You are here:
MODELING TOOLS FOR ASSESSING LANDSCAPE-LEVEL DETERMINANTS OF FISH PRODUCTION: EXAMPLES FROM WESTERN OREGON
Citation:
BURNETT, K., J. L. EBERSOLE, D. P. LARSEN, A. STEEL, D. L. STEVENS, D. MILLER, P. LAWSON, B. GRESSWELL, AND C. TORGERSEN. MODELING TOOLS FOR ASSESSING LANDSCAPE-LEVEL DETERMINANTS OF FISH PRODUCTION: EXAMPLES FROM WESTERN OREGON. Presented at Alaska Yukon Kuskokwin Sustainable Salmon Initiative, Anchorage, AK, February 06 - 09, 2007.
Description:
Most studies addressing relationships between salmonids and factors that affect their freshwater production have focused on small areas and short time frames. Limits of understanding gained at fine spatiotemporal scales have become obvious, and aggregating fine-scale information from disparate sources does not offer decision makers the means to adequately manage fish populations. Consequently, the Alaska Yukon Kuskokwim Sustainable Salmon Initiative recognizes the need for approaches to characterize determinants of salmon production at broad scales. Here we discuss modeling tools that have been applied in western Oregon to understand how landscape features and processes may influence juvenile and adult salmonids in freshwater. The primary objectives of the modeling tools are to characterize landscape features and processes and then relate these to fish. Models that are contributing to salmon recovery in Oregon include: 1) expert-opinion models to characterize habitat potential from digital elevation data; 2) statistical models that characterize spatial patterns in and/or relationships among fish, habitat, and landscape characteristics; and 3) dynamic process-based models that propagate disturbances into and through streams then predict effects on fish and habitat across a channel network. The modeling tools vary in many aspects, including input data (random samples vs. spatially contiguous surveys, reach vs. watershed scale, and field data vs. digital elevation models), analytical sophistication, and empirical foundation, and so can accommodate a range of situations. In areas with a history of salmon-related research and monitoring in freshwater, models in the three classes may be developed simultaneously. In areas with less available information, expert-opinion models may be developed first to organize existing knowledge and to generate hypotheses that can guide data collection for statistical and dynamic process-based models.