December 01, 2006

Columbia Basin Bulliten

The past decade's variable climate can provide environmental clues to better predict just when Columbia River basin spring chinook salmon will make their spawning surge upriver, according to a study conducted this year by University of Idaho researchers.  

A review of Bonneville fish count data over a 20-year span showed strong correlations between high flow and/or low river temperature and late arriving spring chinook runs.  

The study results were presented Nov. 16 at the U.S. Army Corps of Engineers' Anadromous Fish Evaluation Program annual review of research conducted on fish survival.  

The study, "Timing of Adult Spring Chinook Salmon Returning to the Columbia River : Is There A Pattern?," was conducted by Christopher Peery and Matthew Keefer.  

Predicting the timing and size of the spring chinook return has often bedeviled fishery managers, particularly during the past few years. Late arriving chinook caused the closure of fisheries in both 2005 and 2006 for fear that the anticipated numbers of fish would not appear. Fisheries were ultimately reopened both years when more fish began to cross Bonneville. The 2006 upriver run came closing to matching preseason expectations; the 2005 return fell far short.  

The forecasts are necessary to better plan the timing of mainstem sport and commercial harvests, and set quotas for that harvest. Preseason forecasts rely in large part on the size of past years' age classes of chinook and estimations of how many more from each brood year will return the following season.  

The upriver spring chinook run comes from hatcheries and natural spawning grounds in tributaries above Bonneville Dam in Idaho , Oregon and Washington . They begin entering the Columbia River in March on their spawning journey. On their return they range in age from 2 to 6 years old.  

During the run, dam counts are used to continually update the size of the salmon return. The rate of salmon flow through Bonneville's fish ladders is compared with historic fish timing to estimate the overall size of the run, and thus figure out how many fish can be harvested.  

"During 1985-2006, median and peak count dates were significantly positively correlated with total April discharge, while from 1995 to 2006 (years with available water quality monitoring data) median and peak dates were most significantly correlated with March water temperature," according to the study abstract presented at AFEP.  

"Using these correlations, the late run in 2006 was reasonably predictable: the 2006 environment was characterized by temperatures well below average throughout March and in late April and by April discharge well above average," the abstract says.  

"Further analyses, using additional river environment data, ocean environment data, and more precise measures of run composition may yield a useful, predictive model for spring chinook salmon timing in future years," the report says.  

Peery said he and Keefer are finalizing a report that will describe such a model, and incorporate ocean temperature variables.  

He says that identifying those correlations between environmental factors and run timing may be even more valuable in current times.  

A 30-year period that ended in the mid- to late-1990s witnessed relatively low salmon productivity, mostly blamed on poor ocean conditions created by an apparent cyclical, long-term climate pattern. The pattern then turned positive, spawning some record returns early in the new millennium.  

Ocean returns have been much improved since 2000 overall, but have fluctuated more wildly up and down than earlier. So have basin environmental conditions over the past 10 years, Peery said.  

"They've been all over the board," Peery said of the timing and temperature of runoff, and return of the spring fish.  

The dates at which 50 percent of the annual upriver spring chinook run had passed Bonneville ranged from April 17 to May 10 during the 1985-2006 period that the study analyzed, and peak counts ranged from April 10 to May 9.  

But the five latest-timed years all occurred in the past 10 years, as did the earliest-timed run (2001).  

"In fact, timing during 1996-2006 was far more variable than during 1985-1995, apparently in response to substantial among-year differences in river environment and possibly run characteristics like composition and size," the abstract says.  

Nature has not been following form, with early runoff, late runoff, and water temperatures and flows more often deviating from historic patterns.  

"One way of looking at it is we're getting very unstable," Peery said.  

The study used data from PIT- and radio-tagged fish to examine relationships between fish return timing and temperature and flow variables. Eleven of 21 stocks of upriver showed a correlation with temperature and all showed later arrival in colder years. Timing of nine of the 21 stocks was correlated with flow and all showed later arrival in high-flow years.