期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：2010
卷号：XXXVIII - Part 8
页码：925-929
出版社：Copernicus Publications
摘要：ASTER TIR data is useful to detect spatial distribution of surface temperature on wide area in high quality. According to surface temperature data around big cities from remote sensing, surface temperatures on urban centers are often lower than those on surrounding suburbs though urban heat island effect occurs in air temperature at near surface. Generally speaking, such patterns of surface temperature are caused by higher heat capacity of buildings, vegetation and water such as street trees, rivers and parks, and shadows of tall buildings. However, 90-m spatial resolution of ASTER TIR limits to analyze contribution of these factors because there are less pure pixels containing building, vegetation or shadow in urban areas. In this study, we tried to reveal the influences of material and condition of land surface to surface temperature data using ASTER data in Tokyo, Japan selected as a typical large city. First, we classified shade, vegetation and water body from ASTER VNIR data with 15-m spatial resolution, and estimated areal fraction of these surface categories in each pixel of TIR data on Tokyo in several seasons. Areal fractions of each surface type were compared with surface temperature in order to figure out influence of surface condition and material to surface temperature. The areal fraction of shade is up to 60% and up to 100% in spring and in winter, respectively, reflecting seasonal change of solar zenith angle. There was no clear relationship between vegetation fraction and surface temperature because surface temperature is less affected by area of vegetation than vegetation types. Areal fraction of water body also has no clear relationship with surface temperature since water temperature depends on depth as well as width of water body. On the other hand, shaded area affected surface temperature though there is not clear correlation between them. The deviation of surface temperature was smaller in the places with larger shaded fraction. The deviation of surface temperature is caused by surface coverage in sunlit areas. These results imply that shadow controls surface temperature equally in spite of the difference of surface coverage. Because surface temperatures in shaded areas are quite similar throughout a given scene, we estimated surface temperature in sunlit area in one pixel assuming the temperature is constant on shaded surface and long-wave radiation is proportional to the sunlit and shaded areal fractions. Surface temperatures on sunlit areas in mixed pixel around city center are as high as those on surrounding suburbs except for in winter. In winter time, estimated surface temperatures on sunlit areas are 5 to 10°C higher than those on suburbs because the assumption which this study applied inflates sunlit surface temperature unrealistically in the places with small sunlit ratios in one pixel. These results suggest that sunlit roofs in city center have higher surface temperature as much as those on residential areas, but the limitation of spatial resolution attenuates spatial pattern of surface temperature in urban areas on remote sensing data
关键词：ASTER; Surface temperature; Shade; Vegetation; Impervious surface; Urban area
