The California Department of Fish and Game's Trinity River Project conducted tag and recapture operations from May 1991 through December 1991 to obtain chinook salmon (Oncorhynchus tshawytscha), coho salmon (O. kisutch), and fall-run steelhead (O. mykiss) run-size, in-river harvest, and spawner escapement estimates in the Trinity River basin. We placed weirs in the Trinity River near the towns of Junction City and Willow Creek, and trapped 372 spring-run and 1,443 fall-run chinook salmon, 826 coho salmon, and 741 fall-run steelhead.
Based on tagged fish recovered at Trinity River Hatchery and on the return of reward tags by anglers, we estimate 2,381 spring-run chinook salmon migrated into the Trinity River basin upstream of Junction City Weir and that 336 (14.1%) were caught by anglers, leaving 2,045 fish as potential spawners. We estimate 9,207 fall-run chinook salmon migrated past Willow Creek Weir and that 7,231 of these fish continued up the Trinity River past Junction City Weir. Anglers harvested an estimated 1,271 (13.8%) of the fall-run chinook salmon that passed Willow Creek Weir, leaving 7,936 fish as potential spawners.
The coho salmon run in the Trinity River basin upstream of Willow Creek Weir was 9,124 fish, of which 3,996 continued their migration past Junction City Weir. Anglers harvested an estimated 109 (1.2%) of the coho salmon that migrated past Willow Creek Weir, leaving 9,015 fish as potential spawners.
An estimated 11,417 adult fall-run steelhead entered the Trinity River basin upstream of Willow Creek Weir, and 2,285 continued their migration upstream of Junction City Weir. Anglers harvested 2,340 (20.5%) of the adult fall-run steelhead that migrated past Willow Creek Weir, leaving 9,077 fish as potential spawners.
1. To determine the size, composition, distribution and timing of adult chinook and coho salmon, and steelhead runs in the Trinity River basin.
2. To determine the angler harvest and spawner escapements of Trinity River chinook and coho salmon, and steelhead.
The California Department of Fish and Game's (CDFG) Trinity River Project (TRP) conducts annual tagging and recapture operations for adult chinook and coho salmon, and fall-run steelhead in the mainstem Trinity River. This effort determines the composition (species, race, and proportion of hatchery-marked/ or Project-tagged/ fish), distribution, and timing of the chinook and coho salmon, and fall-run steelhead runs in the Trinity River basin. Recaptures of hatchery-marked or Project-tagged fish are used to develop run-size, angler harvest, and spawner escapement estimates for each chinook and coho salmon, and steelhead run.
This is a continuation of studies that began in 1977 with the trapping, tagging, and recapture of fall-run chinook salmon (fall chinook), coho salmon (coho), and fall-run steelhead (steelhead) in the Trinity River in order to determine run-size and angler harvest rates. In 1978, similar studies were added to include spring-run chinook salmon (spring chinook). Steelhead were dropped from the program in 1985 through 1989 and reinstated last year (fall 1990).
The earlier studies were funded variously by the U.S. Bureau of Reclamation (USBR), and with Anadromous Fish Act funds administered by the U.S. Fish and Wildlife Service and National Marine Fisheries Service. The USBR (PL 98-541) has funded the program from 1 October 1989 through the present.
Prior to the current program, all efforts to measure salmon and steelhead populations in the Trinity River basin had been restricted to portions of the upper mainstem Trinity River and certain of its tributaries, or the South Fork Trinity River and some of its tributaries (Gibbs 1956; La Faunce 1965a, 1965b, 1967; Miller 1975; Moffett and Smith 1950; Rogers 1970, 1972, 1973a, 1973b, 1982; Smith 1975; Weber 1965). These earlier efforts did not include fish which use the mainstem and tributaries of the lower Trinity River, or attempt to determine the proportion of hatchery fish in the runs and the rates at which various runs contribute to the fisheries. To develop a comprehensive management plan for the Trinity River basin, all salmon stocks utilizing the basin must be considered.
Trapping and tagging operations were conducted by TRP personnel from May through December 1991 at the same temporary weir sites near the towns of Willow Creek and Junction City in the mainstem Trinity River that were used in 1989 and 1990. The downstream site, Willow Creek Weir (WCW), was located 6.7 km upstream of the town of Willow Creek, 46.8 km upstream of the Trinity River's confluence with the Klamath River, and 136.4 km downstream from Trinity River Hatchery (TRH) (Figure 1). The upstream site, Junction City Weir (JCW), was located 6.4 km upstream of the town of Junction City, 133.2 km upstream from the Klamath River confluence, and 45.5 km downstream of TRH (Figure 1).
The WCW is used to obtain run-size and angler harvest estimates of fall chinook and coho, and steelhead in the Trinity River basin as far downstream as possible. The JCW is used to obtain run-size and angler harvest estimates of spring chinook as far downstream as is feasible during periods of high spring flows. We continue to operate the JCW through December to obtain run-size estimates of fall chinook and coho salmon and steelhead in the upper Trinity River basin.
We trapped at the JCW from 21 May through 13 December 1991, except from 28 May through 5 June when high flows prevented operation. We trapped at WCW from 24 August through 13 December 1991.
At both sites, we attempted to trap four to six nights per week, mid-afternoon on Monday through Friday or Sunday morning. We trapped and tagged fish only at water temperatures <21BC to avoid severely stressing the fish.
As in the previous two seasons, we used the Bertoni (Alaskan) weir design at both weir sites (Figure 2). The weir was supported by wooden tripods set 2.5 m apart. The weir panels were composed of 2.4-m X 2.54-cm (8-ft. X 1-in.) electrical conduit with the centers spaced 5.4 cm apart. The conduit was supported by three pieces of aluminum channel arranged 0.92 m apart, that connected to the supporting tripods. We anchored the tripods with 1.8-m stakes driven into the stream bottom. The weir panels were angled, with the top of the weir standing 1.8 m above the river bottom (Figure 2).
The weir guided fish toward a fyke leading to a trap which measured 2.4 m square and 1.2 m high, and was covered with wood panels to prevent the fish from jumping out of the trap. The trap sides and fyke leading into the trap consisted of 2.54-cm (1.0-in.) electrical conduit welded into panels. The conduit centers were spaced 5.4 cm apart, the same spacing as in the weir panels. The trap entrance was created by elevating the weir conduit allowing fish to enter the fyke and trap.
A gate, inserted between two weir panels, was used to allow boat passage at both weirs. It was modified from the previous years' design by reducing the conduit center spacing from 7.0 cm to 5.4 cm, so that it was similar to the weir and trap. The overall gap between the conduit was reduced to 2.54 cm. The conduit spacing on the gate was reduced to prevent salmonids from 35 to 50 cm fork length (FL) from passing through the gate. During the previous seasons, we noted that salmonids <50 cm were passing through the weir at the location of the gate. (Heubach et al. 1992a, 1992b).
At both weirs, we identified all trapped salmonids to species, measured them to the nearest cm (FL), and examined them for hook and gill-net scars, hatchery marks (fin clips) and tags. All untagged salmonids judged in good condition or unspawned were tagged with a serially numbered FT-4/ spaghetti tag (Project-tagged). To determine angler harvest rates upstream of JCW, various proportions of each species received $10-reward tags at rates inversely related to the numbers of each species we expected to effectively tag during the season. Therefore, all spring chinook, 62% of the fall chinook, 82% of the coho, and all adult steelhead received reward tags, the remainder non-reward tags. This was the second year of a three-year experiment to determine the relative return rates, by anglers above Willow Creek Weir, of the non-reward, $10-, and $20-reward tags. We attempted to tag equal, one-third proportions of the fall chinook, coho and steelhead at WCW with each of the three spaghetti tag types (non-reward, $10-, and $20-reward tags).
We applied non-reward tags on 26% of the fall chinook, 31% of the coho salmon, and 31% of the steelhead trapped. We applied $10-reward tags on 36% of the fall chinook, 33% of the coho, and 33% of the steelhead trapped. While 35% of the fall chinook, 33% of the coho and 32% of the steelhead trapped received $20-reward tags. Our objective was to recover a sufficient number of tags to statistically compare the return rates of the three tag denominations.
To determine tag shedding rates, we removed one-half of the left ventral fin from all spring chinook tagged at JCW. We gave all fall chinook and coho tagged at WCW a single 6.4-mm diameter puncture on the left operculum, while those tagged at JCW received two. The tagged steelhead did not receive a secondary mark at either weir. We released all fish at the respective capture sites immediately after processing.
Each year there is a temporal overlap in the annual spring and fall chinook runs in the Trinity River. Since the timing of each run varies between years, we assign a specific date each season separating the two runs so that numbers of spring and fall chinook can be determined for the run-size and angler harvest estimates. This year we could not use the recovery of fish which were both hatchery-marked and Project-tagged to separate the runs, as was done in 1989-90 and 1990-91 (Heubach et. al. 1992a, 1992b). Too few hatchery-marked salmon were captured and tagged, and consequently too few double marked (Project-tagged and hatchery-marked) fish were recovered. Therefore, we used the entry date of Project-tagged salmon into TRH and the coloration of the chinook salmon at the weir as a subjective indicator of the length of time the fish had been in the river. During the transition period of the run at the weir from spring to fall chinook, dark-colored fish were considered to be late-migrating spring chinook while light-colored fish were considered to be recently migrating fall chinook.
We determined that the spring run was over at both weirs when the light-colored chinook salmon clearly outnumbered the dark-colored chinook salmon. We verified this occurrence by comparing the date that Project-tagged chinook entered TRH to the date that known-race, hatchery-marked chinook entered the hatchery.
As at the weirs, there is an overlap in the migration of spring and fall chinook into TRH. To estimate the respective numbers of spring and fall chinook entering TRH, we expanded the numbers of tags recovered from each returning CWT group by the ratio of tagged to untagged chinook salmon that occurred when they were originally released (same strain, brood year, release site and date). For example, 97,569 fall chinook of CWT code 6-61-46, plus 968,475 unmarked fall chinook were released directly from TRH in September 1987. Since there were 9.9 unmarked chinook salmon released for every CWT chinook salmon released (968,475 unmarked/97,569 marked = 9.9), we multiplied the total number of CWT chinook salmon of code group 6-61-46 by 9.9 to estimate the number of unmarked chinook of that release group that returned to TRH. In doing so, we assumed that return rates to TRH of both CWT and unmarked salmon were the same.
If more chinook salmon entered the hatchery on a particular sorting day than could be accounted for by the expansion of all of the coded-wire-tag groups, we assumed the additional fish were naturally produced. We designated these fish as spring- or fall-run in the same proportions that were determined by the expansion of the coded-wire-tag groups on that day.
We designated the size separating an adult fish from a grilse for spring and fall chinook, and coho based on length frequency data obtained at the two trapping sites and at TRH, evaluated against length data obtained from groups of CWT fish that entered TRH whose exact age was known. Daily chinook salmon FL data from TRH were assigned to either spring or fall chinook when the coded-wire-tag extrapolations indicated >90% of the chinook salmon entering TRH were either spring-run or fall-run fish. Daily FL data from TRH were not used when coded-wire-tag extrapolations indicated the chinook salmon entering TRH were <90% of a specific run.
The length data collected at the weirs and TRH were smoothed with a moving average of five, 1-cm FL increments to determine the nadir separating grilse and adults.
All steelhead >41 cm FL were considered adults, steelhead <41 cm captured at the weirs were assumed to be half-pounders (assumed to have migrated to the ocean) while steelhead <41 cm FL that entered TRH were classified as sub-adults, since we did not know whether they had migrated to the ocean or were resident steelhead.
River surveys were not conducted in the 1991-92 season because very few dead, tagged fish were recovered during river surveys in the previous seasons. We continued to recover dead, tagged fish at the weirs. We examined dead salmonids for tags, fin clips, and spawning condition, and measured them to the nearest cm FL. Heads of adipose fin-clipped (potentially hatchery-marked) fish were removed for the recovery of the coded-wire tag. After examination, the carcasses were cut in half to prevent recounting.
We defined all tagged salmonids recovered dead at the weir or reported by anglers as tagging mortalities, if there was no evidence they had spawned and they were recovered dead <30 days after tagging. Tagged fish recovered dead >30 days after tagging, or those that had spawned, regardless of the number days after tagging, were not considered tagging mortalities.
We processed Project tags returned by anglers to assess sport harvest rates. If not provided with the original tag return, we requested anglers to provide the date and location of their catch in a follow-up thank-you letter. The letter informed them of the fish's tagging date and location.
The Trinity River Fisheries Investigation Project (TFIP), another element of CDFG's Klamath-Trinity Program, conducted salmon spawner carcass surveys in the mainstem Trinity River and its spawning tributaries from Lewiston to the confluence of, and including the North Fork Trinity River (Figure 1). Staff of the TFIP routinely provided us records of the species, tag number, date, and recovery location of Project-tagged fish seen during surveys from 16 September through 19 December 1991.
The TRH fish ladder was open from 16 September 1991 through 27 March 1992. Hatchery personnel conducted fish sorting and spawning operations two days per week through December, and up to seven days per week and twice daily from 2 January through 27 March 1992. More frequent sorting beginning in January was an attempt to reduce suspected predation by river otters (Lutra canadensis) of steelhead in the fish ladder and holding raceway. We considered the initial day a fish was observed during sorting as the day it entered the hatchery.
On all sorting days, salmon and steelhead entering TRH were identified to species, sexed, and examined for tags, fin clips, and secondary tagging marks. We measured all salmon to the nearest cm FL, except those that were Project-tagged fish from the weirs. Project-tagged salmon and steelhead recovered at TRH were assigned the original FL recorded for them at the weir where they were originally tagged.
We removed Project tags from unmarked (no Ad+CWT) salmon on the initial sorting day, while Project tags were removed from hatchery-marked (Ad+CWT) salmon the day they were spawned. On each sorting day, we gave a distinguishing fin clip to hatchery-marked salmon that were placed in ponds to ripen, so the day it initially entered the hatchery (i.e. was sorted) could be determined when it was spawned. Salmon with a secondary tagging mark and no tag were measured to the nearest cm FL and sexed. At the end of the season, we assigned these secondary-marked salmon which had shed their tag, a tag number from a fish of the same species, FL, sex, and weir location where they were originally tagged and released. Tag numbers of the recovered Project-tagged steelhead were recorded the initial day the steelhead was sorted but the tag was not removed.
On the day they were spawned, we removed the heads of all hatchery-marked salmon and placed each in a zip-lock bag with a serially numbered tab noting the date and location of recovery, species, sex, and FL. Salmon heads were given to the CDFG's Ocean Salmon Project for tag recovery and decoding. The Ocean Salmon Project provided us with a computer file of the coded-wire tags recovered for editing and analysis.
We estimated the number of 'effectively tagged' fish by subtracting tagging mortalities of unspawned fish recovered at the weir, dead, tagged fish reported by anglers, and tagged fish recovered or reported downstream of the tagging site from the total numbers of each species tagged at the respective tagging sites.
We determined the run-size estimates for salmon and steelhead migrating into the Trinity River basin above WCW and JCW in 1991-92 by using Chapman's/ version of the Petersen Single Census Method (Ricker 1975):
N = (M+1) (C+1) , where
(R+1)
N = estimated run size, M = the number of `effectively tagged' fish, C = the number of fish examined at TRH, and R = the number of tags recovered (including fish with a secondary tagging mark and no tag) in the hatchery sample.
We attempted to effectively tag and recover enough tagged fish to obtain 95% confidence limits of +10% of the run-size estimate. Confidence limits were determined according to the criteria established by Chapman (1948). In this analysis, the type of confidence interval estimate used is based on the number of tags recovered and the ratio of tagged to untagged fish in the recovery sample.
Each year, we examine the grilse and adult composition of the effectively tagged salmon, the sample of Project-tagged salmon recovered at TRH, and the untagged sample of salmon at TRH to determine if the run-size estimate should be stratified by grilse and adults. Run-size estimates are stratified by grilse and adult salmon when: 1) the proportions of grilse and adult salmon in the effectively tagged sample, the Project-tagged sample of salmon recovered at TRH, and the untagged sample of salmon at TRH are significantly different statistically; and 2) there are sufficient grilse and adult salmon recovered in the tagged sample at TRH to obtain 95% confidence limits of +10% of each of the stratified portions of the run-size estimate.
If we do not stratify the salmon run-size estimate by grilse and adults, we use the proportions of grilse and adult salmon trapped at the respective weirs to estimate the numbers of grilse and adults comprising the run.
All steelhead run-size estimates are for adults only.
For the run-size estimates, we assumed: 1) fish trapped and released from the weir were a random sample representative of the population; 2) tagged and untagged fish were equally vulnerable to recapture (entering TRH); 3) all Project tags and secondary tagging marks were recognized upon recovery; 4) tagged and untagged fish were randomly mixed throughout the population and among the fish recovered at TRH; and 5) we accounted for all tagging mortalities.
Only $10- and $20-reward tags returned by anglers were used to determine angler harvest rates. Each angler harvest rate estimate was computed as the number of reward tags returned by anglers divided by the number of effectively reward-tagged fish released.
The assumptions for the numbers of effectively reward- and non-reward-tagged fish released are the same as those for determining the run-size estimate (See "Run-size Estimates" above). In addition, the numbers of effectively reward-tagged fish released were corrected for tag shedding by multiplying the total number of reward-tagged fish by the percentage of tagged fish recovered at TRH that had not shed their tag.
The confidence limits surrounding the point harvest rate estimate were determined from tables for the binomial distribution. We attempt to effectively reward tag enough fish to obtain 95% confidence limits of "10.0% of the harvest rate.
We estimated the numbers of fall chinook, coho, and steelhead upstream of WCW, and spring chinook upstream of JCW harvested by anglers by multiplying the run-size estimate above the respective weir site by the harvest rate estimate.
The absolute number of fall chinook, coho, and steelhead harvested by anglers in the Trinity River upstream of JCW was determined by multiplying the respective percentage of WCW-tagged fish reported caught upstream of the JCW by the total angler harvest estimate upstream of WCW/.
The mean FLs of samples were compared statistically using a Student's t-test. We did not conduct tests for sample sizes <20 fish and differences in such cases were not considered statistically different. We analyzed the percentages or ratios of adults and grilse, marked and unmarked fish, and the angler return of non-reward and reward tags in samples by Chi-square. A continuity correction (Yates correction) was used for contingency tables of one degree of freedom (Dixon and Massey 1969).
Weekly sampling data collected by Project personnel at the weirs are presented in Julian week (JW) format. Each JW is defined as one of a consecutive set of 52, weekly periods, beginning 1 January, regardless of the day of the week on which 1 January falls. The extra day in leap years is included in the ninth week (Appendix 1). This procedure allows inter-annual comparisons of identical weekly periods.
