Land use and land cover change impact on characteristics of surface evapotranspiration in semi-arid environment of Western Rajasthan, India

The development of the Indira Gandhi Nahar Pariyojana (IGNP) in western Rajasthan has significantly impacted land use and land cover (LULC) changes, which have hugely transformed the hydrologic cycle in recent decades. Hence, we investigate the surface evapotranspiration (ET) response to LULC changes in the IGNP for the period from 1990 to 2020. The Surface Energy Balance algorithm (Surface Energy Balance Algorithm for Land) was applied to the Landsat dataset along with local weather data obtained from the Indian Meteorological Department. A machine learning-based random forest classifier was applied to develop the LULC maps. The higher values of ET are mainly distributed in the northwest. Area covered by croplands has shown an increasing trend over the decades at a rate of 734.76 km2·a−1, urban settlements/ built-up areas increased at a rate of 14.67 km2·a−1, barren area decreased gradually at a rate of 751.67 km2·a−1, and area covered by the few water bodies showed minor fluctuations. Overall, the LULC changes have a smaller influence on ET rates but has a considerable impact on the total ET in the IGNP region due to an increase in the area covered by vegetation as a result of increased water availability. HIGHLIGHTS This is the first of its kind study to assess the changes in the region's overall evapotranspiration along with changes in land use and land cover.; Land use land cover classification using the machine learning-based random forest classifier gave convincing results with high accuracy.; The present study demonstrates spatiotemporal variations of evapotranspiration at a regional scale using an improvised Surface Energy Balance Algorithm for Land method for regions experiencing low wind speeds.;