Scale Issues and GIS-MCDA

Abstract

Decision making is a process. In many decision situations, the process operates across geographic, temporal, and organizational scales. Every change in scale brings about a new decision/evaluation problem, and consequently, the results of GIS-MCDA are scale dependent. The scale consists of extent (observational or geographic scale) and resolution (measurement scale). Each of these scales has spatial and temporal dimensions. Multiscale GIS-MCDA approaches tend to focus on issues of spatial scale and, in particular, the influence of resolution on the results of GIS-MCDA. This can be attributed to two interrelated issues of scale in spatial analysis. First, geographic (areal) datasets collected for analyzing decision situations (e.g., spatial patterns) may have little or no connection to the underlying decision making processes. The GIS-MCDA procedure executed on such data is not independent of how these areal units are configured. This situation is referred to as the modifiable areal unit problem (MAUP). MAUP has two components: (i) the scale effect; that is, the sensitivity of GIS-MCDA results due to the changing number (size) of spatial units of analysis, and (ii) the zoning effect; that is, the results of GIS-MCDA are influenced by the changing shape of the spatial units.Second, a major source of scale effects is spatial heterogeneity. Accordingly, the methods for analyzing the effects are based on the premise that the variability of geographic data is scale dependent and the processes underlying a spatial pattern are best represented at the scale characterized by maximum variability. In the context of decision/evaluation problems, the measure of variability should be related to the main components of multicriteria analysis: the value function (the criterion standardization procedure) and the criterion weights. We suggest that the criterion range value can provide a measure for identifying the most appropriate scale of GIS-MCDA. This Chapter will discuss the two issues of scale in the context of GIS-MCDA. It will also provide an overview of approaches for tackling multiscale decision situations using GIS-MCDA.