Abstract A groundwater flow model has been built for the Chobe region in Botswana. Due to the scarcity of conventional data, alternative sources of information have been explored. METEOSAT and NOAA-AVHR images have been used to estimate the spatial distribution of the water balance. This spatial distribution has been correlated to local chloride measurements from which the recharge rate could be estimated. A digital elevation model has been used as a constraint for the maximum local piezometric head values. Finally, geomagnetic data were used as an indication for the presence of faults. These data have been used together with a limited number of traditional data (hydraulic head and transmissivity measurements) in an inverse calibration procedure. In the inverse conditioning, equally likely realizations, consistent with all the measurements (transmissivity, hydraulic head, digital elevation model, satellite images, chloride measurements and geomagnetic data) have been generated. The objective function contains an extra constraint on a statistical basis that guarantees that the calibrated recharge rate pattern does not deviate too much from the estimated water balance using the satellite image. The study demonstrates the importance of the digital elevation model and the satellite information in improving the groundwater model of the site.

Keywords conditioning; digital elevation model; geomagnetic data; inverse modelling; remote sensing; spatial patterns