Analysis of the Spatial and Temporal Variations of Land Surface Temperature Based on Local Climate Zones: A Case Study in Nanjing, China
Weizhong Su,Yong Wang,W. Zhan,W. Ma,Qin Zhu,Xin Pan,Yingbao Yang,Jia Hu
DOI: https://doi.org/10.1109/JSTARS.2019.2926502
IF: 4.715
2019-07-30
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Abstract:The concept of local climate zone (LCZ) has standardized the calculation of urban heat island (UHI) intensity (UHII) and established the connection among urban morphology, surface property, and UHI. In addition, LCZ has provided a new insight into the studies on urban thermal environment. This study selected Nanjing, China, as the study area and utilized a combined method that comprised remote sensing based and geographic information system based methods based on random forest classifier for LCZ classification. Overall accuracy reached 92%, and kappa coefficient was 0.911. In addition, the seasonal and diurnal differences of land surface temperature (LST) were analyzed via LST retrieval from Landsat data and computational fluid dynamics model simulation, respectively. Results confirmed the warmest and coldest zones in four seasons. The LST distribution characteristics of land cover and built types were basically the same during the four seasons. Moreover, the UHII difference of an LCZ class in various seasons and the UHII difference of a season in various LCZs were investigated. The daily LSTs of the simulated LCZs (1 to 6) within 24 h showed the same variation law but different variation extents in a day. The LST variations of built types were related to building elements, such as building height, building density, building layout, and green ratio. This study identified an existing relation between LST and LCZ and analyzed LST on the basis of LCZs from seasonal and diurnal scales, which provides guidance for future researchers.
Computer Science,Geography,Environmental Science