Using Sentinel-2 time-series for land-use and evapotranspiration mapping in semi-arid Mediterranean areas: Tensift catchment, Morocco

The primary aim of this study is to estimate crop evapotranspiration (ETc) and develop a land use map to support the sustainable management of water resources in the irrigated Tassaout perimeter of Tensift. This is achieved by combining a time series of high spatial resolution Sentinel-2 satellite images, meteorological data, and field data. The method is based on the relationships between normalized difference vegetation index (NDVI) and crop coefficient (Kc). Remote sensing data were used for crop type mapping and deriving (Kc) based on NDVI data. Meteorological data were also used for reference evapotranspiration (ETo) calculations. The obtained land use was adequately simulated with the field data acquired by the Regional Office of Agricultural Development during the 2015-2016 cropping season. The results showed a kappa coefficient (K) and an overall accuracy of 0.91 and 93.3%, respectively. The land use in the studied area is mainly dominated by cereals crops and olive trees which represent 33.54 and 22.44%, respectively. It was also observed that monthly ETc varied over time and is greater than 170 mm from April to July, reaching a maximum value of 303 mm in August and reducing to approximately less than 60 mm from October to March. Such maps of ETc can constitute an important indicator of how water resources should be managed over the area for improving irrigation scheduling policy.