干旱区生态环境知识资源中心

In semi-arid, groundwater-dependent regions of South Africa, allocation of additional water resources can become problematic in the absence of quantified regional groundwater recharge values. In this study in the northern Sandveld, remote-sensing-data products for precipitation (P) and evapotranspiration (ET) are used to quantify groundwater recharge and guide the conceptualization of the hydrogeology of the study area. Data from three ET models (ETMODIS, MOD16, Pitman rainfall-runoff) are compared; these models concur best in years of average rainfall, with model results deviating up to 30 % in wet years. The MODIS data product (MOD16) is used in conjunction with gridded precipitation data to calculate spatial regional recharge. The long-term precipitation minus evapotranspiration (P-ET) budget closes on a positive 13 +/- 25 %; however, when correcting ET (20 % underestimation determined using the chloride mass balance method), the catchment average potential recharge is reduced to -4 +/- 30 %. The use of P-ET clearly identifies potential recharge zones at higher elevation and discharge zones, highlighting irrigated agriculture. The usefulness of identifying recharge zones is demonstrated in the value added to conceptualizing the hydrogeology. Since some uncertainty around the accuracy of ET data still remains, it is recommended that the MODIS data product be validated more comprehensively in semi-arid environments.