- Data from the coho salmon migration were examined from three streams of the Alsea Watershed Study, conducted on tributaries to Drift Creek, Oregon, from 1959 through 1973. With the migration season defined as November 1 through May 31, the mean seasonal smolt count was 2022 fish in Deer Creek, 665 in Flynn Creek, and 277 in Needle Branch. A general decline in numbers in the later years of the study was countered by a significant increase in mean length. The peak migration time was in late March and early April in each stream. A general approach to correlate physical and biological factors to the total seasonal smolt count in each stream through single regression analyses was mostly unsuccessful. A hypothesized positive relationship between summer discharge and smolt yield could not be supported, either by using mean monthly or mean maximum flows. A negative relationship between winter discharge and smolt count was hypothesized, due to
increased mortality from higher flows. This hypothesis was supported by significant regressions that included mean January flows and mean highest flows. Summer temperatures (mean monthly and mean maximum for 7, 15, and 30 days) showed little correlation with the smolt output. Suspended sediment concentrations were negatively correlated with the number of smolts, as expected from the close relationship between sediment and streamflow. September biomass of juveniles in Flynn Creek was significantly related to the smolt output, though the relationships were weaker in the other two streams. The number of spawning females was positively related to the resulting smolt count but the correlations were not significant. Cutthroat trout biomass showed very little correlation with the coho smolt output. Combinations of variables in multiple regression analyses proved generally more successful than the single correlations with smolt count, but interpretation of the variables that proved significant was not always clear. For Deer Creek, 63.9% of the variability in smolt yield could be explained by the regression with September biomass and mean May maximum temperatures. September biomass, January discharge, and mean sediment concentration accounted for 80.7% of the variability in Flynn Creek smolt output. Only 53.8% of the variability in Needle Branch smolt count could be adequately explained (from January discharge and timing of the mean lowest discharge for 15 days), perhaps due to changes in stream variables following logging. When total smolt weight (biomass) was used as the dependent variable instead of count, only the Flynn Creek regression was significant (R²=0.888, with September biomass, mean lowest discharge for 15 days, and spawning female count). It is evident from these analyses that the juvenile populations in each stream have different relationships with various physical and biological factors. Other studies have shown strong positive relationships between
commercial catch of coho salmon off Oregon and Washington and streamflow two years previously, corresponding to the freshwater residence period of the juveniles. In this study, however, negative relationships between winter discharge and smolt output were found. It seems that this negative correlation is masked by downstream, estuary, or ocean factors that are inherent in the relationships between catch and flow.