Quantitative classification as a tool to show change in an urbanizing watershed

Abstract

Quantitative classification of Landsat Thematic Mapper (TM and ETM+) imagery can provide a very useful medium for showing urban change at a watershed scale. It shows both the nature and the degree of urbanization with a resolution equivalent to that of the original TM image. Estimates of urbanization-related factors such as surface imperviousness and canopy cover can be coupled to illustrate the process of urban growth in the watershed using change vector analysis. When used together and combined with standard GIS data sets such as digital line graphs and digital elevation models, quantitative classification can provide the parameters needed to run standard hydrological models that can be used to describe or to forecast the effects of urbanization on area streams. Watersheds provide a meaningful scale for the analysis of the connections between urbanization and the environmental changes it brings. The study area for this paper is the Tinker's Creek watershed in northeast Ohio, USA. Tinker's Creek is one of the largest and most varied watersheds in the area, and it is subject to significant urbanization stress. Urbanization in the watershed is reflected in four TM images of northeast Ohio over a 15-year period (1984-1999). Quantitative classification provides a rich and detailed picture of urbanization over the time period of the study.