Estimates of juvenile salmon survival are important data for fishery managers in the Yakima River Basin. Radiotelemetry studies during 2012–14 showed that tagged juvenile Chinook salmon (Oncorhynchus tshawytscha) that passed through the fish bypass at Roza Dam had lower survival than fish that passed through other routes at the dam. That study also identified flow-survival relationships in the reaches between the Roza Dam tailrace and Sunnyside Dam. During 2012–14, survival also was estimated through reaches downstream of Sunnyside Dam, but generally, sample sizes were low and the estimates were imprecise. In 2016, we conducted an evaluation using acoustic cameras and acoustic telemetry to build on information collected during the previous study. The goal of the 2016 research was to identify areas where mortality occurs in the fish bypass at Roza Dam, and to estimate reach-specific survival in reaches downstream of the dam. The 2016 study included juvenile Chinook salmon and coho salmon (O. kisutch).

Three acoustic cameras were used to observe fish behavior (1) near the entrances to the fish bypass, (2) at a midway point in the fish bypass (convergence vault), and (3) at the bypass outfall. In total, 504 hours of acoustic camera footage was collected at these locations. We determined that smolt-sized fish (95–170 millimeters [mm]) were present in the highest proportions at each location, but predator-sized fish (greater than 250 mm) also were present at each site. Fish presence generally peaked during nighttime hours and crepuscular periods, and was low during daytime hours. In the convergence vault, smolt-sized fish exhibited holding behavior patterns, which may explain why some fish delayed while passing through the bypass.

Some of the acoustic-tagged fish were delayed in the fish bypass following release, but there was no evidence to suggest that they experienced higher mortality than fish that were released at the bypass outfall or downstream of the dam. Most of the tagged fish that were released in the fish bypass moved downstream and re-entered the river within 12 hours, but 9.8 percent of the Chinook salmon and 15.7 percent of the coho salmon remained in the bypass for 2.5–17.4 days. We developed a set of models for Chinook salmon and coho salmon and used model selection to determine if release site was an important predictor of survival of tagged fish. The models that provided the best fit to the Chinook salmon and coho salmon datasets did not include release site as a covariate. Furthermore, survival estimates for groups of fish from the various release sites were nearly identical for both species. Based on these observations, it appears that passage through the fish bypass did not result in increased mortality relative to groups of fish released downstream of the bypass.

Juvenile Chinook salmon migrated downstream faster than juvenile coho salmon and survival for each species varied with release timing. Median travel time from release at Roza Dam to arrival at a detection gate located at river kilometer (rkm) 527.8 on the Columbia River was 15.4 days for Chinook salmon and 37.4 days for coho salmon. Cumulative survival from Roza Dam to the Columbia River detection gate ranged from 0.299 to 0.678 for Chinook salmon, and from 0.321 to 0.627 for coho salmon. Survival was highest for both species when tagged fish were released in mid-April and lowest when tagged fish were released in early-May. Reach-specific survival estimates were standardized to create estimates that described survival per 100 rkm, which showed that survival was very low (less than 0.500) for some release groups, particularly in the Roza, Sunnyside, and Chandler diversion reaches. A more extensive analysis of reach-specific survival is planned for this dataset, which should provide insights into covariates that affected survival during 2016.