Robustness of satellite-derived land surface parameters to urban land surface temperature

Investigating the relationship between land surface temperature (LST) and other satellite-derived land surface parameters is beneficial to exploring adaptation and mitigation strategies for urban thermal environment. The article indicates the key parameters influencing LST at regional scale and object scale using comparison of the relationship between LST and land surface parameters based on ordinary least squares (OLS) and Kendall robust linear fit. The results showed that normalized difference moisture index (NDMI) is the most effective and robust parameter interpreting LST, followed by normalized difference buildup index (NDBI). Although normalized difference vegetation index (NDVI) and the albedo of visible light can strongly affect LST, they showed a lack of robustness because of the large gaps between the determination coefficients of the two regression methods at regional scale. At object scale, the average temperature of factories was the highest, followed by business services construction. NDVI and NDMI are the important factors affecting LST for these points of interest (POIs).