Testing the Idea that Flow Augmentation Increases Salmon Survival

Harvest agencies have steadfastly refused to conduct any competent experimental testing of whether releases from upstream reservoirs will increase the survival of salmon. Instead, they simply assume that releasing water from a reservoir can mimic the constellation of natural events that correspond with high river flow, including (typically) lower temperatures, greater rainfall, and better ocean conditions. The available data suggests that these other factors are much more important to the survival of salmon than the speed of the water they are swimming in.

Given the state and tribal opposition to such research, until quite recently, no one had conducted much formal analysis of week-by-week changes in survival and flow. No one really knew whether increasing river flows by upstream reservoirs decreases the travel time of fish, much less by how much. All of the major reviews of salmon science conducted in the past couple of years have ignored the importance of within-year measurements of flow and survival.

The first analysis I ever heard of came when my clients induced the Oregon Department of Environmental Quality to require the National Marine Fisheries Service to make a statistical evaluation of the available PIT-tag data to determine week-by-week survival changes (see Chapter 12). The results? High variability in the daily estimates showed no correlation between survival, flow, or several other variables. But when the data were “smoothed”, “fairly strong negative correlations were found between survival and flow”.37 Biologist Steve Cramer has found the same inverse relationship between flow and survival from PIT-tag data in 1995, 1996 and 1997 in his studies of spill discussed in Chapter 12. The available data are utterly inconsistent with any survival benefits from flow augmentation.

In September 1997, the National Marine Fisheries Service finally released (in draft form), a more comprehensive analysis of the PIT-tag data they had been collecting since 1993. Notwithstanding a general inclination to give every benefit of the doubt to flow theorists, the scientists concluded that "despite a large data base collected over several years using contemporary techniques, relationships between flow and survival . . . were not strong or consistent".38

There is some evidence that after spikes in river flow, more juvenile salmon are detected downstream at the dams. Below is a graph of Snake River flow in 1994 versus the number of juvenile salmon detected at Lower Granite Dam.

Figure 3: Flow and Juvenile Counts at Lower Granite (1994)39
The plateau of flow (dotted line) was caused by releases from Dworshak Reservoir. You can see that when flow rose rapidly, so did the numbers of salmon detected (solid line). But this could be a coincidence; the flow peak could have been timed coincident with a natural peak in migration. After all, when flow dropped sharply in early August, the number of detections also rose. When flow was high and flat, the number of detections dropped. Even assuming that the first pulse of water had any effect on fish, it looks as if most of the water after the first pulse was wasted. Some biologists look at data like this and speculate that changes in flow cause fish to move.

But this does not mean that increasing flows will have any measurable effect on the overall population of salmon. A reservoir release may be like blowing air at a tree full of robins: the robins will fly away, but they will not necessarily start flying south for the winter. Natural flows in the summer tend to become warmer, but the flows released from reservoirs are colder. No one knows whether the signal of a blast of cold water is sending the message to juvenile salmon that it is time to migrate. The true signal probably comes from the salmon’s own natural development—an internal alarm clock. Nevertheless, leading promoters of the Great Salmon Hoax, like Northwest Power Planning Council member Ken Casavant, continue to promote the idea that “more natural spring flow” will “better trigger the juvenile migration impulse”.40

There have been many studies that, while not focusing on actual changes in migration speed in a single year, looked to see whether salmon tended to migrate earlier in high flow years. The 1993 Supplemental Environmental Impact Statement (SEIS), which was prepared to rationalize the beginning of the government's foray into large-scale flow augmentation, acknowledged that "only 5 of 117 tests of linear correlation of migration timing to flow quantity had a significant positive relationship".41 Dr. McNeil’s work suggests that “the data tend to favor . . . an alternative hypothesis that migration is advanced by low flow and delayed by high flow. This is the antithesis of the NMFS theory which influences public policies on flow management during juvenile migration”.42

When my clients and others filed formal comments on the SEIS to point out the lack of any scientific basis for believing that increasing flow would increase salmon survival, the government responded:

"The comments on this issue demonstrate that opinion within the region on the existence of such a relationship is sharply divided, and that the debate cannot be resolved in the SEIS. . . . no attempt in this SEIS has been made to resolve the issue and render a firm conclusion as to the existence and strength of a flow/survival relationship . . ."43

The basic rationale for spending hundreds of millions of dollars? "[T]he cumulative weight of the research does not demonstrate that absolutely no relationship exists."44 In short, rather than relying on factual support for the flow/survival hypothesis, the government announced that so long as no one could prove beyond doubt that it didn't work, it would go forward with the program. Since 1992, that policy has continued. It is striking testimony to the political power of the flow theorists.

Most recently, the conservation biologists forming the Council’s Independent Science Group have come out against efforts to try and measure any sort of flow/survival relationship. “[W]e suggest abandonment of the search for the elusive ‘correct’ or ‘optimum’ flow and instead we advise focusing on the restoration of a riverine velocity structure as close as possible to the pre-impoundment hydrograph.”45 In short, we should all just assume that higher velocities will advantage salmon. Like conservation biologists generally, the ISG members have abandoned the scientific process in favor of the marketing slogan: just do it!

37 NMFS, 1996 Annual Report to the Oregon Department of Environmental Quality, Jan. 24, 1997, at 8 (emphasis added).

38 S. Smith et al., "Survival Estimates for the Passage of Juvenile Salmonids Through Snake River Dams and Reservoirs, 1996", at 50 (DOE Contract DE-AI79-93BP10891) (Sept. 1997) (review draft).

39 From K. Whitty, “Migration Responses of Juvenile Salmonids to Pulses in Flow” (Figure 3) (1994).

40 K. Casavant, “New thinking about Columbia salmon”, Seattle Times, Jan. 30, 1997.

41 SEIS at 4-11.

42 W. McNeil, “Water Velocity and Migration of Juvenile Chinook Salmon in the Columbia River”, Sept. 26, 1994, at 9 (paper prepared for Hydro Review; later published).

43 SEIS App. H, at H-10.

44 Id.

45 ISG, Return to the River, at 55.

Previous PageTable Of ContentsNext Page

This Web page was created using a Trial Version of HTML Transit 3.0.