Assessing the Vulnerability of Sundarbans Mangroves to <scp>LST</scp> Variability: Utilising 26‐Year Spatiotemporal Data

ABSTRACT The Sundarbans, the world's largest mangrove forest and a UNESCO World Heritage Site, is becoming increasingly susceptible to climate‐induced stress. This study examines 26 years of satellite‐derived land surface temperature (LST) data to evaluate seasonal trends, land use/land cover (LULC) changes and the spatial correlation between LST variations and predominant mangrove species in the Sundarbans of Bangladesh. LST fluctuations were most evident in the pre‐monsoon and monsoon seasons. February and December exhibited a statistically significant cooling trend over the study period. Concurrently, forest cover has decreased at an average annual rate of 5.78 km 2 , whereas coastal water bodies have increased by 5.17 km 2 , triggering microclimatic shifts that reinforce a positive feedback loop where deforestation intensifies surface heating, further accelerating forest degradation. Cluster analyses reveal sharp monthly temperature shifts outside of the pre‐monsoon season, suggesting climatic instability that could push the system toward ecological thresholds. SARIMA modelling demonstrated 95.21% accuracy in temperature forecasting for 5 years, underscoring the predictive significance of temporal analysis for future stress thresholds. Species‐specific clustering showed Ceriops decandra dominating hotter zones (26.69°C) and Heritiera fomes preferring cooler zones (25.88°C), indicating potential future redistribution or decline of sensitive species under climate extremes. This study is the first to combine species‐specific LST analysis, cluster analysis and SARIMA forecasting in the Sundarbans, offering high‐accuracy predictions of thermal stress and advancing species‐informed, adaptive management strategies.