GEOMORPHIC, HYDROLOGIC, AND BIOTIC FACTORS AFFECTING ABUNDANCE OF COHO SALMON IN CENTRAL OREGON COAST RANGE STREAMS

The relative abundance of salmonids is affected by physical habitat, water quality, and biotic factors. We collected data at 244 sites in 10 watersheds in the central Oregon Coast Range to evaluate the factors at various scales that affect the relative abundance of juvenile coho salmon (Oncorhynchus kisutch). All sites lie within the Waldport Ranger District of the Siuslaw National Forest. At each site, the number of coho per 30-meter long transect were counted by ocular estimate and seining; relative abundance was classified as absent (no fish in the transect), low (1-5), medium (6-50), and high (>50). The statistical significance of variables was judged with the chi-square test (for ordinal variables), the Kruskal-Wallis test (for continuous variables), and canonical correspondence analyses (in multivariate analysis).

At the watershed scale, relative abundance depends on the geomorphic landtype, with abundance generally greater in watersheds of moderate relief underlain by marine sandstones and siltstones than in watersheds with high relief underlain by extrusive basalts. Coho abundance also depends on position in the watershed, and highly significant relationships were observed with drainage area, kilometers downstream from the headwaters, stream link magnitude, and stream order. At the reach scale, coho abundance was related to low flow hydraulic conditions, when the field data were collected, more than on bankfull conditions, which control channel dimensions. Channel morphology (entrenchment, channel slope, roughness, percent of transect in pools, effective pool cover) and water temperature were more important than food supply. Substrate and competition from other fish species were not significant. Substrate does affect the abundance of various macroinvertebrates.

The models developed in this study are realistic (based on physical relations between independent variables and coho abundance), accurate (account for a large percent of the variation in coho abundance), precise (use variables that can be measured with consistency), and robust (applicable to well-defined but broad geographic limits). The models can be used to assess the impact of timber harvest activities over time on physical habitat, water quality, and biotic factors.