Pressure on water resources is increasing worldwide due to constantly growing population. This study aims to develop a spatial model and demonstrate its utility for water productivity mapping (WPM) in a smallholder informal irrigation system around Lake Bam (Burkina Faso) with emphasis on tomato. The study involved three major steps leading to WPM: (1) Sentinel-2 (S-2) data were used for crop productivity mapping involving crop classification with random forest algorithm, crop yield modelling with remote sensing (RS) indices, and yield model extrapolation to a larger area; (2) crop water use estimation by multiplying the irrigated surface area by the actual seasonal evapotranspiration developed through the surface energy balance model ‘METRIC’ with Landsat8 (L8) data; and (3) WPM produced by dividing raster layers of the two steps above. An overwhelming 89.46% (769.16 ha) of irrigated tomato area for the season 2016–2017 falls in low WP category of 2.5 kg.m-3 or less. Only 10.5% of the tomato cultivated area had a WP value of 2.5 kg.m-3 or higher. About 82.05% of the tomato area had values lower than 1.63 kg.m-3. The results imply that there is significant scope for increasing WP without having to increase cultivated area or quantity of water utilised. The areas of low WP are spatially pin-pointed and can be used as focus for WP improvements through better land and water management practices.
Keywords: water productivity, evapotranspiration, remote sensing, food and water security