Abstract
A model was developed to estimate fishing-induced mortality and exploitation rates for unmarked salmon Oncorhynchus spp. in mark-selective fisheries. This model is analogous to a pipeline, and the basic idea is similar to that of the change-in-ratio method used for estimating animal abundance. During spawning migration, a unidirectional movement occurs from the ocean to near shore and into freshwater. A series of nonselective and selective fisheries consequently catch these fish on their migration path. The model functions as follows: First, a cohort analysis is completed for the coded-wire-tagged fish to obtain their initial abundance and fishery-specific exploitation rate. Second, fishery-specific exploitation rates of unmarked fish are derived. Third, the initial abundance of unmarked fish is obtained from the total exploitation rate and escapement and then the ratio between the abundances of unmarked and marked fish is estimated for each fishery. Finally, the fishery-specific mortality of unmarked fish is estimated. The method assumes that the release mortality rates in selective fisheries are known. Methods for correcting potential bias are suggested. Sensitivity analyses show that the estimated exploitation rates of unmarked fish are less sensitive to bias in the release mortality rate than to bias in the estimated cohort size of marked fish. Bias in release mortality rates results in a diminishing bias of exploitation rates in sequential selective fisheries, but bias in the estimated cohort size results in an increasing bias. An example is given for Sandy River coho salmon O. kisutch in the lower Columbia River. This method does not require double index tagging and can presumably be applied to most fisheries. Some caveats in applying this model to salmon fishery management are given.