Model output predicting palustrine wetlands locations in the Cache la Poudre watershed, Northern Colorado, generated from three boosted regression trees models using Landsat 5 Thematic Mapper (TM) reflectance and derived indices (Multiple scene time series July 2003 - Nov. 2010), as well as ancillary GIS layers.
Team Location: North Central Climate Science Center, Fort Collins, Colorado
Authors:
Stephen Chignell (Colorado State University)
Brenda Kessenich (University of Colorado Boulder)
Sky Skach (Colorado State University)
Amber Weimer (Colorado State University)
Mentors/Advisers:
Paul Evangelista, Ph.D. (Natural Resource Ecology Laboratory, CSU)
Jeff Morisette, Ph.D. (North Central Climate Science Center, CSU)
Past/Other Contributors:
Tony Cheng, Ph.D. (Forest and Rangeland Stewardship, CSU)
Stephanie Kampf, Ph.D. (Department of Ecosystem Science and Sustainability, CSU)
Dave Merritt, Ph.D. (U.S. Forest Service)
Jeremy Sueltenfuss (Colorado Natural Heritage Program, CSU)
Nicholas Young (Natural Resource Ecology Laboratory, CSU)
Melinda Laituri, Ph.D. (Geospatial Centroid, CSU)
Abstract:
The Cache la Poudre watershed is one of the most important headwaters on the Colorado Front Range and provides important ecosystem and economic services to the region before flowing into the South Platte and ultimately, the Missouri River. Wetlands are an important component of watershed health and water quality but are largely unmapped in the Cache la Poudre watershed. Because of the watershed’s importance to the surrounding area, disasters like the 2012 High Park Fire are of particular concern to community members. Establishing immediate baseline data to assess impacts of disasters as well as long-term impacts of development and ecological change over time is critically important for natural resource stewards. Utilizing remote sensing, geographic information system (GIS) layers, and boosted regression trees modeling, the team conducted the second stage of a multi-term investigation into riparian and palustrine wetland modeling within the region. By modeling the watershed in three distinct elevation zones, and including additional predictor variables, the previous baseline model wasåÊ both expanded upon and refined. The project provided important data for land managers and created a modeling framework that can be reproduced throughout the Intermountain West. Data and end products from the project will be managed and disseminated by the Geospatial Centroid at Colorado State University.