Downscaling of diumal land surface temperature cycles for urban heat island monitoring

For urban heat island (UHI) monitoring diurnal cycles of land surface temperatures (LST) in a spatial resolution of about 100 m would be a major asset. Hence, a recently proposed method was used to downscale an entire diurnal cycle of coarse resolution LST from the Spinning Enhanced Visible Infra-Red Imager onboard the geostationary Meteosat Second Generation satellite. This method underlined that parameters incorporating multitemporal thermal infrared data are very suitable as downscaling predictors. Here it is further shown that the predictors are also appropriate for different times of the day. Compared to Landsat, the downscaling resulted in an RMSE of 2.53 K and an R2 of 0.73, which is slightly worse than in a previous study. Compared to in situ data the downscaling did not improve the coarse measurements but reproduced elemental characteristics of the diurnal cycle. © 2013 IEEE