In recent decades,the significant economic development witnessed in China has been accompanied by the acceleration of urbanization.However,the urbanization process characterized by land-use change has posed a major threat to the green sustainable development of the ecological environment.Especially,the social inequality caused by the imbalance of the urban ecological environment has seriously restricted the high-quality development of livable cities in the future.Previous studies on urban ecological quality mainly focused on general evaluation at the city level,while ignoring the spatial heterogeneity of urban ecological quality inside cities and failed to investigate at different scales.The heterogeneity of urban ecological quality is strongly related to urban land-use.Therefore,taking the downtown area of Xi’an as an example,based on multi-source remote sensing data and social sensing data,we aim to quantify the influence of urban land-use on urban ecological quality at different spatial scales,including impact intensity,impact characteristics,impact rules,etc.,which can provide a scientific basis for sustainable development in rapidly urbanizing areas.Firstly,based on Point of Interest(POI)data,Open Street Map(OSM)data,and Gao Fen-2(GF-2)PMS imagery,three layers of semantic objects(ie point,line,and polygon semantic objects)are constructed,so that a data-driven hierarchical semantic object network(HSON)model for urban land-use classification is proposed.In the HSON model,based on the term frequency-inverse document frequency(TF-IDF)algorithm,a semantic matching-based category mapping(SCM)model is constructed to integrate point,line,and polygon semantic objects.Thus,an application-oriented high-precision urban land-use map is obtained.Secondly,the urban ecological quality is characterized based on the Sentinel-2 MSI image,Landsat 8 TIRS image,and the improved remote sensing-based ecological index(RSEI)combining with principal component analysis.Thirdly,the relationship between urban land-use patterns and urban ecological quality is quantified based on the random forest(RF)regression model at different spatial scales,including 4 administrative scales(i.e.city scale,district scale,street scale,and parcel scale)and 25 geographic scales(i.e.100 m x 100 m,200 m x 200 m...2500 m x 2500m).Finally,integrating all the research results,some policies and suggestions for the coordinated development of sustainable urbanization and ecological environment are proposed to deepen the understanding of the problem of rapid urbanization.Overall,the main conclusions of this article can be summarized as follows:(1)Firstly,the accuracy assessment based on the confusion matrix found that the overall accuracy and kappa coefficient of the HSON model was 90.62%and 0.8817,respectively,indicating that the HSON model had an excellent effect on urban land-use identification.Secondly,the downtown area of Xi’an,there was a common phenomenon of the interaction of various social and economic functions in the parcels,indicating the complexity of urban land-use.And the mixed degree of land functions in the southern area of Xi’an was higher than that in the northern area,which indicated that the development center of Xi’an was still inclined toward the southern area.Thirdly,different land-use types showed significant spatial heterogeneity,showing a development pattern radiating from the center to the surrounding areas.Among them,the top three land-use types with the largest area were residential land(275 km~2),green space and open space land(250 km~2),cultivated land(137 km~2),accounting for 29.13%,15.89%,and 14.5%of the total area of the study area,respectively.In addition,different urban administrative districts showed the different focus of urban development,such as industry,commerce or education.(2)In 2018,the mean value of RSEI in the downtown area of Xi’an was 0.47,indicating the moderately low level of its urban ecological quality.From the perspective of spatial distribution,most of the regions in the study area showed a state of low urban ecological quality,while the regions with better urban ecological quality were mainly distributed in the southeastern and southwestern fringes area.For different urban administrative districts,the urban ecological quality of Baqiao District was best,followed by Chang’an District.The Lianhu District had the worst urban ecological quality,followed by Xincheng District and Beilin District.In the downtown area of Xi’an,the area with a low level of urban ecological quality(i.e.good or excellent level)was as high as 521.83 km~2(56.41%),while the area with a high level of urban ecological quality(i.e.poor and bad level)was only 245.32%km~2(26.52%).In general,the urban ecological quality of Xi’an presented a spatial pattern that the closer it was to the city center,the worse the urban ecology was.(3)At different spatial scales,the impact of land-use patterns on urban ecological quality showed significant spatial differences.At the 4 administrative scales,the land-use pattern at the street scale had the strongest influence on urban ecological quality;At the 25 geographic scales,the land-use pattern within 1900 m both showed an important impact on the local urban ecological quality.For different land-use types,residential land,commercial land,green space and open space land have the strongest impact on urban ecological quality,while special land and wetlands have the weakest impact.In addition,the influence rules of land-use pattern on urban ecological quality can be visualized through the tree structure in RF,thus can assist urban land-use planning to balance and improve urban ecological quality.Finally,altogether the results,we provided some reasonable policies and suggestions for the coordinated development of sustainable urbanization and ecological environment,which was expected to promote regional coordinated sustainable urbanization and the overall development of the ecological environment. |