The Tragedy of the Northwest Power Planning Council: New Lows in Junk Science
When the foundation be destroyed, what can the righteous do? Psalms 11:10
The foundation of science is data. Science is the process of refining theories of how the world works to fit the data, not vice versa. But the foundation of science in the Pacific Northwest is now destroyed, because government "science" driving salmon recovery decisions is no longer based on data at all. This process has reached its nadir in the Northwest Power Planning Council's "Framework Process", a process that started with great promise.
Unlike the Federal "All-H" process, which refuses even to consider sensible salmon recovery alternatives, there is actually a proposal on the table in the Framework Process that makes some sense: Alternative #7. An analysis of the Framework alternatives based on data would probably conclude that Alternative #7 would produce more salmon at lower cost than any other alternative, because only Alternative #7 would stop commercial harvest of the wild salmon we are supposed to be protecting, control salmon predators and choose only cost-effective recovery measures based on real data. Alternative #7 would also leave salmon habitat under local control, to avoid immense (and generally ignored) economic losses associated with tying up productive land and water for no measurable gains in salmon production.
The Council had promised participants in the Framework Process that it would be "transparent". The Council promised that participants that the data used would be released for public comment. The Council promised that the analytical methods and models used to evaluate the Alternatives would be released for public comment. And the Council promised that there would be an "iterative process" through which preliminary results would be released to participants, followed by review and re-analysis. The Council broke every one of these promises.
Instead, the Council released its preliminary results to the press before anyone could object to the Council's extraordinary analytical methods. The results? The radical dam breaching alternative (with continued heavy salmon harvest) produces 306% more wild salmon, while Alternative #7 produces "only" 41% more wild salmon. Oregon Council Member Eric Bloch, leading the charge against Snake River Dams for Governor Kitzhaber, quickly declared that the results "debunk the notion that you can get more fish back by not breaching dams". The Oregonian, which has evolved into the Northwest's Pravda on salmon recovery issues, quickly issued an editorial declaring that "this rates as the best work yet done in the region to help us make the right decision".
What sort of work was done? Few details on the Council's "Ecosystem Diagnosis and Treatment" approach are available. We know enough to conclude, however, that the Council's analysis is based on assumed and unreliable environmental data, assumed and unsupported "rules" for the relationship between environmental data and salmon survival, and assumptions on how government programs will actually change environmental variables. With additional refinement, the Council's approach might someday produce usable results. What the Council has rushed to release so far is junk science, pure and simple.
For example, the Council disclosed that every tributary in the Columbia River Basin was graded from 0-4 based on the extent of water withdrawals, but the Council did not disclose what the gradings mean, much less how particular tributaries are graded. Nor is there any disclosure of which data really exist, and which are assumptions. Much of the data does not exist, and serious salmon researchers, including those within the National Marine Fisheries Service, scoff at the Council's analysis, calling it "nine million points of made-up data".
Worse still, even where there is data, it is unreliable. For example, recent events in the Methow Valley have demonstrated that the water withdrawal data maintained by the Washington Department of Ecology are useless, because WDOE lists active withdrawals that have not been used for decades. The Council could serve a valuable function by circulating its proposed data to localities throughout the Region, and building a database of real data. But it doesn't.
The Council made a preliminary disclosure of some of its analytical assumptions, called the "Book 'o Rules". For example, the December 7th draft, now withdrawn, declared that 50% of juvenile salmon in tributaries rated 2, 3, or 4 for water withdrawals will die. Assuming that levels 2 and 3 allow some water to remain in the tributary, this is silly. According to sworn testimony by a National Marine Fisheries Service biologist and research on the Hanford Reach, the vast majority of juveniles commonly escape downstream in situations where tributaries are dewatered.
The Council's scientists appear to assume that habitat that is closer to "natural" will produce more salmon. But recent research in the Yakima Basin suggests, for example, that salmon concentrations are highest in the degraded segments and lowest in the pristine ones, a result precisely opposite to the rules. Model rules at odds with empirical evidence, though common in the salmon recovery context, make a mockery of the Council's charge to use good science. Reality is much more complicated than 0-5 numerical gradings of habitat.
The Council also has not disclosed how the "rules" interact. Discussions with Council staff confirm that the rules are not additive, or more precisely, multiplicative. In other words, if water withdrawals and ten other habitat factors are each imagined to kill off half the salmon, the computer model does not raise half to the tenth power to compute total survival (0.001). But no one, outside anonymous and inaccessible computer programmers, knows how the rules interact.
Perhaps the worst part of the entire analysis is how the Alternatives are imagined to change real-world conditions. The Council's scientists assume that salmon survival through the dams is much lower than it actually is, and assume, contrary to three decades of research based on real data, that smolt transportation doesn't help adult returns much, thus grossly overstating the benefits of dam removal alternatives. Undisclosed assumptions, such as the unfounded assumption that John Day Dam destroyed huge spawning beds, attribute great productive power to "back to nature" measures, and no little or no power to scientific management.
The Council doesn't get the costs of the various Alternatives right, understating the enormous savings that would result from adopting Alternative #7 by hundreds of millions of dollars a year. Yet citizens can draw some small comfort that the Council was willing to disclose publicly that according to their analysis, under Alternative #7, we could spend $120 million less a year, get the federal and state governments out of the habitat business (which would save even more money), and get 41% more wild salmon (112% overall, including hatchery fish)--all without breaching the dams. Again, if the analysis were done correctly, it would show that better harvest and hatchery management would make this Alternative put more salmon in the rivers than any other, while saving staggering sums of money now wasted. But don't hold your breath expecting the Council to adopt Alternative #7. As far as the insiders are concerned, it isn't even a starter.
James Buchal, February 11, 2000
Link: A scanned page describing Alternative #7 from a recent Council publication (36 Kb).
Return to Other Salmon Materials
Return to www.buchal.com