Quantifying the impact of land surface temperature on vegetation moisture for drought monitoring in Tamil Nadu

Drought significantly threatens agriculture, water resources, and ecosystems, particularly in Tamil Nadu, India. This study explores the link between Land Surface Temperature (LST) and the Normalized Difference Water Index (NDWI) to evaluate their effectiveness for drought monitoring across Tamil Nadu's districts from 2014 to 2023. Utilizing MODIS MOD11A2 for LST and MOD09A1 for NDWI, the analysis examines the influence of temperature variations on vegetation moisture levels. The Pearson correlation analysis identified substantial spatial differences, with strong correlations in districts such as Perambalur, Namakkal, and Dindigul (up to 0.91), suggesting higher temperatures are closely associated with reduced weaker correlations in regions like the Nilgiris and Tirunelveli suggest that temperature exerts a lesser influence on vegetation moisture in those areas. Further quantification was achieved through linear and polynomial regression models. The linear model explained 52.7% of NDWI variation due to LST (R-squared = 0.527) and was validated as the most robust model via cross-validation. While polynomial models accounted for slight nonlinearities, they offered limited predictive improvement, confirming that a linear relationship generally describes the NDWI-LST dynamics adequately. The results indicate that LST is a valuable indicator for drought monitoring in strongly correlated areas. In contrast, additional variables like rainfall and soil moisture may be essential for accurate predictions in regions with weaker correlations. Overall, this study demonstrates the potential of remote sensing for drought monitoring and emphasizes the need to consider local environmental factors to refine predictive models across