I've had good luck using a vegetation index as a surrogate for imperviousness, at least in urban areas. The IDRISI GIS manual has a brief discussion of how to do it. I used a leaf-on LANDSAT image, converting it to the NDVI. Then I found the NDVI for several areas of known 0% and 100% imperviousness, averaged the values, and stretched the image between the two averages. There are, of course, lots of assumptions buried in the process. Bare earth and sand are going to come across as impervious, for example. Not a problem in ubanized Arlington, but it won't do very well in farmland.
E-mail for details. One of these days I'll get this written up nicely.
William H. Frost, P.E.
Water Resources Planner
Arlington County, VA DPW
[log in to unmask]
>>> Paul Susca <[log in to unmask]> 02/16/99 03:18PM >>>
I am looking for a practical way to estimate percent impervious area for
wellhead protection areas and water supply watersheds in our state, if
possible from remote sensing data we already have or can easily get.
We have LANDSAT-derived landcover at 30 m resolution (some
landcover categories have imperviousness ranges associated with
them) and high-altitude SPOT black and white images with 10 m
resolution (we could overlay a grid and count squares). Could someone
point me to other means of developing imperious area estimates?
Paul Susca
Drinking Water Source Protection Program
NH Department of Environmental Services
Concord, New Hampshire
USA
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|