Research On The Temporal And Spatial Variation Process And Prediction Of Urban Comfort In The Main Urban Area Of Kunming

With the rapid urbanization of our country,the urban heat island phenomenon that comes with it becomes more and more severe.For the phenomenon that the urban thermal environment is aggravated seriously,people begin to pay more and more attention to the improvement of the quality and comfort of the urban environment.As an important ecological resource,the city is also the carrier of a series of human activities.The quality of the urban ecological environment is closely related to the sustainable development of a city.At present,in the field of urban ecological environment research,urban comfort has gradually attracted people's attention.Human activities are the most important factor affecting and changing the urban environment.Therefore,improving the comfort of the city and making it meet people's thermal comfort needs of the environment can effectively improve the quality of the urban environment.Urban comfort more or less affects and changes the local urban microclimate,and is also related to the living environment of human beings.Therefore,realizing the accurate prediction of urban comfort level in the future will provide valuable information for macro-control of urban environment,and provide more reference for the implementation of urban planning and management strategies.This study selected the main urban area of Kunming City,Yunnan Province,known as the"Spring City",as the research area.Four meteorological factors(temperature,wind speed,relative humidity,sunshine duration)from 1990 to 2018and urbanization data from 1990 to 2018(land type use,population)as the basic data.Using DI index,K index,CI index and extreme climate index,a comprehensive evaluation index(Hybrid Index,HI)of urban comfort was constructed by principal component analysis,and the urban comfort in the main urban area of Kunming was analyzed and evaluated.Through three different time series forecasting methods(ARIMA model,NAR neural network model,long short-term memory network model),the urban comfort prediction model is constructed,and the significance test and Hurst index are used to analyze the urban comfort from 1990 to 2018.Two aspects of significance were analyzed for change characteristics.The performance of the three models is measured using three evaluation metrics.The results show that since 1990,the urban comfort index(DI<21;-2702.58,p<0.01)and autumn(z>2.58,p<0.01)are the most prominent among the four seasons,with an increasing occurrence of extreme climates.The model results show that the absolute error(MAE),root mean square error(RMSE)and determination coefficient(R~2)of the ARIMA model are1.1024,1.4203,0.055 respectively;The absolute error(MAE),root mean square error(RMSE)and determination coefficient(R~2)of the NAR neural network model are1.6092,0.3425 and 0.9775,respectively;The absolute error(MAE),root mean square error(RMSE),and determination coefficient(R~2)of the LSTM network model are1.6092,0.4476,0.94,respectively,The NAR neural network model performs quite well,The prediction results of NAR neural network model show that there is no significant change trend in HI index,However,the HI index varies to varying degrees,The average annual HI index will still rise slightly,It can be judged that the urban comfort level of Kunming city will continue to decline in the future.