the film forum library tutorial contact DONATE

Gloves Come Off Over Salmon Models by Bill Rudolph NW Fishletter, October 6, 1998

BPA scientists and other consultants involved in the PATH process complained loud and long about a report unveiled at last Thursday's IT meeting that provided little support for the agency's salmon passage model, CRiSP.

Their complaints brought to the surface of salmon recovery a long-simmering dissatisfaction with the PATH process. They told regional salmon policy makers that the review had left out much important information, notably research conducted over the past few years that has failed to show a relationship between increased river flows and improved survival of juvenile chinook in the Snake River.

Some regional scientists have been confounded since the word leaked out a week ago that the four independent reviewers who "weighed the evidence" over seven key certainties found the CRiSP model less convincing than the state and tribal model, FLUSH, which relies on a strong flow/survival relationship. The CRiSP model predicts higher system survival and better results from transporting fish than FLUSH, which assumes survival rates are pretty much the same as they were in the late 1970s when the dams on the Snake were completed. The reviewers decided that CRiSP results were too optimistic.

None of the reviewers gave the CRiSP model better than a 35 percent weight (out of 100 percent), while one gave the FLUSH model the nod by 90 percent. But both passage models were criticized for being too complicated, especially CRiSP, which has been developed to reflect the latest juvenile survival studies with PIT tags. These show little if any correlation between increased mainstem flows and improved fish survivals. The FLUSH model maintains that a relationship does exist and that the rate of fish mortality increases the longer a salmon takes to pass through the hydro system.

Newest Research Ignored by Review Panel

It was evident from the panel's written remarks that most were not even aware of PIT tag results from the past few years that show no correlation between flows and survival of spring chinook--results that were reviewed later by NMFS scientists at the same meeting. One earlier PATH reviewer, who had suggested using the most recent survival data in the deliberations, did not take part in the weighing of evidence exercise.

Panel members included Carl Walters from the University of British Columbia, Jeremy Collie and Saul Saila from the University of Rhode Island, and Steve Carpenter from the University of Wisconsin.

Salia said FLUSH had a stronger empirical basis because of evidence of a strong flow/survival relationship based on work with Atlantic salmon.

The panel also felt there was little evidence of periodic cycles in the ocean affecting productivity. And they felt that the benefits from removing predators was only a short-term answer. Walters wrote that "CRiSP needs a 'demon in the ocean' to explain decline in smolt adult returns (SARs), but declines in SARs of other stocks that live in the same place in the ocean are not as severe and start later than declines in the Columbia."

After PATH facilitator David Marmorek talked about the process that weighed the models' credibility, based on "evidence related to key uncertainties," he described modeling results from the weighted evidence to assess possible outcomes of three potential future strategies that PATH modelers looked at: the present BiOp, another option that maximizes transportation, and a third that would breach the four lower Snake dams.

Dam removal came out as the best chance for achieving the 100-year NMFS jeopardy standard. But both present BiOp and max transport strategies came out with about 70 percent of the model runs achieving the standard, compared to nearly 90 percent achieving the standard from dam removal.

None of the options quite reached the 24-year jeopardy standard, with dam removal slightly higher than the other options.

Both the present BiOp and maximum transport options failed to meet the 48-year recovery standard, but results were much better for the dam removal option, with about 80 percent of the model simulations passing the 48-year standard. Marmorek explained these results came about because the weighting process had downgraded the climate regime hypotheses, which in PATH modeling calls for good conditions between the years 2005 and 2035. He pointed out that the last eight years of the 48 would be in a "bad regime."

CRiSP modeler Jim Anderson of the University of Washington told the packed room that he agreed the models should be simpler, but he said the FLUSH model was not even available to all PATH participants. Though the panel did a "fairly good job," he charged that they were "misinformed on many elements of the process." He said he hoped comment would be allowed to address the inconsistencies in the panel's conclusions.

FLUSH modeler Howard Schaller of the Oregon Department of Fish and Wildlife disagreed, saying the panel had been looking at different pieces of PATH information for the past two years.

"They are aware of different points of view of the different PATH participants," Schaller said.

But BPA's Jim Geiselman, another PATH member, said that additional information needed to be provided to the review panel. He said his agency objected to the weight of evidence process, partly because it failed to include effects of harvest, habitat or hatcheries.

He cited an earlier panel review of the PATH process, which called its focus on passage through the hydro system "shortsighted," because of a total preoccupation on a single life stage of the salmon.

Advice on Experimental Options Criticized

The weight of evidence panel also suggested options for experimental management actions that offered two directions--incrementally--either by removing dams, shutting down hatcheries and transporting fish, one strategy at a time, or by implementing what they called the "reverse staircase" alternative. This means breaching the dams, shutting hatcheries down and stopping transport now, "then turn dams, hatcheries, or transportation back on one at a time."

The panel said this was the more likely way to lead to stock recovery, but would have higher up-front costs.

Consultant John Pizzimenti of Harza Engineering told IT members about the projected costs and timelines of some options, like the fact that construction costs would run $967 million just to implement drawdown at John Day, while tearing out the embankments at the four lower Snake dams would run between $600 million and a billion dollars.

"This won't even pass the laugh test in the outside world," said Pizzimenti of the panel's recommendations. "What about changing hatchery practices, rather than closing them?" He asked, noting that some were successful, others not.

Marmorek replied that participants are facing tight deadlines; many are working on fall chinook issues and have no time to work on the weighting issues. But he flashed a hastily scribbled overhead that showed BPA consultants had literally weighed in with a pile of documents that weighed about as much as that of the FLUSH camp, around a pound a piece. He also said factors such as hatcheries and bird predation had already been considered as other mortality in the PATH process.

Geiselman pressed for another iteration before the report is finalized "to include our concerns," but others pointed out that PATH deadlines are designed to mesh with development on the Corps' draft EIS on the Snake River Drawdown Feasibility Study.

Marmorek agreed that there was disagreement over the weighting process, but he said he thought the overall process was "healthy."

Ironically, the next item on the IT's agenda was a presentation by NMFS statistician Steve Smith, who discussed preliminary results from the 1998 PIT tag work. These added to earlier results that go back to 1994 and show that within years, there is no correlation between flow and survival of spring chinook in the Snake River. He did add that there is a weak correlation among different years that is consistent with data from the 1970's. Smith said steelhead showed the same basic trends, but fall chinook showed a flow survival relationship in the Snake from their release point to Lower Granite Dam. But he said the fall chinook picture could be confounded by other variables, such as water temperature and turbidity. Below Lower Granite, he said the picture is less clear.

Smith's review was a perfect introduction for the next presentation. Consultants Darryll Olsen and Harza's Pizzimenti described various elements of their white paper on flow augmentation. Their report, also based on the new PIT tag information, says there is no biological benefit from augmenting spring flows, which cost between $50 million and $250 million a year. They suggested saving water for fall chinook may make more sense, although the biological benefits of that strategy are still unclear as well.

Pizzimenti reminded the group of the late Don Bevan's remarks, who as chair of the Snake salmon recovery team said, "If there's something we can do for salmon, it could be in the late summer or fall." The Bevan Team, in its own 1995 report, had found spring flow augmentation of dubious value.

Later, the PATH controversy flared up again. As part of Olsen's review, Anderson described the history of fish survival studies in the region and pointed out a number of reports, including Smith's, that had not been "looked at in any detail in the PATH process."

FLUSH modelers said they had received the new data, but when pressed, said they did not have the "within-year data."

NMFS' Smith said he had sent the FLUSH folks the new work, but had not included the within-year findings. After a bit of discussion, state and tribal modelers admitted their FLUSH model is constructed in such a way that it is unable to use data that breaks timelines down to lengths of less than a year.

NMFS policymaker Brian Brown seemed interested and called for a full review of all flow augmentation alternatives. The meeting ended with a report by Idaho water policy manager Karl Dreher, whose own study has found that augmenting spring flows in the Snake does little to actually speed up water particle travel time, something fish managers have long felt improves survival of chinook. Smith's earlier report showed PIT tag research has found a correlation between flow and travel time, but cannot relate it to actual survival of spring chinook.