Issue comments, feedback, suggestions
NW Fishletter #229, April 16, 2007
 Quit Trying To Measure Latent Mortality, Says Science Panel
A panel of scientists charged with looking into the long-standing debate between Northwest salmon researchers over latent mortality says the region should quit wasting its time trying to quantify the "absolute" numbers of salmon deaths attributable to passage through the hydro system once the fish have left it.
That's bad news for some folks like IDFG's Charlie Petrosky and USFWS' Howard Schaller, who have pushed the hypothesis since the early 1990s that the stress of dam passage and barge transport leads to significant fish mortality. Since this mortality occurs beyond the last counting station, it can't be directly measured. But Petrosky and others have said for years that it could be estimated by comparing returns of upriver and downriver stocks.
But the Independent Scientific Advisory Board released a report April 6 that says the panel concluded that the hydrosystem "causes some fish to experience latent mortality, but strongly advises against continuing to try to measure absolute latent mortality. Latent mortality relative to a damless reference is not measurable."
They recommended development of a single model with a merged data set used to evaluate it--with a statistical analysis that helps in selecting hypotheses. The ISAB said this was the most rigorous scientific approach.
The panel said the focus should be on the total mortality of inriver and transported migrants, "which is the critical issue for recovery of listed salmonids." They said it made more sense to look at in-river versus transport mortality "that can be measured directly."
They also said more acoustic tags should be developed and used to partition mortality in the lower river, the estuary, and the continental shelf.
For years, state, tribal and some federal scientists have argued that upstream/downstream comparisons of fish survival prove that latent mortality is a good reason to breach lower Snake dams. In their latest analysis, they told the panel that latent mortality for the Snake spring chinook is in the 60-percent range based on comparisons with downriver stocks from the John Day River, and that accumulated stress from passing those four dams on the Snake is the primary cause.
They also supported another hypothesis that related higher latent mortality to higher water-particle travel time, noting that it now takes two weeks for a fish to complete the journey downriver that took only two days before any dams were in place.
But others presented alternative hypotheses for the ISAB review--one argued that there is no direct way to measure latent mortality, and other processes can readily explain it, including stressors from life stages outside the hydro system--and that latent mortality may be simply an artifact of overly complicated models.
That was the take-home message from consultants Rich Hinrichsen, Tim Fisher, and Charlie Paulsen, whose analyses suggested that climate factors could account for trends in upriver and downriver populations. Hinrichsen also suggested that using a more common productivity parameter for stocks from different regions could greatly decrease estimates of delayed mortality.
But state and tribal biologists fired back in a memo from last summer that said, "Given differences in stream geology and productivity and the range of anthropogenic effects on spawning and rearing habitats that existed for these index populations during this period (Petrosky et al, 2001), there does not seem to be a strong biological basis for an assumption of a common Ricker-α for all populations."
In their report, the ISAB questioned whether the continuing discussion over stock and recruitment analyses, environmental covariates and the value of upriver (Snake River) versus downriver (John Day) stocks "is productive."
They said, "compared to the value of PIT-tag information, stock/recruit (S/R) analyses are a blunt instrument for assessment of annual delayed mortality." The ISAB said "numerous" authors are now using PIT tag data to support S/R findings, estimating smolt-to-adult survival rates and recruitment to age classes after fish have entered the ocean.
The panel's coolness toward the Petrosky-Schaller hypotheses may have also had something to do with the contents of a January report received from ex-ISAB member, Dan Goodman, a statistician from Montana State University. Goodman pointed out some measurement issues to do with stock-recruit estimates, "Wild smolt production is almost unknown; estimates of wild spawning populations are very uncertain; "brood" tables are estimates, not data, with uncertain age assessments and stray assignments; uncertainty in brood tables is usually ignored in subsequent statistics."
Goodman also noted that upstream downstream comparisons were "confounded by stock effects," a point NMFS scientists had been making for years, that natural variabilities in freshwater survivals of stocks and data constraints created enough uncertainty to render such estimates of latent mortality misleading and essentially useless (NMFS Technical Effects Memo, 2005).
One hypothesis by NMFS scientists Mark Sheuerell and Rich Zabel suggests that the difference between post-Bonneville Dam smolt-to-adult returns is a function of arrival time below the last dam, and a year-effect. Earlier arriving fish survive better, they say.
Such differences between barged and inriver fish might also be due to differences in fish size--barged fish tend to be smaller and more vulnerable to predators. That factor, combined with timing of ocean entry could explain plain much of the mortality, especially that which occurs early in the season, according to a hypothesis from NMFS researcher Bill Muir.
Other hypotheses suggested that annual ratios of barged to inriver SARs, called 'D,' can be developed in a variety of ways. Looking at a finer time-scale, another hypothesis suggested these changing ratios within a season show that fish should be barged later than in previous years to survive to adulthood at higher rates.
The ISAB also examined a hypothesis from Canadian researcher David Welch, who is leading an effort to track smolts in the ocean with a series of acoustic arrays along the continental shelf. Welch's hypothesis suggests that the mortality of barged fish shifts to a time when the fish are beyond the hydro system, given the assumption that they experience a fixed rate of mortality--and that "culling" is the primary cause for in-river morality experienced through the hydro system.
The ISAB didn't think much of the fixed rate of mortality.
The panel also said it had little use for any hypothesis that used a SAR that remained constant for a full year, or even season, or for a particular project, which cut to the heart of the state/tribal hypotheses. But they said they should be able to convert sets of project operations into changes of water travel time for any week or season, "so it should be possible to assess the impact of changing WTT on SARs."
As for the state/tribal hypothesis based on upriver/downriver comparisons, the ISAB said it wasn't helpful in weighing operational alternatives. The panel said it could be tested only if enough adult fish return from tagged groups over the next several years.
However, it will take even longer to unravel most of the mysteries of fish mortality below Bonneville Dam, in the Columbia estuary, and the ocean, and how it varies with environmental conditions. The panel said it will take many more years of data collection from both acoustic and PIT tags "before this question can be assessed further." -B. R.
The following links were mentioned in this story:
THE ARCHIVE :: Previous NW Fishletter issues and supporting documents.
NW Fishletter is produced by Energy NewsData.
Check out the fastest growing database of energy jobs in the market today.