城市地表形态对热环境的影响——以上海市为例

基于遥感与GIS技术,利用Landsat7-ETM+影像反演地表温度,用社会经济统计数据、土地利用现状数据和道路交通网络数据计算城市景观形态参数以表征地表特征,从行政区(县)、5km间距同心环带和局部区块3个水平上划分空间单元建立数据样本,分析城市地表形态对热环境空间分布格局的影响。结果表明:(1)景观混合度和景观分裂度对地表温度有恒定的负向影响,区块连通性与地表温度负相关;(2)景观分裂度对热环境的影响取决于地类属性:分裂度大的增温地类,地表增温效应弱;分裂度大的降温地类,地表降温效应强;(3)人口密度和经济密度可对地表温度产生恒定正向影响;(4)人口密度、建设用地比例和房屋建筑比例是分布在区(县)尺度、同心圆环尺度和典型区块尺度上影响地表热环境最显著的地表形态要素。

IMPACTS OF URBAN LANDSCAPE FORM ON THERMAL ENVIRONMENT AT MULTI-SPATIAL LEVELS

During rapid urbanization, urban thermal environment as an important aspect of humans living environment has attracted wide attentions in the fields of sustainable urban planning and landscape designing, climate change, geography, etc. Over the past few decades, research methods of urban thermal environment have transitioned from the statistics of the meteorological observation data, regional climate mode, to analysis based on remote sensing technology; the spatial level also gradually changed from the large scope from urban-rural area to urban internal space, analysis topic has turned from urban thermal environment pattern to the driving factors of urban internal thermal environment. However, urban thermal environment fields are still subject to spatial resolution of remote sensing images and limited by the shortage of surface parameters.Most previous researches on this field analyzed the problem only on one spatial scale, and the multi scale features of the correlation are scientific issues with insufficient discussion and confidential conclusion. We selected Shanghai as an case study to analyze the correlation between land surface thermal environment (STE) and urban landscape form (ULF). We used Landsat7 ETM+ images to retrieve urban land surface temperature; socio economic statistics data, land use maps and urban traffic network maps were used for characterizing urban landscape features; and there are three spatial statistic levels in this study,such as administrative district (county) level, 5 km concentric rings level and street blocks level. We employed ERDAS Imagine software to process remote sensing images, used ArcGIS software for spatial data management, and used Fragstats software to calculate urban landscape parameters. The results showed that: (1) the impact of ULF on STE had a certain scale relative dependence, the correlation between ULF and STE at different level presented the same direction but different numeric degree; (2) population density and economic density had a positive and constant effect on land surface temperature (LST); (3) both landscape mixture and landscape division had constant negative effects on LST, the correlation between block connectivity and LST was negative; (4) the impact strength of landscape division on thermal environment was dependent on landscape attribute: the higher division that warming class had, the lower strength land surface warming effect was; the higher division that cooling class had, the more strength land surface cooling effect was; (5) urban population density, the developed land proportion and housing land proportion were produced the most significant effect on urban thermal environment scale at district/county level, concentric ring level and typical blocks level, respectively. It can be concluded that: urban population density mainly determines the macro distribution pattern of urban thermal environment; in the belt of population density, the distribution pattern of urban thermal environment is mainly affected by the proportion of developed land; and housing land proportion is the landscape factor that effects local land surface thermal environment pattern when the proportion of developed land are comparative