Abstract
A 'power law' describing a relationship between mean and variance made from count data was found by Taylor (1961), and developed by Hughes and Madden (1993) for studies of phytopathology. This power law was applied for the analysis of land use map, a new statistical method for analysing it was developed in this study. Suppose that N cells (e.g. 100 cells) with equal area were set on a map and each cell was divided into n smaller cells (e.g. 4 cells) with equal area. Then, record all land uses occurred in each smaller cell and estimate the occurrence per larger cell for each land use. When the occurrence,the actual variance per smaller cell and variance per smaller cell calculated under an assumption of random distribution were estimated for each land use, let the logarithmic values paired the assumed and actual variances be plotted on abscissa and ordinate, respectively. Then, we could find a fine linearity between these variables. This relationship is called as 'power law' and indicates the total tendency of geographical heterogeneity inherent in the region on the map. The deviation of each land use from the straight line indicates the geographical heterogeneity inherent in respective land use. An example in Ibaraki Prefecture was analysed using this method.
