Paper No. 20
Presentation Time: 2:00 PM
WATER IN THE WEST: UTILIZING LANDSAT IMAGERY, GIS, AND STATISTICAL MODELS FOR MAPPING CRITICAL WETLANDS IN NORTHERN COLORADO’S CACHE LA POUDRE WATERSHED
CHIGNELL, Stephen1, SKACH, Sky
2, KESSENICH, Brenda
3, WEIMER, Amber
4, LUIZZA, Matthew
2, BIRTWISTLE, Amy
2, EVANGELISTA, Paul
5 and YOUNG, Nick
5, (1)Ecosystem Science and Sustainability, Colorado State University, 815 W. Magnolia St, Fort Collins, CO 80521, (2)Graduate Degree Program in Ecology, Colorado State University, Fort Collins, CO 80523, (3)Chemistry/Geology, University of Colorado, Boulder, Boulder, CO, (4)Botany, Colorado State University, Fort Collins, CO 80523, (5)Natural Resource Ecology Laboratory, Colorado State University, Fort Collins, CO 80523, steve.chignell@colostate.edu
The Cache la Poudre River and watershed is one of the most important headwater systems on the Colorado Front Range, providing numerous ecosystem and economic services to the region before flowing into the South Platte and ultimately the Missouri River. Because of the watershed’s significance to the surrounding area, monitoring the effects of climate change, land development, and disasters like the 2012 High Park Fire are paramount. However, wetlands and riparian areas—both vital components of watershed health and water quality—have remained largely unmapped in this key watershed. This is particularly true of high-elevation areas, where an estimated 0% of fens and alpine wet meadows have been recorded. The establishment of baseline data to assess disaster impacts and long-term effects of development and ecological change over time is critical for researchers and natural resource stewards in the area.
This study utilized Landsat imagery, ancillary GIS layers, and boosted regression trees modeling for predicting riparian and palustrine wetlands across the Cache la Poudre watershed. Because wetlands and riparian flora vary along an elevation gradient, the study also tested the effectiveness of modeling the watershed within three elevation zones. This proved successful, generating model accuracies of up to 91.3% and Kappa statistics of up to 0.8261. The models successfully differentiated between open water and palustrine areas, and also identified many previously unmapped wetlands, including those in alpine zones. These results will be field-validated and serve as a baseline dataset for local researchers and land managers. Furthermore, the study developed an efficient and reproducible methodology for modeling wetlands that can be applied to other watersheds across the Intermountain West.
