Identifying irrigated areas in the Rhine basin using land surface temperature and hydrological modelling

In recent years, the Rhine has experienced summer drought which led to extremely low water availability throughout the basin. Additionally, combination of high temperature and low precipitation may increase irrigation demand, putting even more pressure on water availability. Identifying where irrigation occurs and how it evolves over time offers improved insight water use for sustainable water resources planning and management. However, high-resolution maps of irrigated areas for basin-scale studies on water use are often lacking. Here, as part of the HorizonEurope project STARS4Water, we aim to develop a methodology for identifying irrigated areas in the Rhine basin at a 1 km resolution for the period of 2010-2019. This involves utilizing a combination of the hydrological model wflow_sbm to produce land surface temperature (LST_sim) and thermal observations data (LST_obs) from the Moderate Resolution Imaging Spectroradiometer (MODIS) sensor. To provide consistent annual irrigation maps, we employed random forest classification model to further identify irrigated areas from the LST difference between LST_sim and LST_obs. In the absence of ground information data, the irrigated maps are evaluated against national agricultural statistics and compared with existing developed irrigated maps. The results can be used to comprehend the interannual variability in the extent and location of irrigated croplands in the Rhine basin and are a start to assess and model agricultural water use in the Rhine basin.