Study On The Characteristics And Prediction Of Urban Heat Island Effect In The Capital Cities Of Five Northwestern Provinces

Global temperature rise and rapid urbanization have led to changes in urban climate,and the urban heat island effect is one of the most typical urban climate changes.As an important part of the "Belt and Road Initiative" construction,the capital cities of five provinces in Northwest China,The rapid development in recent years has made its urban heat island effect increasingly prominent,and the sustainable development of the city has been seriously challenged.Therefore,this study takes the capital cities of five provinces in Northwest China as the object of study.Based on 2005 and 2020 Landsat image series and 2005-2020 MODIS geothermal data.Firstly,the land surface temperature characteristics for 2005-2020 are obtained by pre-processing,and secondly,the surface urban heat island intensity for 2005-2020 is estimated,and the spatial and temporal variation characteristics of heat islands in the capital cities of five provinces in Northwest China are analyzed.And analysis of urban heat island trends through Sen+Mann-Kendall trends;Correlation is used to analyze the directionality of urban heat island influences and a geodetector model is used to reveal the relationship between surface urban heat island intensity and drivers.Finally,the Cellular Automata-Artificial Neural Network model was used to predict the surface urban heat island intensity in 2035,then by comparing the heat island effect of five cities.The results show that:(1)The LST in Xi’an,Yinchuan and Urumqi is on a downward trend,while Lanzhou and Xining are on an upward trend;The five cities with the largest proportion of LST area in 2005 are sub-high temperature zones,and Xi’an and Xining transform into medium temperature zones in 2020.(2)The multi-year average of five provincial capitals in Northwest China from2005 to 2020,Xi’an and Urumqi,showed no heat island,Yinchuan and Lanzhou showed a stronger heat island,and Xining showed a stronger cold island.The highest percentage of heat island classes in 2005 Xi’an was a strong cold island,Yinchuan and Lanzhou were strong heat islands,Xining was no heat island,and Urumqi was a weak heat island;The highest percentage of heat island class in 2020 Xi’an is a strong cold island,Yinchuan and Xining are strong heat islands,and Lanzhou and Urumqi are stronger heat islands.The largest area share of the SUHII trend analysis area from 2005 to 2020 was no significant change in Xi’an,Xining,and Urumqi,a non-significant increase in Yinchuan,and a non-significant decrease in Lanzhou.The area with an increasing trend in Xi’an,Yinchuan,and Urumqi is larger than the area with decreasing trend.(3)NDBI,UVAI,GDP,and population density are positively correlated with SUHII.Five urban building sites and bare land among land cover types are positively correlated with urban heat island effect,and woodland and grassland are negatively correlated.Factor detection showed that the factors with greater influence on SUHII were DEM,UVAI,and land cover type in Xi’an,NDBI,DEM,and land cover type in Yinchuan,NDBI,UVAI,and DEM in Lanzhou,NDVI,NDBI,and DEM in Xining,and DEM,NDBI,and UVAI in Urumqi.Interaction detection showed that the most influential interaction factors on SUHII were NDVI and DEM in Xi’an,DEM and land cover type in Yinchuan,NDBI and DEM in Lanzhou,NDVI and DEM in Xining,and NDBI and DEM in Urumqi.(4)The overall correctness of the overall value of Kappa notation for the five provincial capitals of Northwest China in the simulation of the SUHII class distribution in 2020 are high and can be used for future SUHII prediction.The forecast results show that 2035,the largest share of area Xi’an is a strong cold island,Yinchuan and Xining is a strong heat island,and Lanzhou and Urumqi is a stronger heat island.Lanzhou,which has the smallest area share,is a weak cold island,while the remaining four cities are all stronger cold islands.(5)The surface temperature and heat island effect in the capital cities of the five northwestern provinces are Yinchuan City > Lanzhou City > Xi’an City >Urumqi City > Xining City in descending order.The five cities are located in different environments,geographical locations,etc.,and their temperature and heat island effects are distributed in different ranges.In comparison to the five cities with the smallest heat island area share in 2005 and 2020,Yinchuan City has the largest heat island area share.