부산대학교 도서관

We improved our predictions of fall chinook salmon (Oncorhynchus tshawytscha) habitat use by analyzing spawning habitat at the spatial scale of redd clusters. Spatial point pattern analyses indicated that redd clusters in the Hanford Reach, Columbia River, were consistent in their location from 1994 to 1995. Redd densities were 16.1 and 8.9 redds·ha in 1994 and 1995, respectively, and individual redds within clusters were usually less than 30 m apart. Pattern analysis also showed strong evidence that redds were uniformly distributed within the clusters where interredd distances ranged from 2 to 5 m. Redd clusters were found to occur predominantly where water velocity was between 1.4 and 2 m·s, water depth was 2–4 m, and lateral slope of the riverbed was less than 4%. This habitat use represented a narrower range of use than previously reported for adult fall chinook salmon. Logistic regression analysis determined that water velocity and lateral slope were the most significant predictors of redd cluster location over a range of river discharges. Overestimates of available spawning habitat lead to nonachievable goals for protecting and restoring critical salmonid habitat. Better predictions of spawning habitat may be possible if cluster-specific characteristics are used.