Modeling the Environmental Protection Agency's Level IV Ecoregions within the Klamath Mountains of Southern Oregon and Northern California: A Geographic Information System Approach

Master of Science, Geography, Oregon State University, Fall 2004
Emphasis in Geographic Information Science, Minor in Physical Geography

Graduate committee: D. Wright, A.J. Kimerling, D. White (EPA), A. Woods (EPA)

Chris Zanger
Dept of Geosciences, Oregon State Univ
Corvallis, OR 97331-5506
zanger-at-gmail.com

Abstract.
Ecoregions are regions of relative homogeneity with respect to specific ecosystem variables (Bailey 1976; Omernik 1995). There has been an increasing awareness that effective management of environmental resources must be undertaken with an ecosystem perspective (Omernik, 1995). Ecoregions serve as a spatial framework for assessing, managing, and monitoring ecosystems that have gained recognition among scientists and resource managers as a more effective boundary for natural resource assessment and management compared to the more arbitrary nature of political boundaries (Bryce et al, 1999). Each region can be viewed as a discrete system that is representative of the interaction between geology, landforms, soils, vegetation, primarily produced by qualitative analyses for boundary delineations, which are often derived by consensus with indistinct weighting of input variables, and are essentially impossible to replicate by others. This paper presents the process and results of using multiple spatial analysis techniques within a geographic information system (GIS) as a potentially more quantitative and transparent tool for delineating ecoregions. Due to the qualitative nature of prior EPA ecoregion delineations and in an effort to quantify accuracy, this research focused on emulating the existing large-scale ecoregions nested within the Oregon portion of the EPA's Klamath Mountain Ecoregion, straddling southern Oregon and northern California. Results indicate that replicating qualitative (EPA) ecoregions with a more quantitative process such as those within a GIS has potential, but is currently problematic, with accuracy that is low (~37%) when compared to the original delineations. Increases in GIS data quantity, accuracy, resolution, and attribute richness will improve quantitative modeling potential. Additionally, quantitative EPA Level IV ecoregion replication that is highly similar to existing ecoregions will require extensive collaboration between researchers and the original manual delineation geographers, with particular attention focused on delineation process in areas with broad ecotones.

Download Thesis (19.3 Mb PDF file)
Also available in the ScholarsArchive@OSU permanent collection

Thesis Defense Presentation

Poster of Cartographic Model (GIS Analytical Flow Chart) (140 Kb PDF file)

Link to EPA Western Ecology Division (Corvallis), Ecoregion Data - FTP Site

Go to DJL to Voyages...