A Study On Urban Heat Island Effect And Causes Supported By Google Earth Engine

As a national central city,Zhengzhou City has experienced significant changes in the structure and type of urban substrates in recent years due to the accelerated urban expansion and the increase in high-rise buildings,which has led to an increase in urban land surface temperature and a serious urban heat island effect.In this paper,the spatial-temporal changes in land use types and land surface temperature were investigated using the Google Earth Engine in the study area,using Landsat,ERA5＿Land,NCEP datasets from 2000 to 2020,to find the correlation between land use types and the urban heat island effect,and give measures to mitigate the urban heat island effect.(1)Using Landsat images,the random forest algorithm provided by GEE was used to obtain the spatial-temporal changes of various land types in the study area between2000 and 2020,in which the area of grassland increased from 16.21km~2 to 131.54km~2,while forest,water bodies and unused land changed less.The change in land type is mainly reflected in the roll-in of built-up land and the roll-in of cropland.(2)Three land surface temperature retrieval algorithms were implemented based on Landsat images,ERA5＿Land reanalysis dataset and NCEP dataset,the most suitable inversion algorithm was selected after accuracy verification,the thermal landscape was classed using the mean-standard deviation method based on the retrieval results,and Overlay analysis of land use types with land surface temperature and thermal landscape classes.The average summer temperatures of the study area from 2000 to 2020 were the highest for built-up land and unused land,and the lowest for water bodies;water bodies and forest were distributed in the strong cold island zone,the stronger cold island zone and the weaker cold island zone,showing the"cold island effect";built-up land and unused land were concentrated in the weak heat island zone and above,cropland was distributed in the weak cold island zone and the normal zone,while the thermal landscape classes of grassland remains consistent with forest.(3)Statistics on the area of heat landscape classes in each administrative district for five years between 2000 and 2020,and continuous analysis of the migration of heat island areas and the centre of gravity for construction land,led to the following conclusions:in2000,the spatial distribution of heat island classes was dominated by Erqi District,Guancheng Huizu District and Jinshui District,while from 2004 to 2020,the distribution changed to Erqi District,Guancheng Huizu District and Zhongyuan District.The centre of gravity of construction land continues to shift from the urban centre to the periphery in the administrative district,with the centre of gravity of the heat island area shifting from the urban periphery to the centre from 2000-2004,and from the urban centre to the periphery from 2004-2013.In the period 2013-2020,the centre of gravity of the heat island area shifted from the urban periphery to the centre again,indicating that the heat island effect in the study area has diminished during this period.(4)The factors affecting the spatial heterogeneity of land surface temperature were investigated using geographic detector,and it was found that single-factor exploration of built-up and vegetation explained more than 40%of the spatial heterogeneity of land surface temperature;while the results of multi-factor interaction exploration showed that the interaction between human activities and built-up land explained 54.6%of the spatial heterogeneity of land surface temperature,the interactive collocation with the participation of built-up land had more explanatory power on the study area The interaction between human activities and built-up land explained 54.6% of the spatial heterogeneity of land surface temperature.