GIS Based Chemical Fate Modeling: Principles and Applications

Chapter 6

Spatial Statistics and Neighborhood Modeling in GIS

Chemical pollution patterns in space are important in identifying and understanding the sources and sinks of contamination, areas most exposed to substances, and the driving mechanisms of environmental fate and transport. Van den Perk [1] introduces the study of soil and water contamination with specific attention to the issue of spatial patterns. In order to characterize patterns, one needs to understand properties such as spatial self-correlation, the local statistics of a continuous spatial distribution of a variable, and the different values that it assumes in different zones of a region of interest, as well as the density in space of certain elements (points or lines).

Spatial statistics include a broad range of statistical analysis tools implemented in most GIS software. Within spatial statistics, a special role is played by geostatistical techniques. The demarcation between spatial and geo-statistics is often vague; in this chapter, which aims at providing a first introduction, we always speak about spatial statistics in general, including in this category typical geostatistical analyses such as variography and kriging. For further information, the reader is referred to the broad literature available and particularly to the introductory book by Clark [2], to Wackernagel [3] for more advanced topics, and to the excellent resources of Hengl [4] for more practical aspects of geostatistics.

6.1 Variograms: Analyzing ...