Wednesday, September 15, 2010: 3:20 PM
320 (Convention Center)
Aimee H. Fullerton
,
Fish Ecology Division, NOAA Northwest Fisheries Science Center, Seattle, WA
Steve Lindley
,
Fisheries Ecology Division, NOAA Southwest Fisheries Science Center, La Jolla, CA
George Pess
,
Environmental Conservation Division, NOAA Northwest Fisheries Science Center, Seattle, WA
Blake Feist
,
Fish Ecology Division, NOAA Northwest Fisheries Science Center, Seattle, WA
Ashley Steel
,
Statistics, US Forest Service PNW Research Station, Olympia, WA
Paul McElhany
,
Conservation Biology Division, NOAA Northwest Fisheries Science Center, Seattle, WA
We used a graph-theoretic approach to evaluate possible metapopulation structure of anadromous salmonids in the Lower Columbia and Willamette Rivers, USA, under several scenarios. Specifically, we evaluated how spatial structure is affected by (1) fish hatcheries, (2) migration barriers, and (3) catastrophic disturbance events. For spring and fall Chinook salmon (Oncorhynchus tshawytscha), and for winter and summer steelhead (O. mykiss), we identified how these scenarios differ from present-day and historical scenarios. We found dramatically increased connectivity under the hatchery scenario; connectivity was higher than historical levels, but spatial structure was inherently different. For each species, we identified which currently extirpated populations would most improve metapopulation connectivity if they were re-established, and which populations would cause the greatest decrease in metapopulation connectivity if they were lost due to disturbance. These analyses suggest places where passage facilities show the most promise, and populations where conservation should be focused in order to bring levels of connectivity closer to those found historically. Additionally, we conducted uncertainty analyses to evaluate modeling assumptions. We found metapopulation connectivity to be highly sensitive to dispersal rate and to which metric we used to represent population size. These are parameters for which improved population monitoring could provide better estimates.