Using Long-term All-weather Land Surface Temperature Data To Study The Urban Heat Island Effect

Satellite thermal infrared land surface temperature(LST)data are susceptible to cloud cover and other factors,resulting in large areas of data deficiencies.The existing surface urban heat island(SUHI)studies based on satellite thermal infrared LST data are limited to clear-sky conditions and cannot represent the real results under all-weather conditions due to the limitation of spatially and temporally continuous data.To solve these problems,this thesis introduces the newly released 1 km all-weather LST product(TRIMS LST)into the SUHI study to investigate the real spatial and temporal patterns of surface urban heat island intensity(SUHII)under all-weather conditions based on its spatiotemporally seamless characteristics.The main research work of this thesis is as follows.(1)To evaluate the accuracy of TRIMS LST under clear-sky and non-clear-sky conditions and the applicability of applying it to examine the SUHI effect: firstly,TRIMS LST was compared with MODIS LST for consistency;secondly,TRIMS LST/MODIS LST was compared quantitatively with measured LST;and finally,the differences of SUHII calculated based on TRIMS LST and MODIS LST under clear-sky,partial-cloudy and cloudy conditions were quantified.Results show that TRIMS LST has high accuracy under both clear-sky and non-clear-sky conditions.Due to the lack of valid pixel data of MODIS LST,SUHII is significantly overestimated or underestimated based on MODIS LST under partial-cloudy and cloudy conditions,annual mean deviation values ranged from-1.83 °C to 0.10 °C,and the frequency of underestimation increases with the number of missing pixels.(2)To quantify the influence of clouds on the all-weather SUHI effect and the magnitude of the clear-sky bias existing between the clear-sky SUHII and the all-weather SUHII: firstly,the average value of the all-weather SUHII within each delineated cloud coverage interval was calculated,and secondly,the difference between the monthly-scale SUHII calculated based on MODIS LST under clear-sky conditions and TRIMS LST under all-weather conditions was used to characterize the clear-sky bias to assess the seasonal and diurnal characteristics.Results show that during the daytime,when clouds accumulate to a certain level,they have a significant mitigating effect on the SUHI effect,e.g.,the SUHI in Guangzhou under cloudy conditions(1.70 °C)is only 60% of that under clear-sky conditions(2.66 °C);during the nighttime,for most cities,their SUHII shows a non-linear decrease with increasing cloudiness.Meanwhile,the clear-sky bias is mostly positive during the nighttime due to the significant mitigation effect of clouds on the SUHI effect at night.(3)To compare the differences in the spatial and temporal patterns of the all-weather SUHI effect based on TRIMS LST and the clear-sky SUHI effect based on MODIS LST:firstly,based on the average value of the SUHII from 2003 to 2020,the difference in the spatial pattern of the clear-sky and all-weather SUHII were analyzed;secondly,the differences in the temporal trend of clear-sky and all-weather SUHII from 2003 to 2020 were analyzed by combining Pearson correlation analysis and linear regression analysis method.The results show that there are significant differences in the spatial patterns and temporal trends of SUHII between clear-sky and all-weather conditions,clear-sky bias values between-4 °C and 1 °C,and the differences are more significant during the daytime than during the nighttime.