Estimation of crop evapotranspiration and crop water stress index using google earth engine in Victoria basin of Uganda

Abstract

Evapotranspiration (ET) is the sum of soil evaporation and plant transpiration, playing a key role in optimizing water use and boosting crop yields thereby contributing to global food security. The Crop Water Stress Index (CWSI) complements ET by quantifying water stress experienced by crops. However, large-scale estimation of CWSI remains challenging due to its sensitivity to crop and environmental variables. The study assessed the suitability of Google Earth Engine (GEE) in estimating crop ET and CWSI across Uganda's Victoria Basin. ET was estimated using the MOD16A2 product in the GEE catalog, focusing on the ET band. Time series charts and spatial maps were generated for the years 2016 to 2020. CWSI was computed using Land Surface Temperature (LST), Normalized Difference Vegetation Index (NDVI), and Land Cover data. Hot and cold pixels were identified in GEE to derive CWSI values. Ground weather data was used to compare daily GEE-derived ET with ground-based ET to obtain correlation coefficients. The results for CWSI were compared with previous studies (Soni et al., 2023). High ET values were recorded in January, June, and December, while low values occurred in April, October, and November. Maps revealed spatial variation in crop stress, with values aligning with vegetation health and temperature patterns. The CWSI values were within the theoretical range of 0 to 1. The correlation coefficients between Masaka, Mbarara and Kabale were 0.513, 0.460 and 0.531 which were weak and positive. GEE effectively captured seasonal and spatial ET and CWSI variations across the Lake Victoria Basin. The relatively low correlation between GEE-ET and ground ET data can be attributed to the coarse spatial and temporal resolution of the MODIS product. The CWSI values matched known seasonal crop water stress trends and were consistent with previous studies, affirming the method's reliability. This study successfully demonstrated the potential of Google Earth Engine in estimating crop ET and CWSI using freely available satellite data. While GEE provides useful insights for regional scale monitoring, local calibration is essential for improving accuracy.