The Hanford Reach Fad, and Other Drawdown "Science"

The Hanford Reach of the mid-Columbia River supports the single largest naturally-spawning run of chinook salmon in the Columbia River Basin. In 1986, the run peaked at over 200,000 fish. The ISG report pointed to the Hanford Reach run of salmon as an example of what “natural conditions” can accomplish. Environmental advocates claim that the Reach is “the only free-flowing stretch of the Columbia River unaffected by reservoirs”,59 and imply that salmon runs everywhere would be like the Hanford Reach runs if we just removed the dams.


The Hanford Reach is not “free-flowing”. It is regulated by several upstream dams, and it is regulated in particular to maximize salmon production by maintaining higher-than-natural (pre-dam) flows to protect salmon redds at critical times. Indeed, there is a formal agreement among the dam operators and fishery managers to do this, called the Vernita Bar Agreement.60


The river regulation has other beneficial effects. As Dr. Don Chapman has pointed out, winter flows are now higher from storage releases, and probably warmer as well, with positive effects on salmon from the elimination of freezing conditions on redds in the shallows and improved conditions for the incubation of juveniles.61 And, of course, the regulation of the reach prevents massive flooding that would scour out salmon redds.


Gregg Easterbrook has pointed out that with water regulation rules in place to protect salmon, “rivers of the Northwest Northwest may become more friendly toward salmon than before genus Homo, as humanity’s dams and reservoirs smooth out the swings of natural flood and drought.”62 In the case of the Hanford Reach, that has already happened. There is every reason to believe that thanks to human development, the Hanford Reach now has better-than-natural conditions for salmon.


One of the things that the Mid-Columbia dam operators did not agree to do was prevent daily fluctuations of river levels associated with changes in electric power generation. Dr. Williams, in his lectures promoting the Return to the River report, shows slides of juvenile salmon trapped in the gravel by receding water below these dams, blaming power generation. He does not mention the extensive studies below the lowermost dam, Priest Rapids, showing little or no effect on fall chinook spawning or abundance from such changes in flows.63


There is some possibility that lower fishing pressure could explain the abundance of the Hanford Reach stock. The harvest managers express no interest in examining this hypothesis; Dr. Al Giorgi has suggested that the stocks “may differ in locations of their ocean residence, which could affect overall population success”.64


In the late 1950s, before the John Day and McNary pools were filled, the Hanford Reach was estimated to provide spawning habitat for only about 15,000 fish.65 Some biologists believe that when the John Day and McNary pools were filled, groups of salmon gradually moved up to the Hanford Reach.66 The Hanford reach run did not increase for many, many years, and suddenly shot up in the 1980s, along with most other salmon runs in the Columbia River Basin.



Figure 1: Interdam Run Size in the Hanford Reach67
At one time, this rise in population was proclaimed to be an accomplishment of the Northwest Power Planning Council’s programs. The fishery scientists do not really understand why the populations increased, and to what extent “natural” conditions had anything to do with it.


In fact, it is possible that the population rebounds on the Hanford reach are really a story of hatchery success. A 1988 report by two University of Washington biologists states that “[a]ttempts to artificially enhance the production of chinook salmon in the Hanford Reach have gone on for over twenty years—first as a spawning channel that apparently failed and now as a hatchery that is apparently succeeding”.68 The operators of the Priest Rapids Hatchery believe that its success has something to do with rebounds in wild Hanford Reach fall chinook.69 Sam Penney of the Nez Perce Tribe claims that "the supplementation projects being developed by his tribe 'are really the same as what is being accomplished at Vernita Bar on the Columbia River—a place, Penney said, that the Council's own science panel 'likes so much'."70


The current fascination with the Hanford Reach runs is best understood as a fad. Conservation biologists such as those on the Council's Independent Science Group like the Hanford Reach because they think it is "more natural" than a reservoir. Since there are lots of salmon returning to the Hanford Reach, they assert that if we drain reservoirs to produce more "natural" conditions, we can have lots of salmon where we drain the reservoirs.


Conservation biologists are so eager to promote drawdown using the Hanford Reach example that they even propose to take action that could well destroy the productivity of the Hanford Reach. Consider the following line of reasoning:


“. . . the Hanford fall chinook spawn only in the upper two-thirds of the reach, probably because interstitial flow pathways are nonfunctional in the lower third of the reach due to the elevated water table created by virtual continual maintenance of the full pool elevation of McNary Reservoir. Lowering McNary pool likely would lower the water table in the alluvial reaches upstream, significantly increasing the size of the river reach at Hanford containing both surface and ground water habitat components.”71

But lowering the water table in the most productive spawning ground around might have some adverse effects, like reducing the subsurface flow that is critical to keep chinook eggs alive. The very same conservation biologists criticize irrigation for reducing water tables and destroying spawning grounds in the Yakima and other rivers, yet they want to reduce the water table in the most productive part of the Hanford Reach.


Drawing down John Day and McNary pools might cause the salmon to tend to move back down the River. But the net gains from this process could be zero—or less than zero. We could wind up reducing the Hanford Reach populations without restoring them in John Day and McNary pools.


Another major question posed by the “normative river” theory is why the regulated reach below the mid-Columbia dams is so productive, while the regulated reach below the Hell’s Canyon Complex isn’t. If, as some believe, “[p]roductive populations spawning in large alluvial mainstem reaches may have functioned as critical core populations”,72 it may be impossible to increase abundance of endangered Snake River fall chinook salmon without a core population in the mainstem reach below the Hell’s Canyon Complex.


Of course, it is always possible that the gravel reaches within Hell’s Canyon were never particularly good spawning habitat, which is consistent with the idea that the Snake River chinook stocks have always been variable. It is awfully hot there. Archaelogical evidence from Hell’s Canyon shows that salmon bones make up only 7-8% of the remains, with the balance coming from suckers and other warm water fishes.73


Yet another problem with the “normative river” theory is that some research suggests salmon may do worse in the more “complex” or “normative” habitat that the ISG would purportedly promote through dam removal. Juvenile Snake River fall chinook get trapped in a series of sloughs and wetlands on the Snake River side of the Columbia River that last for about 10 miles downstream from the confluence of the Snake and Columbia. The ISG speculates that this phenomenon “could be responsible for a disproportionate loss of Snake River fall chinook at this point compared with the Hanford stock coming down the Columbia channel at the same time”.74 Sometimes more complex habitat helps salmon; sometimes it kills them. The ISG offers no reason that "complex" habitat in reservoirs would be any worse for salmon than "complex" habitat in a more natural river.


One of the main lines of reasoning supporting the “normative river” theory is that reservoir shorelines tend to be eroded soil or rock, with little vegetation, while the shorelines of the Hanford reach (and other natural rivers) have more vegetation. The ISG theorizes at one point in its report that rising spring river waters flood vegetation and promote insect populations, which they suggest are an important food source for migrating juveniles.75 They call for additional research into the question of how juvenile fish migrating seaward feed in the reservoirs, and whether they have enough food.



Figure 2: The Mainstem Snake Before Dams (ca. 1890)76
The ISG report is now commonly cited for the proposition that the dams destroyed “riparian cover for resting and predator avoidance”.77 In fact, there was very little such habitat along the lower Snake. The most comprehensive assessment of habitat along the Snake River was prepared in connection with the Lower Snake River Compensation Plan. The U.S. Army Corps of Engineers reported that “riparian habitat existed as scattered, narrow strips along the river”. Most of the inundated area “consisted of rocky cliffs and rather steep hillsides covered mostly with sagebrush and dryland grasses”. There was only “about 1,123 acres of brush and tree-type vegetation backed by fertile bottom lands” along the river shore.78 Destroying four huge dams to bring back 1,123 acres of property seems senseless. It should also be remembered that the Corps acquired 24,124 acres in “mitigation” for the lost 1,123 acres.79


Older fisheries scientists like Dr. Chapman remember what the riverbanks looked like back in the 1950s, before McNary Dam was built. He remembers the edges of the river as being “mostly rock rubble and gravel, with some sandbars”. He thinks “the stream fluctuated so much from freshet to low flow that scouring took away seedlings of willow and other edge species in most margins that juvenile fall chinook might use”.80 The conservation biologists, however, have no institutional memory, and are untethered by historical facts.


Historical photographs of the Snake River reaches now inundated by the reservoirs do not show rich vegetation. They show desert scrub, mud and rock. The ISG recognized that pre-dam photos of the mainstem Columbia “show large sand dunes along the river”.81 How these sand dunes were supposed to produce a large food supply for migrating fish is not explained.


Ultimately, the ISG simply transplanted research findings on the “ecology of regulated streams” to the mainstem Columbia and Snake Rivers based on the assertion that “principles” from the stream literature “directly apply” to one of the larger rivers in the world running through desert country.82


One member of the ISG, Dr. Charles Coutant from Oak Ridge National Laboratory, promotes the idea that reservoirs are so short of food that migrating juvenile salmon starve to death. This hypothesis is not supported by either theory or data.


The ISG report stretched to cite a master’s thesis of an Idaho Fish and Game employee, Tom Curet, as data for the notion that fall chinook smolts were starving in the reservoirs.83 When contacted by Bill Rudolph, a reporter trying to track down the source of the new “starving fish” theory of dam harm, Mr. Curet was reportedly “surprised” to see his work cited by the ISG, since it represented only two years of data.84 Mr. Rudolph even had the perspicacity to contact Mr. Curet’s thesis advisor, who said that “there is no lack of food in [their] stomachs” and that fall chinook had five or six different species of food to choose from in Lower Granite Reservoir.85


Other studies confirm that migrating salmon are not starving. Measuring stomach contents on a scale from 1 (empty) to 7 (distended) at Lower Granite, McNary, and Bonneville Dams, researchers Bill Muir and Travis Coley found averages of 2-3 at Lower Granite and 3-5 at the other two dams.86 One would not expect wild fish to have full stomachs all the time—like any wild animals, they spend a lot of time looking for food. The two researchers suggested that increased hatchery production, coupled with the tendency of migrating fish to delay in the forebay of Lower Granite Dam (where there is not much food), may have accounted for the results at Lower Granite. Muir and Coley suspect that Lower Granite, being the first of the mainstem reservoirs, has finer sediments that reduce feeding opportunities by reducing the diversity of food sources; downstream reservoirs have coarser sediments.87 So much for data showing that juvenile salmon are starving.


As for theory, reservoirs have a larger surface area than rivers, and thus absorb more solar radiation and breed greater quantities of algae. The lower velocity of reservoirs also encourages algae growth. These algae in turn feed larger organisms up the food chain. These points have recently been emphasized by Portland State University researcher Ralph Vaca and the Montana Department of Fish and Game, seeking to prevent destructive drawdowns of Montana reservoirs. The ISG does acknowledge that as some sources of food for salmon have declined, other species of salmon food originally native to the lower river estuary have become established further and further upstream.88 However, they speculate that these newer species are not as nutritious.


In short, conventional wisdom on feeding and reservoirs appears to be dead wrong. Studies on the mid-Columbia River have found that the mean size of subyearlings juvenile chinook sampled in the 1980s were substantially larger than the mean sizes in the 1960s, when reservoirs there were either nonexistent or new, so that food chains in them had not fully developed. Nine biologists, including Don Chapman, who reviewed the evidence in the course of an exhaustive review of the status of summer and fall chinook, believe that subyearling chinook may be “achieving more rapid growth in reservoirs”.89


Indeed, they speculate that the reservoirs may be compensating for the lost of estuarine habitat, so that any delay in reservoirs may provide a net benefit for subyearling chinook.90 If that is true, then draining reservoirs will positively injure fall chinook salmon—the result predicted by computer models, albeit for different reasons.


Notwithstanding the work of Muir, Coley, and Chapman, the media has begun to broadcast starving juveniles in the reservoirs as the newest aspect of the Great Salmon Hoax.


Yet another claim promoted by the Independent Science Group is that draining reservoirs will increase the amount of spawning habitat available in reservoirs. But no one measured how much spawning habitat there was before the dams were constructed. And no one measures how many salmon spawn there now.


To the extent that "groundwater upwelling areas" in the mainstem were used by salmon,91 it is possible that such areas persist inside the reservoirs and are still used by salmon to spawn. It is also possible that the “complex island, point and eddy reattachment bars composed of sand, gravel and cobble” whose inundation is cited by the ISG92 may have been useless without adequate subsurface flows of water.


Like all elements of the Great Salmon Hoax, it requires careful deconstruction to discover the origin of claims of lost spawning habitat. The Independent Science Group now claims that the “second largest” group of fall chinook salmon, some 34,000 strong, “spawned in the mainstem Columbia River in the area now inundated by John Day Dam.93 This number comes from a single 1968 article prepared by biologist Leonard Fulton for the U.S. Fish and Wildlife Service in 1968.94 I first heard about the Fulton Report when a reporter told me that federal fishery biologists had told him that the 34,000 number was a “wild-assed guess”, made by someone trying to count underwater salmon redds from planes.


Sitting down to read the report, I became somewhat suspicious of Mr. Fulton when, following a brief reference to overfishing, he attributed all reductions in upriver spring and summer chinook runs to dam construction.95 Mr. Fulton informs us that


“The areas below the confluence of the Snake River [including the area inundated by John Day Dam] are more turbid, and it has been difficult to distinguish redds and spawning salmon in this reach of the Columbia River. Evidence indicates, however, that a large population of fall chinook spawns in the 160 km. stretch of river below McNary Dam”.96

What is the evidence? Reading further to Table 5, the reference listed is “unpublished information”.97 The ISG reports that the original quantitative data has all been lost by the fishery agencies, and all that remains is “summaries or brief, qualitative accounts”.98 Reading further, one comes to a Table 8, called “Average number of fall-run chinook salmon entering sections of the Columbia River and its tributaries”, and there is the claimed number: 34,000 fish entered the section of the Columbia River from “John Day damsite to McNary Dam”.99 Note that Mr. Fulton does not claim that 34,000 fish spawned there; that is an invention of the ISG.


There is a footnote to the 34,000 number, which claims that “[e]stimates of population using this reach are based on aerial surveys”.100 In other words, even though the water is so turbid that seeing fish or their nests is hard from the ground, these counts are made by airplane. Right.


Then, on the very same page, the 34,000 number appears again, in another context. According to Fulton, “[a]verage counts of fall chinook salmon at Bonneville, The Dalles, and McNary Dams were 163,000, 90,000 and 56,000, respectively, for 1957-60”.101


34,000 just happens to be the difference between 90,000 and 56,000. It is a remarkable coincidence that the total number of fish disappearing in this reach just happen to be the exact same number of fish that the ISG claims were spawning in the reach now inundated by John Day Dam. In other words, every fall chinook salmon that disappeared after being counted at The Dalles Dam, as measured by the lower counts at McNary (the reach that includes John Day Dam), is supposed to be a salmon spawning above the John Day damsite.


If that were true, we would have to assume that no fall chinook salmon stopped before the John Day damsite to spawn. We would have to assume that none turned out of the mainstem Columbia River at the Deschutes River, John Day River, Willow Creek, the Umatilla River, or any of the other tributaries of the mainstem in that reach.102 We would have to assume that no fall chinook salmon were caught by fishermen, or died of any other causes.


By the same logic, we could calculate the number of fall chinook salmon supposedly spawning today in the reservoir behind Bonneville Dam by subtracting counts at The Dalles Dam from counts at Bonneville Dam: 163,000 minus 90,000 or 73,000 fish. But that would be wrong.


Just for fun, I compared more recent fall chinook dam counts at The Dalles and McNary for 1991-93. The results: an average 98,000 fall chinook were counted at The Dalles and 70,000 were counted at McNary.103 The difference? 28,000 fish. This number is probably low by a couple thousand fish because the McNary counts include jacks and The Dalles counts don’t.


Thus the evidence suggests that before and after the construction of John Day Dam, roughly 30,000 adult fall chinook salmon vanished in that reach, for reasons that remain essentially unknown. The net effect from construction of John Day Dam on spawning habitat for fall chinook salmon? Surely salmon spawned there, and perhaps fewer spawn there now. Perhaps the reason that 30,000 salmon now disappear has to do with the rise of the tribal gillnet fishery in the John Day reservoir. But there is not really any scientific evidence to support the claim of a large lost spawning ground.


Yet the Fulton report is cited over and over again by the ISG and others to support the claim that “[s]ome of the most valuable spawning areas were in the mainstem of the Columbia River, nearly all of which were inundated by construction of dams”.104 This is a leading argument for drawdown, but it is not based on scientific data. Maybe the truth is that “inundating” spawning areas doesn’t bother the salmon much. They spawn in the tailraces of dams; they can probably spawn in “inundated” gravel as well. After all, they live underwater.


The same sort of claims about inundated spawning habitat were made more than twenty years ago when fisheries agencies complained about the inundation of habitat alleged to support 5,000 spawning chinook in the Snake River below the mouth of the Clearwater River. The Corps of Engineers pointed out that “accurate counts of the actual numbers of fish spawning in this stretch of river have not been made because the water was too turbid . . . Estimates appear to have been made, at least in part, on the basis of early surveys to catalog areas possessing necessary spawning ground requirements such as gravel availability and proper depths, water velocities, and temperatures.”105


But we now know that chinook salmon never use most of what looks to human beings like suitable habitat, instead sniffing out the few areas where water flows in and under the gravel. Without actually counting salmon redds, biologists have no idea whether valuable habitat has been lost or not.


The ISG also claims that “. . . riparian forest canopy, undercut banks, and large woody debris accumulations in the vicinity of spawning habitat can be critical for survival and successful reproduction of migratory salmonids . . .”.106 There certainly wasn’t much of that along the mainstem Columbia and Snake Rivers as compared to headwaters tributaries. Ultimately, the ISG admits that “[i]t is unknown to what extent reservoirs replace the ecological functions of these lost riverine habitats . . .”—but this is not what they tell the media.107


It is surely true that when John Day Dam was constructed, fishery biologists believed that spawning grounds would be lost. The decision was made to enlarge Spring Creek and Bonneville Hatcheries to compensate for the loss,108 and those hatcheries have been highly successful at producing large quantities of fish for harvest. As far as I know, no one has even tried to figure out whether that enlarged production compensated for alleged losses in the John Day pool. The question is simply ignored.


If we want more spawning grounds in the vicinity of the reservoirs behind John Day and McNary Dams, another approach is simply to dig artificial spawning channels adequately irrigated with cool water and lined with gravel. At one time, before spawning channels fell from favor, they provided 42 percent of the summer chinook juvenile outmigrants in the Columbia River Basin.109 I have heard that some of the channels failed because the operators failed to consider the importance of subsurface water flow, but have seen no data one way or the other. Some irrigators and I were once talking about this, and all agreed that it was worth trying to run drainpipes off the bottom of irrigation dams, burying them in a load of gravel, and creating artificially-irrigated spawning beds with perfect upwelling for salmon.


Spawning channels may not be “natural”, but neither are birdhouses. They may represent a reasonable compromise between hatcheries and natural production; hatcheries might serve to seed the channels for greater consistency of returns. This kind of “supplementation” probably wouldn’t interfere at all with traits needed for survival in the wild. Hatching channel operators would be able to select from adults to improve the size and fecundity of the breed, and seed the gravel with the results; the emerging juveniles would have to learn to fend for themselves. Unfortunately, these sort of solutions are no longer politically-correct. With a decentralized approach to habitat restoration and hatchery management, communities in the Columbia River Basin that wanted salmon to return could experiment with approaches like this.


59 Big River News (Fall 1996), at 13.

60 ISG, Return to the River 205.

61 D. Chapman (pers. comm.)

62 G. Easterbrook, A Moment on the Earth 565.

63 NRC, Upstream at 195. (Prepub. ed.).

64 Cited in id. at 211.

65 B. Rudolph, “NMFS to Army Corps of Engineers: Study Deep Drawdowns”, NW Fishletter, Jan. 21, 1997.

66 Id.

67 From ISG, Return to the River, at 31A (Figure 2.6)

68 D. Rogers & R. Hilborn, “Impact of Redd Loss at Vernita Bar on Hanford Reach Chinook Salmon Production”, Final Report 1988, BPA Contract No. DE-AM79-87BP35885, at 3 (Oct. 1988).

69 See Letter, D. Godard (Manager of PUD No. 2 of Grant County) to M. Walker, April 15, 1997.

70 Quoted in B. Rudolph, "Tribes Complain about Upcoming Project Scrutiny", NW Fishletter, Oct. 14, 1997.

71 ISG, Return to the River 268 (emphasis added).

72 ISG, Return to the River 77.

73 B. Rudolph, “Archaelogist takes long look at salmon recovery”, NW Fishletter, Jan. 29, 1997.

74 ISG, Return to the River, at 212.

75 Id. at 211.

76 This picture, by a photographer named Towne, shows a grain chute running from the hills down to a dockside warehouse, and is from the collection of the Oregon Historical Society (#OrHi 1727). A collection of pictures with other pre-dam views of the Snake River can be found in Steamboat Days on the Rivers (Oregon Hist. Soc. 1969).

77 See, e.g., “Measures to Enhance Salmon and Steelhead Migration Success During 1997” , at 4 (Idaho Governor’s Office Mar. 25, 1997).

78 “Special Report: Lower Snake River Fish and Wildlife Compensation Plan, Lower Snake River, Washington and Idaho”, at 66-67 (USACE Walla Walla Dist. June 1975).

79 “Interim Report, Supplement to Special Report, Lower Snake River Fish and Wildlife Compensation Plan, Lower Snake River, Washington and Idaho, June 1975”, at 8 (USACE Walla Walla Dist. April 1996).

80 D. Chapman (pers. comm.)

81 ISG, Return to the River, at 130.

82 ISG, Return to the River, at 147 (emphasis added).

83 Id. at 157.

84 Quoted in B. Rudolph, “Snake River Chinook far from starved, scientists say”, Clearing Up, Dec. 9, 1996, at 6.

85 Id.

86 W. Muir & T. Coley, “Diet of Yearling Chinook Salmon and Feeding Success During Downstream Migration in the Snake and Columbia Rivers”, Northwest Science 70(4) (1996) (Figure 4).

87 Id.

88 ISG, Return to the River 155-57.

89 D. Chapman et al., “Status of summer/fall chinook salmon in the mid-Columbia region”, Feb. 28, 1994, at 85 (D. Chapman Consultants).

90 Id.

91 ISG, Return to the River 136-37.

92 Id. at 147.

93 Id. at 91.

94 L. Fulton, Spawning areas and abundance of chinook salmon (Oncorhynchus tshawytscha) in the Columbia Basin—past and present”, USFWS Spec. Sci. Rpt.—Fisheries No. 571 (1968).

95 Id. at 4.

96 Id. at 16.

97 Id. at 19.

98 ISG, Return to the River 431.

99 L. Fulton, Spawning areas and abundance of chinook salmon (Oncorhynchus tshawytscha) in the Columbia Basin—past and present”, USFWS Spec. Sci. Rpt.—Fisheries No. 571, at 22 (1968).

100 Id.

101 Id.

102 Fulton’s report contains estimates for fall chinook turning out of the mainstem river into tributaries between Bonneville Dam and The Dalles Dam, but simply ignores the tributaries between The Dalles and McNary. It may well be true that fall chinook have never been a significant presence in these tributaries.

103 USACE, “Annual Fish Passage Report 1993” (Table 40 (The Dalles); Table 67 (McNary))

104 ISG, Return to the River 91; see also id. at 147.

105 “Special Report: Lower Snake River Compensation Plan, Lower Snake River, Washington and Idaho”, at 18 (USACE Walla Walla Dist. June 1975).

106 Id. at 135.

107 Id. at 147-48 (the ISG speculates that the status and trend of fish populations suggests that reservoirs do not replace lost riverine habitats).

108 J. Kincheloe, “Panel 8: Compensation for Fishery Resource Damage`: The Federal Background”, in E. Schwiebert (ed.), Columbia Basin Salmon and Steelhead 166, Spec. Pub. No. 10 (Am. Fish. Soc. 1977).

109 This tantalizing fact is reported in “Compilation of Salmon and Steelhead Losses in the Columbia River Basin”, at 212, with no further explanation.

Previous PageTable Of ContentsNext Page

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