HR: 0830h
AN: B41B-0876 INVITED [PDF]
TI: Stream Health Linkages with High Resolution Land Cover and Landscape Configuration Metrics
AU: * Goetz, S J
EM: sgoetz@whrc.org
AF: Woods Hole Research Center, PO Box 296, Woods Hole, MA 02543-0296 United States
AU: * Goetz, S J
EM: sgoetz@whrc.org
AF: University of Maryland, Dept of Geography, College Park, MA 20742-8225 United States
AU: Wright, R
EM: rwright@geog.umd.edu
AF: University of Maryland, Dept of Geography, College Park, MA 20742-8225 United States
AU: Snyder, M
EM: msnyder@geog.umd.edu
AF: University of Maryland, Dept of Geography, College Park, MA 20742-8225 United States
AB:
The amount of impervious surface area (ISA, e.g. roads, parking lots, buildings) within a watershed has long been known to
impact the quality of nearby streams and waterways, as measured by their chemical and biotic composition. Vegetation in
riparian zones can reduce the negative impacts of ISA by buffering runoff and reducing flow velocities that incise stream
channels. Methods to map impervious areas and stream buffers are therefore of great interest to a broad community of natural
resource managers. Traditionally, stream buffers and ISA were mapped using visually interpreted aerial photographs and, for
example, assigning coefficients to land use categories - a type of "classify and multiply" approach. We have developed
methods, using decision tree algorithms, to map these critical landscape variables using high resolution satellite imagery
(4m Ikonos). Our approach provides highly accurate maps of impervious surfaces and tree cover. We report on analyses of the
links between these maps and stream health for 246 small watersheds within a 1300 km2 area of the mid-Atlantic region - an
area of highly altered land cover and rapid land use change. Impervious surface area was found to be the primary predictor
of stream health, followed by tree cover in riparian buffers, and total tree cover within entire watersheds. A number of
issues associated with mapping using Ikonos imagery were encountered, including differences in phenological and atmospheric
conditions, shadowing within canopies and between scene elements, and limited spectral discrimination of cover types. We
report on both the capabilities and limitations of Ikonos imagery for these applications, and considerations for extending
these analyses to other areas.
DE: 1640 Remote sensing
DE: 1860 Runoff and streamflow
DE: 1884 Water supply
DE: 1894 Instruments and techniques
DE: 6605 Education
SC: Biogeosciences [B]
MN: 2003 Fall Meeting