the film
Commentaries and editorials

River Operators Discuss Refinement
of Spill Decision Making

by Mike O'Bryant
Columbia Basin Bulletin - October 24, 2003

The Fish Passage Center indicated this week that changing the basis of ending spill at Snake and Columbia River dams from the planning date of Aug. 31 -- contained in the NOAA Fisheries biological opinion -- to fish passage data may not always lead to spill ending in August.

According to information from the FPC, the date on which 95 percent of juvenile salmon and steelhead have passed down through the system of dams from Snake and mid-Columbia River locations to Bonneville Dam more often than not occurs after the end of the BiOp's summer spill planning date of Aug. 31.

The Bonneville Power Administration had asked in August to stop spill at federal dams two weeks early. This year, fish passage data showed, by Aug. 15, 95-98 percent of juvenile salmon and steelhead had already passed downstream from Lower Granite Dam, causing the agency to ask for an end to spill. Bonneville representatives said enough of the juvenile run had passed through the system and that the summer spill operation was costing as much as $1 million per day of lost generation revenue in August.

The request was eventually denied by the federal Executive Committee. But this issue, along with Montana's proposed spill study, triggered a process that called on the Implementation Team to determine if decisions on when to initiate and end spill could be refined from the BiOp planning dates to a decision based on fish passage information, or some combination of the two. The final decision could trigger changes in the 2000 BiOp.

IT, along with its in-season management arm, the Technical Management Team, began that process at a joint meeting this week that facilitator Donna Silverberg said was designed to "let the data talk."

At the meeting, Michelle DeHart of the FPC showed that relying on passage data is fraught with problems, including incomplete data for some individual runs. To get data good enough to make in-season decisions on when to start and stop spill based on fish passage, she said, would require much more tagging and monitoring of fish than is now being done.

Still, with the information that is available, she said that in the years 1991 through 2003, the average date at which 95 percent of fish had passed Lower Granite Dam was Sept. 3, but that's not necessarily representative of when spill should end. Planning dates do not incorporate travel time of juvenile fish into the decision, DeHart said.

"For example, the BiOp sets the planning date at Lower Granite at Aug. 31, but the actual (summer) spill at that time is at Ice Harbor," she explained. Some 95 percent of juveniles generally pass the dam before the Aug. 31 planning date in just four of 13 years. That means that if the end-of-spill decision is based only on a percentage of passage and if that percentage is 95 percent, then spill at lower Columbia River dams and at Ice Harbor Dam would continue into September or later in most years, she said

"What we offered today was how well we did when operating to a planning date," DeHart said. "Sometimes the run is earlier, sometimes it is later."

There are advantages to sticking with the BiOp planning dates, DeHart said. The planning date offers certainty in implementing hydro operations, it lessens the controversy involved in deciding when to stop spill as happened this summer, and the region wouldn't have to spend more money in tagging and monitoring juveniles. On the other hand, cutting off spill according to the planning date would often leave late migrating stocks without spill protection.

By basing the decision on a percentage of fish passing the dams, protection for the juveniles would often be extended beyond the current planning date, she said, and there still is a potential economic advantage in years like 2003 when passage occurs earlier than Aug. 31.

The problem is the difficulty in predicting passage. DeHart said that's because the current data collection system is inadequate. Hatchery releases skew data as they are released in big chunks over just a few days, while wild fish migrate over long periods of time. And the groups of fish that are now being marked don't support the passage approach to managing spill.

She added that it's already too late to begin an expanded marking program for next year. "The smolt monitoring program is already locked in," DeHart said. "Besides, most marked groups that were proposed this year were turned down" for funding. "We're beyond the point where we can mark fish."

"The question for us is how to achieve what protection is supposed to be in the BiOp?" said Howard Schaller of the U.S. Fish and Wildlife Service. "Or, do we get to the same protection levels with other measures? We learned today that there are groups of stocks that are woefully underrepresented in the data. IT needs to make some recommendations about filling the holes in this data."

"We can't do this without a better marking program," said Bill Tweit of the Washington Department of Fish and Wildlife. He suggested that the FPC identify groups of fish that would need marking by facility and to get that information peer reviewed as soon as possible.

One key, according to DeHart, will be a final determination about the benefits of transporting juvenile fish.

Ritchie Graves of NOAA Fisheries said that Douglas and Chelan public utility districts that own three mid-Columbia River dams, along with fishery agencies and tribes, have devised a solution that uses both planning dates and fish passage information for when to begin and end spill at Rocky Reach, Wells and Rock Island dams. The agreement was established in the Habitat Conservation Plans for the dams approved by NOAA Fisheries in late August. The HCPs call for a no net impact on salmonids at the dams and set both fish passage requirements and planning dates for spill.

It set a combined adult and juvenile passage survival standard of 91 percent, but it requires the PUDs to make up for the potential 9 percent loss by increasing hatchery production by 7 percent and by funding tributary habitat enhancement projects that mitigate for the other 2 percent. The $46.5 million price tag is to be spent over the 50-year license period.

The HCPs set planning dates for the when to begin and stop spill at the dams, but they also call for spill at the dams until 95 percent of juvenile salmon have passed. While the planning dates are the default dates for spill, the HCPs also allow for adjustments based on passage information, according to Graves.

Douglas PUD relies strictly on planning dates at its Wells Dam, which lacks monitoring facilities. Generally, it begins spill April 10 and continues without stop to Aug. 15, planning dates that capture 95 percent of the juvenile run in 12 of 21 years, Graves said.

Spring spill at Chelan PUD's Rocky Reach Dam, which has monitoring capabilities, generally begins April 20 and ends June 15, and summer spill begins July 1 and ends Aug. 15. The PUD's Rock Island Dam, also with monitoring equipment, begins spring spill April 17 (it is downstream of Rocky Reach) and ends that spill June 15. It starts summer spill July 1 and ends summer spill Aug. 15.

"We have agreed to use the planning dates at these dams unless the (passage) information compels us to do something else," Graves said. This year, summer spill was extended to Aug. 26 at Wells Dam based on information about juvenile migration from tributaries and to Aug. 16 at Rock Island Dam. Rocky Reach summer spill ended one day early on Aug. 14.

Using this method, Graves said spill operations this year protected 95 percent of the run, with one exception. Spring spill began a little late to capture enough of the early spring chinook run.

Determining what a reasonable end point for spill is a policy issue IT will have to deal with, said Eric Braun of the U.S. Army Corps of Engineers.

"This is a policy question. When is enough enough?" he said. "Is it 95 percent or is it 98 percent? And then a percentage of what? The total run or by ESU?"

Tweit noted that the BiOp no jeopardy decision is based on 100 percent fish passage, not the 95 percent assumed in the FPC data presented at the meeting. "We need some insight into what the jeopardy standard becomes when dropping below 100 percent. Fundamental to that is what is the change to risk?"

"We assumed the RPA (Reasonable and Prudent Alternative) for spill covered 100 percent of the migration," said Jim Ruff of NOAA Fisheries. "We didn't decrement by any percentage of the migration. When we redo the BiOp under the remand, that will be something we will have to look at."

IT will continue the discussion on how to determine when to stop summer spill at its next meeting Nov. 6.

Related Sites:
Implementation Team:
Technical Management Team:
Fish Passage Center (the Historical Review of Fish Migration Data is available here):
Mid-Columbia River dam HCPs:

Mike O'Bryant
River Operators Discuss Refinement of Spill Decision Making
Columbia Basin Bulletin, October 24, 2003

See what you can learn

learn more on topics covered in the film
see the video
read the script
learn the songs
discussion forum
salmon animation