Spatiotemporal Variation Characteristics Of Surface Thermal Environment In The Context Of Urbanization

Human activities are the most important factors that affect and change the natural environment,among which the deterioration of the surface thermal environment in urban areas is an important indicator reflecting the frequency and intensity of human activities.In the past 30 years,with the continuous expansion of cities,the increase of man-made surface coverage has led to the emergence and deterioration of a series of urban thermal environment problems.It is the basis of improving urban thermal environment to master the temporal and spatial variation process of different surface temperature.Therefore,based on the data of surface temperature at different time and space scales in Chenggong area in recent 19 years,the paper obtains the spatial variation range and spatial distribution of different underlying surface temperature in Chenggong District in the past 19 years.Meanwhile,the paper selects typical urban communities in the research area to carry out microscopic scale verification and analysis,designs the surface thermal environment monitoring system,and conducts continuous observation on typical areas,and obtains the results The surface temperature observation data and meteorological factor data(temperature,wind speed,wind direction,solar radiation,etc.)of typical areas from July 2019 to June 2020 are presented.The surface temperature estimation model is constructed by BP artificial neural network(BPANN),and the surface temperature missing in typical urban areas in 2019 is estimated.The results show that the coefficient of certainty(R2),absolute error(MAE)and mean square error(MSE)of the water underlying surface estimation model are 0.8657,1.2285 and 4.8297,respectively;the coefficient of certainty(R2),absolute error(MAE)and mean square error(MSE)of the grass underlying surface estimation model are0.8992,1.1582 and 4.4498,respectively;the coefficient of certainty(R2),absolute error(MAE)and mean square error(MSE)of the cement underlying surface estimation model are 0.8992,1.1582 and 4.4498,respectively The absolute error(MAE)and mean square error(MSE)are 0.8854,1.9866 and 8.0945 respectively;the established estimation model has the advantages of high accuracy and low error,and can estimate the historical surface temperature more accurately.The results of time analysis show that the annual average surface temperature in Kunming area from 2002 to 2012 shows a relatively stable trend,with a slight increase in 2004 and 2014,and no significant change in other years.In 2018,it reached an all-time low.The change process of Di image in one day of monthly mean value in summer is presented by superimposing Di value on remote sensing image.The image shows that Di has obvious spatial variability.From the perspective of space,the thermal comfort level of the building and cement underlying surface is significantly lower than that of the surrounding vegetation and water.Because the thermal comfort level of the whole research area is represented by the visual Di image,it is more meaningful in the application than using a single Di value in the traditional method.Therefore,the comfort level can be distinguished in detail in space.