate the rainfall over West Africa during the rainy season using regional climate

mode (RegCM) and;

iv. evaluate the impact of the land use change on West African monsoon using RegCM.

Chapter 2

LITERATURE REVIEWS

2.1 Rainfall Spatio Temporal Variability over West Africa

2.1.1 West African 1970s Drought Causes

Land surface is an important component of the climate system. It is one of the major driving forces of the regional climate. Therefore, Changes in surface energy budgets resulting from land cover changes can have a profound influence (Sivakumar, 2007).

This can be through the modification of the evapotranspiration, sensible heat flux, the flux of moisture to the atmosphere etc. However, reports on the physical causes of the long drought period in the Sahel region in West Africa during 1970s up to 1980s has progressed along two parallel directions (Giannini et al., 2003; Sanni et al., 2012).

The first group was motivated by the belief that humanity was irreversibly impacting the environment and climate. That, by way of land cover and/or land use changes associated with the expansion of farming and livestock herding into marginal areas, ultimately linked to the pressure of rapid population growth. This will emphasized the role of the feedback between the atmospheric circulation and land surface processes. The fa

(Charney, 1975) some years ago linked the long drought period which occurred during 1970s to 1980s to the deforestation over the West African region (Nicholson et al., 1998; Giannini et al., 2003; Zeng, 2003; Dai et al., 2004; Lebel and Ali, 2009 and Lebel et al., 2010). This phenomenon has had an impact on the rainfall spatio-temporal variability over the region.

This argument is sustained by the positive feedback effects induced by the relation between

rainfall and albedo. The human activities have decreased the forest cover. The absence of vegetation increases the surface albedo therefore, the surface energy budget which creates a subsidence over the region and decreased the precipitation. Finally to close the cycle, the process will affect the vegetation cover (Charney, 1975; Hernandez et al., 2000) as there is no rainfall to support growth.

The second group was revitalized by the initial successes in dynamical seasonal prediction of the mid-1980s, and pointed to the atmospheric response to temperature changes in the global oceans as the leading cause of African climate variability. They found a link between the rainfall spatio-temporal variability over Sahel and the sea surface temperature (SST) (Vizy and cook, 2001; Balas et al., 2007; Odekunle and Eludoyin, 2008 and Oueslati and Bellon,

2015). They concluded a link between the mentioned drought and change in SSTs over

Tropical Atlantic Ocean. Their findings were mainly based on case studies. However, a full demonstration of that has not been done (Giannini et al., 2003; Hamatan et al., 2004). The convergence of precipitation axis is often aligned close to the zone of maximum SST, but is not anchored to it. Indeed, the maximum SST located within the equatorial counter-current is a result of the interactions between the trade winds and horizontal and vertical motions in the ocean surface layer (Giannini et al., 2003).

photosynthetic activities at seasonal and interannual scales (Anyamba, 2005). Nicholson et

al. (1990) found strong relationship between NDVI and rainfall in areas with rainfall amount

ranging between 200 and 1200 mm. Giannini et al. (2003) opined that the land-atmosphere

feedback acts to amplify the ocean forcing of the Sahel precipitation signal.

2.1.2 The Intertropical discontinuity and Intertropical convergence zone over West

Africa

The Intertropical convergence zone (ITCZ) is a zonal band of low atmospheric pressure and thunderstorms caused by converging Trade Winds, rising air and intense thermal heating oscillating on both sides of the equator ( Williams and Kniveton, 2011); the location of the

ITCZ shifts throughout the year defining the wet and dry seasons in countries located in the tropics hence, numerous interest for its study. Barry and Chorley (2003) defined it as the tendency for the trade wind systems of the two hemispheres to converge in the equatorial

(low-pressure) zone and referred to it as referred as the ITCZ over Ocean and ITD over land.

It is one of the direct drivers of rainfall variability in West Africa through perturbations in its strength and position (Odekunle, 2010). The ITD forms the ascending branch of the Hadley cell, thus what affects the strength of the trade winds from either hemisphere will also affect the position and strength of the ITD (Fig. 2.1). The North-South seasonal march of the ITD can bring two rainy seasons to the South in West Africa; however the regions far from the equator to the North will only experience one pronounced dry and wet season (Williams and

Kniveton, 2011). It should be noted that the position and intensity of the ITD does not always imply the position of the rainbelt (East-West band of intense localised rainfall), as observed by Nicholson (2009). However the strong moist ascent associated with the ITD means its position and intensity will be an important factor in rainfall variability. It is therefore important to investigate how the mechanisms could perturb the large-scale circulation patterns in the region. Nevertheless, West Africa is a climatologically diverse region shown to vary extensively in both its seasonality as well as its inter-annual and inter-decadal rainfall.

Furthermore, the rainfall is dictated by three rain-bearing processes (MCS, ITD and monsoon). The onset stage of the West African summer monsoon is linked to an abrupt latitudinal shift of the ITD from a quasi-stationary location at 5 N in May-June to a second

quasi-stationary location at 10 -August. This stage corresponds to major changes in the atmospheric circulation over West Africa linked to the full development of the summer monsoon system (Sultan and Janicot, 2003). This abrupt shift occurs mostly between

10 W and 5 E where a meridional land-sea contrast exists and it is also generally characterized by a temporary decrease of convection over the whole of West Africa. During the 1968-2005 period, the mean date of the monsoon onset was 24 June with a standard deviation of 8 days. This has been focal point of numerous studies applying different methods to determine the rainfall onset over the region (Omotosho, 1990; 1992).

Figure 2. 1: Schematic depicts relative latitudinal positions of the ITCZ/ITD, TEJ, AEJ-N, AEJ-S and the WAJ during wet and dry years for 0 10 N (top) and 0 -12 S (bottom) (Refer online version for color images) from Williams and Kniveton, (2011).

After rather weak convective activity in winter, the first rainy season over the Guinea Coast

(the coast line is located at 5 N) begins around mid-April with an intensification of convective activity during the second half of April and most of May. This is followed by a temporary weakening around the end of May and a recovery of convective activity in the first half of June. Convection weakens again from approximately 25 June to 10 July, which corresponds to the typical transition phase of the monsoon onset (Sultan and Janicot, 2003), followed by the installation of the monsoon and convection over the Sahelian latitudes with the centre of gravity of the ITD located between 10 N and 12 N during the whole summer.

This transition phase is said to be centred on 3 July, ten days after the mean onset date, and corresponds to a cumulative probability of occurrence of 10% meaning that 10% on the onsets occurred on and after 3 July.

Views on the exact nature of the ITD have been subject to continual revision (Barry and

Chorley, 2003). From the 1920s to the 1940s, the frontal concepts developed in mid-latitudes were applied in the tropics, and the streamline confluence of the northeast and southeast trades was identified as the Intertropical front (ITF). Over continental areas such as West

Africa and South Asia, where in summer hot, dry continental tropical air meets cooler, humid equatorial air, this term has some limited applicability. Sharp temperature and moisture gradients may occur, but the front is seldom a weather-producing mechanism of the mid- latitude type. Elsewhere in low latitudes, true fronts (with a marked density contrast) are rare.

Recognition of the significance of wind field convergence in tropical weather production in the 1940s and 1950s led to the designation of the trade wind convergence as the Intertropical

Discontinuity (ITD). This feature is apparent on a mean streamline map, but areas of convergence grow and decay, either in situ or within disturbances moving westward, over periods of a few days. Moreover, convergence is infrequent even as a climatic feature in the doldrum zones. Satellite data show that over the oceans the position and intensity of the ITCZ varies greatly from day to day. The ITCZ is predominantly an oceanic feature where it tends

to be located over the warmest surface waters. Hence, small differences of sea surface temperature may cause considerable changes in the location of the ITCZ. A sea surface temperature of at least 27.5°C seems to provide a threshold for organized convective activity; above this temperature organized convection is essentially competitive between different regions potentially available to form part of a continuous ITCZ. The convective rainfall belt of the ITD has very sharply defined latitudinal limits. For example, along the West African coast the following mean annual rainfalls are recorded: 12°N 1939 mm, 15°N 542 mm, 18°N

123 mm.

In other words, moving southwards into the ITD, precipitation increases by 440% in a meridional distance of only 330 km (Barry and Chorley, 2003). As climatic features, the equatorial trough and the ITD are asymmetric about the equator, lying on average to the north. They also move seasonally away from the equator in association with the thermal equator (zone of seasonal maximum temperature). The location of the thermal equator is related directly to solar heating, and there is an obvious link between this and the equatorial trough in terms of thermal lows. However, if the ITD were to coincide with the equatorial trough then this zone of cloudiness would decrease incoming solar radiation, reducing the surface heating needed to maintain the low-pressure trough. In fact, this does not happen.

Solar energy is available to heat the surface because the maximum surface wind convergence, uplift and cloud cover is commonly located several degrees equatorward of the trough. In the

Atlantic, for example, the cloudiness maximum is distinct from the equatorial trough in

August. Convergence of two trade wind systems occurs over the central North Atlantic in

August and the eastern North Pacific in February. In contrast, the equatorial trough is defined by easterlies on its poleward side and westerlies on its equatorward side over West Africa in

August and over New Guinea in February (Barry and Chorley, 2003). The dynamics of low- latitude atmosphere ocean circulations are also involved. The convergence zone in the

central equatorial Pacific moves seasonally between about 4°N in March to April and 8°N in

September, giving a single pronounced rainfall maximum in March to April. This appears to be a response to the relative strengths of the northeast and southeast trades.

2.1.3 Impact of climate change and the additional radiative forcing

Since 1990, every 5 years a group of approximately 2000 natural and physical scientists, economists, social scientists, and technologists assemble under the auspices of the United

Nations-sponsored Intergovernmental Panel on Climate Change (IPCC). These scientists spend 3 years reviewing all of the information on climate change and produce a voluminous report following a public review by others in the scientific community and by governments.

Climate is defined as the 30 to 40 year average of weather measurements, such as average seasonal and annual temperature; day and night temperatures; daily highs and lows; precipitation averages and variability; drought frequency and intensity; wind velocity and direction; humidity; solar intensity on the ground and its variability due to cloudiness or pollution; and storm type, frequency, and intensity. Understanding the complex planetary processes and their interaction requires the effort of a wide range of scientists from many

fluxuations. The current average rate at which solar energy strikes the earth is 342 watts per square meter (W m-2). It is found that 168 W m-2reaches the earth light: 67 W m-2 is absorbed directly by the atmosphere, 77 W m-2 is reflected by clouds, and

30 W m-2

to emit back into space. The net effect is that instead of being a frozen ball averaging -19 C,

Earth is a relatively comfortable 14 C. This difference of 33 C arises from the natural greenhouse effect. Human additions of greenhouse gases appear to have increased the

temperature an additional 0.6±0.2 C during the 20th century. The transmission of visible light from the sun and the trapping of radiant heat from the earth by gases in the atmosphere occur in much the same way as the windows of a greenhouse or an automobile raise the temperature by letting visible light in but trap outgoing radiant heat. The analogy is somewhat imperfect since glass also keeps the warm inside air from mixing with the cooler outside air. The

called radiative forcing. The units radiative forcing from human additions of carbon dioxide since the beginning of the industrial revolution is 1.46 W m-2. Methane and nitrous oxide addition have provided relative radiative forcings of 0.48 and 0.15 W m2, respectively. Other gases have individual radiative forcings less than 0.1 W m-2. The total radiative forcing of all greenhouse gases added by human activity to the atmosphere is estimated to be 2.43 W m-2. This should be compared to the 342 W m-2that reaches the earth from the sun. Hence, the greenhouse gases

which is enough to cause the global temperature to increa .

2.2 Vegetation Dynamics and Rainfall Variability over West Africa

The early 1970s to middle 1980s period has been considered generally as the long drought period over the West African regions, specifically the Sahel has been the most affected among the regions (i.e. agricultural, hydrological, meteorological droughts etc.). Many debates have focused on the impact on rainfall spatio-temporal variability by changes in vegetation cover.

So, this section reviews the synthesis of research based on respectively rainfall recovery and vegetation re-greening over the two last decades in West Africa.

Many studies have underlined the rainfall recovery in the 90s-00s over West African region.

For instance, Nicholson (2005) used rainfall estimates from the Tropical Rainfall Measuring

Mission (TRMM) to deduce a rainfall recovery mainly over western Sahel and dry conditions over northern Sahel. Nowadays, researches are bringing out the causes of this recovery. Some have linked it to the increase of SST and changes in atmospherics parameters. More recently, the rainfall recovery over Sahel has been linked to the Saharan Heat Low (SHL) by Evan et al. (2015). They found an upward trend of the SHL due to greenhouse warming which is caused by the presence of water vapor. Thus, they conclude a consistent recovery form drought with the warming process of the SHL.

Regarding the vegetation re-greening, desertification has been of major concern that affects the land cover changes. Desertification is defined by the United Nations as land degradation occurring in dry land caused by a range of factors including climate variations and human management. Some years ago numerous researches were focusing on it over West Africa

Sahel region. The world has witnessed unprecedented changes in the pace, magnitude and spatial extent of changes in the land surface use (Sivakumar, 2007). So that extended droughts in certain arid lands have initiated or exacerbated desertification. In the past 25 years, the Sahel has experienced the most substantial and sustained decline in rainfall recorded anywhere in the world observed from available records (Hulme and Kelly, 1997).

Some findings have indicated a re-greening of the vegetation over some specific areas of the region. Thus based on the observation that the desertification and revitalization of dunes were phenomena associated with the period between the early 70s and the mid-80s as observed by

Rasmussen et al. (2001) from satellite data Herrmann (2005) also associated a recovery from the great Sahelian droughts to the recent increase in seasonal greenness over large areas of

West Africa. Furthermore, Fenshol and Rasmussen (2011) investigated the relationship between the vegetation productivity and rainfall Sahel-Sudanian zone of Africa and showed an increase in NDVI over Sahel-Sudanian zone during the period 1982-2007. Some other authors have used sequential approach to investigate this. Thus, Anyamba and Turcker (2005) divided the time series of rainfall and NDVI in two periods 1982-1993 and 1994-2003 by

defining respectively the persistence of drought on NDVI and the conditions with region-wide above normal NDVI conditions. More recently, Dardel et al.

(2014) analysed the re-greening of Sahel after some experimentation on two sites in Mali and

Niger using both satellite and observation data. Apart Anyamba and Turcker (2005) who made sequential study to separate the different periods in the NDVI trends, over studies have used continuous approach to investigate the phenomenon of rainfall recovery.

2.3 The Relationship Rainfall and Vegetation

Surface water balances reflect the availability of both water and energy. In regions where water availability is high, evapotranspiration is controlled by the properties of both the atmospheric boundary layer and surface vegetation cover (Bates et al., 2008). Changes in the surface water balance can feed back on the climate system by recycling water into the boundary layer (instead of allowing it to run off or penetrate to deep soil levels). The sign and magnitude of such effects are often highly variable, depending on the details of the local environment. Hence, while in some cases these feedbacks may be relatively small on a global scale, they may become extremely important at smaller space or time-scales, leading to regional/local changes in variability or extremes. The impacts of deforestation on climate illustrate this complexity. Zheng and Eltahir (1998) showed that the meridional distribution of vegetation plays a significant role in the dynamics of West African monsoons. The response of the atmosphere to any perturbation in the distribution of vegetation depends critically on the location of this perturbation. Some studies indicate that deforestation could lead to reduced daytime temperatures and increases in boundary layer cloud as a consequence of rising albedo, transpiration and latent heat loss. However, these effects are dependent on the properties of both the replacement vegetation and the underlying soil surface and in some cases the opposite effects have been suggested. The effects of deforestation on precipitation are likewise complex, with both negative and positive impacts being found, dependent on

land surface and vegetation characteristics (Bates et al., 2008). A number of studies have suggested that, in semi-arid regions such as the Sahel, the presence of vegetation can enhance conditions for its own growth by recycling soil water into the atmosphere, from where it can be precipitated again. This can result in the possibility of multiple equilibriums for such regions, either with or without precipitation and vegetation, and also suggests the possibility of abrupt regime transitions, as may have happened in the change from mid-Holocene to modern conditions. Soil moisture is a source of thermal inertia due to its heat capacity and the latent heat required for evaporation. For this reason, soil moisture has been proposed as an important control on, for example, summer temperature and precipitation. Feedbacks between soil moisture, precipitation and temperature are particularly important in transition regions between dry and humid areas, but the strength of the coupling between soil moisture and precipitation varies by an order of magnitude between different climate models, and observational constraints are not currently available to narrow this uncertainty (Bates et al.,

2008).

2.4 Vegetation Indices over West Africa

The utility of remote sensing data especially satellite images have been proven in climate monitoring and prediction. Also historical baselines of forest cover are needed to understand the causes and consequences of recent changes and to assess the effectiveness of land-use policies (Kim et al., 2014). McGuffie (1994) showed that through the use of satellites it is possible to monitor many aspects of the surface and atmosphere of the Earth. Meteorological satellites have enhanced our understanding of the synoptic processes and now form a routine part of weather information which is distributed to the general public. These satellites have also provided increased understanding of many smaller-scale processes which were not resolved by the surface synoptic network. Indeed, satellite observations provide more spatially and timelier continuous input data coverage sources than ground gauge station

observations. Because of the characteristics of the spatial coverage of climate, satellite images enable us to understand manifestation of drought in larger area in less time consuming way than conventional method. When drought exists, due to reduction in precipitation, the capacity to carry out the chlorophyllian by the vegetation is notably reduced (Kim et al.,

2014; Reiche et al., 2015). The response of the green vegetation is characterized by maximum absorption radiation in the red region and large reflection in the neighbouring infrared region. it also has been observed that in unhealthy, ageing or subject to condition of vegetation stress, the reflectance in red region increases while in near infrared region decreases. The normalized difference vegetation index (NDVI), developed by Trucker in

1979 is the most popular vegetation index used to monitor vegetation at regional to global scales. It is calculated as in the following equation.

NDVI NIR RED (2.1) NIR RED

where NIR and RED are the reflectance in the near infra red (NIR) and in the Red bands respectively, their values vary between [-1, +1].

As NDVI is not sensitive to influences of soil background reflectance at low vegetation cover and log vegetation response to precipitation deficient, NDVI itself does not reflect drought or non-drought conditions. But severity of drought may be defined as NDVI anomaly from its long-term. The anomaly NDVI of drought may be defined as NDVI at current time step, such as month, and a long-term mean NDVI of the same time step for each pixel. When the

NDVI anomaly is negative, it indicates the below-normal vegetation condition and maybe a drought situation. The larger the negative departure, the greater drought severity may be suggested.

(2.2)

where NDVI i is the value for time step i, NDVI i,mean is the long-term mean value of same time step i

Remote sensing has been widely applied in many studies (Browning and Roberts, 1994; Marx et al., 2008; Henke et al., 2012 ; Mito et al., 2012 ; Stow et al., 2014 and Yuan et al., 2015).

Therefore, when compared with other methods, remote sensing has obvious superiority in estimating a real sensible heat flux over different surface conditions. However, a major challenge is in retrieving accurate and reliable values of sensible heat flux by this technique.

The challenge consists of two tasks: the first is to retrieve accurate estimates of surface temperatures from satellite data and to extrapolate them temporally and the second is to relate estimates of sensible heat flux obtained over relatively large surfaces from satellite data, 1-16 km 2, to ground measurements usually representative of less than 0.1 km 2.

Mangiarotti et al. (2012) have used AVHRR-NDVI data to study the predictability of vegetation cycles over the semi-arid region of Gouma in Mali. For that study a model based on a reconstruction approach in which the NDVI signal is taken as a proxy of the system's dynamics was used and also as a model will be for this reason, the corresponding models will be referred to as proxy models here. Forecasts are obtained from these models. And it will base of on the growth of the forecast error. As a result, it indicated a rapid increase in error with regard to the horizon of prediction and shows large interannual variability. And the degree of forecasting error clearly decreases as the aggregation scale increases, revealing the higher predictability of the behaviour of vegetation at the scale of large regions.

Unfortunately, most of the existing land cover products were produced based on remotely sensed data that were acquired during a single year or non-consecutive years, which has largely constrained their contribution to long-term or up-to-date applications. Therefore,

timely updating of the existing land cover products based on the large and growing body of satellite data is still urgently required for numerous applications (Chen et al., 2015).

2.5 The West African Monsoon

The basic drive for the monsoon circulation is provided by the contrast in the thermal properties of the land and sea surfaces (James and Gregory, 2012). Because the thin layer of soil that responds to the seasonal changes in surface temperature has a small heat capacity compared to the heat capacity of the upper layer of the ocean that responds on a similar time scale, the absorption of solar radiation raises the surface temperature over land much more rapidly than over the ocean. The warming of the land relative to the ocean leads to enhanced cumulus convection, and hence to latent heat release, which produces warm temperatures throughout the troposphere. The basis of all transport of air masses in the atmosphere is movements of air from areas with high pressure to areas with lower pressure. Unequal

flux is much higher in the tropics compared to the Polar Regions. Therefore, heat is transferred by air movement as well as ocean currents (such as the warm Gulf Stream) from the equator to the poles. Also, the characteristics of the surface (sand is heated very quickly, whereas ocean surfaces are not) are very important and lead to phenomena that differ very much in scale, from a local see breeze to the distribution of high and low pressures over

the equator prevents the straightforward movement of air from low- to high pressure areas and induces the so-called Coriolis force, resulting in circular movements of air around high-pressure and low-pressure areas. The friction induced by mountains and smaller objects, such as cities or forests, is another factor that influences wind direction and speed.

The annual climatic regime over West Africa has many similarities to that over South Asia, the surface airflow being determined by the position of the leading edge of a monsoon trough.

This airflow is southwesterly to the South of the trough and easterly to northeasterly to its

North. The major difference between the circulations of the two regions is due largely to the differing geography of the land to sea distribution and to the lack of a large mountain range to the North of West Africa. This allows the monsoon trough to migrate regularly with the seasons. In general, the West African monsoon trough oscillates between annual extreme locations of about 2°N and 25°N. In 1956, for example, these extreme positions were 5°N on

1 January and 23°N in August. The leading edge of the monsoon trough is complex in structure and its position may oscillate greatly from day to day through several degrees of latitude. The classical model of a steady northward advance of the monsoon has recently been called into question. The rainy season onset in February at the coast does propagate northward to 13°N in May, but then in mid-June there is a sudden synchronous onset of rains between about 9°N and 13°N. The mechanism is not yet firmly established, but it involves a shift of the lower tropospheric African Easterly Jet (AEJ).

Figure 2. 2: Schematic of the atmospheric circulation in the West African monsoon system during the boreal summer. Closed solid lines represent the isotachs of the African Easterly Jet (AEJ), which lies around 600 hPa. The red arrows show the thermally direct meridional monsoon circulation, and are typical of the time-mean winds in the peak monsoon season (Lebel et al., 2010).

In winter, the southwesterly monsoon airflow over the coasts of West Africa is very shallow

(1000 m) with 3000 m of overriding easterly winds, which are themselves overlain by strong

(>20 m s 1). North of the monsoon trough, the surface northeasterlies (i.e. the 2000 m deep

Harmattan flow) blow clockwise outward from the subtropical high pressure centre. They are compensated above 5000 m by an anticlockwise westerly airflow that, at about 12,000 m and

20 to 30°N, is concentrated into a subtropical westerly jet stream of average speed of 45 m s

1. Mean January surface temperatures decrease from about 26°C along the southern coast to

14°C in southern Algeria.

With the approach of the northern summer, the strengthening of the South Atlantic subtropical high pressure cell, combined with the increased continental temperatures, establishes a strong southwesterly airflow at the surface that spreads northward behind the monsoon trough, lagging about six weeks behind the progress of the overhead sun. The northward migration of the trough oscillates diurnally with a northward progress of up to 200 km in the afternoons following a smaller southward retreat in the mornings. The northward spread of moist, unstable and relatively cool southwesterly airflow from the Gulf of Guinea brings rain in differing amounts to extensive areas of West Africa. Aloft, easterly winds spiral clockwise outward.

2.6 An Overview of Synoptic Scale Atmospheric Features over West Africa

Burpee (1972), one of the AEWs study pioneer tried to improve the knowledge about the origin and the structure of AEWs. Since then, studies have been going on to mastery the contour of this phenomenon. To continuous in the AEWs origin, Albignat (1980) gave more details about it. He used the spectrum and cross-spectrum analysis to locate Waves observed on a time period covering 23 August to 19 September 1974. He found like his predecessor

stward. He found that the mid-tropospheric easterly jet stream plays an important role in the development of the waves

but a possible additional source of energy should come from the release of latent heat by

organized cumulus convection. Mekonnen et al. (2006) used the brightness temperature to

examine the association between convection and AEWs. Adedoyin (1997) has shown that the

warming up of the Indian, Pacific and South Atlantic Oceans strengthens the AEJ. A stronger

AEJ, on the other hand, leads to a southward shift of the zone of squally activities in tropical

North Africa thereby resulting in rainfall deficits north of latitude 12. The change in climate

ributable to fact that the SSTs of the three influencing oceans have persistently warmed up in July,

August and September of the 18year period 1969-1986. Omotosho et al. (2000) used the upper wind to predict the rainfall onset and cessation two months ahead using. He found that agriculturally reliable rainfall commences about 70 days after the first sudden changes in wind direction from westerly to easterly at the 400 hPa level and above. For that he used simple empirical schemes for predicting. Some years later Diedhiou et al. (2002) used the reanalysis data from NCEP/NCAR have studied the energetic of 3-5-day and 6-9-day African

Easterly wave regimes. He confirmed that these energies are due to baroclinic and barotropic instability. According to the source of energy which gives them birth and maintains them, the

3-5-day waves grow between Paeth et al. (2005) focus on the mechanisms of interannual rainfall fluctuations at the synoptic scale during a 25 year hind cast period extending from 1979 to 2003. They used SSTs and large-scale atmospheric circulation as prescribed at the lateral boundaries of the regional climate model sector as key factors in precipitation variability.

2.7 Climate Models

2.7.1 Climate and general-circulation models

Climate models have been put to a variety of scientific and regulatory uses. Primarily the models are used to predict meteorological events, to simulate long term climatic parameters and to estimate the atmospheric concentration field in the absence of climate monitored data.

In this case, the model can be a part of an alert system serving to signal when atmospheric pattern potential is high and requiring interaction between control agencies and emitters. The models can serve to locate areas of expected high concentration for correlation with health effects. Real-time models can also serve as official guides in cases of nuclear or industrial accidents or chemical spills. Here the direction of the spreading cloud and areas of critical concentration can be calculated.

A current popular use for atmospheric diffusion models is in air quality impact analysis. The models serve as the heart of the plan for new source reviews and the prevention of significant deterioration of air quality. Here the models are used to calculate the amount of emission control required to meet ambient air quality standards. The models can be employed in preconstruction evaluation of sites for the location of new industries. Models have also been used in monitoring network design and control technology evaluation.

The basic purpose of climate models originally was to explain features of current climate and general circulation in terms of the geometry of the earth-sun systems and basic physical principles. Once this was accomplished, it was possible to experiment with climate models by changing various input parameters to account for the features of past climates and to

with explaining the temperature distribution; the effect of circulation, if included at all, is prescribed in terms of the temperature distribution. Such climate models, range from simple zero dimensional (averaged over the whole atmosphere) to models allowing for vertical and

horizontal variations of input parameters. The models differ in the way boundary conditions and physical processes are handled. Sometimes clouds are prescribed, or they may be generated by the models. Ocean temperatures may be given, or the atmospheric model may be combined with an oceanic model. One of the important effects in many climate models is the ice-albedo feedback: If the model predicts a cooling, the ice sheets expand, causing increased albedo and more cooling- s so severe that only a small cooling led to an ice covered earth. In spite of such excesses, climate models have explained most features of the vertical temperature distributions and effects of land-water differences and topography. General-circulation models (GCMs) attempt to duplicate the distribution of wind as well as of temperature and moisture. They are usually three dimensional; however, the earliest models had little resolution in the vertical; for example, the pioneering model by Norman (1956) consisted of only two layers. In the meantime, much more complete models have been developed, many requiring the fastest computers available.

In the models, generally seven basic equations are solved for the seven basic variables of meteorology: pressure, density, temperature, moisture, and three velocity components. The seven equations are the gas law, the first Law of thermodynamics, equations of continuity for

modelled explicitly, ozone concentration must be added as a variable and at least one equation must be included to describe the ozone budget. Since the ozone budget depends on concentrations of many other trace gases, many more variables and equations are sometimes added. The location of the Inter-Tropical Discontinuity (ITD) represents the confluence zone of the south-westerly monsoon winds with the north-easterly dry Harmattan winds. The monsoon winds are controlled by the pressure gradient between the low pressures of the

Saharan heat low centred along the ITD and the oceanic high pressures of the Santa Helena anticyclone. The Harmattan winds are controlled by the pressure gradient between the

Saharan heat low and the Libyan and Azores anticyclones. In June the ITD is centred over the southern coast of West Africa, South of 10 N and corresponds to the last part of the first rainy season over the Guinea coast region. At this stage the spatial extension of the monsoon winds is limited and the ITD is positioned between 15 N and 20 N with its northernmost latitude between 0 W and 5 E. In July the ITD is shifted to the north reaching a quasi-stable state around 10 N and the area of westerly winds extend over land and over the tropical Atlantic between 5 N and 15 N, the ITD reaching 20 N, its northernmost latitude. The westerly wind speed increases over West Africa. In August the monsoon is fully developed consistent with the highest pressures in the southern tropical Atlantic and the northernmost location of the

Saharan heat low over land. The ITD is still around 10 N but with increased precipitation.

The westerly wind area has its largest extension, especially over the northern tropical Atlantic where westerly moisture advection inland reaches its seasonal maximum. In September the westerly wind area does not change significantly but the westerly wind speed decreases drastically. This is the last part of the fully-developed monsoon season.

The AEJ maintenance is mainly controlled by the low/mid-levels transverse circulation induced by the heat low (Thorncroft and Blackburn, 1999). This circulation, as well as the

AEJ, has its highest intensity and spatial extension in June before the monsoon onset (Sultan and Janicot, 2003). The core of the jet is located between 5 N-10 N and 10 W-10 E with a mean highest speed of 14 m s 1. In July and August the AEJ moves to the north, and is oriented along a southeast-northwest axis around 15 N over the western coast of West Africa.

Its core speed at this time decreases to 10 m s 1. It retreats southward in September and its speed increases to 12 m s 1. Another core is noticeable south of the equator in September independent of the AEJ and has been described previously by Grist and Nicholson (2001).

At 200 hPa the high-level anticyclone structure, which is the sign of the Indian and African monsoons induces an easterly wind field on its southern flank with a core speed greater than

20 m s 1 centred over the Indian Ocean. This jet, the Tropical Easterly Jet (TEJ), enhances in

July and August in parallel with the monsoon activity. In particular its westward extension over Africa is reactivated and a maximum of 14 m s 1 is evident between 10 W and 30 W. It decreases significantly in September over its whole domain.

In 2006 the TEJ was weaker over the Indian sector during the whole summer compared to the mean (the Indian monsoon was a bit more active over north-eastern India but a bit weaker over central India and highly weaker over the equatorial Indian Ocean and Indonesia compared to the 1979-1999 mean). Over West Africa, after a period of weaker wind speeds in June, the TEJ extended further to the West in July and displayed higher speeds in August and September, consistently with a bit higher active monsoon season compared to the 1979-

1999 mean.

2.7.2 Regional climate models: RegCM

Global modeling can provide (spatially, temporally, and spectrally) complete and consistent datasets for all aerosol properties. Concerns exist, however, as to the accuracy of the underlying assumptions (emissions, transport, and water uptake) and parameterizations

(aerosol processing, interactions with clouds). With rather general constraints (column and component-integrated data from remote sensing), there is significant diversity in aerosol global modeling, especially at modeling sub steps (Textor et al. , 2006). To counteract this diversity and to establish characteristic particle properties from global modeling, aerosol simulations of more than twenty different models were considered. All of these models employed advanced aerosol modules, which distinguished between aerosol components of dust, sulphate, sea salt, organic carbon, and black carbon (Kinne et al. , 2006). Simulated monthly averages were re-gridded to a common 1°×1° latitude/longitude horizontal resolution. The local median value of monthly averages suggested by all models at any grid

point was picked; thereafter, these median values were combined to define monthly fields

from global modeling. These median fields have the advantage that extreme behaviour

(outliers) of individual models is suppressed. In addition, these median model fields tend to

score better when evaluated than individual models (Schulz et al. , 2006).

Regional Climate Models (RCMs) (Giorgi and Mearns, 1999) are now extensively used to downscale large scale climate information to regional scales in order to account for fine-scale processes that regulate the spatial structure of climate variables.

In this study the RegCM 4 will be used. It has been wisely applied through the world for both dynamical and statistical downscaling. Sylla et al. (2012) have used the ICTP regional climate model, RegCM3, nested in NCEP and ERA-Interim reanalyses (NC-RegCM and

ERA-RegCM, respectively) to explore the effect of large-scale forcings on the model biases over a southern Africa domain at 25 km grid spacing. It was discovered that the RegCM3 shows a generally good performance in simulating the location of the main rainfall features, temperature and synoptic scale circulation patterns, along with cloud cover and surface radiation fluxes; it also has some wet and dry biases. used ICTP-RegCM3 to examine the coastal effects over the Eastern Mediterranean region has downscaled to a 10 km resolution over the EM with a 50 km driving nest. As result, the high-resolution simulation captures strong temperature gradients of resolving the steep topography over the eastern

Black Sea and Mediterranean coasts of Turkey, as well as the Ionian coast of Greece.

It indicates that the seasonal temperature biases for 10 km simulation are <1°C and the frequency of dry and wet spells is well reproduced by the model. It must be noted that though it has been wisely applied for temperature and rainfall variability and trend, some also couple it with others atmospheric parameters. This the case of Zanis et al. (2012) who coupled

RegCM3 with aerosols to investigate the direct shortwave effect of anthropogenic aerosols on the regional European climate over a 12 year period (1996-2007). Aerosol feedback induced

small changes in the yearly averaged near-surface temperature over Europe during this period and the greatest negative temperature difference of -0.2°C was observed over the Balkan

Peninsula.

Chapter 3

RESEARCH METHOD

3.1 Study Area

3.1.1 Study Area Location

The investigation was done over West African region shown in Figure 3.1. The climate of the region is controlled largely by two dominant air masses. These are the dry, dusty, continental air mass (which originates from the Sahara desert), and the warm, maritime air mass (which originates from the Atlantic Ocean) (Imo and Ekpenyong, 2011). The influences of both air masses are determined by the movement of the ITCZ over ocean and the ITD over land. The interplay of these two air masses gives rise to two distinct seasons within the sub-region, namely: the dry and wet season. The wet season is associated with the tropical maritime air mass, while the dry season is a product of the tropical continental air mass. The influence and intensity of the wet season decreases from West Africa coastal region towards the North.

Based on the rainfall seasonal distribution, two main sub-regions are observed. South of 8°N, rainfall in the Guinea region is bimodal, the region also presents four seasons (Le Barbé et al., 2002 and Konaté and Kampman, 2012): a long dry season occurs between December and

February, and two rainy seasons from March to July and from September to November are separated by a little dry season in August. Over the Sahel region, rainfall pattern is unimodal with amounts decreasing northwards with a gradient of 1 mm km -1 (Lebel et al. , 2003), and is characterised by a long dry season between November and March and a wet season between

April and September. The precipitation of the rainy season is associated with thunderstorm activity which occurs along disturbance Thus, about 80 % of the

total annual rainfall for most places is associated with line squall activities which are prevalent between June and September (Adefolalu, 1986 and Mathon and Laurent, 2001).

Dietz et al. (2004) have shown that over West African drylands, rainfall data for the period

1960-1990 reveals a decline in average rainfall indicating changes in aridity between 1930-60 and 1960-1990. Some of the regions in the northern zone with semi-arid conditions in 1930-

60 had clearly become arid (on average) in the 1960-1990 period, with unsuitable conditions for millet or sorghum production in most years. A considerable part of the sub-humid zone in the period 1930-1960 had become semi-arid in 1960-1990 with considerable drought risks, certainly for crops which are less adaptable to drought stress (maize, cotton).

The ITD, also known as the monsoon trough or the doldrums, is formed near the equator by the meeting of the north-east and south-east trade winds. These winds force moist air upwards, causing water vapour to condense out as the air rises and cools. The ITD follows

north and south through the year, as the earth tilts on its axis, relative to the sun. In West

Africa, the start of the monsoon depends on the northward progression of the ITD over the period June to August, when the Sahel and the southern Sahara receive most of their rainfall.

Even small shifts in the position of the ITD rain belts result in large local changes in rainfall, bringing severe droughts or flooding (Camilla, 2009).

3.1.2 The Vegetation over West Africa

West Africa hosts a rich variety of forests and each type of forest and woodland plays an essential role in supporting and regulating the ecosystems on which people and plants depend. The vegetation of West Africa presents a simple picture compared to other part of tropical Africa (Konaté and Kampmann, 2012). Due to its low-lying terrain the zones of vegetation largely reflect both the basic climatic zones and soil, essentially exhibiting the same longitude as rainfall. This results in a series of vegetation zone running in roughly parallel bands from the southern Guinea coast with high and evenly disturbed rainfall

throughout the year to zones of increasingly drier vegetation until the Sahara desert is reached in the North.

For the delimitation and description of these vegetation zones various classification approaches exist based on climatic and/or phytogeographic parameters. The most applied and accepted vegetation classification for the whole Africa (White, 1983) is taken into account.

His delimitation of vegetation zones is principally based on patterns of species distribution and distinguishes regional centres of endemism (with>50% of their flora being endemic) and transition zones between them. For each vegetation zone several main vegetation types are recognized, which are characterized by their physiognomy from the North to South the structure of the vegetation change progressively desert in the North, savannah in the Centre

(Fig. 1.3) and forest over coastal regions. Specifically, based on BIOTA WEST programme four vegetation types zones, are mainly identified namely Guinea-Congolian, Guinea-

Congolia/Sudanian zone, Sudanian zone and Sahel zone.

Over 75% where it is filtered and purified. Forests provide the habitats for many thousands of plant and anim

Forests influence the climate through a range of physical, chemical and biological processes that affect the atmosphere, the water cycle and global energy balance (Bonan, 2008), and play two very different, but equally important roles on the global climate stage. Their first role is that of carbon storage. Forests buffer the planet against global warming by absorbing carbon dioxide, thereby helping to stabilize the atmospheric levels of this greenhouse gas. Second, they regulate local and global weather patterns by storing and releasing moisture.

Figure 3. 2: Wooded savannah landscape in the Lamto reserve located in central part of Cote It is the type of vegetation met generally over Sudanian savannah regions of West Africa.

As well as the risk to forest life posed by climate change, forests are also under attack from humans. Around four million hectares of forest are felled or burnt in Africa each year, an area equivalent to roughly twice the size of Rwanda. There are large regional differences in deforestation, with Togo having one of the highest rates, not just in Africa, but in the world, having lost 44 % of its forests since 1990. At a global level, average annual rates of deforestation were around 8.9 million hectares per year in the 90s.

In West Africa, as in other parts of the world, forests are cut in favour of pasture, crops, settlements and infrastructure, and for extraction of fuel and timber, much of which is uncontrolled or under-regulated in the region.

3.1.3 Regional climatology of West Africa

The West African savannah belt with its tropical location and geomorphology, incoming solar radiation is relatively constant as are the temperatures. The southern Sudanian savannah zone is characterised by a night day variation of 20°C (Bagayoko, 2006 and Schindler, 2009), but northwards in central Burkina Faso and near the Sahel, average temperatures of 25°C in

January and 32°C in April and relative humidity of 6% during the dry season and 95% in the rainy season are common. It is found that the temperatures can oscillate strongly, from 15°C during the night to more than 40°C during the day (Sandwidi, 2007). Recent analysis have found a rise in the average temperature at about 1°C between 1960 and 1990 (Ouédraogo,

2004 and Sandwidi, 2007). In general, the region is poor in water resources. The main constraints are the distance to the sea, the unimodal rainfall regime (Fig. 3.3), and groundwater table of crystalline rock with poor aquifer conditions, therefore groundwater levels vary greatly. For example, in the Atankwidi basin in the Upper East Region (UER) in

Ghana, variations between 1 m and 29 m have been reported, which strongly influence the

Schindler, 2009). Similarly, in south eastern

Burkina Faso the groundwater table reduces on average 0.6 mm day -1 in the cropping season, thus the water withdrawal (76 l per capita per day (l/c/d)) greatly exceeds the provision

(20l/c/d), and according to the calculated recharge capacity of the aquifers (2% per year), the projected demand will overtake the supply in 2030 (Sandwidi, 2007). Also the rainfall follows a decreasing gradient from the South to the North (see Fig. 4.1). In the UER, Ghana the monomodal rainfall regime of 3 to 5 months is from April to October, with between 900 and 1000 mm; the remaining seven months are dry (Kpongor, 2007 and Sanwidi, 2007). The onset of the rainy season is generally stormy, but the effective rainfall for agriculture is low, especially due to the high run off and evaporation. The latter can be exacerbated through the

Harmattan (Ouédraogo, 2004; Yilma, 2006 and Kpongor, 2007). Recurrent dry spells are also observed, which are especially harmful during the planting season (June and July), as well as recurrent droughts (Sanwidi, 2007). In Sudanian zone, annual rainfall ranges from 400 to

1100 mm from North to South with high spatial and temporal variability. For instance, in the

Kompiega basin in south eastern Burkina Faso, average rainfall 1959 2005 reaches 830 mm year -1. In general, evaporation exceeds rainfall except during the rainy season when the basin is recharged (Ouédraogo, 2004; Bagayoko, 2006 and Sandwidi, 2007).

Figure 3. 3: Temperature and rainfall monthly mean in Bawku, Upper East Region of Ghana (1993 2011).

Lamto (5.02 W; 6.13 Baoul Ivoire (Fig. 3.4).

It is surrounded by forest and hills and borders the southern rainforest. It covers an area of 27 km 2 comprising 80% of moist savanna and 20% of forest. These characteristics make Lamto not belonging to the northern Sudanian savannah, but to an intermediate vegetation type called Guinea savannah (Diawara et al., 2014). Four seasons characterize the climate of

Lamto: a long dry season occurring in December-February, a long wet season during March-

July, a short dry season in August, and a short wet season in September-November. The mean annual rainfall amount (~1212 mm) is less than that of the neighbouring synoptic stations mainly located in the southern rainforest or at the same latitude, as, for instance, Gagnoa where it reaches 1382 mm. Several variables (rainfall, temperature, sunshine, etc.) that influence the climate are recorded daily at Lamto. The consequences that could involve deforestation and bush-fire occurrences on its climate require particular attention and assessment.

Figure 3. 4: Lamto Geophysical station where many climate and seismic parameters are measured; The station was created in 1962 by Maxime Lamotte and Jean-Luc Tournier.

Figure 3. 5: Temperature and rainfall monthly mean at Lamto station, Guinea region in Cote (1965 2004).

February, March and April are the hottest months at Lamto station. The mean temperature is included ). The period corresponds to bush fire period

are part of the rainier which are April, May, June, July, and August. The area is characterized by bimodal evolution of the temperature over the year.

Figure 3. 6: Intra seasonal Temperature mean at the Lamto station averaged over 1971-2000.

An increase in temperature is clearly seen through the temperature anomaly at Lamto station in addition to the interannual variation of the temperature. The linear trend calculation is giving a yearly increase of temperature abou decade (Fig. 3.7). So as it has been observed in the Burkina Faso, from 1964 to 2004 the temperature has increased at

observation made over the globe. The anomalies are computed based on 1971-2000 average.

Thus compare to the considered normal (1971-2000) the period from 1964 to 1985 the anomaly is mainly negative this period correspond to the coldest of the series. Then from

1986 up to 2004 the anomaly is positive corresponding to the warmth period. Beside global and regional climate impacts on temperature over the region some local phenomenon such as

deforestation, bush fire and aerosol emission could explain the increase in temperature. As it is shown by Jones et al. (1999) mean).

Figure 3. 7: Temperature Anomaly at the Lamto Station plotted with regard to climatology of 1971-2000.

3.2 Data Collection

3.2.1 Rainfall from station data

Rainfall measurements from rain gauge stations are conventionally considered the most accurate and reliable source of rainfall data (Herrmann et al., 2005). However, this is only true for point measurements or areas with a sufficiently dense network of rain gauges. So to perform this study daily rainfall data have been collected data from two stations over Cote

which are Daloa and Lamto station covering the period 1964-2011and collected respectively at the Societe de Developpement et Exploitation Aeroportuaire et Maritime

(SODEXAM) and at the Lamto station.

3.2.2 Climate Research Unit data

The choice was drown on the updated gridded climate dataset from Climate Research Unit

(CRU) version TS3.10.1, which presents a larger time cover with a wide spatial representation over West Africa. In the worries of accurate representation of the mean state and variability of the present climate which is important for a number of purposes in global change research, coarse resolution datasets such as temperature and precipitation have been adequate for monitoring and detection of climate change and GCM evaluation. Capturing temporal variability is as important as the representation of spatial detail (New et al. , 1999).

These include monitoring and detection of climate change; evaluation of General Circulation

Models (GCMs) and regional climate simulations; ground truthing, calibration, or merging with satellite climatology; understanding the role of climate in biogeochemical cycling; and construction of climate change scenarios (Carter et al. , 1994).

The main sources of data are: National meteorological agencies, WMO 1961-90 global standard normal, CRU global dataset of station time series, CIAT South America database

The principal sources used for the routine updating of the Climatic Research Unit (CRU) monthly climate archives come through the auspices of the World Meteorological

Organisation (WMO) in collaboration with the US National Oceanographic and Atmospheric

Administration (NOAA, via its National Climatic Data Center, NCDC). The monthly products were accessed through the Met Office Hadley Centre in the UK and NCDC in the

USA. Web links to these sources can be found in the supporting information (Harris, 2014).

These data comprise 1224 grids of observed climate, for the period 1901-2009, and cover the global land surface at 0.5×0.5 degree resolution. The precipitation data (Eischeid et al., 1991;

Hulme, 1994) have been compiled by the CRU over the last 20 year. The original data have been subjected to comprehensive quality control over the years. Updates for more recent years and additional station data collated by the CRU have also been checked for homogeneity and outliers. The correction of individual records requires detailed local meteorological and station met information, which are not readily available (New et al. ,

2000).

3.2.3 Global Precipitation Climatology Project

The Global Precipitation Climatology Project (GPCP) monthly precipitation analysis is a globally complete, monthly estimate of surface precipitation at 2.5° x 2.5° latitude longitude resolution that spans the period 1979 to the present (Adler et al., 2003; Huffman, et al.,

2009). However, the covered period for the model settlement was 2005 to 2007 It is a merged, monthly analysis that employs precipitation estimates from low-orbit satellite SSM/I and SSMIS microwave data to perform a calibration, that varies by month and location, of geosynchronous-orbit satellite infrared (IR) data in the latitude band 40°N-S. These multi- satellite estimates are combined with rain-gauge analyses (over land) in a two-step process that adjusts the satellite estimates to the large-scale bias of the gauges and then combines the

adjusted satellite and gauge fields with weighting by inverse error variance. The monthly product is typically produced about two months after the end of the observation month.

3.2.4 Vegetation Indices Data form GIMMS

The vegetation indices implanted for this study is the NASA AVHRR NDVI, covering the period 1981-2012 with horizontal resolution of 8 km. It is derived from National Oceanic and

Atmospheric Administration (NOAA) satellites, and processed by the Global Inventory

Monitoring and Modeling Studies group (GIMMS) (Zhu et al., 2013; Donghai et al., 2014 and Jamali et al., 2015) at the National Aeronautical and Space Administration (NASA).

Spectral vegetation indices are usually composed of red and near-infrared radiances or reflectance (Tucker, 1979), sometimes with additional channels included. According to

Cracknell (2001), these indices are one of the most widely used remote sensing measurements. They are highly correlated with the photosynthetically active biomass, chlorophyll abundance, and energy absorption (Myneni et al. , 1995). The use of spectral vegetation indices derived from AVHRR satellite data followed the launch of NOAA-6 in

June 1979 and NOAA-7 in July 1981 (Gray and McCrary, 1981). The AVHRR instruments on NOAA-6 and NOAA-7 were the first in the TIROS-N series of satellites to have non- overlapping channel 1 and channel 2 spectral bands. Overlapping red and near infrared spectral bands precludes calculating a NDVI. The NDVI is calculated as NDVI5 (channel

22channel 1)/(channel 2 + channel 1) (Tucker, 2005). The Normalized Difference Vegetation

Index (NDVI) has become the most used product derived from NOAA AVHRR data

(Cracknell, 2001), largely from the use of NDVI datasets formed via maximum value compositing (Holben, 1986).

The first generation NDVI data (NDVIg) from AVHRR sensors onboard the National

Oceanic and Atmospheric Administration (NOAA) 7 to 14 series of satellites have been

processed by the Global Inventory Modeling and Mapping Studies (GIMMS) group to a

consistent time series of NDVI and is made available to the research community. The latest

version, termed the third generation NDVI data set (GIMMS NDVI3g) has been recently produced for the period July 1981 to December 2011 with AVHRR sensor data from NOAA 7

to 18 satellites. This data set specifically aims to improved data quality in the high latitudes

where the growing season is shorter than 2 months. It has also improved calibration that is

tied to the Sea-Viewing Wide-Field-of-View Sensor, as opposed to earlier versions of

GIMMS NDVI data sets that were based on inter-calibration with the SPOT sensor. The

availability of this new improved NDVI3g data set and its overlap with the Terra MODIS

LAI and FPAR products for the period 2000 to 2009 provides an opportunity to design and

implement a neural network algorithm to generate and evaluate the corresponding LAI and

FPAR data sets. These data sets will be termed LAI3g and FPAR3g henceforth and have the

following attributes: 15-day temporal frequency, 1/12 degree spatial resolution and temporal

span of July 1981 to December 2011 (Zhu et al., 2013).

3.2.5 Forcing Parameters

The experiment is based on the use of ICTP Regional Climate Model RegCM4 coupled with

land surface state model BATS. As vegetation degradation is an ongoing process, the principle of this study is to anticipate the phenomenon by attributing a short grass to the

transition zone stated above which is normally arboretum savannah. The model will be set base on rainfall from GPCP, CRU (New et al. , 1999) and RegCM3 simulated for the

CORDEX experimentation and the temperature from CRU and RegCM3. The second

challenge is the period covered by the simulation which is four months (June-July-August-

September) over two years (2005 and 2006). Can RegCM4 be able to capture the impact of

land surface change on rainfall and atmospheric patterns (TEJ, AEJ and West Africa monsoon

flux) at this range of time? For some authors, the intensity of these systems and their

latitudinal location influence not only the amount of rainfall but also its variability over West

Africa (Sylla et al., 2010). Some particularities of the region lead to make events difficult to forecast. The instability the interconnection between different patterns, land-sea, surface state atmosphere however, many studies have been realised with the previous version of RegCM over the region. The model have been subject to some progresses and updates thus the choice of the last version RegCM4 (Tompkins et al., 2005) the African Easterly Jet (AEJ) is one of the key elements of the West African Monsoon (WAM) system, a major climatic feature of sub-Saharan North Africa that also has important impacts on the tropical North Atlantic

(Sultan et al., 2003 and Sultan and Janicot, 2003).

The ECMWF Archive contains all observational data acquired in real time from the World

Global Telecommunications System (GTS) since the beginning of daily operations in 1979. An archive of First GARP Global Experiment (FGGE) level II-b data was also on-site at ECMWF. To provide a comprehensive set of input data for

ERA data were acquired from a number of additional sources (Simmons et al., 2007; Uppala et al. , 2008).

The initial and lateral boundary conditions used in this experience with RegCM4.4 is ERA

Interim 15 years reanalyse form ECMWF version TS3.10.1. simulation are obtained from the new ERA Interim 2 .5°×2 .5° gridded reanalysis (Simmons et al., 2007; Uppala et al. , 2008), which is the third generation ECMWF reanalysis product. The main advances in this reanalysis compared to ERA-40 are a higher horizontal resolution (0.75°×0 .75° but available also at 1°×1° and 2 .5°×2 .5°), four-dimensional variational analysis, a better formulation of background error constraint, a new humidity analysis, an improved model physics, variational bias correction of satellite radiance data, and an improved fast radiative transfer model. ERA-

Interim uses mostly the sets of observations acquired for ERA-40 with a few exceptions: acquisition of a new altimeter wave-height that provides data of more uniform quality, use of

reprocessed Meteosat data for wind and clear-sky radiance, and new ozone profile information from 1995 onwards.

3.3 Data Analysis

3.3.1 Observation data processing

3.3.1.1 Significance t-test of Differences

The method used to calculate the significance t-test values of differences between two different decades (1980s, 1990s and 2000s). It was assumed that the variance of the two samples is the same.

(3.1)

(3.2)

Where and are number of observations for each experiment and the degrees of freedom is df .

= average, s = standard deviation

(3.3)

(3.4)

This average ) is used to compute the decadal mean of rainfall and NDVI

The Pooled variance ( PV ) and Standard Error ( SE ) are given by (3.5) and (3.6)

(3.5)

(3.6)

The mean difference and the significance t-test are given by equation (3.7) and (3.8):

(3.7)

(3.8)

The cutoff t value is calculated based on df and the significance level.

3.3.1.2 Standardized Precipitation Index

The Standardized Precipitation Index (SPI) was developed by McKee et al. (1993, 1995) to provide a spatially and temporally invariant measure of the precipitation deficit (or surplus) for any accumulation time scale. The SPI is a probability index that considers only precipitation. It is computed by fitting a parametric cumulative distribution function (CDF) to a homogenized precipitation time series and applying an equiprobability transformation to the standard normal variable. This gives the SPI in units of number of standard deviations from the median. It is negative for drought, and positive for wet conditions. It is given by:

(3.9) where is the precipitation of the year i, is the precipitation averaged over thirty years and

is the standard deviation of the series.

Table 3. 1 Drought categories from SPI (McKee et al ., 1993)

SPI Drought category Mild drought Moderate drought Severe drought Extreme drought

3.3.1.3 Correlation between rainfall and NDVI

The correlation between rainfall and NDVI is plotted for the whole region. Rainfall and NDVI monthly data over 29 years (1982 - 2010) the sample size N is equal to 348. No time lag was observed between the two parameters.

To pick up the seasonal climatology, NDVI and rainfall have been averaged over months from 1982 to 2010 over six sites throughout the region: two sites over the Guinea region

) two sites over Sudan region

(Mali and Benin) precisely, in Burkina

Faso and Niger

3.3.2 Model Setting and Simulation

3.3.2.1 Model Description and Simulation

Figure 3.8 shows the simulation domain with its relief shown by the contour lines. The study area over West Africa is bordered by the red box.

Figure 3. 8: Model simulation domain with the topography (in contour) the zone of interest in red box.

ICTP-RegCM4 is the last version of the regional climate modeling system developed by

ICTP. It is an improved evolution of its previous version RegCM3 (Giorgi et al., 2012). The model description and the changes between RegCM4 and the previous versions are described by Giorgi et al. (2012). However, it is a hydrostatic, compressible, sigma-p vertical coordinate model run on an Arakawa B-grid in which wind and thermo dynamical variables are horizontally staggered. The land surface model used to couple RegCM4 with is

Biosphere-Atmosphere Transfer Scheme (BATS) which will define the interaction land surface-atmosphere (Dickinson et al. , 1993). Convective precipitation is calculated with the scheme of Grell et al. (1994) applying the Fritsch and Chapell (1980) closure assumption.

Resolvable precipitation processes are treated with the sub-grid explicit moisture scheme

(SUBEX) of Pal et al. ( 2000), which is a physically based parameterization including subgrid scale cloud fraction, cloud water accretion, and evaporation of falling raindrops. Table

3.1 shows briefly the summary of the simulation description. The projection is Normal

Mercator (NORMER), the SST type (OI_WK) and Era interim 15 (EIN15) for the boundary conditions.

Table 3.2: Summary of the model configuration

Stanzas Parameters Dimensions Domain Number of points in N/S direction 100 Geo-parameter Map projection NORMER Boundary conditions SST type OI_WK Convection scheme - Grell over land and ocean Simulation periods - 2005-2006-2007 Subex Cevapland* 0.003

Cevapland: parameter of the SUBEX moisture scheme, raindrop evaporation rate coefficient over land Cevapoce: parameter of the SUBEX moisture scheme, raindrop evaporation rate coefficient over ocean

The modelling aspect of the thesis is approached through this section. It is a case study of the impact of land use change on rainfall and atmospheric parameters over two years 2005 and

2006. The experiments are based on (1) simulating the rainfall and atmospheric parameters initially without any changes on land surface cover as a control; then (2) the land cover over

West Africa savannah zone is replaced by the short grasses before performing the simulation of the rainfall and atmospheric parameters. The model is evaluated with CRU and GPCP.

Comparison was made between atmospheric patterns before changes and after changes in

June-July-August-September (JJAS) of 2005 and 2006 and same with the rainfall.

3.3.2.2 Model Evaluation

The evaluation of the model was based on two parameters which rainfall and temperature.

The rainfall was validated with respect to two observation data namely CRU and GPCP. And

CRU was used to evaluate the temperature. The model has been set based on above parameters during JJAS of 2005, 2006 and 2007.

3.3.2.3 Land Surface Model

For the soil moisture- wo regions

(1) oceanic regions and (2) continental regions with and without snow cover (our case). For the nonsea-ice-covered oceanic regions the surface temperature Tg1 is prescribed from observational data in the standard model . For other regions, the computation of T g1 depends on the current conditions of snow cover soil moisture, type of surface and temperature of the first layer of the atmosphere.

The vegetation part of the code is only executed for grid squares with vegetation cover greater than 0.001. A mean wind within the canopy is obtained from the mean wind outside the canopy times the square root of the drag coefficient which equals the friction velocity u .

The coefficient of transfer of heat and momentum from leaves is calculated (Fig. 3.9) .

Foliage water is modified by intercepted rainfall . The temperature of the foliage (leaves) is

calculated . Any rain or snow intercepted by leaves in excess of their maximum capacity is determined as falling to the ground and saved for soil-water or snow-budget calculations .

Figure 3. 9: Schematic of individual physical process (Dickinson, 1993)

3.3.2.4 Structure of BATS

In the BATS scheme (excluding the snow sub-model), there are 3 soil layers and 1 vegetation layer, accounting for 7 prognostic variables: canopy temperature ( ), surface soil temperature ( ), subsurface soil temperature ( ), surface soil water ( ), root-zone soil water ( ), total soil water ( ), and canopy water store ( ). There are 18 surface- cover types which are based on Olson et al. (1983), Matthews (1983) and Wilson and

Henderson-Sellers (1985). The soil type data are based on Wilson and Henderson-Sellers

(1985). For each vegetation type, there are about 27 derived parameters which determine the morphological, physical and physiological properties of vegetation and soil.

Rawls et al. (1993) have provided a detailed review of the theory of soil water movement or internal soil water fluxes. The rate of soil water movement is important in surface runoff, groundwater recharge and evapotranspiration. Computation of the internal soil water fluxes within the soil column of 1-10 m thick that is coupled to the atmosphere is difficult to implement in climate models. The reasons are due to the coarseness of the grids, usually 2-5 levels, as limited by practical considerations, and the heterogeneous distribution of topography, soil and vegetation types across the GCM grid-square.

The soil surface evaporation and the internal soil water fluxes in the BATS model are parameterized based on the multilayer soil model integrations. The capillary movement of water from the rooting zone into the surface soil layer is given by

(3.9)

And from the total column into the rooting zone is

(3.10)

Where = ratio of soil water content within the total column to its maximum amount. The gravitational drainage from the surface soil layer to rooting zone is

(3.11)

And from the rooting zone to the total column is

(3.12)

Where

And

3.3.2.5 Experimentation: Change in Land Surface State

The effect of deforestation on rainfall variability in the region was evaluated using the regional climate model (RCM) RegCM version 4. The model was to simulate deforestation by attributing short grasses to the transition zone located between the Guinea and Sudanian region ( - - ) which was originally characterised by woodland savannah with tall grasses. Two experimentations were applied. For the first, a simulation (1) was made without any change in vegetation cover and it was referred as the control (CTL).

Then, for the second, a simulation (2) was a sensitivity test. It was done after the change in vegetation cover (change mentioned above) and it is referred as the sensitivity of the model to the changes in vegetation cover (Sens). The period is detailed in the next sub section

(3.3.2.5). Figure 3.10 shows the change made in the vegetation cover. Countries crossed by

the change are Guinea, , Ghana, Benin, Togo, Nigeria and Cameroon. Based on the initial land cover distribution, the area was mainly occupied by the mixed woodland and evergreen broadleaf tree. Some characteristics attributed to the short grasses will be found into the general description of the vegetation and soil types in BATS (Dickinson et al., 1986).

For each of the land grid points, three other variables are defined in subroutine ALBEDO- visible solar albedo of vegetation ( <0.7 m), near-infrared albedo of vegetation ( >0.7 m), and soil albedo. The land use type in BATS scheme defined the vegetation classes are mentioned in Table 3.2 based on RegCM user manual version 4.4 (Giorgi et al., 2014).

Because of the lack of some types of vegetation in our region all the vegetation type represented in the user manual are not shown in the Table.

Table 3.3: Land cover/vegetation classes

Short grass Evergreen needle leaf tree Deciduous broadleaf tree Evergreen broadleaf tree Desert Irrigated crop Semi-desert Bog or marsh Ocean Evergreen shrub Mixed woodland Forest/Field mosaic

The values for the albedo of vegetation were determined from a variety of sources, in particular but also with reference to Monteith (1959), Barry and Chambers (1966), Federer

(1968), Oguntoyinbo (1970), Stewart (1971), Tucker and Miller (1977), Rockwood and Cox

(1978), Kriebel (1979), Fuller and Rouse (1979) and Kukla and Robinson (1980).

(3.13)

Where is the albedo for a saturated soil and where the increase of albedo due to dryness of surface soil is given for <0.7 m as a function of the ratio of surface soil water content

to the upper soil layer depth ,

(3.14)

Table 3.4: BATS vegetation/land-cover Dickinson, 1993

3.3.2.6 Case study of 2005 and 2006

The simulation covered three years 2005 to 2007. However focus has been on the years 2005 and 2006. The two years were selected due to the contrasts between some patterns which are

tudies have revealed differences between 2005 and 2006. For example during the AMMA campaign the sub period called

Enhanced Observing Period (EOP) covering 2005-2007 is designed to provide a detailed documentation of the annual cycle of the surface and atmospheric parameters from convective scales of a few kilometres up to regional scales. The year 2005 was found to be

The regional SST field is one of the key factors in West Africa rainfall (Lebel et al., 2010). The ocean plays a key role in the WAM dynamics, especially during the onset phase. Warmer Gulf of Guinea

SSTs induces stronger surface sea breeze convergence in the Guinea Coast region and abundant rainfall over the southern part of West Africa. The modification of the TEJ/STJ leads to more humid atmospheric conditions throughout western Africa region and possibly to changes in the triggering and behaviour of AEWs. The large-scale response is still accompanied by changes in land use and vegetation cover. Furthermore, the monsoon has started early in 2005 when in the following year (2006) it started later. One fundamental goal of AMMA in terms of process understanding was to document the various atmospheric structures described above at the appropriate scale in order to better understand their interactions (Lebel et al., 2010).

Figure 3. 11: Diagram showing the differences on SST in Gulf of Guinea and monsoon onset between 2005 and 2006 over West Africa.

Chapter 4

RESULTS AND DISCUSSION

4.1 Mean NDVI and Rainfall over Last Three Decades

4.1.1 Vegetation and Spatial Distribution of Rainfall over West Africa

The NDVI spatial distribution reveals a vegetation gradient increasing from the North to the

South (Fig. 4.1). With rainfall amount of about 600 mm yr-1 and monomodal rainfall pattern

(Konaté and Kampman, 2012), vegetation indices was low (below in extreme North of Burkina Faso, Niger, Mali and Mauritania. It corresponds to the Sahel with vegetation composed of semi arid grasslands, savannah, steppes and thorn shrublands.

However, droughts and dust storms are frequent, and major constraints are overgrazing and soil erosion and desertification (Boateng, 2013).

-savannah mosaic scattered with agricultural zones stretch up to northern part of coastal countries go,

Benin and Nigeria up to the central part of Africa (Congo watershed). The NDVI values are in between 0.4 and 0.6. These regions are characterized by monomodal rainy season alternated with pronounced dry season (Konaté and Kampman, 2012). The annual rainfall varies between 800 and 1200 mm yr-1. The vegetation is dominated by grassland and trees with low density. Lowland rainforest wetter, drier and mixed types with NDVI values above

0.6 are distributed over coastal regions from Gambia up to Cameroon. The landscape is mostly flat with remnants of dense humid forests, with annual rainfall higher than 1500 mm distributed in a bimodal pattern (Konaté and Kampman, 2012). The vegetation distribution is consistent with rainfall spatial distribution.

4.1.2 NDVI Decadal Variability over West Africa during 1981-2010

4.1.2.1 Decadal Mean of NDVI

The decadal means of rainfall and NDVI are shown in Figure 4.2. Decades 80s, 90s and 00s, are shown respectively in Figures 4.2(a-c) for the rainfall and Figures 4.2(d-f) for the NDVI.

No significance level of changes could be depicted the level of from one decade to another.

However, the spatial distribution of the rainfall shows that the rainfall is mainly concentrated over two zones in coastal region of West Africa located from Bissau Guinea to the Southwest of and from Southeast of Nigeria to Cameroun due to high relief in these regions. Ogungbenro et al. (2015) found intense mesoscale convective system (MCS) over these regions of West Africa. This is explained by the effect of local climatological features on MCS occurrences and the abundance of rainfall in these localities. It shows that NDVI high values (above 0.6) are located in area with high rainfall concentration. This highlights the positive relationship between rainfall and NDVI.

Figure 4. 2: Mean annual Rainfall (mm yr -1) shown in [a]; [b] and [c] for respectively decades 80s 90s and 00s and mean annual NDVI shown in [d]; [e]; [f] for respectively decades 80s; 90s and 00s.

4.1.2.2 Seasonal Variability of NDVI

Figure 4.3 shows the seasonal distribution of the NDVI averaged over three decades. Figures

4.3(a-d) indicate the seasonal mean of 80s corresponding to December-January-February

(DJF), March-April-May (MAM), June-July-August (JJA) and September-October-

November (SON) respectively. DFJ is characterized by low values of NDVI over Sudan savannah, the NDVI does not exceed 0.3(Fig. 4.3a). Only the extreme southwest of the region has some high values of NDVI.

The NDVI values in MAM that is equal to 0.3 have decreased in the southern Sudan, while it shows increasing values towards the Northern region (Fig. 4.3b) but remains unchanged in the extreme North.

In JJA an important shrinkage of areas with 0.3 and 0.2 are observed for the benefit of the growing vegetation over Sudan savannah meanwhile, there is reinforcement of NDVI over

Guinea region with appearance of areas with NDVI above 0.7 (Fig. 4.3c).

In SON the most observed change is the enlargement of the area with 0.7 as NDVI value over western and southern part of the region.

The decades 90s (Fig. 4.3(e-h)) and 00s (Fig. 4.3(i-l)) show the same spatio-temporal distribution scheme of NDVI over the region as obtained in decade 80s. However, the main seasonal change from decade to decade is the increase of NDVI values (above 0.7) and enlargement of area with high NDVI values during the last decade 00s over Guinea Coast.

Figure 4. 3: NDVI decadal mean showing changes in vegetation cover with progressive southward increase in decade 80s (a, b, c and d); decade 90s (e, f, g and h) and decade 00s (i, j, k and l).

4.2 Spatio-Temporal Distribution of the Rainfall over Three Last Decades

This section describes of the rainfall over thirty year period (1981-2010). The whole period is split into three decades. First of all, all the decades were compared to the normal. Then an inter-comparison was made between the different decades. Furthermore, the spatial distribution of the SPI over some sites is given. Finally, the section was concluded by observing the seasonal changes between decades.

4.2.1 Changes compare to thirty years climatology

The comparisons between each decade and the thirty years climatologic mean (1981-2010) are shown in Figure 4.4. Compare to the last three decade means (the normal), decade 80s indicates the driest years (Fig. 4.4a). A rainfall deficit was observed all over the region with significant decrease in northern part of Nigeria and Liberia while the western and eastern

Sahel were the most affected regions. Decade 90s was wetter compare to the normal. The rainfall has increased in the regions where the deficit was observed namely western and eastern Sahel (Fig. 4.4b). A slight but not significant deficit was observed in the central part of Ivory Coast. In 00s, the rainfall was observed to have increased over some areas. Thus, the wet condition was maintained over Liberia and only extended to Ivory Coast in 00s, whereas it has decreased over Guinea regions. Some deficits are mainly observed over Benin and

Nigeria (Fig. 4.4c).

Among the three decades, decade 80s was the driest. The whole region suffered below normal rainfall. This result is in agreement with earlier studies in the West African region

(Nicholson et al., 1998; Giannini et al., 2003; Zeng, 2003; Dai et al., 2004; Lebel and Ali,

2009 and Lebel et al., 2010). However, there was a recovery in 90s over many areas of the region (Nicholson, 2005; Evan et al., 2015 ) this drought was confirmed in 00s over some

because some of areas are going back to deficit conditions.

Figure 4. 4: Spatial distribution of rainfall significant changes in decade [a] 80s; [b] decade 90s and [c] decade 00s compare to 30 years average (1981-2010) over West Africa at a level of 95%.

4.2.2 Decade to decade changes

The rainfall evolution from decade to decade was computed using the significance t-test at a level of 95% between the decadal mean of the considered decades (Fig. 4.5). Figure 4.5a shows the difference between mean rainfall of 90s and 80s. Positive changes were observed over large area of the region with strong differences (greater than 300 mm yr-1) in Northern

Nigeria, Liberia and Guinea. This result confirmed the positive anomaly observed in sub section 4.2.1 in 90s. Over western and eastern part of West Africa, the rainfall increase in

90s is significant at 95%. Finally, a slight decrease in the rainfall is observed over some parts of and Ghana. Also, Figure 4.5b shows the differences between decade 00s and

80s, similar to 90s, decade 00s indicates wetter than 80s over large area of West Africa except in some part of Ghana and Benin where the rainfall had decline. It is still noted some narrow zones where the change is positive but not significant in Mali and Mauritania.

Figure 4.5c presents the difference between 00s and 90s. This was the most contrasted in terms of spatial distribution of positive and negative differences. A decrease in rainfall was shown in many areas mainly over Nigeria, Benin, Togo, Ghana and the Republic of Guinea during 00s after a recovery period in the 90s. On the other hand, the rainfall kept increasing mainly over , Senegal and Mauritania. The areas which were going back to drought condition in 00s (see sub section 4.2.1) are still wetter than 80s. So the level of the drought in 80s is not achieved.

Figure 4. 5: Decadal changes in rainfall seasonal spatial distribution over West Africa at a level of 95%. Blue lines are areas with significant changes.

4.2.3 Rainfall distribution

The time latitudinal diagrams of the rainfall decadal mean and the rainfall decadal anomalies in 80s, 90s and 00s are shown in Figure 4.6.

Decade 80s was marked mainly by weak magnitude of rainfall (Fig. 4.6a) hence the negative anomaly all over the study period (Fig. 4.6d). The rainfall amount above 240 mm month -1 is reached over Sudanian region later in September.

In Decade 90s an increase in rainfall in August over Sudanian region was observed (Fig.

4.6b). Furthermore, an enlargement of the area with the rainfall amount above 240 mm month -1 was also observed. However the anomaly shows a clear contrast between the long rainy season and the short rainy season (Fig. 4.6e). The long rainy season was wetter compared to the short rainy season. The third decade at some slight difference was similar to the second decade (Fig 4.6c and Fig. 4.6f).

The differences between the three decades are mainly about the rainy period and the rainfall magnitude. From 80s to 90s the rainfall has shifted on time. Therefore, the extreme North was reached by the rainfall belt in August in 80s however it is reached earlier around July in

90s and 00s. In terms of rainfall magnitude, it was more intense in 90s and 00s contrary to

80s.

4.2.4 Upward Trend of the Rainfall over the Region

To monitor the severity of the drought, the Standardized Precipitation Index (SPI) is computed at different sites over Sahelian, Sudanian and Guinea regions. It shows at these sites an alternation of dry and wet sequences (Fig. 4.7). The deficit in rainfall occurred mainly at the beginning of 80s till the middle or the end of decade 80s for some sites over the

Sahel and Sudanian region. Numerous studies referred to early 70s as the beginning of the drought over the West African region. Sanni et al. (2012) evaluating the drought severity in the Sudano-Sahel Region (SSR) of Nigeria, they found that most of the drought severity with the highest magnitude occurred between the 70s and 90s. However, contrary to number of finding, no severe drought sequences were noted during the period of this study. Thus, this can be explained by the selected climatology period which is 1981-2010. However, it was observed that since early 90s and 00s the recurrence of positive value of the SPI over the region. More wet years are observed in decade 90s at all the considered sites. This observation is in agreement with Dai et al. (2004) who found that large multi-year oscillations appear to be more frequent and extreme after the late 80s than previously (before

1980). Thus for them, that may mean become more unstable and prone to droughts after the prolonged severe droughts from the early 70s to late 80s (Dai et al., 2004).

4.2.5 Changes over Seasons

The differences between seasonal rainfall in 80s, 90s and 00s were shown in Figure 4.8. The first differences between decade 90s and 80s. The result for DJF shows very narrow areas with significant positive changes in Liberia and Guinea. However, a deficit was observed over the southern part of Nigeria (Fig. 4.8a). For MAM there was an enlargement of the area with significant positive changes over Guinea coast and central Sudan region (Fig. 4.8b). JJA period shows positive changes and moved northward in the Sahel region (Fig. 4.8c) with an increase in rainfall trend in 90s. Some decreases have been observed in the central part of

little rainy season (SON), extreme West and East of the region were the wettest (Fig. 4.8d). These regions with significant positive changes coincided with the areas where rainfall had significantly increased between 80s and 90s.

Almost the same observation was obtained when compared 00s and 80s (Fig. 4.8(e-h)), this implies that decade 00s and 90s was not much different in terms of rainfall spatial distribution at a seasonal scale. This could explain the scatted distribution of significant differences between the two decades. However, the difference between 00s and 90s shows an important mitigated distribution of the rainfall in term of quantity. The rainfall has decreased over some areas during the last decade namely from Ghana to Nigeria and Guinea to Sierra

Leone in MAM and JJA (Fig. 4.8(i-l)).

There was consistent increase in rainfall over some areas during the three decades, for instance nd Liberia where an increase in rainfall was observed during the dry season (DJF). The result was opposite over southern part of Ghana where rainfall amount shows a decline in JJA. However, in most cases it is fluctuating between wet and dry conditions. The differences showed that 90s was wet

than 80s over a large area of West Africa except and Nigeria where 90s was dry in JJA and also in a southern part of Ghana for SON.

However, the observed increase in rainfall between 80s and 90s over large area could be explained the fact that from 90s till now, more wet years are observed (Fig. 4.8). High rainfall amount and a greater number of wet periods were observed contrary to the 80s within which there were many dry sequences of rainfall over many areas throughout West Africa

(Nicholson, 1993 and Omotosho, 2008). The extreme and severe droughts occurred in 1983 and 1984 in many areas of West Africa. This result supported previous work on extreme and severe drought in the region based on Palmer indices (Dai, 2011 and Hua et al., 2013).

Compare to 80s, there was an increase in rainfall in 00s over the western part of West Africa

Studies have revealed a rainfall recovery over Sahel regions starting in early 90s and still going on in some areas

(Nicholson, 2005; Herrmann et al., 2005; Ali et al., 2008 and Lebel and Ali, 2009).

The whole Guinea zone have also registered a slight increase in the rainfall amount from

, the same result was found in the northern part of Nigeria by

Herrmann et al. (2005). However, in the last decade there was appearance of some areas with negative changes but very small within Nigeria and Benin, this implies non-coherence of the spatial distribution and Changes in surface energy budgets resulting from land cover change

4.3 Spatio-Temporal Distribution of the Vegetation over Three Last Decades

This section describes the vegetation dynamics through the Normalized Difference

Vegetation Index (NDVI). The decadal anomaly was firstly computed, followed by the inter comparison between decades. Finally, a spatio temporal variability and the seasonal change over decade were analysed.

4.3.1 NDVI Decadal Anomaly Variability over West Africa

Compare to the whole period, decade 80s was widely negative in term of vegetation growing all over the region. This was highlighted throughout Guinea and Sudanian region. However some narrow areas with positive greenness were observed over central part of Ghana and

Cot a).

In 90s the re-greening has started in central West African region and over West Sudanian

is period the change in vegetation indices is still negative over Guinea region (Fig. 4.9b).

In the last decade 00s the rate of re-greening is the highest over Guinea, Sudanian regions and over western Sahel. So contrary to 80s, in addition to some areas over Sudanian region, the positive change has extended to the coastal region, meaning that the vegetation has greened progressively southward when the decreases are becoming more significant over Sahel region

(Fig. 4.9c). In the Gulf of Guinea, the re-greening is the highest all over the West African region.

Above this region it is still observed significant. Furthermore, an enlargement of the area with negative differences is observed

kina. These refer to a decrease of the vegetation cover over these areas.

4.3.2 NDVI Decadal Variability over West Africa

The difference between 90s and 80s shows positive increase in vegetation cover over

Sudanian 4.10a). After the severe drought occurred in 80s, the vegetation has been significantly greened over the Sudanian region

ian region during this period could be due to the rainfall recovery which has started in decade 90s and may have induced some vegetation cover recovery.

It was not the case in Guinea region where there were no changes or even negative changes were observed during this period. When over Guinea region significant increase in vegetation

compare to 80s.

central West African region and to Sudanian region of West Africa (Fig.

4.10b). This positive difference is highlighted over coastal region by significance difference.

So contrary to 80s, the positive change has extended to the coastal region, meaning that the vegetation has greened progressively southward when the negative changes are becoming more significant over Sahel region. The NDVI provides information about the green leaf area index (Myneni and Williams, 1994). Anyamba and Tucker (2005) using the same set of data but over 23 years found a gradual recovery from drought conditions.

The comparison between 00s and 90s shows the persistence of significant positive changes in the coastal region still observed some positive changes up

however, not significant (Fig. 4.10c). Furthermore, an enlargement of area with negative differences is observed over Sah over Niger, Mali and Burkina. These referrers to a decrease of the vegetation cover over these

areas. Therefore, the opposite phenomenon is observed over Guinea coast with the greening process going on until 90s.

4.3.3 Decadal Change on NDVI

The time latitudinal diagrams of the NDVI decadal mean and decadal anomalies in 80s, 90s and 00s are shown in Figure 4.11. First, in 80s it is shown that

April period corresponding to the dry season (Fig. 4.11a). Then there was a narrowing of this area tow , they are concentrated in Guinea zone between April and November

Sudanian region during August-September. However, the anomaly does not show significant difference in 80s (Fig. 4.11d). In 90s the same spatio-temporal distribution of the NDVI was observed (Fig. 4.11b) like in 80s, except the fact that in the South - s a decrease in NDVI from mid July up to September. The anomaly shows a contrast between the first eight months of the year (January to August) and the last four months (August to

December). The positive anomalies are observed during the first eight months contrary to the last four months where negative anomalies are observed (Fig. 4.11e). In 00s the NDVI value has increased in the South the 0.5 NDVI is observed early apart from January and the 0.6 starts before April (Fig. 4.11c). Furthermore, there was an appearance of 0.7 NDVI value between October and November in the Sudano-Guinea region. However, the decrease has extended from the Guinea up to Sudano-Guinea region between July and September. The anomaly shows the same contrast as in 90s (Fig. 4.11f).

Figure 4. 11: Time latitudinal diagrams of seasonal NDVI shown in a, b and c for decade 80s, 90s and 00s respectively and NDVI seasonal anomaly - shown in d, e and f for decades 80s, 90s and 00s.

4.3.4 NDVI seasonal variability over West Africa

The difference between decade 80s and 90s shows narrow significant positive changes scattered mainly and Ghana in DJF (Fig. 4.12a). In the following season

MAM positive changes (Fig. 4.12b), have slightly moved northwards in the sub Sudanian region with some significance in Guinea Republic. However, it is noted some negative changes in the South of voire, Ghana and Guinea in JJA (Fig. 4.12c). Then in SON

(Fig.

4.12e).

The difference between decades 90s and 00s is positive and significant in the Guinea region from Senegal to Nigeria (Fig. 4.12f). In MAM there is not upwards movement of the positive change however, the negative change has highly increased over Sahel region (Fig. 4.11g).

The following seasonal difference in JJA is slightly positive in some narrow places over

Sahel band. When in the Guinea region the change has become significantly negative (Fig.

4.12h). The difference between DJF NDVI of decade 00s and 80s shows significant positive change over Guinea region (Fig. 4.12i). In MAM the change is still positive over Guinea region negative change is observed above 12 4.12j). For JJA, the positive change is scattered over Sahel region when over Guinea region the negative change is widely significant (Fig. 4.12k). In SON the positive change is mainly significant and located between

4.12l).

From these analysis, it is observed that the difference between decade 80s and 90s show northwards seasonal increase in vegetation covert. But this difference seems to be less significant. Compare to the previous test, the positive difference between the second and third decades is almost sedentary for the two first seasons DJF and MAM then widely negative in

JJA. The difference between 90s and 00s is better highlighted than the previous difference thus more significant. The last test is almost similar to the second meaning that 90s compare to 80s is quite similar to 00s compare to 80s this induces that there was not important spatial changes between 90s and 00s. The vegetation covert seems to be constant.

With respect to the NDVI spatial distribution, it is noted that the test of means difference between the three decades shows a strong seasonal variability from a season to season and a decade to another (Fig. 4.12). The difference between decade 80s and 90s is not showing in general an expended significant changes. However, the positive changes occur in Guinea

. The most important changes occur in SON over sub Sudanian and Sahel zone negative changes throughout Guinea region. The difference between decade 00s and 90s is significantly positive all over Guinea region in DJF and MAM contrary to the Sahel region where the changes are negatives mainly in Niger, North Burkina Faso and Mali. JJA is relatively negative over Sahel and Guinea region. The last test which is computed between

00s and 80s shows

DJF and MAM. In JJA, it is identified some negative changes over the Sahel and Guinea regions. Except the southern part of , Liberia and some regions in Nigeria, the whole and 00s.

4.3.5 Frequency distribution of the NDVI

The level of vegetation greenness was defined based on the method applied by Lim and

Kafatos (2002) who divided vegetation

Over Niger site, the high frequency is located in between 0.24 and 0.25 (Fig. 4.13) which

ur case the Sahel region of West Africa where the mean rainfall is about 469.76 mm yr-1.

Figure 4. 13: Annual frequencies and distribution of NDVI at Niger (left) and Burkina Faso (right) sites over Sahel region.

The vegetation of West Africa is characterised by a combination of factors related to both climate and soil, essentially exhibiting the same longitudinal zonation as rainfall. The structure of the vegetation also changes progressively from North to South. The further

South, the taller the vegetation, the greater the proportion of woody species (trees, shrubs, bushes), and the higher amount of ground cover.

occasional woody species of small trees or shrubs. Grasses are perennials, generally not taller than 80 cm, and animals; woody species are often thorny, like the typical Acacia. This is the

Sahel proper. Here the surface is to a large extent bare soil and the vegetation tends to be

clustered in sites of favourable conditions of soil or runoff, creating a mosaic pattern

(Nicholson, 1993).

The higher frequency was obtained between 0.32 and 0.34 at Burkina Faso site this range belongs to medium green in Lim and Kafatos (2002) classification (Fig. 4.14).

Figure 4. 14: Annual frequencies and distribution of NDVI at Mali (left) and Benin (right) sites over Sudanian region. The Sudanian Savanna is characterized by the coexistence of trees and grasses. Dominant tree species are often belonging to the Combretaceae and Caesalpinioideae some Acacia species are also important. The dominant grass species are usually Andropogoneae, especially the genera Andropogon and Hyparrhenia, on shallow soils also Loudetia and Aristiada. Much of the Sudanian Savanna region is used in the form of parklands, where useful trees, such as shea, baobab, locust-bean tree and others are spared from cutting, while sorghum, maize, millet or other crops are cultivated beneath.

Figure 4. 15: Annual frequencies and distribution of NDVI at (left) and Ghana (right) sites over Guinea region .

The Guinea Forests of West Africa hotspot encompasses all of the lowland forests of political

West Africa, stretching from Guinea and Sierra Leone eastward to the Sanaga River in

Cameroon. This includes the countries of Liberia, , Ghana, Togo, Benin, and

Nigeria, which maintain remnant fragments of the forests. The hotspot also includes four islands in the Gulf of Guinea: Bioko and Annobon, which are both part of Equatorial Guinea, and Sao Tomé and Principe, which together form an independent nation. Bioko is a continental-shelf island, whereas the remaining three are oceanic.

The Guinea forests consist of a range of distinct vegetation zones varying from moist forests along the coast, freshwater swamp forests (for example, around the Niger Delta), semi- deciduous forests inland with prolonged dry seasons. Of all West African countries, only

Liberia lies entirely within the moist forest zone, although a substantial portion of Sierra

Leone also falls within the boundaries.

4.4 Relationship between Rainfall and NDVI over West Africa

This section focused on the relationship between rainfall and vegetation. It is first of all based on statistical analysis of the relationship between the two parameters at different sites over

West Africa.

4.4.1 Intra Annual Variability of Rainfall and NDVI

The statistical analysis of the rainfall and NDVI data was performed over three major climatic zones in West Africa (see Figure 4.1). The behaviour of the vegetation cover was analysed from 1982 to 2010. Parameters like rainfall and NDVI mean and standard deviation are given in Table 4.1. The correlation between rainfall and NDVI is also shown. The interannual variability over Guinea region is the highest in term of both rainfall and NDVI compare to Sahel and Sudan where most correlation are the Sahel region 0.56 (Niger) and

0.58 (Burkina) followed by the Guinea region 0.45 correlation is weak for the Sudanian region 0.24 and 0.21 at respectively Mali site and Benin site. The low correlation could be due to the time lag between rainfall and vegetation growing which is not always systematic as it is assumed in this work.

Table 4.1: Descriptive Statistics for the rainfall and NDVI time series

Rainfall NDVI Sites Mean STD. Mean STD. Corr Sahel Niger site 469.76 117.04 0.24 0.012 0.56** Burkina site 595.48 123.49 0.33 0.018 0.58** Sudan Mali site 1126.6 136.47 0.50 0.016 0.24 Benin site 923.54 119.88 0.49 0.018 0.21 Guinea 1244.4 181.37 0.63 0.036 0.45* Ghana site 1198.3 141.25 0.63 0.040 0.40*

**. Correlation is significant at the 0.01 level (2-tailed) ; *. Correlation is significant at the 0.05 level (2-tailed).

The rainfall seasonality could be responsible for the strong variability of rainfall and NDVI over Guinea region where the region is characterised by four seasons two rainy seasons and two dry seasons. Also the land cover type is an important factor to Wang et al. (2003) found an average of 0.85 for grassland and 0.79 for forest.

4.4.2 Relationship between Rainfall and NDVI over Guinea Region of

The correlations between NDVI and rainfall vary strongly from one month to another (Table

4.2 and 4.3). The rainfall of some months is significantly correlated with its NDVI. It is the case of April, May, August and September at Lamto and January, February, April and

December at Daloa. But it is not always the case in some cases it is observed 2 or 3 months gap between rainfall and NDVI. That is the case of July rainfall and NDVI of August, August rainfall and NDVI of October and November at Lamto. However, the correlation is significantly negative in November (-0.60).

DECn JANn FEBn MARn APRn MAYn JUNn JULn AUGn SEPn OCTn NOVn -0.27 JANp 0.29 0.13 -0.22 -0.04 0.03 -0.04 -0.10 -0.31 -0.23 -0.22 -0.19 -0.23 FEBp 0.39 -0.17 -0.01 0.17 -0.26 -0.27 0.05 0.07 -0.45* -0.23 -0.30 MARp 0.16 0.08 0.41* -0.18 0.09 0.24 0.08 -0.09 -0.18 0.13 APRp 0.56* 0.16 -0.25 -0.01 -0.17 0.03 -0.28 0.34 -0.14 MAYp 0.41* -0.26 0.22 -0.32 -0.31 -0.18 0.44* -0.06 JUNp -0.03 0.13 0.24 0.15 0.18 -0.24 0.01 JULp 0.22 0.55* -0.07 0.04 -0.04 -0.17 AUGp 0.59** 0.27 0.46* -0.60** 0.02 SEPp 0.43* -0.25 -0.15 -0.06 OCTp -0.07 -0.25 -0.24 NOVp 0.01 0.12 DECp

The situation at Lamto and Daloa is almost similar but with some slight differences.

Therefore at Lamto the February to February correlation is low 0.39 contrary to Daloa where the two parameters are significantly correlated 0.54, and then in May the correlation is significant between rainfall and NDVI at Lamto (0.41) when it is too weak at Daloa (-0.01).

For both localities June and July are not showing any correlation between the two parameters on time. The rainfall of March and April is negatively correlated with NDVI of April and positively correlated with NDVI of October at Daloa.

This could be explained by the rainfall regime and the vegetation types over the two sites.

Lamto is located in the Sudano Guinea savannah zone where the rainfall has much influence on vegetation contrary to Daloa site where there is still some forest and the vegetation is less influenced by the rainfall. However, the effect of rainfall can be seen during the dry season like in December, January and February (Table 4.3).

100

4.4.3 Rainfall Intra-seasonal Variability

The intra annual variability of the rainfall at the six selected sites over the three main climatic zones of West African region is differently observed. The selected sites over Sahel region are

and are located in Niger (Fig. 4.16a) and Burkina Faso (Fig. 4.16b). The rainfall regime is mono modal with the peak in August. Over the two Sahel sites the difference between the three decades is mainly observed at Niger site with 90s as the rainiest

(about 200 mm in August) and 80s the driest (about 100 mm in August). This finding is also seen by Hagos and Cook (2007). At Burkina Faso site the visual analysis does not show considerable changes.

Over Sudanian region the rainfall regime is still mono modal with the peak in August 350 mm and 250 mm respectively in Mali and Benin. But the rainfall amount is higher compare to the rainfall in Sahel. The important change noted was the shift of the rainy period during the last decade at Mali site (Fig. 4.16c). This is due to a late onset and late secession of the rainfall and less precipitation in the core of the rainy season finding in agreement with Louvet et al. (2015). In Benin the rainy period did not change only some slight change has been observed in rainfall amount in August (Fig. 4.16d).

At last the two selected sites over Guinea regions (Fig. 4.16e) and Ghana (Fig.

4.16f) show a bimodal regime of the rainfall. The two peaks are obtained respectively in

May-June and October due to the presence of two rainy seasons. Detailed description of this regime could be found in Konate and Kampmann (2012). The change occurs in the second rainy season with and an increase in the rainfall over the last decade. However many studies and research programmes (African Monsoon Multidisciplinary Analysis) built up over West

Africa have shown the deep implications of the monsoon in the rainfall amount and spatial

101

distribution over the West African region (Pospichal et al., 2010; Gosset et al., 2010 and

Lebel et al., 2010).

Figure 4. 16: Decadal rainfall at the monthly timescale plotted for the six selected sites in the Sahel region ([a] Niger and [b] Burkina Faso), the Sudanian region ([c] Mali and [f] Benin) and the Guinea region ([d] and [e] Ghana).

102

onset and the early cessation. Thus, the rainfall amount increases whereas the length of the rainy seasons decreases. The distribution of the rainfall over the year has changed. The rainfall drops in August due to the monsoon jump (Hagos and Cook, 2007). Hence, during this period, the rainfall is mainly due to locale conditions and some remote parameters. This exposition is supported by Odekunle and Eludoyin (2008). Yaw and Ian (1994) found positive correlations between the high SST in the Gulf of Guinea and high rainfall; same colder oceans are associated with lower rainfall in June-July and September-October in

Ghana. Many studies and research programmes (African Monsoon Multidisciplinary

Analysis) built up over West Africa have shown the deep implications of the monsoon in the rainfall amount and spatial distribution over the West African region (Pospichal et al., 2010;

Gosset et al., 2010 and Lebel et al., 2010).

4.4.4 NDVI Intra-seasonal variability

The intra annual variability of NDVI is analysed over decades 80s, 90s and 00s (Fig. 4.17) at the six selected sites. Like the rainfall, the NDVI has a monomodal evolution over the year.

The NDVI is weak over Sahel (<0.4) at both sites Niger (Fig. 4.17a) and Burkina Faso (Fig.

4.17b). Compare to the three decades, the NDVI has decreased during the first semester in

00s. However, the peak of 80s NDVI is the lowest. The vegetation has increased in July-

August-September during the two last decades (90s and 00s).

In the Sudanian region, no significant change can be observed from January to July. On the other hand, it can be observed an increase of the NDVI during decades 90s and 00s from

August to December. As shown in the previous section the higher amount of rainfall occurs in

August however decade 00s was not the rainiest but the vegetation response was the highest

(Fig. 4.17c and 4.17d).

103

In Guinea region the NDVI seasonal variability follows the rainfall regime with two peaks.

No important visual change is observed between decade 80s and decade 90s. However, it is observed a strong intra annual variability over the last decade. The vegetation starts growing early in the last decade and end the growing process late however it decreased in August (Fig.

4.17e and 4.17f).

104

Figure 4. 17: NDVI decadal mean averaged over months at six different points over Sahel region [a] Niger and [b] Burkina Faso, Sudanian regions [c] Mali and [f] Nigeria and Guinea

105

4.4.5 Rainfall and NDVI Monthly Climatology

Figure 4.18 shows the NDVI and rainfall intra annual variability. These are monthly mean of rainfall and NDVI over 30. Two types of annual distribution of the two parameters are shown. The NDVI shape is following the rainfall seasonal distribution. In the Sahel region the rainfall is mono modal. The peak is achieved in August (around 150 mm) when the NDVI peak occurred one to two months later (Fig. 4.18a and 4.18b).

106

107

The site selected over Sudanian region at the southern part of Mali (Fig. 4.18c) and northern part of Benin (Fig. 4.18d) the rainfall can achieve 300 mm in August and the NDVI 0.7 in

October. In April the NDVI start growing and will continuous up to September. The vegetation over these regions is more sensitive to rainfall. Mali and Benin sites are close to

The Guinea region which is characterized by two rainy seasons, two peaks of NDVI peaks is observed (Fig. 4.18e and 4.18f). Contrary to the rainfall, the second peak of NDVI is the highest as shown in (Fig. 4.18e). The amplitude is higher in the case of mono modal distribution of NDVI namely in Sahel regions. The amplitude of seasonal evolution shows that the saturation occurs early in region with high density of vegetation (Eklundh and

Olsson, 2003) as Guinea region of West Africa contrary to Sahel where vegetation takes some time to accumulate water.

The period of year identified as having a consistent upward trend in time series NDVI corresponding to the beginning of measurable photosynthesis in the vegetation canopy are mid June over Sahel region, April for Sudanian region and early in March over Guinea region.

The response of the vegetation to the rainfall seasonality is mainly seen in MAM and SON after a period of higher rainfall. Lamb (1980) revealed that approximately 83% of the rainfall occurs within this period of the year. In turn, they alter the equilibrium state of the vegetated surfaces (Monteny, 1986). As shown by Fensholt and Rasmussen (2011) and Herrmann et al.

(2005), an annual positive trend in vegetation greenness and rainfall were noted throughout the study over the sub Sudanian region during the last decades. This fact has been assigned to a desertification reverse by some authors. The rate of the increase of vegetation greenness can reach 50% in some areas in parts of Mali, Mauritania and Chad. Some areas where the

108

changes are not significant are observed thus, moving from the south of Benin to the

4.4.6 Relationship between Rainfall and NDVI over West Africa

For the relationship studies, focus was made on correlation between rainfall and NDVI. The correlation between rainfall and NDVI is widely positive (>0.4) over large area of West

4.19). In decade 80s the correlation can reach 0.8 over

Republic of Guinea, Senegal and in the Southern part of Mali, this positive difference is mainly significant at 95% level over the region (Fig. 4.19a). However it is weak over coastal region from Sierra to Nigeria mainly included in between -0.4 and 0.4. The same situation is observed in decade 90s nevertheless with some slight differences (Fig. 4.19b). So contrary to

80s, the area with high values of correlation with significant level has decreased. Furthermore slight enlargement of the zone with weak correlation values northward apart from the coastal region up to sub-Sudanian regions have been noted during this period. In 00s the area with high positive correlation has shrunk with an accentuation of northward expansion of the area with weak values over Sudanian region (Fig. 4.19c). In sum the area with high correlation kept dropping all over the region but precisely over Republic of Guinea, Mali and Senegal during the three decades. The report is that in general the vegetation growing is controlled by the rainfall availability over the region. However in the coastal region which is mainly covered by forest, vegetation is weakly influenced by rainfall. As shown in the previous sections theses region have significantly greened over the two last decades. This finding is similar to Richard and Poccard (1998) result who found weak sensitivity of the NDVI to rainfall over coastal areas, mountain regions and flooded areas they concluded that the interannual variability of the rainfall does not have a significant effect on the photosynthetic activity.

109

Figure 4. 19: Spatial correlation between NDVI and rainfall over [a] decade 1980s, [b] decade 1990s and [c] decade 2000s significance areas at 95% confident level.

110

Our findings are in agreement with those of Yuan et al. (2015) who found that grass vegetation is most sensitive to the changes in precipitation at about 250 mm. According to them, the correlation between rainfall and vegetation decreases with the increase in precipitation when precipitation exceeds 250 mm. It may be due to the limits of grass vegetation RUE, indicating that rainfall might not be the main constraint factor for grass, and the increase of NDVI might be related to other factors, such as temperature, radiation and so on.

111

Figure 4. 20: Scatter plot showing correlation and linear equation between rainfall and NDVI over 1981-2010 at six different points over Sahel region [a] Niger and [b] Burkina Faso, Sudanian region [c] Mali and [f] Benin and Guinea region [d] and [e] Ghana.

112

The correlation between rainfall and NDVI was high in large area of the region mainly over savannah areas it

Nigeria. The high values are mainly observed in region where the annual rainfall is around

1000 mm. So the vegetation growing depends directly on rainfall. This shows some linear

relationship between rainfall and NDVI over these regions. This finding is shared by

Nicholson et al. (1990) who found a linear relationship between rainfall and NDVI in the

Sahel below a rainfall threshold of about 1000 mm per year. At certain amount of rainfall the vegetation becomes less sensitive to rainfall. Using the REMO to simulate the rainfall, Paeth et al. (2005) found that rainfall is associated with large-scale circulation and less sensitive to the annual cycle of vegetation cover.

At a decadal scale, it is noted that the response of the NDVI to rainfall is strong during dry decade the case of 80s, however, this decrease for the wet periods like 90s and 00s. This observation has been also done by Wang et al. (2003) at yearly scale. After a year to year relationship analysis between rainfall and NDVI they conclude that NDVI responded more rapidly to precipitation during dry years, and during a year immediately after four consecutive dry years. By contrast, NDVI responded more slowly to precipitation during wet year, and during year immediately after a wet year.

113

4.5 Changes in Atmospheric Parameters

Compared to CRU data, RegCM4 shows some wet condition over some areas of Nigeria and

over the southern part of Sierra Leon,

GPCP rainfall data; the model output shows that the western part of the coastal area is wet over land and sea

(Fig. 4.21c and 4.21d).

114

Figure 4. 21: Rainfall monthly mean biases computed over June-July-August-September (JJAS) based on CRU a) and b) and GPCP c) and d).

115

Figure 4. 22: Temperature monthly mean biases computed over June-July-August-September (JJAS) based on CRU a) 2005 and b) 2006.

116

Nikulin et al. (2012) found TRMM drier than GPCP over tropical Africa, this point of view is

shared by Sylla et al. (2013) who found GPCP more consistence with gauge based

observations. But, founding some similarity between GPCP and TRMM Diallo et al. (2014)

used both the two observation data to detect the performance of HadGEM3-RA in monsoon

onset studies ov temperature is studied based on CRU temperature data. Compare to CRU the model gives hot biases over Sahel region mainly over Senegal, Mauritania, Mali and Niger and some cold biases in Guinea region Sierra Leon, Southern part of and Nigeria (Fig. 4.22a and 4.22b). Our findings in term of 2 m air temperature biases are similar to those determined by Diallo et al. (2014) using CRU to validate HadGEM3.

Table 4.4: Brief description of observational datasets and model used to set the simulation of RegCM4.4

CRU GPCP RegCM3 Rainfall and Temperature Rainfall Rainfall and Temperature 50 km 50 km 50 km Land only Land and ocean Land and ocean 1901-2006 (rainfall) 1991-2010 1989-2008 Harris et al., 2013 Sylla et al., 2012

117

4.5.1 Upper, middle and lower levels tropospheric winds

The atmospheric Jets have an important role in rainfall spatial distribution over this region.

The three main types were shown in the sub sections 4.3.1, 4.3.2 and 4.2.3. Wind spatial distribution is shown at 200 hPa, 700 hPa and 850 hPa before changes and after making changes on vegetation cover. These different levels represent respectively the location of the

Tropical Easterly Jet (TEJ), African Easterly Jet (AEJ) and the monsoon fluxes averaged over

JJAS which corresponds to the core of the rainy season over the sahelian region in 2005 and

2006.

4.5.2 Tropical Easterly Jet

The experimentation has given in general a TEJ speed ranging from 2 to 8 m s-1 over West

Africa for both experiments and the core of the wind is located over the Indian Ocean. So the comparison between TEJ before (Fig. 4.23a and 4.23c) and after changes in vegetation cover is not showing strong changes. However, only some slight changes on the wind speed occurs in region above Nigeria and Cameroon (Fig. 4.23b and 4.23d). The TEJ is said to be one of the most intense features over equatorial Africa and is one mechanism for the formation of

AEWs. Its average location ranges between 5 -10 N in August and 5 -10 S in January. The

TEJ is also produced from the thermal contrast. Central Africa is in the west exit region of the

Asian branch of the TEJ; this exit region will enhance upper-level divergence and lower-level convergence, promoting convective activity. Variability in the TEJ is associated to perturbations in the Tibetan high and so this feature could have a remote impact upon central

African rainfall (Farnsworth et al., 2011).

118

Figure 4. 23: JJAS mean Tropical Easterly Jet at 200 hPa before changes [a] and [c]; and after changes in vegetation cover [b] and [d] in 2005 and 2006.

119

The TEJ (200 hPa wind) according to Janicot et al. (2008) is the high-level anticyclonic structure, which is the sign of the Indian and African monsoons induces an easterly wind field on its southern flank. Configurations with stronger monsoon winds tend to have a stronger core of the TEJ (Klein et al., 2015). During the AMMA field campaign in 2006 Janicot et al.

(2008) found the same result by studying month per month from May to September. They went beyond the West African region and located the core speed greater than 20 m s 1 centred over the Indian Ocean. Recently Klein et al. (2015) using the WRF model and the

Atmospheric convective Model Version 2 (ACM2) spread ranges from 20 m s 1 and for ERA-

Interim, the maximum winds in the core exceed 20 m s 1 at 200 hPa. And some parallelisms are found with the monsoon activity in July and August.

4.5.3 African Easterly Jet

The highest speed of AEJ (4 m s-1) is located in the Western part of the region along Senegal,

rence between AEJ before and after changes al area boarding Liberia and Republic of Guinea. The averaged ITD was

4.24a and Fig. 4.24 in 2006.

The intensity of the African Easterly Jet is said to be a result of the communication of the surface temperature gradient into the lower troposphere (Cook, 1999). According to many authors (Payne and McGarry, 1977; Chen and Ogura, 1982 and Rowell and Milford, 1993), its presence has been associated with the occurrence of African wave disturbances and, arguably, with the modulation or even the instigation of intense, small-scale precipitation events. However, the jet is thought to be hydrodynamically unstable and African wave disturbances may be an expression of this instability (Thorncroft and Hoskins, 1994).

120

Figure 4. 24: JJAS mean Africa Easterly Jet at 700 hPa before changes [a] and [c]; and after changes in vegetation cover [b] and [d] in 2005 and 2006.

121

However, there are opposing views in the literature. Schubert et al. (1991), for example, suggest that the reversed potential vorticity gradients that mark the region of instability between the ITD and the African Easterly Jet are due only to the presence of a well-defined

ITD over West Africa. Thorncroft and Rowel (1998) finds that both the jet and the ITD contribute to the reversal of the potential vorticity gradient and, therefore, the unstable environment. The result in section 4.3.2 about AEJ is different from Janicot et al. (2006) over the region. This could be due to the considered level (600 hPa for Janicot et al. (2006) and

700 hPa in our case) attributed to the AEJ. The effects occur mainly above the region where changes have been made on vegetation cover. Studies have revealed the sensitivity of the AEJ to the land surface (Sylla et al., 2010). The convection has been linked to African Easterly

Waves (AEWs) which development and maintain is based on the presence of AEJ (Berry and

Thorncroft, 2005; Li et al., 2015). According to Diedhiou et al. (2002) over land the waves in the ITD are mainly located in the neighbourhood of the jet and result mainly from barotropic instability of the jet. AEJ is prior to AEWs development and enhanced convection cross West

Africa and Atlantic at later gaps (Alaka and Maloney, 2012). There is a marked increase in

AEW activity in June. For Thorncroft and Hodges (2001) the increased activity over the land in June is perhaps consistent with the increased solar heating at the surface at this time and the development of a deep well-mixed boundary layer. They also noted that another notable

This equatorward intense activity of the AEWs was even more pronounced in September.

4.5.4 Monsoon fluxes

The zero-isoline of the meridional wind component shown in red is delimitating the monsoon fluxes or easterlies and the northern Sahel winds or harmattan this is the location of the ITD over the continent during the monsoon. The key factor of West African Monsoon rainfall is

122

the position of the ITD and its annual cycle. In addition, the seasonal variation of temperature is associated with the seasonal variation of the ITD. Thus, temperature is at maximum/minimum value during the period when ITD is at its lowest/highest mean position around 5° N/20 22° N (Abatan et al., 2014). Based on experimentation, the ITD was located

(Fig. 4.25a and c) (Fig. 4.25b and d) over

West Africa. This position of the ITD has been observed through many dataset ERA15, regional climate model REMO from the Max-Planck Institute for Meteorology in Hamburg,

CRU (Paeth et al., 2005), ERA40 ( Mekonnen et al., 2006 ), NCEP. The moisture is brought toward the continent during the rainy period by the wind at 925 hPa.

123

Figure 4. 25: JJAS mean Monsoon fluxes at 850 hPa before changes [a] and [c]; and after changes in vegetation cover [b] and [d] in 2005 and 2006.

124

4.5.5 Zonal Wind, Convection and Wind Velocities

4.26a).

These areas are the zones where the changes in vegetation cover seem to affect. The areas

easterlies. The wind movement is dominated by the westerlies over the region during JJAS.

Strong zonal wind is observed when changes are made in vegetation cover. Thus in 2005 after changes 1 m s-1 is widely represented (Fig. 4.26b) when in 2006 it is observed the appearance and persistence of 1.5 m s-1 and 2 m s-1 over the region. Referring to the findings of Diro et al. (2011) who found consistent wind anomaly at 850 mb during deficit rainfall years, our results is prelude to a probable drop of rainfall due to changes in vegetation. Consistent low level wind deals with deficit rainfall.

125

Figure 4. 26: Time latitudinal variability of JJAS zonal wind mean at 850 hPa averaged along vegetation cover (c and d) in 2005 and 2006.

126

Figures 4.27 and 4.28 show the vertical cross-section of JJAS mean wind velocity averaged

- respectively for 2005 and 2006. Negative (positive) values of the velocity correspond to upward (downward) motion (Fontaine et al., 2002). According to the wind vertical speed different phases are noted. The negative wind speed shows convection areas.

Base on the negative velocity location and height, generally three convection levels are observed -

between 700 hPa and 200 hPa. The third

positive wind speed corresponding to subsidence zone in the Sahara desert is observed. The first convection zone correspond the ITD location in May-June before its abrupt jump (Sultan and Janicot, located in the abrupt jump of ITD corresponding to the monsoon onset over Sahel region.

127

Figure 4. 27: Vertical cross- - of 2005 showing convergence and divergence zones a) before the change and b) after the change.

Figure 4. 28: Vertical cross- - of 2006 showing convergence and divergence zones a) before the change and b) after the change.

128

When it comes to the difference between winds velocities before change has been made on vegetation cover (Fig. 4.27a and 4.28a) and after making change on vegetation cover (Fig.

4.27b and Fig. 4.28b), it is not too much perceptible. However, the weakest convection values are always high before the change in vegetation cover.

The ensemble plots in Figure 4.29 show the time latitude diagram of wind velocity at 925 hPa. The progression of the velocity on time over the latitude positive and negative vertical wind speed as the previous. The negative wind represents the ascension zones over the region. But the convection time progression shows continuous ascension before changes in vegetation cover when after changes the convection is discontinuous in time in 2005 (Fig. 4.29a and Fig. 4.29b). And in 2006 the convection is the stronger before the change (Fig. 4.29c and Fig. 4.29d). The change in convection occurs at area where the vegetation has been changed. Strong motion ascension occurs at the end of June till the end of August before the change and after the change the strong ascension occurs late in middle of July and stay only a month. These mean that the change has affected the convection on time. The abrupt jump could be explained by the strongest of the convection.

129

Figure 4. 29: Time latitudinal variability of JJAS wind velocity mean at 925 hPa averaged anges in vegetation cover [c] and [d] in 2005 and 2006.

130

4.5.6 Relative humidity and Convection

Figures 4.30 and 4.31 show the vertical cross-section of relative humidity and wind velocity

- Fig. 4.30a) and 2006 (Fig. 4.30b) for the control. Negative (positive) values of the velocity correspond to upward (downward) motion

(Fontaine et al., in general. Furthermore the moisture is well distributed by the TEJ above 200 hPa. Other

hPa) and poor moisture content. However, specifically the year 2005 was the stronger in term of moisture convection it is note N convection speed about 0.08 m s-1 with 60% of moisture when in 2006 it was 0.06 and 50% of moisture. The change in vegetation cover

It is observed a decrease of the moti wind speed which corresponds to subsidence zone in the Sahara desert is observed. The first convection zone correspond the ITD location in May-June before its abrupt jump (Sultan and

Janicot, D

Sahel region.

131

Figure 4. 30: Vertical cross-section of relative humidity in percentage (shaded) and the wind - 2005 [a] and in JJAS of 2006 [b].

132

When it comes to the difference between winds velocities without change has been made on vegetation cover (Fig. 4.30a and Fig. 4.31a) and after making change on vegetation cover

(Fig. 4.30b and Fig. 4.31b), it is not too much perceptible. However, the weakest convection is always high before the change in vegetation cover. In the next the velocity will be observed in time to see how the change in vegetation cover is affecting it. With the initial land cover, strong motion ascension occurs at the end of June till the end of August. However, after the change the ascension occurs late in middle of July and stay only a month. These mean that the change has affected the convection on time. The abrupt jump could be explained by the strongest of the convection.

133

Figure 4. 31: Vertical cross-section of relative humidity in percentage (shaded) and the wind velocity (contour) after changes in vegetation cover averaged along - 2005 [a] and in JJAS of 2006 [b].

134

4.5.7 Change in Surface Temperature, Evapotranspiration Flux and Albedo

Initially some differences are noted between 2005 and 2006 in term of Evapotranspiration, surface temperature and albedo averaged over the experiment band previous to the change in vegetation cover. It is noted that the surface temperature averaged over JJAS was equal to

er s of evapotranspiration the high value 3.84 mm day -1 is observed in 2005 when in 2006 it is computed 3.60 mm day -1. No

no change in surface albedo between the years. However after making change in vegetation cover some differences are released depending on the year, control 2005 and Sens 2005 and

(Table

4.5). The evapotranspiration has been reduced at about 0.52 mm day -1 when change is made in vegetation cover in 2005 and at about 0.42 mm day -1 in 2006. The change has increased the albedo at about +0.03 in both 2005 and 2006.

Table 4.5: Changes in Evapotranspiration, Temperature and Albedo due to vegetation cover change in JJAS of 2005 and 2006 over changed band.

Evapotranspiration Temperature Albedo Initial difference CTL 2006-2005 -0.24 mm/day 0.6 0

Impact of land Sens-CTL 2005 -0.52 mm/day +0.03 Sens-CTL 2006 -0.45 mm/day +0.03 surface change

135

Therefore, the drop in evapotranspiration is about 0.43 mm day -1 and 0.37 mm day -1 namely

13% and 12.50% respectively in 2005 and 2006 over the area where changes have been made in vegetation cover. This is equivalent to a loss of about 53 mm and 45 mm during the whole season JJAS.

4.5.8 Rainfall Seasonal Variability

The distribution of the intraseasonal variability of the rainfa

Guinea region is studied in this section. It is based on rainfall daily and monthly means covering

JJAS. Figure 4.32 and Figure 4.33 show daily and monthly time latitude variability of the rainfall over West Africa in 2005 and 2006 respectively. Northward progression of the ITD is

the end of June in 2005.

Contrary to 2005, in 2006 the rapid jump of the ITD occurs later in July almost at the end of

July (Mounkaila et al., 2014) then, in September starts a southward displacement of the ITD movement. An important feature for the monsoon and ITD progression is the Saharan heat low (Janicot et al., 2008).

The different phases of the monsoon within the period JJAS are well captured by the

in Sahel. In 2005 the abrupt shift of the monsoon occurs early at the end of June contrary to

2006 where it occurs in July. The change in vegetation has an impact on monsoon onset it seems to delay the monsoon onset over Sahel for instance in 2005 where it is observed an early monsoon onset in Sahel before making changes in surface state. After the changes the

136

time gap between the end of monsoon in Guinea region and its onset in Sahel is considerable more than one months in 2005 and around one month in 2006. In this changes are not clearly seen in the daily data, there are clearly observed in monthly plots. At the end of September there is a southwards trend of the rainfall displacement this is the beginning of the second season of the rainfall over Guinea region. The change in surface state over Guinea region is going to affect the rainfall seasonal variability over Sahel region by delaying the onset of rainfall over there. This finding is in agreement with some previous studies of Charney

(1975). The African Sahel is a transition zone between the arid Sahara Desert and the more humid Gulf of Guinea registering about 80% of annual rainfall between June and September during the south-west monsoon season (Monerie et al., 2013).

137

Figure 4. 32: Time latitudinal diagram of daily and monthly mean rainfall (mm day -1) - b) and d) with change in surface in JJAS 2005.

138

The change in land surface cover will modify the surface albedo by increasing it. This finding is sustained by numerous previous studies realised over Sahel (Charney, 1975). Fuller and

Ottke (2002) at the end of their study on land cover, rainfall and land surface albedo in West

Africa they conclude that albedo and rainfall are related only modestly at short time scales

(monthly and annual). This could explain the observed changes in the monsoon onset over

Sahel.

139

Figure 4. 33: Time latitudinal diagram of daily and monthly mean rainfall (mm day -1) - b) and d) with change in surface in JJAS 2006.

140

The spatial distribution of the rainfall biases computed between the two surface states (before and after changes) is shown in Figure 4.34. In 2005 dry biases are spread out all over the

Fig. 4.34a).

The phenomenon is slightly different in 2006 with an accentuation of dry biases in region

N and wet biases over Guinea region (Fig. 4.34b). So in general, strong dry biases are observed over sub Sudanian

gCM4 used for this study supports previous findings which linked the drought in Sahel to the changes in land surface state over Guinea region.

Figure 4. 34: JJAS rainfall biases between the experiment without changes in vegetation cover and the experiment with changes in vegetation cover in 2005 [a] and 2006 [b].

141

So in general, strong dry biases are observed over sub Sudanian band. The changes in

el zones

previous findings which linked the drought in Sahel to the changes in land surface state over

Guinea region (Charney, 1975). As the sub Sudanian area is the most affected in term of

rainfall biases distribution, in Figure 4.35 based on zonal averaged rainfall time series over

Sudano-Guinea - - es on time. Thus, in 2005 (Fig. 4.35a) after making changes in vegetation cover the rainfall have dropped about 0.42 mm day -1 and 0.23 mm day -1 in 2006 respectively 5% and 3% of the rainfall over the region (Fig. 4.35b). These results are similar with the grass scenario applied by Salih et al. (2013) over Republic of Sudan.

142

Figure 4. 35: Rainfall averaged over sub Sudanian band before changes (CTL) and after changes (Sens) in vegetation cover in 2005 [a] and in 2006 [b].

143

During the whole monsoon period, the rainfall has dropped over the region probably caused by the changes in vegetation cover which in turn induced some changes in surface albedo thus in the energy budget. It is observed that before the monsoon surge there was no difference between the rainfall before and after changes in vegetation cover. However, the difference starts with the monsoon onset. So as soon as the rainfall starts to go beyond 8 mm per day the impact of the changes on vegetation cover start to appear in August with the decrease in rainfall amount after making changes in vegetation cover. Using a short range of time the two experimentations clearly depict the impact of vegetation cover change on West

Africa Monsoon thus in rainfall. This finding is consistent with previous studies linking drought in this area to the changes in vegetation cover over Sudano-Guinea regions of West

Africa. Charney (1975) found about 40% drop in rainfall over Sahel region due to the change in vegetation cover.

144

Figure 4. 36: Rainfall averaged over sub Sudanian band before changes and after changes in vegetation cover in 2005 and 2006.

145

The 0.52 mm day -1 and 0.42 mm day -1 decrease in evapotranspiration observed for respectively 2005 and 2006 seems to be low compared to those found by some authors over the region. For instance, these values are not as high as those of Abiodun et al. (2008) who found a uniform decrease in Evapotranspiration about 2 mm day -1 which contributes to decrease in rainfall about 20%. This difference should have been caused by the assumption made by them. However, our experimentation area was covering only the transition zone between Guinea and Sudanian regions which is widely less than their one covering the entire guinea region. In the other hand, two zones are identified with positive differences these are

Guinea Land evapotranspiration (ET) is a key component of the coupling between the land surface and the atmosphere. This contributes to the decrease in the rainfall observed over Sudano-Guinea and

Sudanian Also the change has increased the albedo at about +0.03 in both 2005 and 2006.

146

Chapter 5

CONCLUSION AND RECOMMANDATIONS

5.1 Conclusion

Decadal variability of the rainfall and the vegetation over West Africa is revisited from 1981 to 2012 using CRU, station observation rainfall data and NDVI from NOAA. From decade

80s to 90s, significant return to wet condition was observed over West Africa, this was sustained during decade 00s except over Central Benin and all the western side of Nigeria where there were observed decreases in annual rainfall magnitudes. From decades 80s to 90s, a re-greening of the Central Sahel and Sudano-Sahel regions was also observed. From decade

90s to 00s, this re-greening belt was observed extended to the South and the Coastal areas, mainly over the Guinea Coast, Sudano-Guinea and Western Sahel regions. Over the Sahel, observed changes in the rainfall pattern are mainly changes in magnitude during the core of rainy season (July, August and September) and length of the rainy season starting sooner during the two last decades. Over the Sudanian region, observed changes in the rainfall pattern are also in the magnitude during the peak of the rainy season and a shift of the rainy period (JJAS - JASO), starting and ending later during the two last decades. Over the Guinea

Coast, the changes were observed mainly during the little rainy season which becomes more intense in magnitude and longer in duration during the last two decades. The NDVI intra- annual variability shows generally the same evolution pattern compared to the rainfall, but during the last two decades, significant NDVI values are found one to two months after the end of the rainy season over the entire region.

147

Correlations between rainfall and NDVI were significant over the Sahel, Sudan and northern part of Guinea Coast, but they become weaker in magnitude Guinea Coast from decade 80s to

00s meaning that in wetter conditions, there is no linear relationship between NDVI and rainfall over this region.

It is quite clear from the result of this study that there is recovery of rainfall over some part of

West African region after the long drought period. The increasing tendency observed in vegetation greenness is moving from the South to North. Although for the decade 90s the re- greening process was mainly below latitude 10 N however, in decade 00s it has significantly reached After the severe drought of the 80s, the vegetation has been significantly re- greened over the Sudanian increase in vegetation over Sudanian region during this period could be due to the rainfall recovery which has started in decade 90s and may have induced some vegetation cover recovery. This is however not the case in Guinea region where there were no changes were observed during this period. The comparison between 00s and 90s shows the persistence of significant positive changes in the coastal region

The RegCM4 model has been able to simulate the early onset of monsoon in 2005 and the late onset in 2006. It also captured the impact of vegetation cover change on rainfall spatial and temporal distribution over the West African region. The model results has also shown that the change in vegetation cover during the peak of the rainy season in the Sahel (JJAS) does not have a clear effect on synoptic dynamic patterns like TEJ, AEJ and monsoon fluxes. Also the mean position of ITD is not affected. However, what has changed is the time lag between the start of rains in the Guinea and Sudan savannah, which was observed to be about a month without vegetation cover and a month and a half with vegetation cover. The reliability of this study resides in the fact that its findings are coherent with some previous findings on this topic.

148

The results of the study have shown the impact of deforestation over the West African savannah zone on rainfall spatial and temporal variability and provided maps of rainfall and vegetation index variability that can guide decision making for policy makers to prevent further deforestation and soil degradation.

5.2 Recommendations

Despite the fact that some important results have been found on vegetation evolution over these last decades over West Africa and the impact of vegetation cover change on rainfall spatio-temporal variability, some gaps still remain. So as recommendation to improve further understanding, (1) other parameters such as the Leaf Area Index and vegetation types should be taken into account in order to characterize the vegetation types. Also, above a given rainfall threshold the NDVI is no longer sensitive to rainfall in wet regions. (2) Modeling aspect should be done over a long period of time in order to achieve more accuracy. (3)

Future studies could involve the use of different models to better understand the model strengths and limitations. (4) The policy makers should really care about the destruction of the forest cover over the region because of its impact on rainfall spatio temporal distribution.

5.3 Limitations of the Study

Our research has some limitations based on the fact that the vegetation over Guinea Coast is heterogeneous and the vegetation over Sudan is more homogenous. The link between rainfall and vegetation is still complex to be clearly defined. So it could be better to add more parameters such the Leaf Area Index, the Photosynthetically Active Radiation Fraction

(FPAR) and the soil types. The vegetation type is not taken into account. The unavailability of climate data from stations is real issue over the region. The simulation period covered only three years, this should be extended to more years. Also the model resolution was 50 km x 50 km this could be increased by some downscaling methods in order to fit well with the

149

vegetation indices. The simulation facilities were not accessible on time that had an impact on the length of the simulation.

150

REFERENCES

Abatan, A. A., Abiodun, B.J. and Omotosho, B.J. (2014). On the characteristics of sea breezes over Nigerian coastal region. Theor Appl Climatol, 116:93 102.

Abiodun, B.J., Pal, J.S., Afiesimama, E.A., Gutowski, W.J. and Adedoyin, A. (2008). Simulation of West African monsoon using RegCM3 Part II: impacts of deforestation and desertification. Theor. Appl. Climatol., 93: 245 261.

Adefolalu, D.O. (1986). Rainfall trends in Nigeria. Theor. and Appl. Climat., 37:205-219.

Adedoyin, J.A. (1997). Instability of squall-inducing waves, global sea-surface temperature anomalies and climate change in tropical North Africa. Meteorol. Atmos. Phys., 62: 79- 89.

Adefolalu, D.O. (1985). Further aspects of the Sahelian drought as evident from rainfall regime in Nigeria. Archive of Meteorology, Geophysics and Bioclimatology , 36: 277-295.

Adler, R.F., Gu, G., and Huffman, G.J. (2010). Estimating climatological bias errors for the GPCP, J. Hydrometeorology , submitted.

Alaka, G.J.J.R. and Maloney, E.D. (2012). The Influence of the MJO on Upstream Precursors to African Easterly Waves. J. Climate, 25: 3219-3235.

Ali, A., Lebel, T. and Abou, A. (2008). sahel Sécheresse. Sciences et changement planétaires/ Sécheresse , 19 (4):227-35.

Albignat, J.P. and Reed, R.J. (1980). The Origin of African Wave Disturbances during Phase III of GATE. Mon. Wea. Rev. , 108, 1827 1839.

Anyamba, A. and Tucker, C.J. (2005). Analysis of Sahelian vegetation dynamics using NOAA-AVHRR NDVI data from 1981 2003. Journal of Arid Environments , 63:596 614.

Bagayoko, F. (2006). Impact of land use intensity on evaporation and surface runoff: Processes and parameters for eastern Burkina Faso, West Africa. Cuvillier Verlag, 2006, Vol. 40, Ecology and Development Series, West Africa, Bonn, Germany.

Balas, N., Nicholson, S.E. and Klotter, D. (2007). The relationship of rainfall variability in West Central Africa to sea-surface temperature fluctuations. Int. J. Climatol., 27: 1335 1349. DOI: 10.1002/joc.1456.

(2012). Effects of coastal topography on climate: high-resolution simulation with a regional climate model. Clim. Res., Vol. 52: 159 174.

151

Barry, R.G. and Chambers, R.E. (1966). A preliminary map of summer albedo over England and Wales . Q. J. Roy . Meteor . Soc., 92: 543-548.

Barry, R.G. and Chorley, R.J. (2003). Atmosphere, Weather and Climate. Routledge , (eds) 8.

Bates, B.C., Kundzewicz, Z.W., Wu, S. and Palutikof, J.P. (2008) Climate Change and Water. (Eds) Technical Paper of the Intergovernmental Panel on Climate Change, IPCC Secretariat, Geneva, 210p.

Bertrand, A. (1983). La déforestation en zone de forêt en Côte d'Ivoire. Rev. Bois et Forêt des Tropiques 202:3-17.

Berry, G.J., and Thorncroft, C.D. (2005). Case study of an intense African easterly wave. Mon. Wea. Rev., 133: 752 766.

Bliefernicht, J., Hamann, I., Kunstmann, H., Legutke, S., Laux, P., Wagner, S., Rhmmler, T., Omotosho, J., Gochis, D., Sarr, A., Konaré, A., Lamptey, B., Annor, T., Waongo, M., Hingerl, L. and further members of the WASCAL team (2012). DKRZ Project 726 - Regional Climate Simulations and Land-Atmosphere Simulations for West Africa, Poster.

Boateng, P.K. (2013). Agricultural Production, Land-use/cover Change and the Desertification Debate in the West African Savannah: An Adapted Political Ecology Approach. Journal of Arts and Humanities , Vol. 2, No. 8: 21-35.

Boko, M., Niang, I., Nyong, A., Vogel, C., Githeko, A., Medany, M., Osman-Elasha, B., Tabo, R. and Yanda, P. (2007). Africa Climate Change 2007: Impacts, Adaptation and Vulnerability, Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Parry, M.L., Canziani, O.F., Palutikof, J. P, van der Linden, P.J. and Hanson, C.E. (eds) Cambridge University Press . Cambridge UK. pp433 467.

Bonan, G.B. (2008). Ecological Climatology, Cambridge Univ. Press, Cambridge, (eds). 2.

Browning, K.A. and Roberts, N.M. (1994). Use of satellite leading to frontal imagery to diagnose events thunderstorms: part I of a case study, Met Apps 1: 303-310.

Browning, K.A. and Roberts, N.M. (1995). Use of satellite leading to frontal imagery to diagnose events thunderstorms: part II of a case study, Met Apps 2: 3-9.

Burpee, R.W. (1972). The origin and structure of easterly waves in the lower troposphere of North Africa. J. Atmos. Sci., 29: 716-726.

Camilla, T. (2009). Climate change in Africa (African Arguments). Zed Books Ltd . 15-31, 50- 69: 145-152.

152

Carter, T.R., Parry, M.L., Harasawa, H. and Nishioka, S. (1994). Intergovernmental panel on climate change technical guidelines for assessing climate change impacts and adaptation. University College London and Center for Global Environmental Research, London.

Charney, J.G. (1975). Dynamics of deserts and drought in the Sahel. Q. J. Roy. Meteorol . Soc. 101:193 202.

Che, M., Chen, B., Innes, J.L., Wang, G., Dou, X., Zhou, T., Zhang,, H., Yan, J., Xu, G. and Zhao, H. (2014). Spatial and temporal variations in the end date of the vegetation growing season throughout the Qinghai Tibetan Plateau from 1982 to 2011. Agricultural and Forest Meteorology , 189 190, 81 90.

Chen, Y.-L. and Ogura, Y. (1982). Modulation of convective activity by large-scale flow patterns observed in GATE, J. Atmos. Sci., 39: 1260 1279.

Chen, X., Yang, D., Chen, J. and Cao, X. (2015). An improved automated land cover updating approach by integrating with downscaled NDVI time series data. Remote Sensing Letters , 6:1, 29-38.

Cook, K.H. (1999). Generation of the African Easterly Jet and Its Role in Determining West African Precipitation. J. Climate , 12: 1164-1184.

Cracknell, A.P. (2001). The exciting and totally unanticipated success of the AVHRR in applications for which it was never intended. Advances in Space Research , 28: 233 240.

Dai, A., Lamb, P.J., Trenberth, K.E., Hulme, M., Jones, P.D. and Xie, P. (2004). Comment the recent Sahel drought is real. Int. J. Climatol., 24: 1323 1331.

Dai, A. (2011). Characteristics and trends in various forms of the Palmer Drought Severity Index during 1900 2008. J. of Geophysical Research , Vol. 116, D12115, doi: 10.1029/2010JD015541.

Dardel, C., Kergoat, L., Mougin, E., Grippa, M. and Tucker, C.J. (2014). Re-greening Sahel: 30 years of remote sensing data and field observations (Mali, Niger). Remote Sensing of Environement, 140: 350-364.

Diallo, I., Bain, C.L., Gaye, A.T., Moufouma-Okia, W., Niang, C., Dieng, D.B. M. and Graham, R. (2014). Simulation of West Arican monsoon onset using the HadGEM3-RA regional climate model. Clim. Dyn. (accepted).

Diawara, A., Yoroba, F., Kouadio, K.Y., Kouassi, K.B., Assamoi, E.M., Diedhiou, A., and Assamoi, P. (2014). Climate Variability in the Sudano-Guinea Transition Area and Its Impact on Vegetation: The Case of th Advances in Meteorology , Article ID 831414, 11p.

153

Dickinson, R.E., Henderson-Sellers, A., Kennedy, P.J. and Wilson, M.F. (1986). Biosphere Atmosphere Transfer Scheme (BATS) for the NCAR Community Climate Model. NCAR Technical Note , NCAR, TN275+STR, 69p.

Dickinson, R.E., Errico, R.M., Giorgi, F. and Bates, G.T. (1989). A regional climate model for the Western United States. Clim Change , .

Dickinson, R.E., Henderson-Sellers, A., and Kennedy, P.J. (1993). Biosphere-Atmosphere Transfer Scheme (BATS) version 1E as coupled to the NCAR Community Climate Model. NCAR Tech. rep. TN-387+STR, 72p.

Dickinson, R.E. (2003). Land Atmosphere interactions, Georgia Institute of Technology, Atlanta, GA, USA, Elsevier Science Ltd .

Diedhiou, A., Janicot, S., Viltard, A. and P. de Felice, (2002). Energetics of easterly wave disturbances over West Africa and the tropical Atlantic: a Climatology from 1979-95 NCEP/NCAR reanalyses. Clim. Dyn., 18: 487-500.

Diedhiou, A., Janicot, S., Viltard, A., P. de Felice and Laurent, H. (1999). Easterly wave regimes and associated convection over West Africa and tropical Atlantic: results from the NCEP /NCAR and ECMWF reanalyses. Clim. Dyn., 15: 795-822.

Dietz, A.J., Ruben, R. and Verhagen, A. (2004). The Impact of Climate Change on Drylands: With a focus on West Africa. (eds) Environment & policy, ISSN 1383-5130, vol. 39, 465p.

Diro, G.T., Grimes, D.I.F. and Black, E. (2011). Large scale features affecting Ethiopian rainfall, in Williams, C.J.R. and Kniveton, D.R. (eds) African Climate and Climate Change. Adv. Global Change Res., vol. 43: 13 50.

Eischeid, J.K., Diaz, H.F., Bradley, R.S., and Jones, P.D. (1991). A comprehensive precipitation dataset for global land areas. DOE/ER-6901T-H1, U.S. Department of Energy, Washington, DC, [Available from National Technical Information Service, U.S. Dept. of Commerce, Springfield, VA 22161] 81p.

Eklundh, L. and Olsson, L. (2003). Vegetation index trends for the African Sahel 1982 1999, Geophysical Research Letters , Vol. 30, NO. 8, 1430, doi: 10.1029/ 2002GL016772.

Eltahir, E.A.B. and Gong, C. (1996). Dynamics of wet and dry years in West Africa, J. Climate , 9(5): 1030 1042.

Evan, A. T., Flamant, C., Lavaysse, C., Kocha, C. and Saci, A. (2015). Water Vapor Forced Greenhouse Warming over the Sahara Desert and the Recent Recovery from the Sahelian Drought, J. Climate , 28: 108-123.

154

FAO (2000). In: The Challenges of Sustainable Forestry Development in Africa: Twenty-First FAO Regional Conference for Africa, Yaounde, Cameroon, February 21 25. Food and Agriculture Organization of the United Nations, Rome, Italy.

Farnswort, A., White, E., Williams, C.J.R., Black, E. and Kniveton, R.D. (2011). Understanding the Large Scale Driving Mechanisms of Rainfall Variability over Central Africa. Advances in Global Change Research , 43: 101-122.

Fensholt, R. and Rasmussen, K. (2011). Analysis of trends in the Sahelian 'rain use efficiency' using GIMMS NDVI, RFE and GPCP rainfall data. Remote Sensing of Environment , 115: 438 451.

Fontaine, B., Philippon, N., Trzaska, S., and Roucou, P. (2002). Spring to summer changes in the West African monsoon through NCEP/NCAR reanalyses (1968 1998). Journal of Geophysical research, Vol. 107, NO. 0,10.1029/2001JD000834.

Fritsch, J.M. and Chappell, C.F. (1980). Numerical Prediction of Convectively Driven Mesoscale Pressure Systems, Part I: Convective Parameterization. J. Atmos. Sci., 37, 1722 1733.

Fuller, S.P. and Rouse, W.R. (1979). Spectral reflectance changes accompanying a post-fire recovery sequence in a subarctic spruce lichen woodland. Remote Sensing of Environment , 8: 11-23.

Fuller, D.O. and Ottke, C. (2002). Land cover, rainfall and land-surface albedo in West Africa. Climatic Change , 54: 181 204.

Giannini, A., Saravanan, R. And Chang, P. (2003). Oceanic forcing of Sahel rainfall on interannual to interdecadal time scales. Science , 302: 1027-1030.

Giorgi, F. (1990). Simulation of regional climate using a limited area model nested in a general circulation model. J. Climate , .

Giorgi, F., Bates, G.T. and Nieman, S. (1993). The multi-year surface climatology of a regional atmospheric model over the Western United States. J. Climate , .

Giorgi, F. and Mearns, L.O. (1999). Introduction to special section: regional climate modeling revisited. J. Geophys Res., 104: .

Giorgi, F., Pal, J.S., Bi, X. Q., Sloan, L.C., Elguindi, N. and Solmon, F. (2006). Introduction to the TAC special issue: the RegCNET network. Theor Appl Climatol , .

Giorgi, F., Coppola, E., Solmon, F., Mariotti, L., Sylla, M.B, Bi, X., Elguindi, N., Diro, G.T., Nair, V., Giuliani, G., Turuncoglu, U.U., Cozzini, S., Güttler, I., , T.A., Tawfik, A.B., Shalaby, A., Zakey, A.S., Steiner, A.L., Stordal, F., Sloan, L.C. and Brankovic, C. (2012).

155

RegCM4: model description and preliminary tests over multiple CORDEX domains, 52: 7 29.

Giorgi, F., Nellie, E., Stefano, C. and Graziano, G. (2014). Regional Climatic Model RegCM

Gosset, M., Zahiri, E.P. and Moumouni, S. (2010). Rain drop size distribution variability and impact on X-band polarimetric radar retrieval: Results from the AMMA campaign in Benin. Q.J.R. Meteorol. Soc., 136(s1): 243 256.

Gray, T.I. and McCrary, D.G. (1981). Meteorological satellite data a tool to describe the -NI-04042, Johnson Space Center, Houston, Texas, 7p.

Grell, G., Dudhia, J. and Stauffer, D.R. (1994). A description of the fifth generation Penn State/NCAR Mesoscale Model (MM5) National Center for Atmospheric Research Tech Note NCAR/TN-398 + STR, NCAR, Boulder, CO, 121p.

Grist, J.P. and Nicholson, S.E. (2001). A study of the dynamic factors influencing the variability of rainfall in the West African Sahel, J. Climate, 14: 1337 1359.

Hagos, S.M. and Cook, K.H. (2007). Dynamics of the West African Monsoon Jump, J. Climate, 20: 5264-5284.

Hamatan, M., Mahe, G., Servat, E., Paturel, J.E. and Amani, A. (2004). Synthèse et évaluation des prévisions saisonnières e , Note de recherche, Sécheresse, 15 (3) : 279-86.

Harris, I., Jones, P.D., Osborn, T.J. and Lister, D.H. (2014). Updated high-resolution grids of monthly climatic observations the CRU TS3.10 Dataset. Int. J. Climatol., 34: 623 642.

Henke, D., Smyth, P., Haffke, C. and Magnusdottir, G. (2012). Automated analysis of the temporal behavior of the double Intertropical Convergence Zone over the east Pacific, Remote Sensing of Environment , 123: 418 433.

Hernandez, M. , Miller, S.N., Goodrich, D.C., Goff, B.F., Kepner, W.G., Edmonds, CM. and Jones, K.B. (2000). Modeling runoff response to land cover and rainfall spatial variability in semi-arid watersheds, Environmental Monitoring and Assessment, 64: 285 298.

Herrmann, S.M., Anyamba, A. and Tucker, C.J. (2005). Recent trends in vegetation dynamics in the African Sahel and their relationship to climate, Global Environmental Change, 15:394 404.

Holben, B.N. (1986). Characteristics of maximum-value composite images from temporal AVHRR data. Int. J. Remote Sensing , 7: 1417 1434.

156

Hua, T., Wang, X. M., Zhang, C. X. and Lang, L. (2013). Temporal and spatial variations in the Palmer Drought Severity Index over the past four centuries in arid, semiarid, and semi humid East Asia, Chinese science Bulletin , Vol.58 No.33: 4143-4152.

Hulme, M. (1994). Global changes in precipitation in the instrumental period. In Desbois, M. and Desalmand, F. (eds) Global Precipitation and Climate Change, Springer-Verlag, 387 405.

Hulme, M. and Kelly, P.M. (1997). links between desertification and climate Chapter 15, In Owen, L. and Unwin, T.B.H. (eds.) Environmental management: readings and case studies, Blackwell, Oxford, UK, 448p.

Imo, J.E. and Ekpenyong, N. (2011). Extreme Climatic Variability in North-western Nigeria: An Analysis of Rainfall Trends and Patterns, Journal of Geography and Geology, 3: 51- 62.

IPCC (2007). Climate change 2007. The Physical Science Basis, Contribution of working Group I to the fourth assessment Report of the Intergovernmental Panel on Climate Change, edited by S. Solomon, J.M., Dyurgerov, and R.S.W. Van de Wal (2007 ) Cambridge Univ. Press, Cambridge, U. K., 996 pp .

IPCC (2013). Climate change physical basis, working group I contribution to the fifth assessment report of the intergovernmental panel on climate change. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 1535pp.

Jamali, S., Jönsson, P., Eklundh, L., Ardö, J. and Seaquist, J. (2015). Detecting changes in vegetation trends using time series segmentation. Remote Sensing of Environment , 156: 182 195.

James, R.H. and Gregory, J.H. (2012). An Introduction to Dynamic Meteorology. Academic Press, vol.88 of International Geophysics (eds) 5, 552pp.

Janicot, S., Thorncroft, C.D., Ali, A., Asencio, N., Berry, G., Bock, O., Bourles, B., Caniaux, G., Chauvin, F., Deme, A., Kergoat, L., Lafore, J.-P., Lavaysse, C., Lebel, T., Marticorena, B., Mounier, F., Nedelec, P., Redelsperger, J.-L., Ravegnani, F., Reeves, C. E., Roca, R., P. de Rosnay, Schlager, H., Sultan, B., Tomasini, M., Ulanovsky, A. and ACMAD forecasters team (2008). Large-scale overview of the summer monsoon over West Africa during the AMMA field experiment in 2006, Ann. Geophys., 26, 2569 2595.

Jones, P.D., New, M., Parker, D.E., Martin, S. and Rigor, I.G. (1999). Surface air temperature and its variations over the last 150 years, Reviews of Geophysics, 37: 173-199.

Kim, D.H., Sexton, J.O., Noojipady, P., Huang, C., Anand, A., Channan, S., Feng, M. and Townshend, J.R. (2014). Global, Landsat-based forest-cover change from 1990 to 2000. Remote sensing of Environment , 155: 178-193.

157

Kinne, S., Schulz, M., Textor, C., Guibert, S., Balkanski, Y., Bauer, S.E., Berntsen, T., Berglen, T.F., Boucher, O., Chi, M., Collins, W., Dentener, F., Diehl T., Easter R., Feichter, J., Fillmore, D., Ghan, S., Ginoux, P., Gong, S., Grini, A., Hendricks, J., Herzog, M., Horowitz, L., Isaksen, I., Iversen, T., Kirkevag, A., Kloster, S., Koch, D., Kristjansson, J. E., Krol, M., Lauer, A., Lamarque, J.F., Lesins, G., Li, X., Lohmann, U., Montanaro, V., Myhre, G., Penner, J.E., Pitari, G., Reddy, S., Seland, O., Stier, P., Takemura, T., and Tie, X. (2006). An AeroCom initial assessment optical properties in aerosol component modules of global models. Atmos. Chem. Phys ., 6: 1815 1834.

Kinne, S. (2009). Climatologies of Cloudrelated Aerosols Part 1: Particle Number and Size From the Strüngmann Forum Report. In, Jost Heintzenberg and Robert J. Charlson (eds) Clouds in the Perturbed Climate System: Their Relationship to Energy Balance, Atmospheric Dynamics, and Precipitation, MIT Press ISBN 978-0-262-01287-4.

Klein, C., Heinzeller, D., Bliefernicht, J. and Kunstmann, H. (2015). Variability of West African monsoon patterns generated by a WRF multi-physics ensemble, Clim. Dyn. , 23pp, DOI 10.1007/s00382-015-2505-5.

Konaté, S. and Kampmann, D. (2012). Biodiversity Atlas of West Africa, (eds) vol. 3: Cote , Abidjan and Frankfurt/Main, pp: 44-47.

Kpongor, D. (2007). Spatially explicit modeling of sorghum (Sorghum bicolor L) production on complex terrain of a semi-arid region of Ghana using APSIM. PhD thesis, Ecology and Development Series 51. Center for Development Research (ZEF), University of Bonn. Bonn, Germany, 132pp.

Kriebel, K.T. (1979). Albedo of vegetated surfaces: Its variability with differing irradiances . Remote Sensing of the Environment , 8: 283-290.

Kukla, G.J. and Robinson, D. (1980). Annual cycle of surface albedo . Mon , Wea Rev ., 108: 56-67.

Lamb, P.J. (1980). Sahelian Drought, New Zealand, Journal of Geography , 68:12 16.

Laux, P., Kunstmann, H. and Bárdossy, A. (2007). with atmospheric circulation patterns, Proceedings of Symposium HS2004 at IUGG2007, Perugia, IAHS Publ., 313pp.

Le Barbe, L., Lebel, T. and Tapsoba, D. (2002). Rainfall Variability in West Africa during the Years 1950 90, J. Climate, 15:187-202.

Lebel, T., Diedhiou, A. and Laurent, H. (2003). Seasonal cycle and interannual variability of the Sahelian rainfall at hydrological scales, Journal of Geophysical Research, 108: 8389 8392.

158

Lebel, T. and Ali, A. (2009). Recent trends in the Central and Western Sahel rainfall regime (1990-2007). Journal of Hydrology , 375 (1 2): 52 64.

Lebel, T., Parker, D.J., Flamant, C., Bourles, B., Marticorena, B., Mougin, E., Peugeot, C., Diedhiou, A., Haywood, J. M., Ngamini, J.B., Polcher, J., Redelsperger and Thorncroft C. D. (2010). The AMMA field campaigns: Multiscale and multidisciplinary observations in the West African region. Q. J. R. Meteorol. Soc., 136(s1): 8 33.

Li, R., Jin, J., Wang, S. Y. and Gillies, R.R. (2015). Significant impacts of radiation physics in the Weather Research and Forecasting model on the precipitation and dynamics of the West African Monsoon, Clim. Dyn., 44:1583 1594.

Lim, C. and Kafatos, M. (2002). Frequency analysis of natural vegetation distribution using NDVI/AVHRR data from 1981 to 2000 for North America: Correlations with SOI. Int. J. Remote Sensing, 23:17, 3347-3383.

Louvet, S., Paturel, J.E., Mahé, G., Rouché, N. and Koité, M. (2015). Comparison of the spatiotemporal variability of rainfall from four different interpolation methods and impact on the result of GR2M hydrological modeling case of Bani River in Mali, West Africa. Theor. Appl. Climatol , DOI 10.1007/s00704-014-1357-y.

Mangiarotti, S., Mazzega, P., Hiernaux, P. and Mougin, E. (2012). Predictability of vegetation cycles over the semi-arid region of Gourma (Mali) from forecasts of AVHRR- NDVI signals. Remote Sensing of Environment , Elsevier, 123: 246-257.

Marx, A., Kunstmann, H., Schuttemeyer, D. and Moene, A.F. (2008). Uncertainty analysis for satellite derived sensible heat fluxes and scintillometer measurements over Savannah environment and comparison to mesoscale meteorological simulation results. Agricultural and forest meteorology, 148: 656 667.

Mathon, V. and Laurent, H. (2001). Life cycle of Sahelian mesoscale convective cloud systems. Q. J. R. Meteorol. Soc. , 127: 377 406.

Matthews, E. (1983). Global vegetation and land use: New high-resolution data bases for climate studies, J. Clim. Appl. Meteorol. , 22: 474-487.

McGuffie, K. (1994). Remote Sensing Input to Climatological Datasets. In, Robin, A. Vaughan, and Arthur, P. Cracknell (eds) Remote Sensing and Global Climate Change, NATO ASI Series. Vol. I, 24pp.

McKee, T. B., Doesken, N. J. and Kleist, J. (1993). The relationship of drought frequency and duration to time scales. Preprints, Eighth Conf. on Applied Climatology, Anaheim, CA, Amer. Meteor. Soc., 179 184.

McKee, T. B., Doesken, N. J. and Kleist, J. (1995). Drought monitoring with multiple time scales. Proc. Ninth Conf. on Applied Climatology, Dallas, TX, Amer. Meteor. Soc., 233 236.

159

Mekonnen, A., Thorncroft, C.D. and Hall, N.M.J. (2006). Analysis of convection and its association with African easterly waves, J. Climate , 19: 5405-5421.

Mito, C.O., Boiyo, R.K. and Laneve, G. (2012). A simple algorithm to estimate sensible heat flux from remotely sensed MODIS data. Int. J. Remote Sensing , 33:19, 6109-6121.

Monerie, P.-A., Roucou, P. and Fontaine, B. (2013). Mid-century effects of Climate Change on African monsoon dynamics using the A1B emission scenario. Int. J. Climatol , 33: 881 896.

Monteny, B.R. (1986). Importance of the tropical rain forest as an atmospheric moisture source, Proc. ISLSCP Conference, Rome, Italy 2-6 December 1985, ESA SP-248:449-454.

Mosley, P.M. and McKerchar, A.I. (1993). Streamflow. In Maidment D.R. (eds) Handbook of Hydrology (p. 8.1), McGraw-Hill, Inc.

Mounkaila, M.S., Abiodun, B.J. and Omotosho, J.B. (2014). Assessing the capability of CORDEX models in simulating onset of rainfall in West Africa. Theor. Appl. Climatol., 18pp.

Myneni, R.B. and Williams, D.L. (1994). On the relationships between FAPAR and NDVI. Remote Sensing of Environment , 49: 200 211.

Myneni, R.B., Hall, F.G., Sellers, P.J. and Marshak, A.L. (1995). The interpretation of spectral vegetation indexes. IEEE Transactions on Geoscience and Remote Sensing, 33: 481 486.

Monteith, J.L. (1959). The reflection of short-wave radiation by vegetation. Q. J. R. Meteor. Soc., 85: 386-392.

New, M., Hulme, M. And Jones, P. (1999). Representing Twentieth-Century Space Time Climate Variability, Part I: Development of a 1961 90 Mean Monthly Terrestrial Climatology. J. Climate, 12: 829-856.

New, M., Hulme, M. and Jones, P. (2000). Representing Twentieth-Century Space Time Climate Variability. Part II: Development of 1901 96 Monthly Grids of Terrestrial Surface Climate. J. Climate , 13: 2217-2238.

Nicholson, S.E. (1993). Sahel, West Africa, Encyclopedia of Environmental biology. Vol. 3, Academic Press, Inc.

Nicholson, S.E., Devenport, M.L. and Malo, R. (1990). A comparison of vegetation response to rainfall in Sahel and East Africa using Normalised Difference vegetation index from NOAA AVHRR. Climatic Change , 17: 209-241.

160

Nicholson, S.E. (1993). An overview of African rainfall fluctuations of the last decade. J. climate , 6:1463-1466.

Nicholson, S.E., Tucker, C.J. and Ba, M.B. (1998). Desertification, drought, and surface vegetation: an example from the West African Sahel. Bulletin of the American Meteorological Society , 79: 815 829.

Nicholson, S.E. (2005). On the question of the recovery of the rains in the West Africa Sahel. Journal of Arid Environments , 63: 615-641.

Nicholson, S.E. (2009). over West Africa. Clim. Dyn., 32:1155 1171.

Nikulin, G., Jones, C., Giorgi, F., Asrar, G., Büchner, M., Cerezo-Mota, R., Bøssing, O. Christensen, Déqué, M., Fernandez, J., Hänsler, A., Meijgaard, E., Samuelsson, P., Sylla, M.B., and Sushama, L.I. (2012). Precipitation climatology in an ensemble of CORDEX Africa regional climate simulations. J. Climate, 25:6057 6078.

Norman, A.P. (1956). The general circulation of the atmosphere: A numerical experiment. Q. J. R. Meteorological Soc., 82: 123-164.

Odekunle, T.O. and Eludoyin, A.O. (2008). Sea surface temperature patterns in the Gulf of Guinea: their implications for the spatio-temporal variability of precipitation in West Africa. Int. J. Climatol., 28: 1507 1517.

Odekunle, T.O. (2010). An Assessment of the Influence of the Inter-Tropical Discontinuity on Inter-Annual Rainfall Characteristics in Nigeria. Geographical Research, 48(3):314 326.

Ogungbenro, S.B., Ajayi, V.O. and Adefolalu, D.O. (2015). Mean state and kinematic properties of mesoscale convective systems over West Africa. Theor Appl Climatol , 9p.

Oguntoyinbo, J.S. (1970). Reflection coefficient of natural vegetation, crops and urban surfaces in Nigeria, Q. J. Roy. Meteor. Soc., 96: 430-441 .

Olson, J.S., Watts, J.A. and Allison, L.J. (1983). Carbon in live vegetation of major world ecosystems. U.S. Dep. Energy, DOE/NBB-0037, No. TROO4, 164p.

Omotosho, J.B. (1990). Onset of thunderstorms and precipitation over Northern Nigeria. Int. J. Climatol., 10:849-860.

Omotosho, J.B. (1992). long-range prediction of the onset and end of the rainy season in the west african sahel. Int. J. Climatol., vol. 12: 369-382.

161

Omotosho, J.B., Balogun, A.A. and Ogunjobi, K. (2000). Predicting monthly and seasonal rainfall, onset and Cessation of the rainy season in West Africa using only surface data. Int. J. Climatol., 20: 865 880.

Omotosho, J.B. (2008). Pre-rainy season moisture build-up and storm precipitation delivery in the West African Sahel. Int. J. Climatol., 28:937 946.

Ouédraogo, E. (2004). Soil Quality Improvement for Crop Production in semi-arid West Africa. Wageningen, the Neederlands: Wageningen University, Tropical Resource Management 51p.

Oueslati, B. and Bellon, G. (2015). The double ITCZ bias in CMIP5 models: interaction between SST, large scale circulation and precipitation. Clim. Dyn ., 44:585 607.

Paeth, H., Born, K., Podzun, R. and Jacob, D. (2005). Regional dynamical downscaling over West Africa: model evaluation and comparison of wet and dry years. Meteorologische Zeitschrift , Vol. 14, No. 3: 349-367.

Pal, J.S., Small, E. and Eltahir, E. (2000). Simulation of regional-scale water and energy budgets: representation of subgrid cloud and precipitation processes within RegCM. J Geophys Res., 105: 29579 29594.

Pal, J.S., Giorgi, F., Bi, X., Elguindi, N., Solomon, F., Gao, X., Francisco, R., Zakey, A., Winter, J., Ashfaq, M., Syed, F., Bell, J.L., Diffanbaugh, N.S., Kamacharya, J., Konare, A., Martinez, D., da Rocha, R.P., Sloan, L. C. and Steiner, A. (2007). The ICTP RegCM3 and RegCNET: regional climate modeling for the developing world. Bulletin of American Meteorological Soc. , 88: 1395 1409.

Payne, S.W. and McGarry, M.M. (1977). The relationship of satellite inferred convective activity to easterly waves over west Africa and the adjacent ocean during phase III of GATE. Mon. Wea. Rev., 105: 413 420.

Pilgrim, D.H. and Cordery, I. (1993). Chapter 9: Flood Runoff. In: Maidment D.R. (eds.) Handbook of Hydrology, McGraw-Hill, New York, USA.

Pospichal, B., Karam, D.B., Crewell, S., Flamant, C., Hunerbein, A., Bockd, , F. (2010). Diurnal cycle of the intertropical discontinuity over West Africa analysed by remote sensing and mesoscale modelling. Q. J. R. Meteorol. Soc., 136(s1):92 106.

Quevauviller, P. (2011). Water Sustainability and Climate Change in the EU and Global Context Policy and Research Responses. Issues in Environmental Science and Technology , 31, Sustainable Water, 24p.

Rasmussen, K., Fog, B. and Madsen, J.E. (2001). Desertification in reverse? Observations from northern Burkina Faso, Global Environmental Change , 11:271 282.

162

Rawls, W.J., Brakensiek, D.L. and Shirmohammadi, A. (1993). Infiltration and soil water movement. In: D.R. Maidment (eds) Handbook of Hydrology McGraw-Hill, New York, pp5.1-5.51.

Reiche, J., Verbesselt, J., Hoekman, D. and Herold, M. (2015). Fusing Landsat and SAR time series to detect deforestation in the tropics. Remote Sensing of Environment , 156: 276 293.

Richard, Y. and Poccard, I. (1998). A statistical study of NDVI sensitivity to seasonal and interannual rainfall variations in Southern Africa. Int. J. of Remote Sensing , 19:15, 2907- 2920 .

Rockwood, A.A. and Cox, S.K. (1978). Satellite inferred surface albedo over north-western Africa . J. Atmos . Sci ., 35: 513-522.

Rodwell, M.J. and Hoskins, B.J. (1996). Monsoons and the dynamics of deserts. Q. J. R. Meteorol. Soc., 122: 1385 1404.

Rowell, D.P. and Milford, J.R (1993). On the generation of African squall lines. J. Climate, 6: 1181 1193.

Salih, A.M., Kornich, H. And Tjernstrom (2013). Climate impact of deforestation over South Sudan in regional climate model, Int. J. Climatol., 33: 2362-2375.

Sandwidi, J.P. (2007). Groundwater potential to supply population demand within the Kompienga dam basin in Burkina Faso. PhD thesis. Center for Development Research, University of Bonn. Ecology and Development Series, 2007. 55p. Bonn, Germany. Available from: http://www.zef.de/fileadmin/template/Glowa/Downloads/Sandwidi_do c_thesis_2007.pdf .

Sanni, M., Odekunle, T.O. and Adesina, F.A. (2012). Spatio-Temporal Variation of Drought Severity in the Sudano-Sahelian Region of Nigeria: Implications for Policies on Water Management. In: W. Leal Filho (eds), Climate Change and the Sustainable Use of Water Resources, Climate Change Management, DOI: 10.1007/978-3-642-22266-5_26.

Schulz, M., Textor, C., Kinne, S., Balkanski, Y., Bauer, S., Berntsen, T., Berglen, T., Boucher, O., Dentener, F., Guibert, S., Isaksen, I.S.A., Iversen, T., Koch, D., Kirkevåg, A., Liu, X., Montanaro, V., Myhre, G., Penner, J. E., Pitari, G., Reddy, S., Seland Ø., Stier, P., and Takemura, T. (2006). Radiative forcing by aerosols as derived from the AeroCom present-day and pre-industrial simulations. Atmos. Chem. Phys., 6: 5225-5246.

Simmons, A., Dee, D., Uppala, S. and Kobayashi, S. (2007). Erainterim: new ecmwf reanalysis products from 1989 onwards. In: ECMWF Newsl, 110. ECMWF, p .

Sivakumar, M.V.K. (1992). Climate change and implications for agriculture in Niger. Climatic change, 20 (4): 297-312.

163

Sivakumar, M.V.K. (2007). Interactions between climate and desertification. Agricultural and Forest Meteorology, 142:143 155.

Schindler, J. (2009). A multi-agent system for simulating land-use and land-cover change in the Atankwidi catchment of Upper East Ghana, PhD thesis, Center for Development Research, University of Bonn. Ecology and Development Series, 68p. Bonn, Germany. Available from: http://hss.ulb.uni-bonn.de:90/2009/1954/1954.htm .

Schubert, W.H., Ciesielski, P.E., Stevens, D.E. and Kuo, H.C. (1991). Potential vorticity modeling of the ITCZ and the Hadley circulation. J. Atmos. Sci., 48: 1493 1509.

Stewart, J.B. (1970). The albedo of a pine forest . Q. J. R. Meteor . Soc ., 97: 561-564.

Stow, D.A., Shih, H.-C. and Coulter, L.L. (2014). Discrete classification approach to land cover and land use change identification based on Landsat image time sequences. Remote Sensing Letters, 5:10, 922-931.

Sultant, B. and Janicot, S. (2000). Abrupt shift of the ITCZ over West Africa and intraseasonal variability. Geophys. Res. Lett., 27: 3353 3356.

Sultan, B., Janicot, S. and Diedhiou, A. (2003). The West African Monsoon Dynamics. Part I: Documentation of Intraseasonal Variability, J. Climate , 16: 3389-3406.

Sultan, B. and Janicot, S. (2003). The West African Monsoon dynamics, Part II: the monsoon. J. Climate, 16(21):3407 3427.

Sylla, M. , A., Ruti, P.M. and Giorgi F. (2010). Simulation of the intraseasonal and the interannual variability of rainfall over West Africa with RegCM3 during the monsoon period. Int. J. Climatol, 30: 1865 1883.

Sylla, M.B., Giorgi, F. and Stordal, F. (2012). Large-scale origins of rainfall and temperature bias in high-resolution simulations over southern Africa. Clim Res., Vol. 52: 193 211.

Sylla, M.B., Diallo, I. and Pal, J.S. (2013). West African Monsoon in State-of-the-Science Regional Climate Models, Climate Variability-Regional and Thematic Patterns, Dr. Aondover Tarhule (eds), ISBN: 978-953-51-1187-0, InTech, DOI: 10.5772/55140.

Textor, C., Schulz, M., Guibert, Kinne, S., Balkanski, Y., Bauer, S., Berntsen, T., Berglen, T., Boucher, O., Chin, M., Dentener, F., Diehl, T., Easter, R., Feichter, H., Fillmore, D., Ghan, S., Ginoux, P., Gong, S., Grini, A., Hendricks, J., Horowitz L., Huang P., Isaksen, T. Iversen, S. Kloster, D. Koch, A. Kirkevag, J. E. Kristjansson, M. Krol, A. Lauer I., Lamarque, J. F., Liu, X., Montanaro, V., Myhre, G., Penner, J., Pitari, G., Reddy, S., Seland, Ø., Stier, P., Takemura, T., and Tie, X. (2006). Analysis and quantification of the diversities of aerosol life cycles within AeroCom. Atmos. Chem. Phys., 6: 1777 1813.

164

Thorncroft, C.D. and Hoskins, B.J. (1994). An idealized study of African easterly waves. I: A linear view. Q. J. R. Meteorol. Soc., 120: 953-982.

Thorncroft, C.D. and Rowell, D. (1998). Interannual variability of African easterly waves in a GCM. Int. J. Climatol., 18: 1305-1323.

Thorncroft, C.D. and Blackburn, M. (1999). Maintenance of the African easterly jet. Q. J. R. Meteorol. Soc., 125: 763-786.

Thorncroft, C.D. and Hodges, K. (2001). African Easterly Wave Variability and Its Relationship to Atlantic Tropical Cyclone Activity, J. Climate, 14:1166-1178.

Tompkins, A.M., Diongue-Niang, A., Parker, D.J. and Thorncroft, C.D. (2005). The African easterly jet in the ECMWF Integrated Forecast System: 4D-Var analysis. Q. J. R. Meteorol. Soc., 131: 2861 2885.

Tucker, C.J. and Miller, L.D. (1977). Soil spectra contributions to grass canopy spectral reflectance. Photogrammatic Engineering and Remote Sensing , 43: 721-726.

Tucker, C.J. (1979). Red and photographic infrared linear combinations for monitoring vegetation, Remote Sensing of Environment , 8: 127 150.

Tucker, C.J., Pinzon, J.E., Brown, M.E., Slayback, D.A., Pak, E.W., Mahoney, R., Vermote, E.F. and Saleous, N.E. (2005). An extended AVHRR 8-km NDVI dataset compatible with MODIS and SPOT vegetation NDVI data. Int. J. of Remote Sensing , No. 20, 26: 4485 4498.

Uppala, S., Dee, D., Kobayashi, S., Berrisford, P. and Simmons A. (2008). Towards a climate adapt assimilation system: status update of ERA-Interim. ECMWF Newsletter, 115: 12-18.

Vizy, E.K. and Cook, K.H. (2001). Mechanisms by Which Gulf of Guinea and Eastern North Atlantic Sea Surface Temperature Anomalies Can Influence African Rainfall. J. Climate , 14: 795-821.

Wang, J., Rich, P.M. and Price, K.P. (2003). Temporal responses of NDVI to precipitation and temperature in the central Great Plains, USA. Int. J. Remote sensing , Vol. 24, 11: 2345 2364.

White, F. (1983). The vegetation of Africa, a descriptive memoir to accompany the UNESCO/AETFAT/UNSO vegetation map of Africa. UNESCO, Natural Resour. Res., 20: 1-356.

Williams, C.J.R. and Kniveton, D.R. (2011). African Climate and Climate Change, (eds) 101 Advances in Global Change Research 43, DOI 10.1007/978-90-481-3842-5_5, Springer Science+Business Media B.V.

165

Wilson, M.F. and Henderson-Sellers, A. (1985). A global archive of land cover and soils data for use in general circulation climate models. J. Climate , 5: 119-143.

Wu, D., Wu, H., Zhao, X., Zhou T., Tang, B., Zhao, W. and Jia, K. (2014). Evaluation of Spatiotemporal Variations of Global Fractional Vegetation Cover Based on GIMMS NDVI Data from 1982 to 2011. Remote Sens., 6: 4217-4239; doi:10.3390/rs6054217.

Yaw, O.-A. and Ian, C. (1994). Atlantic Sea Surface Temperatures and Rainfall Variability in Ghana, J. Climate , 7: 551-558.

Yilma, T. (2006). Modeling farm irrigation decisions under rainfall risk in the Whit Volta Basin of Ghana. A tool for policy analysis at the farmhousehold level. Bonn, Göttingen, Germany: University of Göttingen. 170p.

Yuan, X., Li, L. and Chen, X. (2015). Increased grass NDVI under contrasting trends of precipitation change over North China during 1982 2011, Remote Sens. Letters , 6:1, 69- 77.

Yuen, B. and Kumssa, A. (2011). Climate Change and Sustainable Urban 3 Development in Africa and Asia (eds), DOI 10.1007/978-90-481-9867-2_1, Springer Science+Business Media B.V.

Zhu, Z., Bi, J., Pan, Y., Ganguly, S., Anav, A., Xu, L., Samanta, A., Piao, S., Nemani, R.R. and Myneni, R.B. (2013). Global Data Sets of Vegetation Leaf Area Index (LAI) 3g and Fraction of Photosynthetically Active Radiation (FPAR)3g Derived from Global Inventory Modeling and Mapping Studies (GIMMS) Normalized Difference Vegetation Index (NDVI3g) for the Period 1981 to 2011. Remote Sens , 5, 927-948.

Zanis, P., Ntogras, C., Zakey, A., Pytharoulis, I. and Karacostas, T. (2012). Regional climate feedback of anthropogenic aerosols over Europe using RegCM3. Clim. Res., Vol. 52: 267 278.

Zeng, N. (2003). Drought in the Sahel, Science 3.

Zheng, X. and Eltahir, E.A.B. (1998). The Role of Vegetation in the Dynamics of West African Monsoons. J. Climate , 11: 2078-2096.

166