Methods For Spatial Analysis

Abstract

Creating, manipulating, and storing the GIS data base are rarely the final end-products of GIS in the sciences. Here, we are interested in understanding and modeling the dynamic processes that generated and shaped the information stored in the data base. Very often the researcher will want to make predictions concerning the process to unobserved locations. Where sampling is expensive, much cheaper auxiliary information sampled more intensely than the quantity of interest may be used to provide estimates of the desired entity. Further, we may be interested in testing various ideas or hypotheses concerning the process. In practice we are unable to entirely sample our region of interest and so our GIS is built using data containing some degree of uncertainty. Thus, it is often convenient to model our spatial process as a stochastic spatial process, permitting our models of the process to accommodate the uncertainty in our measurements and to help explain the observed process. It is commonly accepted that measurements made near in time or in space are much more likely to be alike than would be measurements widely separated in time or space. It is upon this assumption that spatial data analysis methods have been developed. In this chapter, we will address several of the common problems encountered in spatial data analysis as it relates to GIS and will illustrate methods for their analysis.