Satellite observations and spatial regression to assess geographical parameters on land surface temperature

Urban and industrial expansion raises the land surface temperature (LST), posing challenges that require understanding for sustainable city planning. This study assesses variables influencing LST at Sharjah University City, UAE, a semi-arid area experiencing the reverse Urban Heat Island (UHI) effect. This study assess key factors like land use land cover (LULC), spectral indices, wind speed, and topography using meteorological records, Landsat 8 imagery, WorldView-3, Shuttle Radar Topography Mission (SRTM), and ERA5. The relationships among these variables are evaluated using the ordinary least square regression model. The study period spanned from October 2020 to January 2021 (winter) and from June 2021 to August 2021 (summer). The results demonstrate that different factors influence LST seasonally. In winter, the normalized difference built-up index (NDBI), normalized difference water index (NDWI), and wind speed significantly impact LST, with an R² of 0.67. Wind speed correlates positively with LST during the winter, with a correlation coefficient of approximately 4.95. During the summer, the normalized difference vegetation index (NDVI) emerges as a crucial factor alongside wind speed, resulting in an R² of 0.73. Conversely, wind speed is negatively correlated with the LST, with a correlation coefficient of −1.45 in the summer. The developed LST model provides high-resolution predicted LST maps that can fill data gaps due to cloud cover. The model discrepancies include overestimation in residential areas and underestimation on roads. This study highlights the importance of wind speed and informs urban planning by adding vegetation that could mitigate temperature rise.