Quantifying human disturbance in watersheds: Variable selection and performance of a GIS-based disturbance index for predicting the biological condition of perennial streams

Ecological Indicators
By: , and 

Links

Abstract

Characterizing the relative severity of human disturbance in watersheds is often part of stream assessments and is frequently done with the aid of Geographic Information System (GIS)-derived data. However, the choice of variables and how they are used to quantify disturbance are often subjective. In this study, we developed a number of disturbance indices by testing sets of variables, scoring methods, and weightings of 33 potential disturbance factors derived from readily available GIS data. The indices were calibrated using 770 watersheds located in the western United States for which the severity of disturbance had previously been classified from detailed local data by the United States Environmental Protection Agency (USEPA) Environmental Monitoring and Assessment Program (EMAP). The indices were calibrated by determining which variable or variable combinations and aggregation method best differentiated between least- and most-disturbed sites. Indices composed of several variables performed better than any individual variable, and best results came from a threshold method of scoring using six uncorrelated variables: housing unit density, road density, pesticide application, dam storage, land cover along a mainstem buffer, and distance to nearest canal/pipeline. The final index was validated with 192 withheld watersheds and correctly classified about two-thirds (68%) of least- and most-disturbed sites. These results provide information about the potential for using a disturbance index as a screening tool for a priori ranking of watersheds at a regional/national scale, and which landscape variables and methods of combination may be most helpful in doing so.

Additional publication details

Publication type Article
Publication Subtype Journal Article
Title Quantifying human disturbance in watersheds: Variable selection and performance of a GIS-based disturbance index for predicting the biological condition of perennial streams
Series title Ecological Indicators
DOI 10.1016/j.ecolind.2009.05.005
Volume 10
Issue 2
Year Published 2010
Language English
Publisher Elsevier
Contributing office(s) National Water Quality Assessment Program
Description 10 p.
First page 264
Last page 273
Other Geospatial Western United States