You are here:
A Fuzzy Set Approach To Classification of Land Cover PatternEPA Grant Number: U915360
Title: A Fuzzy Set Approach To Classification of Land Cover Pattern
Investigators: Harrington, Thomas
Institution: Boston University
EPA Project Officer: Boddie, Georgette
Project Period: July 1, 1998 through December 1, 2000
Project Amount: $77,064
RFA: STAR Graduate Fellowships (1998) RFA Text | Recipients Lists
Research Category: Fellowship - Geography , Ecological Indicators/Assessment/Restoration , Academic Fellowships
The objective of this research project is to develop and demonstrate a framework for environmental monitoring of landscape-scale, landcover spatial patterns using fuzzy set theory.
The present approach combines cluster analysis, discriminant analysis, and fuzzy sets to create a multivariate classification of landscapes by land cover pattern. First, I quantify land cover pattern for a set of landscapes using various pattern indicators. Second, I use cluster analysis to group the landscapes by these pattern variables. Third, I use discriminant analysis to determine which indicator variables are significant for distinguishing among landscape groups. Fourth, I use fuzzy set methods to redefine the hard or crisp boundaries between groups based on general knowledge of pattern-ecology relationships. The result is a classification of landscapes by their degree of membership in different landscape-pattern groups. The way that the degree of membership of landscapes in pattern groups changes through time or between landscapes provides a framework for environmental monitoring. The utility of this framework is demonstrated in a study of an industrial forest landscape in Maine.