Estimation And Influencing Factors Of Anthropogenic Heat Flux In Chinese Cities

Rapid global urbanization can lead to the increases in city size and population growth,with impacts on urban underlying surface and canopy.High-intensity human activities in cities continue to release huge amounts of Anthropogenic Heat（AH）into urban environment.Studies have shown that AH affects the local environment and microclimate of cities and it is a crucial factor contributing to urban heat island effect.Moreover,due to the differences in economic and social development levels and geographical environments,anthropogenic heat emissions in different regions exhibit significant spatial heterogeneity.Therefore,it is important to quantitatively analyze the quantity and spatial distribution characteristics of anthropogenic heat generation and the potential influencing factors for urban thermal environment regulation and sustainable urban development.In this study,200 cities in mainland China were selected as the study area.Firstly,the annual and monthly Anthropogenic Heat Flux（AHF）estimation methods were studied using the energy inventory method and multi-source data.Secondly,hotspot analysis,cluster and outlier analysis,and direction distribution analysis tools were used to explore the spatial and temporal variations of AHF in Chinese cities.Finally,the Multiscale Geographically Weighted Regression（MGWR）model is used to quantitatively analyze the AHF influencing factors and their spatial heterogeneity.Additionally,the variation of AHF influencing factors across regions,seasons,and city sizes were explored.The main findings are as follows:（1）The annual urban AHF estimation results were efficiently obtained by building a model of provincial average AHF with Normalized Nighttime Light(NTL_nor).Meanwhile,the monthly AHF estimation results for 200 cities in mainland China in 2015 were obtained for the first time based on monthly temperature data.Overall,the heating（December to February）and non-heating season（June to August）yielded the highest and the lowest AHF of the year,respectively.In the regional comparisons,annual and monthly AHFs were higher in the northwest,northeast,and north China than in other regions.The annual and monthly AHF in large cities were higher than those in other cities.（2）The annual urban AHF in general yielded a trend of gathering hot spots in the north and cold spots in the south.The differences in urban AHF between southern and northern cities were greater in the heating season.In addition,the standard deviation ellipse of urban AHF in mainland China shows a"northeast-southwest"distribution.（3）Among the influencing factors of urban AHF,two-dimensional（2D）and three-dimensional（3D）building morphology parameters yielded an important impact on urban AHF.In general,the impact of 2D parameters is greater than that of 3D parameters.Among them,the2D building morphology parameters（i.e.,building density）in the northeast region have a significant positive impact on the annual AHF（β=0.42±0.13）.Among the 3D building morphology parameters,the Mean Building Height（MBH）yielded a weaker effect on urban AHF than Mean Building Volume（MBV）.The effect of MBV on urban AHF was mainly concentrated in the northeastern region of China.Among the socio-economic indicators,Per Capita Gross Domestic Product（PCGDP）in general yielded a stronger positive impact on annual urban AHF in the north than in the south of China.Among the natural environment indicators,Normalized Difference Vegetation Index（NDVI）yielded a strong negative effect on urban AHF,especially in East China during the non-heating season（β=-0.59±0.12）.In small cities,PCGDP and NDVI yielded a stronger effect than in large cities.In addition,water body percent,road density and science and technology branch yielded a weak effect on urban AHF.（4）The MGWR model had advantages in quantitatively analyzing the influencing factors of urban AHF and revealing their spatial heterogeneity,yielded better model performance compared with Geographically Weighted Regression（GWR）and Ordinary Least Squares（OLS）models.