Mainstem Passage Strategies In the Columbia River System: Transportation, Spill, and Flow Augmentation (aka "Giorgi Report")

read full document >

Also see Dr. Giorgi's presentation and public comments

Executive Summary

The National Marine Fisheries Service 2000 Biological Opinion (BO) for the Federal Columbia River Power System (FCRPS) prescribes guidelines for smolt transportation, spill and flow augmentation to improve survival of salmonid stocks listed under the ESA. With respect to these strategies the NPPC is concerned about the following issues:

  1. What does the scientific literature inform us regarding the benefits, shortcomings, or risks associated with each passage strategy, and as compared to other passage options?
  2. Which aspects of the scientific information are in dispute?
  3. What are the critical uncertainties attending each strategy?
  4. What is being, or could be done to reduce uncertainty and disputes?

In terms of scope, the NPPC seeks information for both ESA-listed and unlisted salmonid populations across a range of water years. The Council seeks clear, concise and succinct treatment of these issues. Our approach is to review key research and analyses that have appeared in the literature, then distill out the key findings and synthesize the results. The focus is on evaluations conducted under contemporary river operations, which were initiated in the early 1990s and formalized as guidelines in the 1995 and 2000 Federal Columbia River Power System Biological Opinions. Additionally, we identify key uncertainties and gaps in the information base, and identify research that is in place or planned to fill those gaps.

Smolt Transportation

Using general, annual indices of performance, both NMFS and CBFWA analyses showed that the majority of the time, fish transported from Lower Granite and Little Goose dams produced TIRs higher than or equal to corresponding inriver control groups. Bouwes et al. (2001) concluded modest transportation benefits were evident for hatchery chinook, and slight to negligible benefits for wild fish. Sandford and Smith (in press) state that, "once a juvenile fish is entrained in a bypass system at a "collector dam", transporting the fish maximizes the probability of its eventual return as an adult." Based on assessments by those two investigations, it appears that there is a survival advantage associated with transporting Snake River hatchery spring/summer chinook and steelhead, particularly from the upper two dams, Lower Granite and Little Goose dams. However, the rationale for transporting smolts from Lower Monumental and McNary dams is less clear. The benefits of transporting Snake River hatchery fish from those dams are equivocal.

In some years, small sample sizes have resulted in poor or undefined precision for key estimates. This can limit the ability to make statistically defensible conclusions. Authors examining recent estimates do not confidently state that transported fish survive at significantly higher rates than inriver counterparts. Neither Sandford and Smith (in press), nor Bouwes et al. (2001) explicitly tested key hypotheses such as; D > Vc, or TIR > 1.0. Presumably future analyses by these two research groups will do so. In recent times the resurgence in adult returns offers improved precision and opportunities for meaningful statistical tests

Whether or not wild fish respond favorably to transportation is difficult to ascertain at this juncture. Even though the limited numbers` of evaluations indicate higher return rates for transported smolts, the estimates are based on such small sample sizes that the precision for wild fish is particularly poor. Thus, reliance on the point estimate as a representative value is questionable.

Survival from smolt to returning adult (SAR) for hatchery and wild spring summer chinook has increased substantially since 1993, and has been increasing steadily from 1997-1999, reaching SAR levels in 1999 that approach and in some cases exceed the 2% minimum recovery threshold for wild stocks as identified in PATH. This suggests that neither transport nor inriver migration conditions may be a bottleneck to recovery, when marine-based survival is at some adequate level.

No mass transportation study has been conducted that targets Snake River fall chinook. Such evaluations are warranted, and planned for initiation in 2002

There is evidence to suggest that homing fidelity may be impaired for some species of transported fish, including fall chinook, sockeye, and steelhead. Studies that target spring/summer chinook and steelhead require emphasis. Straying may in part contribute to delayed effects associated with transporting smolts. It may be advantageous to ascertain the extent of straying associated with transport of all species to address certain ESA concerns. Excessive straying may result in increased hatchery fish intermingling among wild adults on the spawning grounds. This may not be desirable. Ongoing telemetry/PIT tag-based studies of adult passage should offer additional insight on this matter.

Delayed differential effects relative to inriver migrants are consistently evident for transported fish. However, by-and-large adult return rates to Lower Granite Dam exceed those of inriver migrants designated as controls. In such cases, there would still be a survival advantage to transport Snake River fish from Lower Granite and likely Little Goose dams.


Apart from the Snake River stocks, which can largely be transported, the majority of smolts emanating from the rest of the basin continue to migrate in-river to below Bonneville Dam. Optimizing smolt survival during downstream migration has been a longstanding goal of fisheries managers.

We focus on contemporary passage survival estimates and estimation techniques (balloon-, radio-, and PIT-tag methods) developed during the 1990's that continue to be applied this decade.

The collective information indicates that spillways appear to be the safest passage routes available at dams, even more benign than many smolt bypass systems, particularly those involving the screening of turbine intakes. The magnitude of smolt survival through spillways varies across dams and species. This is particularly evident when total effects are reflected in the empirically obtained estimates. This suggests that species- and dam-specific estimates should be updated for each dam and applied in any future passage modeling analyses. Spillway flow deflectors (gas abatement devices) appear to increase smolt mortality relative to a standard spillbay, by 1-3 percentage points. Even so, survival will typically still exceed the turbine route at most dams. The potential for increased smolt losses at the concrete needs to be balanced against gains associated with gas abatement. It is not clear that passage models currently provide an accurate assessment of this tradeoff.

Studies assessing the direct and total effects associated with spillway passage indicate that survival is related to discharge at some sites, with mortality increasing at excessive discharge volumes. The difference in survival across discharges can range from negligible to nearly 7 percentage points, depending on the dam and species.

In passage modeling analyses, values for model parameters should be periodically updated for each dam and species. The set of empirical estimates that characterize smolt passage survival through spillways, as well as spill efficiency, are being continually expanded. However, that collective information is not being systematically compiled and synthesized on a regular basis for the hydrosystem at large. Notable exceptions include papers by Muir et al. (2001a), Ploskey et al. (2001) and Anglea et al. (2001) for selected sites.

Passage modeling may afford the only practical means to evaluate the relative benefits of various spill scenarios, at the level of the overall smolt population. The other approach requires obtaining reliable empirical survival estimates linked specifically to spill conditions. This requires a well-designed experimental protocol that will likely be very difficult to implement in this complex system of competing uses

The NMFS Total Dissolved Gas standard of a maximum 120% saturation in the tailrace of Columbia River dams is generally achievable by following the dam-specific gas caps identified in the Biological Opinion, and implementing the spill program currently in place in the Mid-Columbia. The exception occurs in higher flow years when spill volumes cannot be effectively controlled due to flows exceeding the hydraulic capacities at the various dams. The standard appears satisfactory for protecting salmonid species within the hydro-system, but it exceeds water quality guidelines established by the Environmental Protection Agency.

The full biological impacts of a spill program have not been evaluated in their entirety. Smolt survival receives emphasis. Model analyses try to predict changes in smolt survival to below Bonneville Dam. Quantitative system analyses have not formally addressed the potential for impacts on adult mortality. Empirical evaluations conducted in situ, have limitations as well. For example those recently conducted by Zabel et al. (in press) and FPC (2001) observed changes over small segments (projects), thus cumulative effects through the system are not evident. Furthermore, results from empirical evaluations are equivocal, because spill effects have not been clearly isolated from other factors.

The effects of spill operations and levels on adult passage behavior as linked to long-term survival are not well understood. Some of the recent adult passage research suggests that higher spill volumes may exacerbate migration delay and fallback. But, convincing quantitative relationships have not been developed. Adult passage studies are continuing and may provide insight on these matters.

Flow Augmentation

Flow augmentation (FA) is the intentional release of water from storage reservoirs for the purpose of increasing flows to enhance migratory conditions for juvenile and adult life stages of salmonids in the Snake and Columbia rivers. Flow augmentation provided to the upper Columbia River (downstream from Chief Joseph Dam) comes from large storage reservoirs such as Grand Coulee Dam and a complex of storage reservoirs that drain into it from Canada and Montana. In the Snake River flow augmentation is provided from Dworshak Dam and through the Hells Canyon Complex in Idaho. The foundation for prescribing such actions is based on two premises:

  1. Increased water velocity  increases migration speed of smolts  increases survival.
  2. Lowering water temperature (summer)  improves migratory and rearing conditions for both juvenile and adult salmonids  results in improved survival.

Information obtained or reported since the early 1990's is the focus of this report, but a brief historical backdrop is provided where needed. Both river operations and the mark-recapture tools and associated analytical procedures have changed markedly from previous decades. Thus, the contemporary information is most applicable today.

Flow effects on smolt migration speed: For most spring-migrating species the evidence indicates that increased flow (water velocity) contributes to swifter migration speed. Information regarding fall chinook is equivocal.

  • River discharge appears to be the most influential variable affecting migration speed of steelhead and sockeye salmon in the Snake and mid-Columbia rivers.
  • Two factors, flow and the degree of smolt physiological development, explain the observed variation in the migration rate of yearling chinook salmon (except in the mid-Columbia where only smolt development has been identified as a predictor variable).
  • At least four variables have been implicated as influencing the migration speed of sub-yearling (fall or summer/fall) chinook; flow, water temperature, turbidity and fish size. However, strong correlations among these predictor variables confound the ability to identify causative agents.

Flow effects on smolt survival: PIT tag based smolt survival estimates acquired since 1993 provide the most relevant data set for characterizing smolt survival dynamics through the impounded mainstem Snake and Columbia rivers.

  • Based on recent PIT tagged based estimates there is little evidence supporting a flow survival relationship across the water years experienced from 1993-2000, for yearling chinook or steelhead.
  • However, in 2001 under the extreme low flow conditions, steelhead survival decreased dramatically to about 63% per project (typically it is near 90%). Slow migration speed and rapidly increasing water temperatures are implicated as causative factors affecting residualization and mortality.
  • A complex of factors is implicated as influencing Snake River fall chinook survival including, flow, water temperature and turbidity. These environmental variables are strongly correlated during the summer migration, confounding the ability to identify the most influential one. Knowing if water velocity or temperature is the most influential could be important when the decision is to use Dworshak or Hell's Canyon for flow augmentation, since the temperature of those water sources differs greatly.

The premise for reducing summer water temperature, particularly in the Snake River, to improve rearing and migratory conditions for juvenile fall chinook and adult salmonids appears sound.

  • The literature indicates that maintaining river temperatures at or below 20?C is advantageous to both life stages of fall chinook, and adult steelhead, all of which are in the river in August and early September.

However, it is not clear that releasing cool water from Dworshak effectively alters the thermal structure of most of the Lower Snake River. The major effect is localized at two upper reservoirs (LGR, LGO) according to results reported by Bennett et al. (1997).

  • When cool water enters the reservoirs it sinks to the bottom. This can provide cool refugia in deeper waters, but not uniform cooling of reservoirs.
  • The greatest change in temperature attributable to FA releases from Dworshak are evident at LGR, where water temperatures under FA are predicted to be as much as 4-8 ?F below base conditions at certain times. At Ice Harbor the difference is on the order of 1-2 ?F.

Flow Augmentation Evaluations are generally lacking. Only a handful of studies have attempted to:

  • Quantify the volume and shape of water provided specifically as FA.
  • Translate that incremental increase in flows to changes in water velocity and temperature.
  • Predict the change in smolt travel time and survival attributable to those increases
  • Identify whether populations of interest (e.g. ESA stocks) have encountered FA events.

The last such evaluation treated information through the 1995 water year, and only for the Snake River. Given the community's sensitivity to this controversial management action, a holistic comprehensive updated evaluation seems prudent, and long overdue. The scope of future evaluations need to more fully address the balance of benefits and risks between anadromous and resident fish resources.