Challenges in groundwater pollution management in transboundary basins

Challenges in groundwater pollution management in transboundary basins in Africa: SUPPORT TO WATER QUALITY IN VOLTA BASIN Dr. Issoufou Ouedraogo, Assistant Professor, University of Fada N’Gourma, Burkina Faso & Dr. Marnik Vanclooster Full Professor, University catholique of Louvain, Belgium Expert Consultation workshop for the Basin, 25 -27/02/2020/Accra

• Special thanks to the UN Environment (UNEP), and the World Water Quality Alliance (WWQA) for the invitation and financial support to attend this Workshop meeting in Accra 1

Introduction: Groundwater resources in Africa 75 % of the African population depends on groundwater for basic water supply Groundwater used for other activities (irrigation, livestock watering, urban and industrial development, etc) Groundwater demands will increase in future due: • Population increase • Climate change • Need to combat growing food insecurity Source: Tindimugaya, 2017 2

Groundwater in Africa: Pollution pressures • Groundwater is an important water resource in Africa • Africa relies very strongly on groundwater use • Pressures on African groundwater resources are important 3

Groundwater in Africa: Pollution pressure in smallholder farming systems Source: Mamadou and Vanclooster, 2011

Groundwater pollution in Africa Rapid urban development has resulted in many informal settlements!!!

Introduction: Groundwater pollution with nitrates in Africa: Case of Abidjan/Ivory Coast Source: Ahoussi et al. , 2013 4

Objectives of the study General objective: To improve knowledge on the state of groundwater quality at the pan-African scale, supporting the monitoring of the implementation of the UN Sustainable Development Goals agenda. Specific objectives: • To map groundwater vulnerability for pollution at the pan-African scale. • To develop a multi-variate statistical model that predicts nitrate concentration in groundwater at the pan-African scale. • To validate continental-scale groundwater nitrate pollution model using regional datasets. 5

1. Mapping vulnerability at the pan African scale Intrinsic vulnerability • Standard DRASTIC methodology Specific vulnerability • Crossing with most recent land use map Originality • Extension: pan African scale • Resolution: 15 km x 15 km • Parametrisation using most recent spatial available data sets 6

Results: Mapping depth to water (D) 7

Results: Mapping net recharge (R) 8

Results: Mapping aquifer media (A) 9

Results: Mapping soil media (S) 10

Results: Mapping topography (T) 11

Results: Mapping impact of vadose zone (I) 12

Results: Mapping hydraulic Conductivity (C) 13

Results: Mapping intrinsic vulnerability 14

Results: Mapping Landcover/land use (L) Land. Cover/Land. Use reclassified Original Land Cover 15

Results: Mapping groundwater pollution risk 16

Results: Mapping groundwater pollution risk under transboundary basins Hotspot zones!!!!! 17

2. Statistical model based on meta-analysis • Systematic reviews were carried out from published litterature (Scopus, Science direct, Google scholar and, …). • A meta-analysis has identified approximately 250 studies of groundwater pollution for Africa. • Results were regressed on ancillary available spatially distributed predictor data using linear and non linear models (random forest). 18

2. Statistical model based on meta-analysis GIS Slope Depth to groundwater table African physical attributes Soil type Land use Population density Precipitatio n Groundwater Recharge Statistical model Multiple Linear Regression Random Forest Regression Nitrate in groundwater at the African scale Nitrate simulation Nitrate observations (Meta-analysis) 19

Results: Distribution of studies considered in the meta-analysis 20

Methodology: Meta-data creation for statistical analysis Search engine Search criteria Groundwater pollution + Africa Nitrate in groundwater + “African country name” or “African capital city name” Groundwater quality + Africa Google, Google Nitrate and agricultural practices in Africa Scholar, and Google Groundwater vulnerability + “African country name” Pollution des eaux souterraines par les nitrates+ ‘’nom du pays Africain’’ (in French) Books Pollution des eaux souterraines + "nom du pays Africain" (in French) Nitrate concentrations under irrigated agriculture + “African country name” Groundwater pollution by nitrate + “African country name” Nitrate in groundwater + “Africa capital city name” Pollution des eaux souterraines par les nitrates + "nom du capital des Web of Sciences, pays"(in French) Scopus and Sciences Groundwater contamination by nitrate + “Africa countries” or “African capital city name” Direct Africa irrigated agriculture + nitrate Groundwater contamination by nitrate + “Africa country name” or “African capital city name” Books Nitrate concentrations under irrigated agriculture + “African country name” Groundwater vulnerability to nitrate contamination + “Africa country name” or “African capital city name” Groundwater pollution in Africa ( Xu and Usher, 2006) 21

Results: Statistics of nitrate in the meta-analysis Statistic Maximum NO 3 ln(NO 3 -) Mean NO 3 - Mean ln(NO 3 -) Minimum NO 3 ln(NO 3 -) Number of samples Minimum 206 82 82 185 0. 08 -2. 52 1. 26 0. 231 0 0 Maximum 4625 8. 43 648 6. 473 180 5. 19 Median 73. 64 4. 29 27. 58 3. 317 0. 55 0. 43 190. 05 3. 99 54. 85 3. 169 8. 91 1. 08 183778. 94 3. 39 163. 92 43. 901 537. 07 1. 78 CV 225. 56 46. 18 8085. 08 1. 935 260. 08 123. 04 Std. Dev. 428. 69 1. 84 89. 91 1. 391 23. 17 1. 33 Kurtosis 60. 24 0. 90 23. 99 -0. 167 25. 57 0. 37 6. 75 -0. 74 4. 31 -0. 294 4. 56 1. 2 Mean Variance Skewness 22

Results: Multiple Linear regression and random forest (RF) model. Mean Error=0. 95 R 2=0. 64 Multiple Linear Regression Mean Error=0. 04 R 2=0. 97 Random Forest Regression 23

Results: Variable importance in RF model 24

3. Can the pan-African model be used at the regional / national scale? Senegal South Africa Burkina Faso 25

Results: Validation of the pan-African model using regional / national data sets Senegal South Africa Burkina Faso 26

Results: Recalibration and validation at the national scale (a) (b) R 2=0. 91 Calibration graph R 2=0. 92 Validation graph 27

Conclusions • The pan-African groundwater vulnerability and pollution risk map showed that areas under very high and high pollution risk are mainly characterised by shallow groundwater systems. • The risk map of groundwater pollution in Africa shows that water resources are under pressure in large agricultural basins and urban environments. • At the African scale, population density, depth to shallow groundwater, aquifer type, rainfall and recharge explains to a large extent the observed mean nitrate contamination. • Non linear Random Forest models were more powerfull than linear models to model nitrate contamination. • At the regional / national scale, recalibration of the models is strongly suggested, but this needs good monitoring data. 28

Perspectives • Evaluate the limitations of data processing methods, subjectivity in assigning weights and ratings by decision-makers, and non-linear relationships between hydrogeological factors; • Integrate best available dataset from a local scale to improve the model’s performance at regional scale (such as, age of groundwater, denitrification, GDP per capita, degree of sanitation, physicalchemical variables, etc. ); 29

Further reading • • • Ouedraogo, I. , Defourny, P. , Vanclooster, M. (2016). Mapping the groundwater Vulnerability for pollution at the pan-African scale. Science of the Total Environment, Vol. 544, p. 939 -953. Ouedraogo, I. , and Vanclooster, M. (2016). Shallow groundwater poses pollution problem for Africa. In: Sci. Dev. Net. Ouedraogo, I. , and Vanclooster, M. (2016). A meta-analysis and statistical modelling of nitrates in groundwater at the African scale. In: Hydrology and Earth System Sciences, Vol. 20, no. 6, p. 2353 -2381. D Ouedraogo, I. , Defourny, P. , and Vanclooster, M. (2017). Modeling groundwater nitrate concentrations at the African scale using Random Forest Regression Techniques. Accepted 24 th April to review in the special issue on Groundwater in Sub-Saharan Africa for Hydrogeological Journal (HJ) (in progress, book expected in December 2017). Ouedraogo, I. , Defourny, P. , and Vanclooster, M. (2017). Validating a continental scale groundwater diffuse pollution model using regional datasets. Submitted 23 rd June to Environmental Science and Pollution Research (ESPR) journal for IAH 2016 special issue. (Under review). . Impact Factor: 2. 741 (2016) Ouedraogo, I. , Girard, A. , Defourny, P. , Vanclooster, M. , and Jonard, F. (2017). Time dynamic pollution risk modelling of groundwater at the African scale. To be submitted. 30

Communicating groundwater information and impacts 31

32

Scientific implication: New pollution risk map for Africa to help with achieving safe water for everyone Media Release: World Water Day 22 March, 2017 Responding to UNICEF/WHO report on Safely managed drinking water From Lapworth et al. 2017 https: //upgro. org/2017/03/13/new-pollution-risk-maps-forafrica/ 33

Challenges in assessment water quality in hotspots in Africa : Case of Volta Basin!!! 34

Introduction: The Volta Basin ü International Basin Organization (6 countries), BENIN BURKINA CÔTE D’IVOIRE GHANA MALI TOGO ü Population of the basin~ 20 million ü Surface: ~400, 000 km 2 ü January 2007 – creation of the VBA through a Convention adopted by The Head of States ü August 2009 – Entry in force of the VBA Convention 35

Introduction: Volta Basin Member Countries Burkina Faso, Bénin, Côte d’Ivoire, Ghana, Mali, Togo Autorité du Bassin de la Volta (ABV) 36

Surface Water Resources in the Volta Basin The Volta River basin; shared by Burkina Faso, Ghana, Ivory Coast, Togo, Benin and Mali. 38

GROUNDWATER IN THE VOLTA BASIN/Hydrogeology The Voltaian system (sedimentary) and the basement complex dominate the hydrogeological systems 38

Transboundary Groundwater Shared Basin Iullimeden Taoudéni Tano Keta Volta Liptako Gourma (not included) Characteristics Countries 3 aquifer basins with River Niger as the southern limit. Various types of sandstone. Algeria, Benin, Mali, Nigeria, Algeria, Burkina, Mali, Mauritania, Coastal sedimentary basin, which includes the Volta estuary. Information is very scanty. Crystalline basement aquifer; partly in the Volta basin? Cote d’Ivoire, Ghana Benin, Ghana, Nigeria, Togo, Benin, Burkina, Togo Burkina, Niger Source: Biney, C. (nd) § All the six riparian countries of the Volta share, at least, one groundwater basin with another country § Two basins, Tano and Volta are shared solely by the riparian countries of the Volta while four basins are shared with other non riparian countries 39

Transboundary Groundwater 3. Basin Taoudéni 4. Basin Iullimeden 12. Basin Liptako Gourma 13. Basin Tano 14. Basin Keta Figure: Shared Aquifer Basins in the Volta Basin (Source: ISARM Africa) 40

Use of Groundwater Vulnerability/case of Volta Basin Skoulikaris et al. 2018 Figure. Transboundary aquifers and groundwater pollution risk in Africa. 41

ACKNOWLEDGEMENTS Tom Gleeson, Mc. Gill University, Victoria University Petra Döll & Felix Portmann, Frankfort University Nils Moosdorf & Jens Hartmann, Hambourg University Arthur Girard , Master student, UCL 1) Alan Mac. Donald, BGS/London; 2) Moussa Cissé, DGPRE/Senegal; 3) Jacqueline Zougrana/DEIE/Burkina; 4) Ashton Maherry/CSIR/South Africa; 5) Etc. Céline Lamarche & Pierre Defourny, UCL Marnik Vanclooster, UCL

Let's make our surface water and groundwater less polluted! Merci pour votre attention ! /Thank you very much ! Email: ouedraogo. issoufou 03@gmail. com