Dataset: Predictions of land use land cover and land surface temperature based on machine learning algorithms to support climate-resilient urban planning in UK cities

This dataset contains derived outputs from a machine learning framework for projecting calibrated land surface temperature (LSTc) and land use land cover (LULC) across Glasgow, Cardiff, and Cambridge, UK. Included are: (1) calibrated LSTc rasters for historical years 1990–2023 for each city, derived from Landsat thermal imagery calibrated to ERA5-Land 2 m air temperature using weighted least squares regression; (2) projected 2031 LSTc rasters generated using a Random Forest model; and (3) projected 2031 LULC rasters generated using a MOLUSCE ANN-CA model. All rasters are in GeoTIFF format, British National Grid (EPSG:27700), 25 m resolution. Input datasets are not included due to licensing restrictions; sources are detailed in README.txt and in the associated publication.