Detecting changing trajectory of urban heat island using Gaussian model in Beijing, China

Author(s):  
Jinling Quan ◽  
Yunhao Chen ◽  
Wenfeng Zhan ◽  
Ji Zhou
Author(s):  
M. Burnett ◽  
D. Chen

Abstract. The urban heat island (UHI) is a common effect caused by urbanization and has been studied to evaluate the thermal condition in cities worldwide. However, most previous UHI analyses are performed in major metropolitan cities. This study conducts a spatiotemporal analysis of UHI in a rapidly expanding low-density suburban centre and determines how bio-productive land covers react and the extent of the disturbance to each land cover based on time series land surface temperatures extracted from Landsat 7 ETM+ images. Two methods applied and compared are the single exponential decay method, which measures UHI footprint (UHIFP) on vegetation phenology, and the two dimensional Gaussian surface, which quantifies the influence based on distance from the local urban perimeter. Three spectral indices (Normalized Difference Vegetation Index (NDVI), Moisture Index (NDMI), and the Enhanced Vegetation Index (EVI)) were extracted and the residuals from the Gaussian model were compared based on these indices in order to better understand the thermal variations of each land cover within a UHI. The results show that the UHIFP of the studied low-density suburban centre is 1.4 times larger than the size of the urban centre, marginally smaller than previous analyses performed within high-density metropolises. All vegetated land covers experienced their maximum cooling effects before reaching the UHIFP perimeter while urban surfaces begin to diverge from the Gaussian model outside of the UHIFP. The residuals of sparse vegetation maintained strong correlations with each index throughout the growing season while NDMI retained the strongest relationships with every land cover. This study has helped us better understand the UHI effects of small communities with varied vegetation phonology based on the distribution of built-up pervious and impervious surfaces within the neighbourhood structure. The similar results from both methods indicate a strong urban cover influence overpowering the dominant distribution of agricultural surfaces throughout the growing season.


Sign in / Sign up

Export Citation Format

Share Document