Remote sensing and GIS-based WetSpass-S model for estimating actual evapotranspiration in Grombalia region, Northeast Tunisia: comparison with ETLook WaPOR model

The assessment of actual evapotranspiration and other components of the water balance is important for the sustainable management of surface and groundwater resources. This work aims to estimate and map the spatio-temporal distribution of the actual evapotranspiration in Grombalia phreatic aquifer for the 2020/2021 agricultural year using WetSpass-S model. The results are then compared with those obtained from the FAO’s ETLook model. WetSpass-S integrates climatic, soil and land cover data to simulate water balance components, while the ETLook model combines remote sensing data with meteorological inputs. Sentinel-2 images were downloaded and processed. Multi-date supervised classifications mapped land cover (LULC) categories (orchards, agriculture, build-up, shrub and water surfaces), following the requirements of the WetSpass-S, with kappa index of 0.68. This map was integrated with hydro-meteorological and biophysical data. The results include seasonal and annual estimates actual evapotranspiration, surface runoff and recharge. Annual means are, respectively, 350, 28 and 36 mm, stating that the water balance is driven mainly by actual evapotranspiration, with 85% of the received precipitation (412 mm). The actual evapotranspiration values are compared with FAO-published data (ETLook model). WetSpass-S gives higher values than FAO in surface water and building areas and lower values in irrigated areas, the difference exceeding 800 mm. The main cause of this variation is the exclusion of irrigation as input for WetSpass-S. This highlights the importance of careful consideration of the input datasets used and a wise interpretation of the results obtained for appropriate surface and groundwater management, as well as hydrological and hydrogeological assessment.