Study On Near Surface Urban Heat Island Effect Based On Human Comfort

With the rapid development of the global economy and the acceleration of city development,the phenomenon of Urban Heat Island(UHI)has gradually emerged and its influence has been expanding.It has become a major global environmental problem of the Heat.Urban heat island effect not only accelerates global warming and changes urban microclimate,but also imperceptibly affects people's life and health.Studying the temporal and spatial characteristics and evolution law of urban heat island effect is of great significance to promote urban green sustainable development.Remote sensing monitoring method has the characteristics of large-scale and comprehensive observation and saving time and labor.It is one of the important means to study urban heat island effect.The traditional method is mainly to obtain the channel brightness temperature or surface temperature through remote sensing data inversion to study the heat island effect.There are few studies on the urban heat island effect based on the near surface temperature(shutter height temperature).Taking Taiyuan as the study area,this thesis selects seven remote sensing image data in 2002,2008,2014 and 2020,inverts the surface temperature LST by using the single window algorithm,calculates the normalized vegetation index NDVI,the improved normalized difference water index MNDWI,the normalized building index NDBI and the surface Albedo,and takes the elevation data DEM as the influencing factor of near surface temperature,using the random forest model Ordinary Kriging interpolation method and multiple regression statistical method are used to estimate the near surface temperature.After comparative analysis,the random forest model with the highest accuracy is selected to obtain the spatially continuous near surface temperature in the study area.Based on this,the urban heat island effect of Taiyuan is studied,and the human comfort index is introduced to further explore the comfort of living environment in the study area.The main conclusions are as follows:(1)Based on the data of spring,summer,autumn and winter in 2020,from the perspectives of mean absolute error(MAE)and root mean square error(RMSE),the overall accuracy of near surface temperature estimation results based on random forest model is the highest and more in line with the actual situation,followed by multiple regression statistical method,and the accuracy of ordinary Kriging interpolation method is the lowest.It is found that the accuracy of estimation results in winter and summer is higher than that in spring and autumn.In addition,the importance of each independent variable parameter of the near surface temperature estimation model is analyzed.The results show that LST has the greatest impact on the near surface temperature estimation model.(2)Based on the results of near surface temperature estimated by random forest model,the urban heat island effect is studied.It is found that the near surface temperature in Taiyuan is increasing from suburbs to urban areas,and the heat island effect is significant.From 2002 to 2020,the heat island effect increased significantly,especially in Wanbailin District and Xiaodian District.The center of gravity of the high temperature area,the sub high temperature area and the medium temperature area all shift to the south,among which the center of gravity of the high temperature area shifts the most,and the offset distance reaches 9862.22 meters.The center of gravity of the sub medium temperature area and the low temperature area shifts slightly to the north,and finally forms a double heat island central pattern of the central heat island of Taiyuan and the southern heat island of Taiyuan distributed in a strip along the Fenhe River section.(3)By introducing the human comfort index and comprehensively considering the impact of near surface temperature,wind speed and relative humidity on the comfort of living environment,the research results show that from 2002 to 2020,the area of cool area and comfortable area in Taiyuan decreased by 12.744 km ^ 2 and 617.376 km ^ 2 respectively,the area of slightly hot area and hot area increased by 389.4km ^ 2 and 143.016 km ^ 2respectively,and the human comfort in Taiyuan decreased as a whole.Taiyuan has high comfort in spring and autumn,dry and cold in winter and hot and humid in summer.In all seasons,the comfort of suburban areas is higher than that of urban centers.(4)The correlation between human comfort and its influencing factors is analyzed by path analysis.The results show that air temperature has the greatest impact on CIHB,while relative humidity and wind speed have little impact on CIHB,but it will have an indirect impact on human comfort through air temperature.(5)In the study area,the temperature of building land is the highest and the temperature of water body is the lowest.NDVI,NDBI,MNDWI and BSI are introduced to analyze the correlation between surface cover types and near surface temperature.The results show that the near surface temperature is negatively correlated with vegetation and water body,and positively correlated with building land and bare soil.