Preferred stream discharges for salmon spawning and rearing in Washington

Abstract

Stream discharges preferred by salmon for spawning were determined from relationships between discharge and spawnable area at 84 study reaches on 28 streams in Washington. Preferred discharges for spawning were found statistically equivalent for chinook, pink and chum salmon. Regression equations developed for estimating discharges preferred by these species for spawning at other stream sites had standard errors of estimate of 40 percent where a relationship with toe-of-bank channel width was used, and 55 percent where basin drainage area was used. Similarly, equations for estimating the preferred discharge for spawning by sockeye and coho salmon (also statistically equivalent) had standard errors of 48 percent using channel width and 61 percent using drainage area. In general, the discharges preferred for spawning by salmon ranged in magnitude from about 0.3 to 11 times the median monthly mean discharges for September and October and about 0.1 to 6 times the median monthly means for November and December--the four months when spawning is greatest. Stream discharges preferred by salmon for rearing were determined from relationships between discharge and wetted perimeter at the study reaches. Those discharges ranged from about 0.7 to 4 times the median monthly mean discharge for September, when low flows are usually most limiting on the rearing capacity of streams. Equations developed for estimating preferred rearing discharges at other stream sites had standard errors of 57 percent using channel width and 81 percent using drainage area. Peak-unit spawnable area, or maximum area per unit length of channel that has preferred water depths and velocities, was similar for the five salmon species. Equations developed for estimating that area at other sites had standard errors of 27 percent using channel width and 47 percent using drainage area. In general, reducing discharge below the preferred spawning discharge by 25, 50, and 75 percent had the effect of reducing spawnable area by about 5, 15, and 40 percent of the peak-unit spawnable area.