Preliminary estimates of survival of wild juvenile salmon and steelhead from the Snake River were so low this year that despite a high degree of uncertainty in the estimates, fish managers should be talking about potential causes of the mortality, NOAA Fisheries representative Claire McGrath told fellow members of the Columbia River Technical Management Team on Nov. 3.
The low survival estimates for wild fish occurred during the second year of extremely high spill rates that are intended to boost survival, but a high uncertainty of the results will make it tough to determine whether the high spill is helping or hurting survival, according to NOAA's 2021 memo on preliminary survival rates of PIT-tagged salmonids.
"Flexible spill operations were conducted for the third consecutive year in 2021 and the second consecutive year using an increased spill cap of 125 percent dissolved gas saturation," the memo says. "There is much interest in the potential benefits of increased spill to fish survival and travel time through the hydropower system. Unfortunately, the precision of the estimates of these metrics has suffered under the high spill operations due to substantial decreases in detections of fish. This makes it difficult to assess whether the new operations provided benefit or were detrimental for fish, especially when focusing on results from individual river reaches."
The annual preliminary analysis with survival estimates for the Columbia River's PIT-tagged juvenile salmonids by scientists at NOAA's Northwest Fisheries Science Center will be followed by a presentation and discussion at TMT's year-end-review meeting on Dec. 1, and a final report in 2022.
While survival of hatchery and wild fish combined for many of the Columbia River runs appears to be statistically similar to 10-year averages, fish scientists raised other concerns about the low 2021 survival of some wild runs and about potential problems in their journey through specific dams and tailraces that could be caused by high spill.
"Also of concern is that survival from McNary Dam tailrace to Bonneville Dam tailrace was below average for Snake River steelhead (hatchery and wild combined), Snake River sockeye, Upper Columbia River hatchery Chinook, Upper Columbia River hatchery steelhead, and Upper Columbia River sockeye," the memo says. "Although these estimates suffered from uncertainty, the consistency of them all being well below average along with those for wild fish from the Snake River could be pointing to a problem that requires further investigation."
Juvenile migration conditions this year included an early heat wave resulting in high water temperatures, and—in the Snake River—very low flows and the highest percentage of spill ever recorded, McGrath noted.
That high spill means that significantly fewer juvenile salmon and steelhead passed through the juvenile bypass systems at each dam, which is the only location at most dams where the PIT tags are detected.
Fewer detections resulted in much higher uncertainty in the results of the 2021 preliminary juvenile survival rates in the Columbia Basin, which NOAA Fisheries has analyzed since 1993.
"Unfortunately for quantity and quality of data, a side effect of increased spill since 2006 has been a drop in detection rates of smolts at most Snake and Columbia River dams," the 2021 memo says. "The exceptionally high spill in 2020 and 2021 resulted in exceptionally low detection rates." Detection rates at Little Goose, Lower Monumental and John Day dams were less than half of the average from 2007 through 2019, it says.
The scientists wrote that without data from a PIT-tag spillway detector installed at Lower Granite Dam by the U.S. Army Corps of Engineers in 2019, very few survival estimates would have been possible this year. Other than detection of the few fish that go through bypass systems at the dams, the analysis also relied on detections from a corner collector at Bonneville Dam, an estuary trawl, an estuary pile dike detector and PIT tags recovered from estuary bird colonies. Some of those sources of data were not available for the preliminary memo, but will be included in the final report on survival estimates.
In addition to reduced detections, the PIT tagging at Lower Granite Dam had reduced capacity in 2021, so sample sizes were significantly smaller than average, the memo notes.
At the TMT meeting, Idaho Department of Fish and Game representative Jonathan Ebel said that tagging opportunities were limited partly because there are fewer fish in the Snake River. He also said his agency's ability to tag fish has been affected by the higher costs of PIT tagging while funding has remained flat.
He added, "Right now, the uncertainty on these estimates makes it very difficult to identify any patterns that may be emerging under this high spill regime."
This year's high spill also resulted in the lowest barge and truck transportation rates on record for juvenile salmonids in the Snake River. In the spring, the percentage of nontagged fish that were transported included about 10 percent of wild Chinook and 6 percent of hatchery Chinook, along with 11 percent of wild steelhead and 4 percent of hatchery steelhead. Including tagged fish, about 15 percent of wild and 6 percent of hatchery Chinook, and 11 percent of wild and 7 percent of hatchery steelhead were transported, the memo states. The transported smolts were collected at the three dams—Lower Granite, Little Goose and Lower Monumental—according to the memo.
In years when conditions are poor, transported juveniles often have higher survival rates compared to those that migrate in-river, so barge or truck transport has been used in the Snake River to spread mortality risks. The objective of the transport program is to "transport juvenile fish when the best scientific information indicates doing so will increase adult return rates," according to TMT's 2021 Fish Passage Plan.
The memo notes that 2021 collection rates were less than half of the already low rates from 2020, and less than a quarter of the rates from 2018 and 2019—primarily due to a combination of extremely high spill and low flow.
Despite uncertainty in the numbers, the preliminary report raised concerns about the survival estimates of wild smolts, which were far below average for both Chinook and steelhead, and among the lowest in the past 10 years. "However, despite the lack of certainty, these low survival estimates are concerning, as they are occurring during a management regime of extremely high spill rates intended to boost survival," the memo notes.
In-river conditions for juvenile salmon and steelhead—especially in the Snake River—included high spill, low flows and periods of high temperatures. Total dissolved gas was slightly higher than the long-term average, the report stated.
While the amount of spill at Snake River dams averaged 38.1 kcfs—similar to the average of 35.6 kcfs since 2006—the spill as a percentage of flow averaged 63.2 percent, far greater than the long-term average of 38 percent, the memo says. "Daily mean spill percentages in 2021 were very consistent and extremely high for the entire migration period."
The average flow at Little Goose Dam was the third lowest since 1993, at 56.1 kcfs, far lower than the 92 kcfs average from 1993 through 2021, the memo notes. The low flow likely contributed to longer travel times between Lower Granite Dam and Bonneville Dam, which may have contributed to lower estimated survival due to increased exposure to predation, dissolved gas and warmer water.
The average water temperature of 11.6 C (52.9 F) during the migration period was only slightly above the long-term average, it says. However, daily water temperatures were about 1 C (1.8 F) higher than average during three time periods lasting a total of 33 days, and about average otherwise, the memo notes.
If high levels of spill are to continue, the region should find ways to compensate for losing the ability to detect most of the juvenile salmon during their migration, the memo says. It points to the success of the new spillway detector at Lower Granite Dam, and concludes, "We believe that the region should place a very high priority on installing similar systems at other dams on the Snake and Columbia rivers, especially McNary Dam and Bonneville Dam. There should also be an emphasis placed on development of alternative technologies that will boost our abilities to detect PIT-tagged fish, such as placing detection barges in the forebay or tailrace of the dams and other detection equipment downstream of Bonneville Dam."