Modelling groundwater nitrate concentrations at the pan-African scale using Multiple Regression and Random Forest Statistical Models

Ouedraogo, Issoufou [UCL] Defourny, Pierre [UCL] Vanclooster, Marnik [UCL]

Groundwater is a crucial natural resource supporting the development of the African continent, but it is subjected to many pressures. According to Xu and Usher (2006), degradation of groundwater is the most serious water resources problem in Africa. Nitrate is a common chemical contaminant of groundwater and the level of contamination also increases in many African aquifers. Statistical models can be deployed to explain the spatial distribution of observed nitrate concentration in terms of available environmental and anthropogenic attributes, or to discriminate sources of contamination. The purpose of this study is to use the non linear Random Forest Regression(RFR) technique to explain and predict groundwater nitrate contamination at the continental scale and compare the performance of RFR statistical model with Multiple Linear Regression (MLR) techniques.