Snake River Fall Chinook Salmon research and monitoring

Abstract

In Chapter 1, we report on development and application of an integrated population model (IPM) for the natural-origin fall Chinook salmon population upstream of Lower Granite Dam. This year’s efforts represent the third update to the model. Initial efforts focused on generating juvenile and adult abundance estimates, with estimates of uncertainty, for informing the life-cycle model and estimating the effects of covariates on key demographic parameters. The goals of this year’s report are to 1) describe the modifications and advances made since the previous report, 2) to annually update and report the abundance estimates and other quantities used in the model, 3) to provide annual estimates of population parameters estimated by the IPM, and 4) to outline the next year’s tasks for advancing and/or applying the model. Since our last report on the life-cycle model, we have made a number of changes including: 1) incorporating jack abundance and age-structure data into the observation model, 2) changing smolt-to-adult survival (SAR) for subyearling and yearling to partial SARs that represent the joint probability surviving and entering the ocean at a given juvenile age, 3) combining age categories for rarely observed ages, 4) using scale data from unmarked fish to estimate age structure, and 5) generating composite life-cycle demographic parameters (cumulative capacity and productivity) from stage-specific parameters. We also generated juvenile abundance estimates, extended the model to include three additional brood years (1992– 2021), and ran the model to forecast returns to Lower Granite in 2022. For posterior medians of life stage-specific parameters, we estimated a mean productivity of 438 natural-origin juvenile recruits per female spawner, a capacity of 1.36 million juveniles, and a mean smolt-to-adult survival (SAR) of 1.2%. We detected strong density-dependent regulation, with juvenile recruits per spawner declining to about 50 juvenile recruits per female spawner at high spawner abundance. Across the entire life cycle, these stage-specific parameters resulted in a median cumulative intrinsic productivity of 1.93 adult female recruits per female spawner and a median equilibrium abundance of 2,851 female spawners (7,842 total spawners). Annual juvenile productivity varied from about 250–1,000 juveniles per spawner but displayed no temporal trends or patterns. For the three most recent brood years added to the model, recruits per spawner were higher than average but well within the range of uncertainty observed over the entire time series. In contrast to juvenile recruitment variability, SAR varied considerably among years and exhibited two periods of high survival (1996–2001 and 2007–2012) when SAR ranged from 2% to 6% and cumulative productivity ranged from 2 to 8 recruits per spawner. Partial SARs revealed that yearling outmigrants contributed substantially to the high SARs in the first high-survival period, but the second period was dominated by subyearlings. Yearlings contributed >30% to SAR in most years prior to 2007, and <30% since 2007. Our two-stage IPM provides a wealth of information about population dynamics affecting two key life-stage transitions (spawner to juvenile, and juvenile to spawner) centered on passage at Lower Granite Dam. By summarizing these stage-specific demographic parameters across the entire life cycle, this information will be useful for informing the recovery status of this threatened population. Whereas previous versions introduced hydrosystem and ocean covariates into the model, this phase of model development focused on solidifying the underlying model structure by introducing the concept of partial SARs and developing composite productivity and capacity as a function of underlying stage-specific parameters. Given this advancement, our next steps are to re-incorporate covariates into the model, specifically to understand how different factors affect partial SARs of subyearling and yearlings. Longer term model developments include:1) incorporating hatchery fish to explicitly estimate their survival as an alternative method for estimating natural-origin age composition, 2) expanding the model’s structure to include the three major spawning aggregates, 3) more explicitly modeling hydrosystem effects including transportation, and 4) using the model to assess retrospective and prospective management actions. In 2022, the U.S. Geological Survey (USGS) focused adult salmon survey efforts in the Snake River on deepwater redd searches and fish collection for parentage-based tagging (PBT) analyses. We use used a boat-mounted underwater video camera to count 99 deepwater redds at 16 of the 29 sites surveyed. Redd depths averaged 4.4 m. In conjunction with the Idaho Power Company, we collected genetic samples from 318 live fall Chinook salmon (Oncorhynchus tshawytscha) and 19 carcasses at 40 unique geographic locations that spanned 91 river kilometers. Eighty fish were collected at three sites (High Range [rkm 332.3], Dug Bar [rkm 315.4], and Three Creek [rkm 384.0]), which accounted for 23% of all collected fish in 2022. Most (333 fish) post-spawned salmon were collected from early to mid-November just after the peak of spawning. A summary of 2021 PBT results produced by the Idaho Power Company can be found in Appendix A.2. Beach seining and PIT tagging of subyearling fall Chinook salmon was conducted in Snake and Salmon rivers to obtain information on population metrics and growth as well as to provide data for ongoing life-cycle modeling. In the Snake River, we collected 7,496 subyearlings, tagged 4,139, and recaptured 502 (12.1%). Using 8-mm tags in 45–49-mm fish allowed us to represent an additional 25% of the juvenile population through PIT tagging beyond just using standard 9- and 12-mm tags. In the Salmon River, we captured 206 natural subyearlings with the majority (52%) of fish being captured at two sites: rkm 20 and 26. We tagged 145 subyearlings and recaptured 9 fish. Many of the subyearlings we tagged in the Snake River were detected passing Lower Granite Dam, but only 4 fish tagged in the Salmon River were detected. In total we detected 484 (11.3%) tagged fish at Lower Granite Dam, and detection rates varied by tag size and passage route. More subyearlings were detected passing via the removable spill weir (RSW) earlier in the season while more fish were detected passing through the juvenile fish bypass system (JBS) earlier in the season while more fish were detected passing via the removable spill weir (RSW) later in the season. In general, fish tagged with 12-mm PIT tags had higher detection rates than fish tagged with smaller tags. Survival to Lower Granite Dam was low and ranged from 0.22 to 0.36. Season-wide, growth of subyearlings was higher in the lower reach than in the upper reach of the Snake River.