A Multiple‐Release Model to Estimate Route‐Specific and Dam Passage Survival at a Hydroelectric Project

Abstract

AbstractPrevious methods of estimating route‐specific passage and survival probabilities for anadromous salmonids past hydroelectric dams have often failed because of faulty assumptions. We present a robust, multiple‐release model that combines release–recapture methods that are known to solve parts of the overall problem. Release 1 allows estimation of route‐specific passage proportions and relative route‐specific survival probabilities. Releases 2 and 3 provide an estimate of absolute survival through a particular route, which gives estimates of absolute route‐specific survival probabilities and dam survival when combined with release 1. Releases 1 and 4 together provide an estimate of project (dam and pool) survival. Combining information from all four releases gives estimates of pool survival. The method is demonstrated through a 2006 study of sockeye salmon Oncorhynchus nerka passing Rocky Reach Dam on the Columbia River in Washington State. We estimate that the majority of tagged smolts passed via the powerhouse (51.41%, = 2.31%) and the fewest over the spillway (1.71%, = 0.60%). Estimates of route‐specific survival ranged from 0.9657 ( = 0.0220) through the powerhouse to 1.0301 ( = 0.0160) over the spillway and through the bypass system. The estimated overall survival was 0.9794 ( = 0.0264) past the dam and 0.9527 ( = 0.0285) through the pool, so estimated survival through the Rocky Reach Project was 0.9331 ( = 0.0121).