“The results are horrible and deadly. In a heavily afflicted fish, bubbles of free nitrogen appear under the skin and in the fins, tail and roof of the mouth. Eyes protrude or hemorrhage, and in extreme cases they are actually blown out of the head. Fish blinded in this dreadful manner have been known to live long enough to beat themselves to death against the concrete barrier of a dam they can no longer see.” Outdoor Life (1972).1

Untold numbers of salmon and other fish in the Columbia and Snake Rivers died in the 1970s because water at the federal dams, then under construction, was directed over the top of the dams, through spillways, causing levels of dissolved gas in the river to rise to supersaturated levels. Absorbed into the bodies of migrating salmon, the supersaturated water produced horrifying symptoms, akin to a severe case of “the bends” in humans.

An entire generation of fishery biologists in the Columbia River Basin fought to make the dams safer for salmon by limiting excessive spill and dissolved gas. The U.S. Environmental Protection Agency produced a national water quality guideline intended to limit excessive dissolved gas from spill. Legislative history during passage of the Northwest Power Act reflects some consensus that spill was something to be avoided.

To the power engineers who designed the federal projects in the Columbia Basin, spilling water over dams instead of generating electricity with it was something to be avoided for reasons entirely independent of salmon. Spill was economic waste, pure and simple. At a single large dam with available turbines, spilling can cost tens and even hundreds of thousands of dollars a day in lost revenues. Ultimately, as more and more turbines were installed, less water had to be diverted through the spillways, and the spill problem of the 1970s went away.

Today, harvest managers have generally forgotten the lessons of the 1970s, and push for spill as an unquestioned good. The question of spill has become entirely politicized, and all the fine points—whether and to what extent more spill might help salmon—have been lost in an overriding political imperative for more spill, whatever the cost to salmon, or humans.

The Superficial Appeal of Spill

Superficially, spill over hydroelectric projects seems to be a sensible way to pass fish over dams. Spill proponents point to several studies showing spillway mortality to be negligible—2% or even less. The problem is that spill at every hydroelectric project is different. It works at some projects and not at others, depending on the configuration of the water and concrete. Sometimes the results are dramatically different at various spillways in a single dam. The results also vary among species.

Juvenile steelhead may have problems with spill. An older study showed mortality as high as 27.5% for steelhead passing over Lower Monumental Dam when the fish went over dams without spill deflectors, called “flip lips”.2 The fishery agencies disregard that study as unrepresentative, but have no coherent criticism of its design or execution. More recent research at Little Goose Dam showed steelhead survival close to 100%, except in spillways with spill deflectors, installed, where survival feel to a range of 93-97%3—within the range of direct turbine mortality.

Spill can also have adverse effects at dams like The Dalles, where spill dumps juvenile fish in an area of shallows and islands where they suffer higher predation than the fish emerging from turbines.4 U.S. Army Corps of Engineers modeling suggests that “40% spill is the maximum that could occur without subjecting passing juveniles to excessive predation risk”.5 Yet the National Marine Fisheries Service has required 64% spill at The Dalles.6 Preliminary data collected on coho salmon during 1997 showed losses of approximately 12% when spill volumes were at the required 64% level—probably significantly higher than turbine losses—and 6% losses for juvenile fall chinook.7

Spill at The Dalles also dumps the smolts onto “energy dissipators” which may cause direct physical injuries to the fish. In 1995 studies using balloon tags, five percent of the fish released above the spilllway were not recovered at all, suggesting high levels of physical stress on spilled fish.8

The Bonneville Power Administration has been trying for quite some time to get the fisheries agencies to do some dam-specific work to see what spill survival really is, and how it is affected by particular structures over which the water flows. Unfortunately, the states and tribes continue to veto the research. When the National Marine Fisheries Service finally proposed to study spillway survival at The Dalles in the summer of 1997, the Executive Director of the Columbia River Inter-Tribal Fish Commission declared that the proposed study was “unacceptable” and threatened to bring the matter before Judge Marsh under his continuing jurisdiction in United States v. Oregon harvest allocation case, purportedly because of tiny losses to hatchery coho salmon from tagging them for study purposes.9

The pro-spill bias is so overwhelming that all problems with spill are minimized. As is the case with turbine mortality, there is a single, politically-correct number for losses from spill: 0-2%. And high spill is required at all projects, even when it is counterproductive.

The most widely-known problem with spill (apart from its cost) is the 1970s problem: spilling water from 60 feet or more over the top of dams tends to dissolve a lot of atmospheric gases in the water. This phenomenon is known as dissolved gas supersaturation, and results in the water becoming like soda water full of gas, which gradually comes out as bubbles. When these bubbles come out inside of a fish, they can kill it. Tiny bubbles forming in the gills of fish can quickly asphyxiate them; larger bubbles can cause blindness and loss of physical control. The same phenomenon happens to human divers when they come up too rapidly from the deep: the nitrogen gas bubbles cause "the bends", a condition which is often fatal.

In the late 1960s and early 1970s, there were huge quantities of spill at federal projects on the Columbia and Snake River, primarily because all of the turbine units had not yet been installed. Large quantities of fish began showing up dead, with then-mysterious symptoms. Federal fish biologists, led by Dr. Wes Ebel at NMFS, began a crusade to identify the problem and cure it.

Laboratory experiments soon proved that salmon and steelhead were acutely sensitive to dissolved gas. Early experiments putting juvenile salmonids in water saturated at levels above 110% of normal killed large proportions of the fish; results varied by species with steelhead being more susceptible than salmon.10 More refined experiments indicated that fish in deeper tanks could tolerate higher saturation levels. Still, at 124% and 127% saturation, juvenile chinook salmon suffered 67% and 97% mortality in water 2.5 meters deep.11 Fish deeper than three to four meters are generally safe. The United States Environmental Protection Agency eventually established a national water quality criterion of 110% as the maximum allowable level of dissolved gas. One of the key considerations in establishing the standard was the effects on salmon. The studies by Dr. Ebel and others on the Columbia River achieved national significance.

Under the federal Clean Water Act, the states were permitted to adopt a more rigorous standard. Oregon adopted the 110% standard, setting a more stringent standard of 105% for waters used by salmon hatcheries. At the time, L.B. Day, the first director of Oregon's Department of Environmental Quality, said that he would have preferred a 105% standard. Since then, only the politics of spill have changed, not the science.

One of the first big battles to gain more spill was fought before the Northwest Power Planning Council in the middle 1980s. Former Council member Kai Lee of Washington has pointed out that in 1985, fishery advocates were complaining that is was not a good year for fish because there was not enough spill. Yet adult returns soared for the following three years.

“Of course, much happens to salmonids after they leave the Columbia Basin, and these data do not show that 1985 was a good year after all. They do illustrate the shaky underpinnings of the fishery managers’ experience, which is usually offered without explanation ex cathedra.[12] . . .
“By 1986, the Council’s impatience with unreliable statements of inscrutable experience had grown to the point that a recommendation for improved spill operations failed on a vote of four to four. The Council’s rationale . . . was that there were no significant biological benefits from increasing spill. More precisely, the support for the fishery managers’ claim of biologically significant benefit had fallen below a bare majority.”13

A decade later, “unreliable statements of inscrutable experience” on spill continue to fill decisionmaking forums in the Pacific Northwest. The 1994 Ninth Circuit decision discussed in Chapter 9 even went so far as to declare that the Northwest Power Act “requires only the best available scientific knowledge, not data14 in arguing that the Northwest Power Planning Council must give a "high degree of deference" to the fishery managers. Fishery managers now tell the Council that their "unreliable statements of inscrutable experience" must be followed as a matter of law.

From the perspective of my clients, the main problem with spill was simple: its cost. Luckily, since the juvenile salmon tend to migrate at night, spill levels tend to be higher at night, when electric power demand is less. When experts told my clients that, in addition to wasting money, the spill would not help salmon, and probably injure them, they thought there was some chance that they could stop increases in spill. They were wrong.


Reprinted in K. Petersen, River of Life, Channel of Death 139.

2 C. Long, F. Ossiander, T. Ruehle & G. Matthews, “Survival of Coho Salmon Fingerlings Passing through Operating Turbines with and without Perforated Bulkheads and of Steelhead Trout Fingerlings Passing Through Spillways with and without a Flow Deflector”, NMFS NW Fish. Sci. Center. (1975).

3 B. Rudolph, "Latest Survival Research Highlighted at IT Meeting", NW Fishletter, Nov. 11, 1997; see also BPA, “Interim Research Monitoring, and Evaluation Program to Support the FCRPS Biological Opinion and Recovery Plan”, at 23 (Nov. 15, 1995 draft) (similar results elsewhere).

4 NMFS, Biological Opinion on FCRPS Operations, Mar. 2, 1995, at 109.

5 Memo, J. Ruff & B. Hevlin to Implementation Team, April 22, 1997, Attachment (3), at 1.

6 NMFS, Biological Opinion on FCRPS Operations, Mar. 2, 1995, at 106.

7 USACE, Memorandum for the Record, Aug. 15, 1997, at 2 (CENWP-PE-E (1146)) (Review of “Relative Survival of Juvenile Salmon Passing Through the Spillway and Ice and Trash Sluiceway of The Dalles Dam”, Admin. Code MPE-P-92-2).

8 Memo, J. Ruff & B. Hevlin to Implementation Team, April 22, 1997, Attachment (3), at 1.

9 Letter, T. Strong to W. Stelle, Jr., Mar. 14, 1997, at 3.

10 See generally, D. Weitkamp & M. Katz, “A Review of Dissolved Gas Supersaturation Literature”, Trans. of the Am. Fish. Soc. 109:659-702 (1980).

11 Id. at 675.

12 This phrase was originally applied to the decisions of the popes from their cathedra, or chair.

13 K. Lee, “Rebuilding Confidence: Salmon, Science, and Law in the Columbia Basin”, 21 Envt’l Law 745, 795 (1991) (footnotes omitted).

14 Northwest Resource Information Center v. Northwest Power Planning Council, 35 F.3d 1371, 1391 (9th Cir. 1994). This case is the subject of extensive commentary in Chapter 9.

Previous PageTable Of ContentsNext Page

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