Columbia River stock declines resulted from long term loss of habitat, commercial harvest, changes in the ocean, and development of the hydrosystem.
The completion of Snake River dams in 1976 changed the river ecology. To mitigate these changes fish are now barged from Snake River dams to below Bonneville dam. Coincident with hydrosystem completion and barging, the ocean went through a fundamental change which decreased survival of fish entering the ocean.
Controversy exists as to the importance of the ocean and dams in producing the recent stock declines. The controversy can be distilled into two competing stories: It's natures fault (ocean) vs. It's our fault (dams).
Our research indicates that, although many factors have caused the salmon decline, barging and improved dam passage have mitigated much of the effects of dams. Poor ocean survival is likely the most important factor in the recent decline.
Flow and spill actions in the recovery plan are analyzed and shown to be ineffective recovery actions. Contrary to the claims of some analyses, the flow augmentation will have little impact on survival. Spill can actually decrease survival because of the resulting gas bubble trauma. Moreover, it is likely that the 1995 spill actions decreased fish survival significantly. Our research suggests that improving barging and dam passage are the most effective salmon recovery actions.
The controversy can be resolved by testing scientific claims in a peer review process and by continued monitoring, experimentation, and evaluation of barging, flow, spill, ocean conditions, and fish condition on fish survival. Open access of information, models and analyses is essential for salmon recovery. The World Wide Web is a our best means of achieving open access.
Habitats used by Endangered Snake River salmon
To determine if these actions are prudent we must consider the reasons for fish declines and the effectiveness of the recovery actions.
Significant events in the Columbia River chinook populations:
A measure of survival, smolts to adults returns (SAR), increases with decreasing travel time through the Snake River (TT). This finding has been used to justify flow augmentation and reservoir drawdown to improve survival to adults. The rationale is that increased flow or reservoir drawdowns will decrease travel time of smolts in the Snake River, thereby increasing survival of adults back to the spawning grounds.
The analysis ignores non-travel time effects including the relationship of seasonal temperature and flow and the ocean regime shift that affected all North Pacific stocks. Both factors can have a significant effect on survival and generate a correlation between SAR vs. TT.
Reservoir drawdown and flow augmentation will have very different, and possibly negative, impacts on temperature and little impact on ocean conditions.
The ocean regime shift involves a switch between two basic patterns of temperature and currents in the North Pacific. In the shift the subarctic boundary moves, north in warm years and south in cold years. The shift occurs every few decades. The last major shift occurred in 1977 when the ocean switched from the cold year pattern to the warm year pattern. We are currently in a warm pattern.
The ocean regime shift is correlated with stock abundances. The warm pattern favors Alaskan salmon. The cold pattern favors west coast salmon.
Decadal shifts in salmon abundance suggest ocean factors play an important role in the Smolt to Adult survival relationship.
The ocean regime shift alters the date of the transition from winter to summer river plume patterns. The timing of the transition and smolt ocean entry may affect smolt survival. Fish entering the estuary prior to the transition appear to have lower survival than fish entering after the spring transition.
The survival/travel time relationship is generated by data from the period prior to the Snake River dams (1976) and the ocean regime shift. Both graphs below can be generated with the ocean survival hypothesis.
Increased spring flows from storage reservoirs is proposed as a recovery action (BiOp2). This requires restricting flows in the winter.
Models of mainstem juvenile survival disagree on the impact of flow. The FLUSH3 model uses older survival data and predicts large benefits from faster travel through the river. The CRiSP1 model uses more recent data and predicts smaller benefits from faster travel through the river.
Survival has improved in the river since the 70's. CRiSP considers the recent passage survival information from 80's onward. The strong travel time survival relationship in FLUSH is biased by early survival studies. The estimates of reservoir survival in FLUSH did not account for high dam passage mortality in the early years. As a consequence, the additional dam passage mortality was included in the reservoir survival biasing that information. Because of steady improvements in the hydrosystem, fish in-river passage survival has improved over the years. This is illustrated by plotting the mortality rate coefficient (lower values equate to higher survival) against year.
Rate coefficient estimates vs. year show improving survival over the years |
CRiSP predictions of improved river survival are supported by independent studies.
Spring chinook survival to Little Goose dam (LGS) in 1993 (CRiSP 70% vs. obs 78%) and Lower Monumental dam (LMO) in 1994 (CRiSP 65% vs. obs 66%)
Mid-Columbia spring chinook survivals from Methow to Priest Rapids dam (PRD). (CRiSP 45.1% and 46.8% vs. obs 47% and 47.3%)
Spring chinook survival from radio tag tracking of fish released below Bonneville dam (BON) (CRiSP 90% vs. obs. > 80%)|
|
Validation sites used in the CRiSP salmon passage model
CRiSP predicted survivals with spill produced gas levels
Information on flow, spill and total dissolved gas are available on the Columbia River Web pages and can be accessed through the Internet from http://www.cbr.washington.edu.
The impacts of the 1995 spill program were evaluated with the CRiSP mainstem passage model.
An analysis of mainstem recovery actions indicates the actions should have little impact on juvenile passage. In terms of additional adults returning to spawn, the actions will yield few additional fish over current operations.
 |
The information present here is in sharp contract to the beliefs held by many salmon advocates. Much of the controversy involves competing models for smolt passage through the hydrosystem. Depending on which model is believed very different approaches should be taken to recover the endangered species. Although there is uncertainty in how to proceed, I believe a significant amount of this uncertainty is in the agencies' and public's understanding of the issues and not in the science. I believe that extensive and peer reviewed model evaluations will advance our understanding, and clarify the real scientific uncertainty. In the current climate much of the public confusion is being attributed to scientific uncertainty. Open access to information can reduce this confusion. The World Wide Web is a vehicle to achieving open access.
Information on which this paper is based is available through the Web page http://www.cbr.washington.edu.
Dr. James J. Anderson is an Associate Professor in the School of Fisheries and Center for Quantitative Science at the University of Washington. His work on salmon issues has been funded by Bonneville Power Administration and the Army Corps of Engineers. The views in this document are a result of that research. The presentation of this paper was supported by the Direct Service Industries, Inc.
1 The University of Washington has developed the Columbia River Salmon Passage model under funding by Bonneville Power Administration. The project began in 1989.
Base Case
Columbia Flow
Drawdown