Document: AMA-3-88-13

Dispersal, environmental variation, and spatial correlation: Effects on extinction risk of Snake River spring chinook.

STANLEY, A.G.*

University of Washington, Seattle, WA 98195 USA ¹

Abstract:
Available time series data for Snake River spring chinook, recently listed as threatened under the Endangered Species Act, indicate not only large declines in population size, but strong coherence between subpopulations as well. Redd counts for 36 subpopulations show strong positive correlation (0.2-0.8) over the past 20-40 years; however the cause of this spatial correlation (dispersal, environmental variation, or some combination thereof) is currently unknown. I developed a spatial model for Snake River spring chinook to determine how assumptions about the cause of the spatial correlation between subpopulations changes estimates of extinction risk. In the model, each subpopulation has the same growth and variance parameters, calculated from values in the literature and fitted to the available trend data. The probability of dispersing from one subpopulation to another is a function of dispersal rate, source population size, and distance between subpopulations. Environmental variance is modeled as a variance-covariance in growth rate. Dispersal rate, amount of covariance, and the number of initial subpopulations was varied and risk of extinction calculated from Monte Carlo simulations. Parameter sets that did not generate observed levels of spatial correlation between subpopulations were discarded. Extinction risk increased with increasing contribution of environmental variation; higher dispersal rates reduced extinction risk through a rescue effect. The model results indicate that the assumptions regarding the cause of spatial correlation can profoundly influence estimates of extinction risk.

Keywords: chinook salmon, extinction risk, spatial correlation, metapopulation, dispersal

This abstract is being presented at: 4:30 PM in session:
Oral Session #68: Dispersal of Seeds and Fruits.