Simulating the impact of land use and land cover change on surface air temperature trends in South-Central Vietnam

Abstract

Vegetation degradation, particularly due to deforestation and urbanization, has substantially increased local surface air temperatures in many regions. This study aims to quantify the impact of land use and land cover change (LUCC) on surface air temperature trends in the South Central Coast region of Vietnam. Monthly mean datasets covering the past 26 years were used, including the Normalized Difference Vegetation Index (NDVI), land surface temperature (LST), and 2 m air temperature (T2m). Statistical analyses and artificial neural network (ANN) modeling were employed. The main explanatory variables included the mean values and temporal trends of NDVI and LST extracted from buffer zones surrounding meteorological stations. The results indicate that the influence of LUCC T2m is most pronounced during the warm season (April–August). Analyses for this period show that the selected remote sensing variables exhibit the strongest relationships with T2m when extracted within buffer radii of 2–8 km, while the prevailing wind direction has no significant influence on temperature trends. After selecting the optimal predictors, the ANN model successfully reproduced the observed temperature trends with high reliability. Relative contribution analysis showed that the NDVI trend and mean NDVI accounted for approximately 65% and 23% of the model performance, respectively, demonstrating that NDVI effectively captures the influence of LUCC on T2m. A scenario assuming no land use change, represented by zero trends in the explanatory variables, produced smaller and more spatially uniform temperature increases among stations, reflecting the background warming associated with climate change. The proposed framework provides a practical approach for assessing the environmental impacts of LUCC and for estimating its contribution to observed temperature trends.