Effects Of Urban Spatial Structure Change On Urban Thermal Environment Of Changchun

The impacts of urbanization on ecological environment has become an important focus in the study of regional human-environment system.Urbanization,especially urban land expansion,has been the most dominant form of urban growth at the local,regional,and global scales over the last several decades.The response to rapid even unplanned urbanization,makes urban spatial structure and its thermal environment change in profound ways.Consequently,urban populations are highly exposed to the effects of regional climate change,including more intensive heat island effect,which poses serious threats to public health.Urbanization most directly contributes to changes of urban spatial structure,which is the most obvious and direct relation between urban system and environmental change.Urban land surface temperature is intrinsically related to location,and is prominently associated with urban surface characteristics,such as surface parameter types,composition,configuration,and so forth,which would have significant impacts on urban thermal environment especially in the course of urbanization.In recent years,Changchun city has experienced rapid urbanization with dramatically urban spatial expansion,and thus it has inevitably resulted in acute issues of urban thermal environment.Therefore,under the background of rapid urbanization,it is of great significance to explore effects of urban spatial pattern change on urban thermal environment for more rational and order expansion of cities and more optimization of urban spatial structure,and to help alleviate urban heat island effect.Land cover map of Changchun city in this study was extracted based on Landsat TM/OLI imagery and urban spatial structure was defined using the segmentation of the built-up density map,which was produced using a moving lkmx lkm grid.Based on GIS platform and landscape analysis,geospatial approaches were used to help promote a systematic analysis on urban spatial expansion as well as urban spatial structure change of the study area.Besides,based on MODIS/Terra 8-day LST data with 1 km2 resolution,spatial distribution and gradient characteristics of daytime/nighttime surface theremal environment of Changchun city were described in this study by calculating the mean LST in summer season of 2000,2008 and 2016 year.The spatial statistical methods including urban-rural gradient analysis,spatial autoregressive model and GWR model were introduced to explore different effects of urban spatial morphology,scale and structure of surface parameter,and landscape mixing and coordination degree on surface thermal environment at 1 km grid units.Furthermore,this study also paid more attention to dynamic process of urbanization and thus was aimed to understand how urban thermal environment responds to it.The results are as follows:(1)The daytime mean LST in 2008 was relatively lower compared with that in 2000.And although the daytime mean LST in 2016 was close to that in 2000,the dispersion degree of temperature was higher with significance in spatial heterogeneity of urban thermal environment.Meanwhile,it exhibited an obvious growing trend of the nighttime mean LST,which rised from 17.43? to 18.76? in 2008 and increased to 19.17? in 2016.Furthermore,day/nighttime UHI intensity generally decreased in period of 2000?2016,while the area of grade ??? heat island continued to expand along with the intensity continued to upgrade,especially the IV grade of nighttime UHI.(2)Changchun city had experienced rapid urban land sprawl with obvious urban spatial expansion.Overall,urban spatial expansion made the day/nighttime LST increase in newly urbanized space,and meanwhile it might make the LST increase slightly in the original urban area as well but weakened in the process of urbanization.Besides,the average annual growth rate of urban built-up land in Changchun city in period of 2000?2008 and 2008?2016 reached 6.72%and 8.08%respectively.The expansion intensity of urban fringe areas was the highest,with its most significant expansion in southwest and northeast directions respectively.According to GWR model analysis,the response of the daytime UHI to the UII varied spatially and significant positive effects were mainly distributed in numerous continuously newly urbanized areas and urban fringe areas,where it was significantly associated with higher UII.(3)The response of daytime LST to the increase of built-up land scale in units changed by phase:On average,a 5%increase in the percentage of built-up areas approximately increased daytime LST by 0.4?.Moreover,instead of evenly rising,the daytime LST increased sharply when the proportion of built-up land in units was around 40%and 70%.However,the response of nighttime LST was not sensitive to the increase of built-up land scale.Furthermore,the increase of built-up density would significantly affect the upgrading of daytime UHI other than night.(4)The study of the spatial gradient change of temperature showed that:from city center to the exurban,the daytime LST variation pattern along the urban-rural gradient was consistent with that of built-up land and vegetation.Urban expansion reduced the temperature difference in suburban area and the daytime LST changes from logarithmic decline to linear decline.However,the nighttime LST was less affected and was always decreasing logarithmically.(5)The results of spatial autoregressive model of day/night surface temperature and explanatory variables of urban spatial pattern indicated that the remote sensing spectral index and landscape composition/configuration index in the explanatory variables had different effects on day/nighttime LST.The percentage of built-up land had the greatest explanatory power for the increase of daytime LST,while the vegetation coverage,percent of water,largest patch index of water and crop land shape index were negatively correlated with it.Beisides,MNDWI index,percent of built-up land,percent of urban vegetation and largest patch index of crop land were all significantly positively correlated with the nighttime LST.That is,the nighttime LST was largely affected by the comprehensive influence of multiple landscape elements,and of which built-up density was not the only high-value factor to explain it.In addition,LST has spatial autocorrelation and SEM model had better fitting results on the relationship between day/nighttime LST and explanatory variables.