摘要：AbstractWeather stations and field observation provide imperfect near-surface air temperature (NSAT) to study canopy urban heat island (UHI), whereas thermal images offer a high spatial resolution of land surface temperature (LST) to investigate surface UHI. Because NSAT is widely recognized as robustly dependent on LST, we suggest a method to construct continuous, high spatial resolution NSAT in Qlyub City, Egypt, derived from LST and other urban variables, such as albedo, thermal inertia, and climatic conditions. The proposed method integrates thermal and optical imageries, field observation, statistical analysis, and geographical information systems (GIS) modeling to perform a multi-scale analysis of NSAT and LST relationships. Linear regression revealed a moderately strong connection between NSAT and LST. Standardized regression embedding multiple times, land uses, and environmental variables suggested an interesting finding of near-perfect correlation. The nighttime connection was stronger than the daytime. Relative importance analysis showed that LST is the most important predictor of NSAT, followed by observation time, thermal inertia, humidity, and normalized difference vegetation index (NDVI). All results were statistically significant, and uncertainties were minimal (mostly <±0.5C). GIS modeling was utilized to extract NSAT from LST. The findings were validated using in-situ observations. It is concluded that the methodology is reliable to provide continuous, high spatial resolution NSAT for dependable canopy UHI detection and thermal structure analysis.
关键词：KeywordsUrban heat islandThermal imageryAir temperature predictionIn-situ measurementRegression analysisGIS modeling
