ESTIMATING SEASON-WIDE SURVIVAL RATES OF OUTMIGRATING SALMON SMOLT INTHE SNAKE RIVER, WASHINGTON

Standard release-recapture models can provide release-specific estimat es of survival probabilities for a group of salmonid smelt released at a particular time and place in the river. However, reliable estimates of season-wide survival for the population of outmigrating smelt are needed in the Snake-Columbia River Basin for careful management of the resource. Alternative estimators are presented to estimate season-wid e survival of spring chinook salmon (Oncorhynchus tshawytscha) smelt. Using daily fish tagging, survival for the middle 95% of the migration was estimated to be (S) over cap(s) = 0.873 (SE = 0.005) from the tai lrace of Lower Granite Dam (RK 695) to the tailrace of Little Goose Da m (RK 635) in 1995. Daily survival estimates were remarkably stable ac ross the migration season with some evidence of decreased survival tow ards the very end of the migration. Sample size calculations suggest g ood precision can be attained (i.e., projected SE = 0.01) with tag rel eases as small as n = 500 fish per day (d = 7) across the outmigration . Less than daily sampling can result in season-wide survival estimate s that are too imprecise for many management purposes.