Cloud Removal Of Surface Temperature And Spatio-temporal Evolution Patterns Of Heat Island

Under the double background of global warming and accelerated urbanization,urban thermal environmental problems are becoming more and more prominent,negatively affecting the quality of human living environment as well as environmental security,and restricting sustainable urban development.Therefore,it is important to accurately grasp the spatial and temporal characteristics of the urban thermal environment and make suggestions for this problem,in order to achieve a win-win situation between economic and ecological construction in cities.The availability of surface temperature data is constrained by the fact that the thermal infrared remote sensing is affected by cloud occlusion and there are missing data in the surface temperature products.To obtain seamless surface temperature data for Zhengzhou City,the MODIS surface temperature from 2011-2020 and the corresponding explanatory variables were used as input data for three machine learning models:Support Vector Machine(SVR),Random Forest(RF)and XGBoost.The optimal model was used to estimate cloud-covered surface temperature values,fill in the vacant values of the MODIS surface temperature products.Afterwards,the impervious surface data was used as the basic data for the extraction of the built-up area based on the distance-weighted method to calculate the cohesion density of impervious surface.In addition,combining seamless surface temperature data and built-up area data,the spatio-temporal characteristics of the heat island effect in Zhengzhou cities(counties and districts)under urban expansion are quantitatively analysed from the perspectives of heat island area,heat island intensity and heat island footprint.Ultimately,the corresponding optimization recommendations are proposed in order to provide a reference for improving the urban thermal environment in Zhengzhou.The results indicate that(1)By training three machine learning models,SVR,RF and XGBoost respectively,the R~2values of the three models were 0.84,0.88 and 0.93,and the RMSE values were 0.50,0.44 and 0.34 respectively.The XGBoost model has the largest coefficient of determination and the smallest root mean square error.The results show that the XGBoost model can effectively fill in the missing surface temperature values in the original MODIS surface temperature data.(2)The built-up areas of the cities(counties and districts)of Zhengzhou show varying degrees of expansion from 2011 to 2020.During the study period,the built-up area of Zhengzhou City increases by a total of 509.19 km~2and in 2020 its area is 1.93times larger than in 2011.The built-up area of the main urban area grew by 330.34 km~2,and the remaining cities’built-up areas are,in descending order:Aeroport District(37.93 km~2),Xingyang City(30.9 km~2),Xinzheng City(23.39 km~2),Zhongmou County(19.21 km~2),Gongyi City(13.84 km~2),Xinmi City(12.63 km~2),Dengfeng City(12.34km~2)and Shangjie District(10.52 km~2).(3)To some extent,the heat island effect has been mitigated in most of Zhengzhou,but the issue of the thermal environment still needs attention.The spatial distribution of higher surface temperature areas remains largely consistent with the built-up area,with the heat island area increasing each year as the built-up area expands.In terms of the characteristics of the change in heat island area,the sum of the proportions of high and sub-high temperature areas decreased in all cities,except for the Shangjie District and the Aeroport District.In terms of the characteristics of changes in heat island intensity,the cities(counties and districts)basically show a trend of first increasing and then decreasing heat island intensity.In terms of the characteristics of the heat island footprint,with the exception of Dengfeng City,the rest of the cities in Zhengzhou show different degrees of expansion.During the study period,the cities of Gongyi,Xinmi and Dengfeng have not yet experienced superposition of the thermal environment,while the remaining cities show varying degrees of overlap with the heat island footprint of neighbouring cities.To improve the thermal environment of Zhengzhou,measures such as reasonable control of urban scale,optimisation of urban spatial layout,enhancement and optimisation of urban green space and water area and spatial layout,application of new energy and new technology,and promotion of green and low-carbon travel are proposed to alleviating the urban thermal environment.