Assessing Vegetation Response to Hydroclimatic Variability in Semi-Arid Regions: NDVI-LST-GWL Interactions

This study examined the relationships between NDVI, EVI, NDWI, land surface temperature (LST), and groundwater level (GWL) from 2001 to 2024. NDVI anomalies showed an overall increase, with more pronounced positive deviations after 2015. A strong correlation was found between NDVI and EVI (r=0.84), and a strong negative correlation between NDWI and LST (r=-0.90). The NDVI–GWL relationship was weak (r=0.22; anomalies r=0.16). Linear and multiple regression analyses had low explanatory power (max R²=0.049; multiple regression 6.9%). Stationarity tests showed that NDVI anomalies were persistent, while LST and GWL anomalies were stationary. Granger causality tests indicated no predictive relationship between NDVI and GWL. As a result, vegetation dynamics were mainly influenced by seasonal climate cycles, with groundwater playing a small but consistent role. The weak linear correlations highlight the need to integrate rainfall, soil moisture, and land-use change using nonlinear and lagged modeling approaches.