Solving spatio-temporal non-stationarity in raster database with fuzzy logic

Abstract

Vegetation dynamic in semi-arid regions is widely affected by climate and hydrological regimes. These regimes create a patchy landscape which, therefore, cannot be treated as stationary in space. Overall complexity significantly increases when the temporal component is added to the analysis. However, despite the complexity involved with variation in time, most of the temporal data storage in raster data-models is done using series of snapshot layers associated with a particular time event. The reason is mainly associated with resolution problems and limitations in controlling large databases with computer resources. Recently, the combination between spatial and temporal databases has become crucial in GIS modeling for solving spatial dynamic problems. We suggest hereby the use of fuzzy theory to apply stationary rules to non-stationary environment in raster database. The method is demonstrated as an analysis of vegetation dynamics in a semi arid environment. The results display the simplicity of the combination between spatially explicit rules and raster database which allow our method to analyze ongoing processes in cell ij at the same time resolution. © Springer-Verlag 2006.