Research On The Thermal Environment Effect Of The Landscape Pattern Of Xi'an City

With the rapid development of the global urbanization process,a large number of natural landscape is replaced by the hard water-impermeable surface,the urban thermal environment is deteriorating,the extreme weather process is frequent,and the urban ecological security and the sustainable development capacity are decreasing continuously.Xi ’an is working on the construction of national central city and national forest city,facing the double pressure of economic development and ecological construction.In this background,the study of the landscape pattern of Xi ’an and the quantitative relation with the urban heat environment can provide a scientific reference for the rational planning of the urban landscape pattern and the relationship between the urban economic development and the construction of the ecological environment.Based on the remote sensing image data of 1994,2006 and 2016 in Xi’an city,multi-temporal landscape types and land surface temperature data were obtained by using supervised classification and single window algorithm.By means of dynamic analysis,transfer matrix,landscape index and moving window,spatial regression model and so on,the temporal and spatial variation characteristics of surface landscape and thermal environment in Xi’an and the quantitative relationship between them are studied.The conclusions are as follows:(1)During the study period,woodland,cultivated land and construction land were the main landscape types of Xi’an.The cultivated land area continues to decrease,showing a trend of contraction;woodland,construction land area continues to increase,the main source of increase is the type of arable land.This is related to the urbanization of Xi’an and the continuous improvement of the level of ecological environment construction.From the point of view of landscape pattern,the fragmentation degree of surface landscape in Xi’an from 1994 to 2016 has been improved,the spatial distribution of various landscape types tends to be balanced,and the complexity of structure composition has been enhanced.The main results are as follows:with the increasing of urbanization level and intensity of human activities,the number of patch in construction land increases,the dominant landscape types such as woodland and cultivated land are segmented continuously,the degree of aggregation of natural surface decreases,and the degree of fragmentation increases.(2)Comparing and analyzing the four common inversion methods of surface temperature-split window algorithm,atmospheric correction method,Qin Zhihao single window algorithm and Jim e nez-Mu n oz universal single channel algorithm,it is found that Qin Zhihao window algorithm has the highest inversion accuracy.The best fitting result with land temperature product data MOD11A1 is the best method for inversion of land surface temperature in Xi’an during this research period.(3)Thermal landscape is a thermal region with high spatial heterogeneity,which can clearly describe the spatial variation characteristics of urban thermal environment.The thermal landscape with different temperature characteristics was divided into different thermal patch types,including high temperature patch(heat island patch,strong heat island patch),moderate temperature patch type,and low temperature patch(green island patch).From 1994 to 2006,the area of high temperature patch type in the study area was mainly increased,while the area of low temperature patch type was mainly decreased.From 2006 to 2016,the area of high-temperature and low-temperature patch types changed little,but the proportion of strong heat island and green island increased significantly.There are obvious transformation characteristics among different thermal landscape types.The conversion rate between high-temperature patch and low-temperature patch was low,but the conversion rate between hot(green)island patch and hot(green)island patch was higher in the interior of high-temperature patch than that of low-temperature patch.The conversion rate between the middle temperature patch and the other four thermal patches is higher,which plays a transitional role in the transformation between the thermal landscape types.The fragmentation degree of thermal landscape in the study area decreased gradually,the agglomeration level of each patch type further increased,and the urban heat island effect increased significantly.(4)The thermal environmental effects of different surface landscape types are obviously different.The order of surface temperature of different landscape types is construction land>cultivated land>unused land>grassland>water area>woodland,in the order of construction land>cultivated land>unused land>grassland>water area>woodland.And the construction land and the thermal landscape composition of the forest land are increasing continuously.This is the main cause of the increase in the thermal level in the study area.According to the change of the construction land and the non-construction land,the surface temperature corresponding to the different change process has a significant difference.In which,the "non-constant construction land" and the "non-construction land" have the highest and the lowest average surface temperature,respectively,and the time sequence of the construction land in the same area will affect the surface temperature,and the landscape type in the early stage will affect the thermal environment of the post-landscape type.(5)Through the comparison between the ordinary least squares model and the geographical weighted regression model,the influence of landscape pattern index on the surface temperature shows that the latter has a better fitting effect in the aspect of spatial non-stationarity in regression analysis.It can more accurately characterize the quantitative relationship between landscape pattern and surface temperature.There is also a significant difference in the interpretation rate of land surface temperature among different landscape pattern indices,among which the various types of indices related to the area index have the best fitting effect with the land surface temperature,and the interpretation rate of the surface temperature is the highest.(6)Under the scale of 1200 m and 1200 m,the fitting superiority R2 of the surface temperature data and the main surface landscape type area ratio is the highest.The area proportion of the construction land is increased by 10%,the surface temperature is about 0.55K,and the area ratio of the forest land is increased by 10%.The surface temperature is about 0.95 K.The cooling effect of the forest land is higher than that of the construction land.