Statistical Programs |
College of Agriculture | University of Idaho |
Seminar Announcement |
"Applied Statistics in Agriculture" |
Modeling absolute survival of
juvenile chinook salmon as a function of
environmental covariates using paired
release-recovery data
Presented By |
Dr. Ken B. Newman |
Statistical Programs College of Agriculture University of Idaho |
Tuesday, March 7 3:30 P. M. Ag. Science 323 |
For over twenty years the U.S. Fish and Wildlife Service has
been making experimental releases of coded-wire- tagged
juvenile chinook salmon at various locations throughout the
Sacramento and San Joaquin river system. A subset of these
releases were paired releases, with one release made at an
upstream location and the other at a downstream location.
For a given release-recovery pair, assuming that recoveries
in different locations and times can be modeled by product
multinomial distributions (Seber 1970), the survival rate
between the upstream and downstream location can be estimated.
Estimates of survival alone are of less interest to fisheries
biologists and water managers, however, than quantifying the
influence of environmental factors on survival. I present an
approach that uses all the paired release-recovery data
simultaneously and models the survival rates as a function of
environmental covariates. Following Cormack (1993) I also
sketch an alternative approach to modeling survival that makes
less stringent assumptions, allowing for fish schooling for
example. Keywords: capture-recapture, survival, logistic regression, quasi-likelihood, overdispersion |
Return to Seminar Listing. |