Comparison of Two Alternative Approaches for Estimating Dam Passage Survival of Salmon Smolts

Traditional single and paired release-recapture models are incapable of providing unbiased estimates of dam passage survival because their assumptions can never be met. Nevertheless, regulatory requirements mandate the estimation of this important performance measure for migratory fish species passing through hydroprojects. We present a new release-recapture model that uses a virtual release of in-river migrants known to have arrived at the dam, combined with a paired release below the dam, to estimate dam passage survival. Analytical comparisons of the sampling precision of the proposed virtual-paired-release model and an established route-specific model found that the new approach was always more precise for equal release numbers of tagged fish. In a field trial at Rocky Reach Dam, Washington, using smolts of sockeye salmon Oncorhynchus nerka, the new approach estimated dam passage survival with more than twice the precision of the alternative method. The proposed virtual-paired-release design has the potential to provide robust and precise estimates of dam passage survival under a variety of hydroproject scenarios.