HR: 1340h
AN: B43C-0299    [Abstracts]
TI: Land cover and Urbanization links with stream Biota in the mid-Atlantic USA
AU: * Goetz, S J
EM: sgoetz@whrc.org
AF: Woods Hole Research Center, PO Box 296, Woods Hole, MA 02536 United States
AU: Fiske, G
EM: gfiske@whrc.org
AF: Woods Hole Research Center, PO Box 296, Woods Hole, MA 02536 United States
AB: The amount of roads, parking lots, buildings and other developed land within a watershed, its impervious surface area (ISA), has long been known to impact the biotic health of streams and waterways. Vegetation in riparian zones can reduce the negative impacts of ISA by buffering runoff, filtering pollution, and reducing flow velocities that incise stream channels and transport pollutants. ISA has traditionally been mapped by assigning coefficients to land use categories, a type of 'classify and multiply' approach. This can be substantially improved using digital imagery capable of discriminating fine scale information of the land surface, including built areas and tree cover in riparian buffer zones. We used a combination of Landsat (30m) and IKONOS (4m) satellite imagery to develop accurate maps of subpixel ISA and tree cover across the Chesapeake Bay watershed, an area of highly altered land cover and rapid land use change. We report on analyses of the links between these maps and stream biotic measurements for a wide range of small watersheds across Maryland for which extensive in-stream monitoring has been established as part of the Maryland Biological Stream Survey (MBSS). A regression tree statistical approach was used to determine the relationship between land cover and indices calculated by the MBSS, focusing on benthic indices of biologic integrity (IBI). We also explored the influence of landscape configuration and distance weighting of land cover relative to stream channels and sampling points. Thresholds of impervious and tree cover within the watersheds and riparian buffer zones established using the IKONOS imagery were found to vary more substantially across the diverse range of watersheds within the state, as assessed using the Landsat subpixel maps. Nonetheless, ISA was found to be the primary predictor of stream health, followed by tree cover in riparian buffers and within the watersheds. This work advances the estimation of stream health characteristics in areas where MBSS measurements do not exist, and aids the development of guidelines for management and restoration. We are now expanding the work to include our model projections of future urbanization trends, and will report on preliminary results from this analysis.
UR: http://www.whrc.org/midatlantic
DE: 0410 Biodiversity
DE: 0480 Remote sensing
DE: 0483 Riparian systems (0744, 1856)
DE: 0493 Urban systems
SC: Biogeosciences [B]
MN: Fall Meeting 2005