Soil Moisture Variations and Their Relationships with Different Vegetation Types in The Upper Blue Nile River Basin

The vegetation–soil moisture relationship is complex and nonlinear. Vegetation activity determines the space–time distribution and availability of soil moisture. In this study, we have used the 3rd generation Global Inventory Monitoring and Modeling Systems (GIMMS3g.v1) Normalized Difference Vegetation Index (NDVI) dataset to represent the vegetation activity in the Upper Blue Nile River Basin (UBNRB). We also used the distributed Water and Energy Processes (WEP) hydrological model to generate the soil moisture data. We used the Global Land Cover (GLC2000) data to identify specific land cover types. To examine the trends in the NDVI and soil moisture information, we employed the non-parametric Mann–Kendal trend test along with the Theil–Sen slope estimation technique. We applied another non-parametric correlation analysis named Spearman correlation to investigate the degree of relationship between soil moisture changes and vegetation responses. We limited our study to growing season (April–October) NDVI and soil moisture values to reflect the vegetation's growth status better. The results show that shrublands in northwestern lowlands along the Ethio-Sudan border and southern Ethiopian highlands exhibit a significantly increasing trend. Overall, 32.3% of the NDVI pixels and 66.6% of the soil moisture pixels show a significant trend. The sparse grassland of the southwestern lowlands shows a decreasing vegetation activity trend. In total, 28.8% of the valid pixels (excluding pixels of no data, cities, and waterbodies) show a higher correlation (ρ > 0.7), whereas 49.7% indicate a correlation between 0.5 and 0.7, mainly in the shrubland, grassland, and forested areas. The correlation between same month (lag0) NDVI and soil moisture is substantially higher than the subsequent previous month's (lag1-lag5) soil moisture values. This quick response relates to croplands, shrublands, and grasslands, confirming their sensitivity to short-term soil moisture changes. Forested areas of the basin did not show an appreciable correlation with soil moisture fluctuations.