ADDENDUM TO THE

February 23, 1996

BIOLOGICAL ASSESSMENT

for

The Fishery Management Plan for Commercial

and Recreational Salmon Fisheries off the

Coasts of Washington, Oregon, and California

as it affects the Sacramento River

Winter Chinook Salmon

National Marine Fisheries Service

Southwest Region

Fisheries Management Division

JANUARY 31, 1997

INTRODUCTION

Sacramento River winter chinook salmon were listed as threatened under the Endangered Species Act in 1989 and as endangered in 1994. In 1995, the U.S. Fish and Wildlife Service (USFWS) estimated spawning escapement of hatchery-origin winter chinook. This data, combined with estimates of recoveries of coded wire tags (CWT) from the 1994 ocean fisheries, allowed, for the first time since the early 1970s, an estimation of the ocean harvest rate on winter chinook. The estimated 1995 spawning run and the 1994 ocean CWT recoveries indicated an ocean harvest rate of 0.54.⁽¹⁾ The new information on harvest rate and the continued low numbers of spawners resulted in the National Marine Fisheries Service (NMFS) initiating section 7 consultation on the ocean harvest of salmon as managed under the salmon fishery management plan. In March, 1996, NMFS issued a Biological Opinion (Opinion). The Opinion concluded that "although harvest was not identified as a factor causing the decline of the population upon listing, the incidental harvest of winter-run chinook salmon as evidenced by recent CWT recoveries represents a significant source of mortality to the endangered population. NMFS concludes that incidental harvest impacts contribute substantially to restraining the population at very low abundances and limiting population growth, such that the risks of extinction are high from environmental, demographic and genetic stochasticities." The Opinion required that the incidental ocean harvest of winter chinook be reduced sufficiently to allow spawning escapement to increase by 35 percent.

At the time the Opinion was issued, the 1996 winter chinook spawning escapement estimate was not available. The large majority of winter chinook that returned to spawn in 1996 left the ocean prior to the beginning of the 1996 fishing seasons and benefited little, if at all, from the harvest restrictions required by the Opinion. The incidental take in ocean fisheries of the 1992 and 1993 broods occurred primarily in the presence of harvest regulations in effect in 1994 and 1995, prior to the issuance of the Opinion. The escapement data, when combined with the CWTs recovered from the 1995 ocean fisheries, permit estimates of ocean harvest rate and cohort replacement rate for the 1993 brood year and may be considered part of the base period data set upon which harvest reductions required by the Opinion are based. It is therefore appropriate to consider the 1996 escapement data with regards to the requirements of the Opinion. This addendum presents the 1996 winter chinook spawning escapement data and revised estimates of ocean harvest rates and population growth rates.

1996 SPAWNING ESCAPEMENT ESTIMATE AND OCEAN CWT RECOVERIES

Natural Spawning Escapement The size of the naturally spawning winter chinook population is estimated by counting fish as they pass through fish ladders at Red Bluff Diversion Dam (RBDD) after the gates are lowered (closed) on May 16. Based on historical counts at RBDD when the gates were in place throughout the year, it is assumed that the sample counted represents the last 15 percent of the run passing RBDD. Some unknown number of winter chinook may spawn down stream from the dam; however because of unfavorable water temperatures below RBDD, it is believed that little winter chinook production occurs in that reach of the river. Estimates of the total wild run size passing RBDD, and its components, are shown in Table 1. The 1996 adult escapement (excluding marked adults) of 612 fish represents a 129 percent increase over the primary parent brood year of 1993 and the jack estimate, a product of the weak 1994 brood year, is the highest jack return since 1988.

Hatchery Spawning Escapement The purpose of the winter chinook artificial propagation program at Coleman National Fish Hatchery is to supplement the naturally spawning winter chinook population. Efforts to avoid producing fish that imprint on and return to Battle Creek have not been successful. Winter chinook produced at the hatchery have returned there to spawn and no hatchery-origin winter chinook have been recovered at survey sites on the Sacramento River spawning grounds or at Keswick Dam. It is assumed that all of the marked hatchery-origin fish observed passing RBDD returned to Battle Creek, where poor spawning conditions prevent them from contributing to the wild population. Hatchery-origin winter chinook are therefore not included in the analysis of adult replacement rates discussed below. The USFWS has placed a moratorium on collecting winter chinook for artificial propagation pending a solution to the imprinting problem and no 1996 brood year winter chinook were released. Suspension of the propagation program will unfortunately affect our ability to assess ocean harvest rates beyond the 1995 brood.

The run size of the marked, hatchery-origin population is estimated by monitoring survey sites on the Sacramento River and Battle Creek. The USFWS estimated that 237 hatchery-origin winter chinook returned to Battle Creek to spawn in 1996. The number of adults, 114, is estimated using the marked adult fraction observed passing RBDD. This estimate is lower than that obtained at RBDD but is believed to be more reliable because the entire run is monitored on Battle Creek. The USFWS was able to successfully video tape approximately 80 percent of the period between March 26 and July 1, 1996, and surveys on Battle Creek encompassed the entire migration and spawning period.

Ocean CWT Recoveries Three tagged winter chinook were recovered in the 1996 recreational ocean fisheries, one from the 1993 brood and two from the 1994 brood. Table 2 summarizes the tag recoveries from the 1993 through 1996 ocean fisheries.

Ocean Harvest Rate The ocean harvest rate on the 1993 brood is estimated to be 0.19, less than half of the estimated rate of 0.54 on the 1992 brood year (Table 3). A weighted average harvest rate of 0.40 is obtained by pooling the ocean tag recoveries and estimated spawning returns for the 1992 and 1993 hatchery-origin brood years. The estimate of the marked escapement at Battle Creek is such that it does not permit the estimation of confidence limits. Were such confidence limits available, the sampling variance associated with the harvest rate estimate would certainly be large, particularly for the 1993 brood, given the small numbers of tagged winter chinook recovered.

COHORT REPLACEMENT RATE

The Opinion estimated cohort replacement rates (CRR) of winter chinook females from the 1989, 1990 and 1991 broods based on the total run size estimate and an assumption that the spawning runs had a stable sex- and age- composition. While the number of females is a better measure of the reproductive potential of a chinook population than the number of spawning adults, it is likely that the large uncertainties associated with the estimated fraction of females in a given year outweigh any advantage in their use.

Winter chinook females return primarily as age 3 spawners. Only 8 percent of a cohort's spawners return as age 4 spawners and 25 percent as age 2 spawners, over 90 percent of which are males. The CRR analysis presented here uses a three year adult replacement ratio as an estimate of the CRR. The three year adult replacement ratio for the 1992 brood year is the number of adults returning in 1995 divided by the number of adults returning in 1992. The use of the 3 year adult replacement ratio of the 1989 through 1993 brood years results in a slightly lower estimate of the mean CRR than the estimate based on the female replacement rate of the 1989 through 1991 brood years reported in the Opinion.

Table 4 shows the adult spawning escapement and replacement rates associated with the 1989 through 1993 brood years. The mean replacement rate for the population is 1.35, that is, the population has been growing at a mean rate of 35 percent per generation since it was listed in 1989. If a time series is considered which includes the period when the population was in steep decline, for example since 1980, the mean CRR is well below one. However, beginning in the mid 1980s, actions were taken to improve the spawning, rearing and migration habitats of the population, and since incremental improvements have continued since listing, these actions may have removed many of the causes of the run's decline.

The probabilities that future adult 3 year replacement rates would be greater than or equal to 1, given varying increases in the observed 3 year replacement rates for the 1989-1993 broods, are shown in Table 5. If the mean adult replacement rate remained the same as the mean rate for the 1989 through 1993 broods (i.e. no changes in natural or fishing mortality), then the expected probability that future replacement rates will equal or exceed 1 is 0.68. If the mean replacement rate increased by 88 percent, then the expected probability that future replacement rates will equal or exceed 1 increases to 0.90.

The probabilities associated with the adult replacement rates were determined as follows. It is assumed that the adult replacement rate R is distributed as a lognormal random variable, with mean and variance ² on the log-scale: R~lognormal(,²). For any given = *, it follows that the probability of positive growth for any particular generation is

Pr{R1} = Pr{log(R)0} = Pr{t_df-*/},

where t_df is a t-distributed random variable with df = n-1 degrees of freedom, ² = s²(1+1/n),

s² = (y_i-)²/(n-1), and = y_i/n for the observed y_i=log(R_i) data.

For the 1989-1993 brood years: n = 5, = .298, s² = .304778, and ² = .3657336. Thus, for example, with * = .569, Pr{t₄-.569/.3657336} = .80. On the unlogged scale exp(.569) = 1.77; a 31% increase over the 1989-1993 geometric mean of exp(.298) = 1.35.

WINTER CHINOOK OCEAN HARVEST MODEL

The Opinion relied on the Winter Chinook Ocean Harvest Model (WCOHM) to estimate the relative reduction in landed winter chinook associated with various relative increases in spawning escapement. The 1996 ocean salmon seasons were developed using the WCOHM and structured to produce a 35 percent increase in relative escapement of the winter chinook population. Winter chinook ocean abundance is not known and escapements cannot be predicted reliably. The model uses a synthetic population reconstructed from the marked winter chinook broods of 1969 and 1970. Fishery impacts and escapement are projected for this population under alternative fishery regimes, and compared with those impacts and escapement expected under baseline conditions. Baseline conditions are a composite of the 1971 - 1973 fishing seasons and effort. Because fishing seasons remained largely in tact from the early 1970s through late 1980s, it is assumed that these baseline conditions are also indicative of fishery conditions just prior to listing. Model-projected increases in the relative escapement are assumed to correspond to increases in the adult 3 year replacement rate for the population. For example, if the winter chinook population were increasing at a mean rate of 36 percent per generation (CRR = 1.36), a 35 percent increase in relative escapements would require the population to increase at a rate of 84 percent per generation (CRR = 1.84). It is expected that the WCOHM will continue to be used to model the effect of fishery impacts on winter chinook escapement, and will be adjusted consistent with any changes in the best scientific information on hooking mortality rates.

CONCLUSIONS

The Opinion estimated that, since listing, the winter chinook population has grown with a mean CRR of 1.36. Inclusion of the 1996 spawning escapement estimate and reconsideration of the best estimator of CRR results in no substantial change to that estimate.

The best available information on the ocean harvest rate on the 1992 and 1993 brood years indicates an average catch/(catch + escapement) rate of 0.40 rather than the rate of 0.54 reported by the Opinion. Fishery impacts on the these two brood years occurred almost entirely during the 1994 and 1995 fisheries.

In an April 12, 1996, letter to the Pacific Fishery Management Council, NMFS clarified that a 35 percent increase in the mean CRR would avoid jeopardy because it provided a high likelihood of positive growth for the population. The Opinion, however, did not provide probabilities associated with achieving positive growth at various increases in the mean CRR. When possible, fisheries management decisions should be based on risk averse criteria and it is important that estimates of risk, if available, be presented for consideration.

1. ¹The February, 1996, Biological Assessment presented data indicating the catch/(catch+escapement) value for the 1992 brood year was 0.54, which was within the range of the estimates available for the 1969 and 1970 brood years of 0.47 and 0.56 respectively. Harvest rate is the fraction of available adults landed by a fishery. The catch/(catch + escapement) value, when calculated using the ocean and inriver recoveries from all year classes of a brood, is an approximation of harvest rate that does not take into consideration the fraction of the population removed by natural mortality. Throughout this document, the term harvest rate refers to catch/(catch+escapement). The Biological Opinion referred to catch/(catch+escapement) as the harvest fraction.