Status of the Yellowtail Rockfish Resource in 2000

Yellowtail rockfish abundance was last assessed in 1997, this report represents the scheduled triennial update on the status of stocks. Data, compiled for four separate stock units, and evaluated for three model configurations are analyzed using the ADMB version of the assessment model

Stocks: The Pacific Fishery Management Council (PFMC) manages the U.S. fishery as two stocks separated at Cape Mendocino, California (40° 30' N Lat.). Within the assessment we evaluate four separate stock groupings: a coast-wide stock (Cst) extending from Cape Mendocino to approximately 49° N. Latitude, and three subarea stocks. Subarea stocks include: the Eureka/South Columbia stock (EUR) extending from 40° 30' N. latitude (Cape Mendocino) to 45° 46' N. latitude (Cape Falcon); the Northern Columbia stock (COL) extending north from Cape Falcon to 47° 20' N. latitude (Cape Elizabeth); and, the Southern Vancouver stock (VAN) reaching north from Cape Elizabeth to approximately 49° N. latitude.

Catches: U.S. yellowtail rockfish catch increased from 1200 mt in 1967 to 9500 mt in 1983 After imposition of specific yellowtail rockfish trip limits in 1985, catch declined markedly. Over the past five years, 1995-1999, annual U.S. catch has averaged 4300 mt. Coast-wide catch (including Canada) increased from 5000 to 7800 mt or 56% from 1991 to 1992, then gradually declined to 6900 mt in 1995. In 1996, landings rose sharply to 8300 mt due to a strong Canadian catch. Severe restrictions on the U.S. fishery dropped coast-wide landing to 2900 mt in 1997. Since then, landings have gradually increased to 4500 mt in 1999.

Data and assessment: Yellowtail rockfish stock abundance was last assessed in 1997, using a customized statistical age-structured model written with AD Model Builder software. The updated assessment uses the same model with minor modifications. Input data include catch from 1997 to 2000 (2000 catch is assumed equal to 1999 catch), fishery catch at age data from 1977 to 1999 (time series depends on the stock unit evaluated), 1977 to 1998 NMFS triennial trawl survey biomass estimates, 1980 to 1998 NMFS trawl survey age data (sexes combined), a domestic trawl logbook CPUE index from 1988 to 1999, and a whiting fishery bycatch index from 1978 to 1999. In addition the model uses year specific estimates of weight-at-age, area specific estimates of proportion mature-at-age, and an age transition matrix to account for ageing error.

Unresolved problems and major uncertainties: There are significant issues regarding the appropriateness of auxiliary abundance indices used in the model. The whiting bycatch index is uninformative, but the presumptive trend in stock biomass has been fairly flat and a strong signal in the bycatch index may not be expected. The logbook CPUE index may contain some signal relative to stock trends, but there are concerns over bias resulting from annual variability in fishery catchability, the effect of regulations on fishery CPUE, and the unmonitored discard of actual catch. The trawl survey biomass estimates are imprecise which can lead to wide fluctuations in the perception of stock trends. However, we feel the survey abundance indices are least likely to have systematic bias.

Reference points: Benchmark values reported here are from the reference model, which uses all three auxiliary indices. Estimates of unfished biomass (B₀) assume a constant level of annual recruitment equal to the long-term (1967-1997) mean of 14.0 million fish. B₀ is 114,700 mt. Unfished spawning biomass (SPB₀) is 32,500 mt. Target level spawning biomass (SPB_40%) is 13,000 mt. The overfishing threshold spawning biomass (SPB_25%) is 8,100 mt. The recommended fishing mortality rate (F_50%) is 0.09.

Stock biomass: Using results from the reference model, estimated coast-wide biomass in year 2000 was 69,400 mt with a 23% CV. Estimated 1997 biomass was 80,800 mt compared with 56,700 mt estimated in the prior assessment. The revised estimates of biomass are higher than those estimated in the 1997 assessment reflecting the sensitivity of the model to the trawl survey biomass index and assumptions about increases in effective effort of the fleet. Biomass trend in the current assessment shows a period of stability in the mid-to-late 1980s extending through 1995, with recent biomass trending down. Current (year 2000) biomass is 50% of the 1967 estimated biomass and spawning biomass is estimated to be 158% of the target spawning biomass (SPB_40%).

Recruitment: There is no obvious spawner/recruit relationship. Median (1967-2000) annual recruitment is 11.0 million fish at age 4, with average recruitment reaching 13.6 million. The 1989 and 1990 year classes (age 4 in 1993 and 1994) were the last to be above average. The 1995 through 1998 recruitment estimates are about half the median estimate. The current assessment suggests that recruitment is more volatile than depicted in the 1997 assessment, particularly for recent years.

Exploitation status: Fishing mortality peaked in 1983 the last year of essentially unconstrained removals. After imposition of trip limits on the "Sebastes Complex", fishing mortality declined to a low in 1985, and then increased until 1996. Following the 1996 stock assessment, more severe catch restrictions were imposed and fishing mortality dropped to a modern era low in 1997. Since 1997, fishing mortality has climbed back to the 1985 level.

Management performance: Over the last 18 years, yellowtail rockfish trip limits have been ineffective at constraining catch within the HG set for the US fishery. In 14 of the last 18 years, total catch (including estimated discard) exceeded the HG by more than 10%. Since 1983, annual estimated catch has averaged 51% greater than the HG. In the last 10 years, catch has exceeded the HG by 26% and over the last 5 years by 11%. Total catch has exceed Council ABC's an average 53%, 29% and 24% over the last 18, 10 and 5 years respectively. Moreover, total catch has always exceeded the assessment's low estimate of ABC and has exceeded the high ABC 44% of the time since 1983.

Forecasts: Average projected coast-wide yield over the next three years (2001-2003), is 4500 mt for the Ref Model and ranges from 3900 to 4700 mt for other models evaluated. Total stock biomass is projected to decline. This is consistent with the low level of recruitment experienced between 1995 and 1998 and the fact that the stock is above the target level biomass. Over the next three years, recruitment from these below average year classes will form the heart of the fishery, which remains dependent on 7 to 14 year old fish. Spawning biomass in year 2003 is projected to be 112% to 125% of the SPB_40% depending on the model used. Based on stock specific yield projections, we estimate that 14% of the yield should be harvested in the Eureka/S.Columbia area, 42% in the N.Columbia area, and 44% in the S.Vancouver area. Council ABC and OY determinations need to account for expected harvest by Canadian fishers. From 1995 to 1999, the U.S fisheries took on average 81% of the coast-wide catch of yellowtail rockfish.

A more risk adverse harvest strategy could be selected by requiring a higher level of certainty in the projected yield. Cumulative probability profiles of projected yield facilitate such an estimate. For the coast-wide reference case model, and the F_50% SPR rate, the 3-year mean projected yield, consistent with a 75% probability that yield is no less than projected, is 3600 mt, and the estimate consistent with a 25% probability that it is no more than projected is 5400 mt. The lower estimate (3600 mt) is approximately 80% of the mean point estimate (4500 mt). If desired, the lower projected yield could be the basis of a more conservative ABC and OY recommendation.

Recommendations: We have identified three areas of research for yellowtail rockfish: 1) updating estimates of the proportion mature-at-age; 2) evaluating the effectiveness of trawl surveys for species like yellowtail rockfish with particular attention to variation in annual catchability, and 3) stock structure studies using microsatellite DNA