Abstract
Seasonal change and passing year change of vegetation area for 20 years are investigated by pattern decomposition method and we make a study of vegetation vigor using pattern decomposition coefficients. Pattern decomposition method enable to reduce multi-dimensional satellite data into three-dimensional data which directly related to spectral pattern of standard land cover objects, namely water, vegetation and soil. For analysis of vegetation area change, vegetation area is extracted by classification analysis of land cover types. Validation of classification results was performed using detailed land use data of Nara city in 1985. It is made clear that 1) ratio of evergreen forest to vegetation area in Kansai district is found to be 38% from the seasonal change analysis on 1987 data, 2) the decrease of vegetation area of Kansai district and Keihanna area in Kansai district from 1975 to 1995 are about 10% and 20% of vegetation area on 1975, respectively, 3) vegetation decomposition coefficient represents information from only vegetation area in a pixel while NDVI is affected by the other land cover such as soil in a pixel, and 4) the vegetation vigor change from 1985 to 1995 is strongly related to average temperature change. Vegetation change and vegetation vigor are well analyzed by pattern decomposition method.
