The Influence Of Geographical Unit Division On The Response Regularity Of Urban Thermal Environment

Scale effect and spatial heterogeneity determine that the effects of the same factors on ecological processes are different in different geographical units.In this paper,urban thermal environment effect,a typical urban ecological process,is selected as the research object.Aiming at the current situation that there are few studies on the response rule of geographical unit division to thermal environment effect,four geographical units,namely land use classification,grid,block and local climate area,are selected.Normalized Difference Vegetation Index(NDVI),Impervious Surface Area(ISA)and Normalized Difference Water Index(NDWI)were selected as the influencing factors of thermal environment.In this paper,the urban area of Beijing redefined by POI data was used as the study area.Land surface temperature was retrieved from summer Landsat data in 2015,2016 and 2017,and thermal environment influencing factors were extracted to explore the differences in response rules of urban thermal environment to landscape components in different geographical units.The research conclusions are as follows:(1)The land surface temperature increased year by year.In 2015,the areas with high land surface temperature were relatively concentrated,while in 2016 and 2017,there were several small heat islands.The surface temperature of hutong residential and commercial areas is relatively high.NDVI and MNDWI values increased year by year,while ISA values decreased year by year.(2)Single factor Pearson correlation analysis showed that NDVI was significantly negatively correlated with land surface temperature in block(0.59,0.65,0.76)and grid(> 0.7)units,and the fitting curve was smooth through significance test.Land use classification and local climate zones do not apply.ISA has a high correlation between the grid and the land surface temperature,and the Pearson coefficient is between 0.37-0.6.ISA ISA geographical unit suitable for predicting temperature changes.In 2015 and 2016,ISA had a high correlation in local climate zone(0.666,0.622)and land use classification(0.672,0.719)units.Block units did not pass the significance test.The 2017 ISA was positively correlated with the surface temperature in the block(0.691).MNDWI was significantly positively correlated with land surface temperature in grid(0.206,0.383,0.618)and block(> 0.4)units.Local climate zones and land use classification units were not applicable.(3)According to the multiple linear regression model,the results of the multi-factor combination model are better than those of the single factor,and all the results can pass the significance test,and the grid element is more universal.Study the effects of vegetation and impermeable or vegetation and water combination on surface temperature by selecting blocks or grid cells.Study impervious and water body or vegetation,water body,impervious interaction,select land use classification or grid unit.(4)Different urban planning and management methods need to choose different geographical units and regression models to alleviate the urban heat island effect.Selecting land use taxon according to land cover type;Selecting grid for uniform management mode;Study specific areas and population selection blocks;Adjust urban building structure and choose local climate area.ISA,NDVI and MNDWI regression models were selected to have the best fitting effect.