A marking, tagging, and recovery program for Central Valley hatchery chinook salmon

Ken Newman, Allan Hicks, Dave Hankin

July 7, 2004

---

EXECUTIVE SUMMARY

We extended statistical procedures that we had previously developed for estimating the production of wild and hatchery chinook salmon in the Central Valley. Corresponding enhancements were also made to a computer program called CFM Sim which simulates marking, tagging, sampling, and production estimation procedures. Using CFM Sim and a factorial experiment design, we evaluated the effects of varying four man-controlled factors on the quality of estimates of wild and hatchery chinook salmon production.

The four man-controlled factors were:

1. Constant Fractional Marking (CFM) rate (f): the percentage of production releases at a given hatchery that received a coded-wire-tag and an adipose fin clip,
2. Catch sampling rate (CSR): the fraction of ocean and freshwater salmon catch being sampled,
3. Escapement sampling rate (ESR): the fraction of in-river escapement being sampled, and
4. Coefficient of variation of escapement estimates (ECV).

Costs were also included in our analysis. In particular we considered the cost of tagging and marking juvenile hatchery-reared chinook salmon, the cost of sampling ocean catches, the cost of estimating in-river escapement, and the cost of recovering and reading coded-wire-tags. Cost calculations were incomplete, however, in that costs for freshwater catch sampling (which, to the best of our knowledge, is not currently being done) were not included. Additionally the costs of escapement sampling, including sampling especially for coded-wire-tagged returns in the in-river escapement, were only approximately calculated.

The association between current escapement sampling efforts, the precision of current estimates of escapement, and the costs of escapement sampling was imprecise and coarsely approximated, too. As a result our recommendations given below, regarding constant fractional marking levels and some of the other manipulable factors, are of a general, relative nature.

General recommendations:

1. Implement a system-wide constant fractional marking program for Central Valley hatchery reared salmon. The recommended constant fractional marking rate for production releases from all hatcheries is at least 1/3 or 33%. If relatively good estimates are wanted for all the individual watershed natural production estimates, then a CFM rate greater than 1/3 may be wanted.
2. Calculate measures of the precision and bias of watershed-specific wild salmon escapement estimates. This means at a minimum calculating standard errors to accompany the point estimates for each watershed (and race, e.g., fall-run, late fall-run, winter run). For one or more watersheds, or portions of watersheds, carry out an estimation procedure which can serve as a benchmark for assessing the accuracy of existing escapement estimation procedures. This could mean installing temporary weirs on smaller streams, for example, which would allow counts of returning spawners to be made in conjunction with implementation of current sampling and estimation procedures on the same streams.
3. Implement a consistent system-wide freshwater catch sampling program and develop, if not already available, corresponding estimation procedures. Standard errors should be calculated along with point estimates of total catches and catches of particular wild and hatchery stocks.
4. In order to evaluate how well a hatchery release group, chosen as a surrogate stock for a wild stock, actually represents the wild stock (in terms of age 3 and older survival rates, harvest rates, and maturation rates, in particular), mark and tag for multiple years, on multiple watersheds outmigrating wild juvenile salmon. Compare the tag recovery patterns for these tagged wild fish with those of the surrogate hatchery fish.