Analysis Of The Temporal And Spatial Characteristics And Driving Factors Of Surface Urban Heat Islands In China Based On Time Series Decomposition

Rapid urbanization process occurs and regional climate change is particularly prominent in China since the 21 st century.The urban heat island effect,as one of the manifestations of regional climate change,poses a serious challenge to the urban ecological environment.The urban thermal environment will be further exacerbated by urban heat islands under the background of climate warming and the frequent occurrence of global extreme climate,which pose a serious threat to human health.In recent years,the monitoring of the spatio-temporal variations of urban heat islands and the evaluation of driving factors have been hot topics for researchers and government agencies.In this study,a simple statistical analysis was used to quantify the spatio-temporal trends of surface urban island heat island intensity and driving factors,based on land cover,surface temperature,albedo,vegetation greenness,vegetation continuous field,night light,temperature,precipitation,and elevation data.The Prophet model is applied to characterize the seasonal variation and temporal trend in surface urban heat island intensity in 102 large cities with 100 million in China from 2000 to 2017,and is used to predict the intensity of heat islands in 2030.Finally,spatial autocorrelation analysis,multicollinearity diagnosis and the establishment of least squares regression,stepwise multiple linear regression and geographic weighted regression were used to analyze the relationship between the surface urban heat island intensity and driving factors.The conclusions of this study are as follows:(1)The spatio-temporal trends in surface urban heat island intensity and driving factors were quantified based on linear regression and statistical analysis.The results show that both the intensity and driving factors of surface urban heat islands in China show significant spatial heterogeneity and complexity.During the day,the surface urban heat island intensity is the strongest in summer and the weakest in winter.The surface urban heat island intensity in the southern cities is greater than that in the northern cities.The seasonal variation in the surface urban heat island intensity is small and the spatial variation is obvious at night.The surface urban heat island intensity is obviously higher in northern cities than that of southern cities.The spatial difference in the temporal trend of the surface urban heat island intensity is significant,showing a significant increasing trend in most cities.The driving factors show significant differences in spatio-temporal trends in different cities.(2)The seasonal variation and temporal trend of surface island heat island intensity were analyzed using the Prophet model.The results show that,in terms of diurnal changes,the annual seasonal amplitude is greater during the day than that at night,which is mainly concentrated at 1 ? to 3 ?(the highest frequency appears around 2 ?),and it is concentrated at 0.2 ? to 0.9 ? at night.In terms of seasonal changes,the maximum heat island intensity of 88% of cities occurs in summer,and the minimum heat island intensity of 80% of cities occurs in winter.In terms of interannual changes,78% and 83% of the cities were found to exhibit increasing surface urban heat island intensity during the day and night,respectively.During the day,more cities in the south showed an increasing trend of heat island intensity than those in the north,and the phenomenon is exactly opposite at night.The surface urban heat island intensity will be stronger during both the day and night in China in 2030.(3)The relationship between the surface urban heat island intensity and driving factors is analyzed using least squares regression,stepwise multiple linear regression and geographic weighted regression based on spatial autocorrelation analysis and multicollinearity diagnosis.The results show that the intensity and driving factors(except the urban-rural difference in built-up intensity and the urban-rural difference in nighttime lights)of urban heat islands have significant spatial autocorrelation.In addition,the determination coefficient of the geographic weighted regression model is greater than those of least squares regression and stepwise multiple linear regression,suggesting that the geographical weighted regression is more effective in quantifying the relationship between the surface urban heat island intensity and driving factors,which implies that the relationship of surface urban heat island intensity with driving factors is spatially variable.In summary,this study quantitatively analyzes the seasonal variation and temporal trend of the surface urban heat island intensity,and analyzes the relationship between the surface urban heat island intensity and driving factors on a national scale.The results of this paper can be used to monitor and manage urban heat islands for reference.