In Search of a Cost‐Effective Approach for Estimating Dam Passage Survival

• Published in: North American journal of fisheries management Vol. 40; no. 4; pp. 865 - 882
• Main Authors: Harnish, Ryan, Skalski, John, Townsend, Rich, Ham, Kenneth
• Format: Journal Article
• Language: English
• Published: United States Taylor and Francis 01-08-2020
• Subjects: BASIC BIOLOGICAL SCIENCES
• ISSN: 0275-5947; 1548-8675
• DOI: 10.1002/nafm.10448

Passage of downstream‐migrating fishes through hydropower dams poses a management issue in rivers throughout the world. As such, regulations have been enacted at many locales to understand and limit fish losses by requiring dam passage survival estimation. However, limiting estimates of survival to the dam and immediate tailrace presents a challenge to researchers. In this study, two alternative release–recapture methods were used to estimate dam passage survival of juvenile salmonids through Lower Granite Dam on the Snake River, Washington, in 2018. One approach, the virtual/paired‐release (ViPRe) model, was extensively used in the Columbia–Snake River basin at federally operated hydroelectric dams during 2010–2014. This existing approach uses three releases of tagged smolts to isolate dam passage survival, defined as survival from the upstream dam face to the tailrace mixing zone 1–2 km downstream of the dam. An alternative approach, the virtual release/dead‐fish correction (ViRDCt) model, uses one live‐release group paired with a release of dead tagged fish at the dam to estimate the same survival parameter. The alternative estimation approaches were tested on two spring out‐migrating stocks, yearling Chinook Salmon Oncorhynchus tshawytscha and steelhead Oncorhynchus mykiss, and on a summer out‐migrating stock of subyearling Chinook Salmon. The alternative estimates for these stocks were all within 1 SE. However, the ViRDCt model produced survival estimates with SEs that were 59% smaller than those of the existing ViPRe model and did so using 42% fewer tagged fish and one less downstream acoustic detection array. Because of the reduced sample size, nearly US$540,000 less in tag costs and 4,000 fewer labor‐hours were required to implement the ViRDCt model compared to the ViPRe model. As such, the ViRDCt model represents a cost‐effective and precise approach to characterizing dam passage survival.