Publication Date
4-9-2007
Abstract
The Impervious Cover Model (ICM) has been identified as a useful tool for evaluating stormwater impaired watersheds in New England. The Center for Watershed Protection and other investigators have established a strong correlation between increasing impervious cover and decreasing quality of stream physical, hydrologic, water quality, and biological function. In 2005, EPA Region 1 successfully conducted a set of pilot TMDL projects demonstrating the utility of the ICM for supporting TMDL assessment. In each pilot watershed, %IC was estimated by subbasin and used to efficiently support TMDL development. Presently, low quality impervious cover data for impaired watersheds limits the accuracy and the value of the ICM for conducting stormwater TMDLs. Impervious cover estimates are typically obtained using out-of-date land cover data and rough empirical estimates of %IC for each type of land cover. The resulting impervious cover maps are on a gross scale with a high level of uncertainty. Remote sensing can provide accurate, site-specific characterization of impervious cover and can support efficient identification, prioritization, and removal of stormwater sources. The authors will demonstrate the use of high resolution QuickBird satellite imagery to accurately map impervious surface cover for a study area in Arapahoe County, Colorado. This multispectral imagery is particularly well-suited for mapping impervious surfaces. It has 2.4 meter spatial resolution for its 4-band multispectral data and 0.60 meter spatial resolution for its panchromatic data. Two image processing techniques for impervious cover characterization will be described. One technique utilized a whole pixel unsupervised classification method and the other used a pan-sharpened classification approach. Results of an accuracy assessment and compelling examples will be provided. The applicability of these remote sensing techniques for improving stormwater TMDL development and implementation in New England will be discussed.