Event probability correlation analysis for comparison of two-phase ecological maps

The use of maps as an important tool for investigating ecological, environmental and geographical phenomena has grown in popularity over recent years, mainly due to the development of computerized cartographic systems and geographic information systems (GIS). With the aid of such systems, spatial simulation can be presented in a map form. It is important to find a way to compare the difference between a simulation result and the corresponding surveyed map. However, so far there is no procedure universally agreed upon to determine whether two maps are statistically different, since it is clear that two-phase mosaic maps are completely different from dot maps. In this paper, we present a new method, event probability correlation analysis, to determine spatial similarities in the two-phase ecological maps. A comparison of subtropical savanna spatial patterns is used to explain our method. Our results indicate that the method is very convenient, useful and effective.