Abstract:
Accurate discrimination between river returns of wild and hatchery anadromous Pacific salmon (Oncorhynchus spp.) is necessary if the status of the two stocks is to be monitored and if the success of hatchery mitigation programs is to be assessed. Usual hatchery management practices, including release of large numbers of unmarked fish and variable fractional marking of releases, prevent such discrimination. Many Pacific-coast salmon hatcheries presently release some fish with an adipose fin clip and a binary-coded wire tag inserted in the nasal region (AD-CWT). If a constant fraction of remaining releases carried a distinctive identifying mark, then the proportion of hatchery fish in subsequent spawning runs could be estimated. Should the majority of hatchery fish in the spawning escapement return to the hatcheries of origin, then the expected change in hatchery fish proportions between river entry and natural spawning grounds, combined with known returns to the hatcheries, may be used to estimate system-wide escapement. Estimators, and their variance, for total escapement are proposed for cases in which: (1) no returning hatchery or wild fish stray from their predestined natural spawning grounds or hatcheries, (2) only hatchery fish stray, and (3) both hatchery and wild fish stray. The variance of an estimated proportion of hatchery fish in the run depends strongly on the fraction of releases marked in excess of AD-CWT releases. The variance declines substantially as this fraction increases from 0.05 to 0.25, but marking fractions above 0.50 gives little further statistical improvement. Sample sizes necessary for precise estimation of the hatchery fish proportion depend directly on the reciprocal of this marking fraction. The variance of escapement estimators becomes inflated when high proportions of hatchery fish stray; when more than half of the hatchery fish in the spawning escapement fail to return to hatcheries, effective application of the estimators becomes very limited. Release of hatchery fish away from the hatchery rearing site may greatly increase later straying rates, and such practices should be avoided if the proposed estimators are to be used in rivers that support substantial wild-fish returns.