PARENT SESSION
Oral Session 7: GIS / Remote Sensing I.
Presiding: E Ellis and P Valko
Monday, August 2, 8:00 AM to 11:30 AM, Meeting Room B 116.
Derivation of remote sensing indicators of forest status for ecological studies.
Townsend, Philip*,1, Kingdon, Clayton1, Seagle, Steven1, Chastain, Robert2, Sturtevant, Brian3, Foster, Jane1, 1 University of Maryland Center for Environmental Science, Frostburg, MD, USA2 USGS Northern Prairie Wildlife Research Center, Columbia, MO, USA3 USFS North Central Forest Experiment Station, Rhinelander, WI, USA
ABSTRACT- An evolving component of landscape and watershed assessment is the derivation of ecological indicators from widely available remote sensing data. The remote sensing approach is especially desirable because it is spatially comprehensive and potentially provides methods that can be (1) repeated through time to assess change and (2) applied to new study landscapes not in the original analysis. However, to be useful the indicators must be integrated with field data for testing and validation. In this study, we developed a set of remotely sensed ecological indicators designed to assess forest condition as related to nutrient cycling and ecological condition within several mostly forested watersheds in the Mid-Atlantic Highlands. Hyperspectral imagery from Hyperion was used to map to foliar nitrogen concentration (MAE <0.25% N), which was found to correspond directly to historical disturbance patterns and the capacity for watersheds to retain nitrogen (several models were developed; cross-validated R2 > 0.7). We derived a pixel-based forest structure model from field measurements, and determined that combined Landsat and synthetic aperture radar data could be used to predict forest height and a forest vertical complexity variable based on the Gini coefficient (several models were developed: 0.65 < R2 < 0.84). These indicators were applied to a larger study area and were strongly associated with diverse measures of forest condition, ranging from nutrient status to site index to diversity of forest interior birds. Therefore the potential exists to develop remotely sensed indicators that relate to multiple components of forest ecosystems.
Key words: ecological indicators, watershed hydrology, remote sensing, forest structure