Empirical analysis for estimating land use transition potential functions—case in the Tokyo metropolitan region

Abstract

Although some CA based models have been applied to the real areas, only a few of them were developed based on the rigorous empirical analysis of the relationships between the factors which might affect the land use transition of a cell and the transition potentials from its present land use to other land uses. Additionally, in 1998 a new data set which includes the land use data of each 10 2 square cell in the Tokyo Metropolitan Region became available to us in Japan. In this study, aiming at developing a CA-based land use model which can depict changes in land use of each 100 m 2 block and will be useful for planning works at the local governments, we carried out two types of statistical analyses, namely regression analysis and discriminant analysis, by using the above-mentioned land use data set in order to identify significant explanatory factors of the transition potentials. The study area is a suburban district of Tokyo, which is a 10 km square area. Additionally, we assumed that aggregated land use state in the neighbourhood of a cell and its accessibility to transport services may affect its transition potentials, and we grouped the active land uses into eight categories. The result of the regression analysis suggested that there is a strong linear relationship between the transition probability of a block from the non-urban land use state to the urban land use state and the number of urban land use blocks within the Moore neighbourhood of the block. By the discriminant analysis we could picked out the factors which primarily affected the changes in land use of each block.