Satellite-Based Assessment of Vegetation Water Stress Across West Africa Using Evapotranspiration Limitation and Independent Rainfall Evaluation from 2001 to 2023

Abstract Assessing crop water stress in tropical and semi-arid regions is complicated by interactions among surface temperature, vegetation condition, evapotranspiration, and atmospheric moisture demand. This study developed a scalable, fully satellite-based framework for monitoring crop water stress across West Africa within Google Earth Engine for 2001–2023. The primary stress metric is an evapotranspiration-based index, defined as 1 − ET/PET, which quantifies the limitation of actual relative to potential evapotranspiration; land surface temperature (LST) is retained as an independent diagnostic of surface energy balance rather than an input. The index was benchmarked against the Temperature–Vegetation Dryness Index (TVDI) and validated against an independent moisture measure, the standardised CHIRPS rainfall anomaly. A strong spatial gradient was evident: mean stress rose from 0.369 in the humid Guinea Coast to 0.940 in the semi-arid Sahel, with LST increasing from 30.94°C to 40.55°C and evapotranspiration declining roughly sevenfold (1125.95 to 156.93 mm) northward. Against independent rainfall, the ET/PET-based index correlated more strongly than the temperature-based TVDI (median per-pixel r = -0.29 versus − 0.05). We further tested whether adding a vegetation term (1 − NDVI) improved the index; it did not. NDVI alone was uncorrelated with rainfall (median r = -0.01), and the multiplicative combination weakened the ET/PET signal (median r = -0.09), so the parsimonious ET/PET formulation was retained. Over the period, NDVI increased significantly (Mann–Kendall p = 0.001), whereas the ET/PET-based stress index showed no statistically significant trend (Mann–Kendall p = 0.13). The framework provides an interpretable, externally validated basis for drought monitoring in data-scarce tropical environments.