CONTENTS Introduction The Fishery Stock Abundance and Biomass Indices Mortality Estimates of Stock Size and Fishing Mortality Biological Reference Points Conclusions Acknowledgments Literature Cited
Northeast Fisheries Science Center Reference Document 06-10
Assessment of the Georges Bank Atlantic Cod Stock for 2005Loretta O'Brien, Nina Shepherd, and Laurel Col
National Marine Fisheries Service, 166 Water Street, Woods Hole MA 02543-1026
Web version posted June 27, 2006Citation: O’Brien L, Shepherd N, Col L. 2006. Assessment of the Georges Bank Atlantic Cod Stock for 2005. Northeast Fish. Sci. Cent. Ref. Doc. 06-10; 148 p.
Information Quality Act Compliance: In accordance with section 515 of Public Law 106-554, the Northeast Fisheries Science Center completed both technical and policy reviews for this report. These predissemination reviews are on file at the NEFSC Editorial Office.
This report presents an updated analytical assessment of the status of the Georges Bank Atlantic cod Gadus morhua stock (NAFO Division 5Z and Subarea 6) for the period 1978-2004 based on analysis of USA and Canadian commercial landings data and research vessel survey data through 2004. Estimates of 2004 fishing mortality and spawning stock biomass, 2005 beginning year stock size, and the precision of the fishing mortality and spawning stock biomass estimates are presented.
Total commercial landings of Georges Bank cod in 2004 were estimated at a record low 4,583 mt, a 42% decrease from the 7,963 mt landed in 2003. The USA fleet landed 76% of the total, and the Canadian fleet landed the remaining 24%. Fishery-independent surveys conducted by the Northeast Fisheries Science Center show similar trends in both biomass and numbers of cod since 1982. The 2004 indices remain below the long term average. Autumn recruitment indices of age 1 cod have been below the time series average since 1989 and among the lowest in the time series.
Spawning stock biomass declined from about 93,000 mt in 1980 to a record low of 17,340 mt in 1995, and had increased to 30,000 mt by 2001 primarily due to the growth of the 1996 and 1998 year classes. The 2004 estimate of SSB was 22,564 mt, a decrease of 25% from 2001. Fishing mortality doubled between 1979 and 1985, increased to a record high of 1.5 (72% exploitation rate) in 1994, and has since declined to 0.24 (19% exploitation rate) in 2004. The 2004 year class (10.4 million age 1 fish) was strong, but below the time series average (14.7 million age 1 fish). The 2003 year class (21.2 million age 1 fish), however, is the first above average year class since 1989. This year class, and continued low exploitation with minimal discarding, offer a realistic opportunity for the stock to recover toward the SSBMSY of 217,000 mt.INTRODUCTION
This report presents an updated analytical assessment of the Georges Bank cod Gadus morhua stock (NAFO Division 5Z and Subarea 6) for the period 1978-2004 based on analysis of commercial landings and research vessel survey data through 2004. This stock was last assessed and peer reviewed in October 2002 (O’Brien et al. 2002) . Landings were 12,769 mt in 2001 and fully recruited F (ages 4-8, unweighted average) was estimated to be 0.38 in 2001, the second lowest F in the time series (1978-2001). Spawning stock biomass was 29,170 mt in 2001 and continued the increasing trend from the record low estimate of 17,375 mt in 1995. Since 1991, recruiting year classes were all below the long term average and the 2000 and 2001 year classes were the lowest in the time series. The NEFSC spring and autumn bottom trawl survey indices continued to remain near record low values. Autumn recruitment indices for age 2 fish from the 1994 through 1998 year classes were all below the time series (1963-2000) average. The most recent above average autumn recruitment index occurred in 1993.
The current assessment presented here is considered an update and the methodology has remained the same as in the 2002 assessment (O’Brien et al. 2002) . The life history of Georges Bank cod and the history of the commercial fishery are described in O'Brien (1999) . An outline of the history of management is provided in Appendix 1, Table 1 and Appendix 1, Table 2.THE FISHERY
The collecting and processing of the commercial fishery and landings data has been conducted using two methods during the time series. Prior to 1994, information of the catch quantity, by market category, was derived from reports of landings transactions submitted voluntarily by processors and dealers. More detailed data on fishing effort and location of fishing activity were obtained for a subset of trips from personal interviews of fishing captains conducted by port agents in the major ports of the Northeast. Information acquired from the interview was used to augment the total catch information obtained from the dealer.
In 1994, a mandatory reporting system was initiated, requiring anyone fishing for or purchasing regulated groundfish in the Northeast to submit either vessel trip reports (logbooks) or dealer reports, respectively (Power et al. 1997) . Information on fishing effort (number of hauls, average haul time) and catch location were now obtained from logbooks submitted to NMFS by vessel captains instead of personal interviews. Estimates of total catch by species and market category were derived from mandatory dealer reports submitted on a trip basis to NMFS. Catches by market category were allocated to stock based on a matched subset of trips between the dealer and logbook databases. Both databases were stratified by calendar quarter, port group and gear group to form a pool of observations from which proportion of catch, by stock, could be allocated to market category with the matched subset. The cross products of the market category by stock proportions derived from the matched subset were employed to compute the total catch by stock, market category, calendar quarter, port group, and gear group in the full dealer database. The USA landings for Atlantic cod for 1994-1996 were derived for Eastern Georges Bank (statistical areas 560, 561, 562, 551, 552) and Western Georges Bank (statistical areas 520-526, 530, 537-539, 600-639) using the proration methodology described above. The 1997-2004 data were also prorated using the same methodology; however, the criteria for matching the data were modified and resulted in a larger data set being available for proration (Wigley et al. 1998). In May 2004, dealers started submitting the dealer report electronically.
Total commercial landings of Georges Bank cod in 2004 were estimated to be a record-low 4,583 mt, 42% lower than in 2003 (Table 1, Figure 1). The USA fleet landed 76% (3,471 mt) of the total, and the Canadian fleet landed the remaining 24% (1,112 mt).
USA cod landings are generally highest in the second calendar quarter (April-June) and are taken predominantly from the western part (statistical areas [SA] 521-522, 525-526, 537-539, and Subarea 6) of Georges Bank throughout the year (Figure 2 and Figure 3). The majority of the landings from the eastern part (SA 561-562) of Georges Bank are taken in the first and second calendar quarter (January to June; Table 2). The Canadian fishery for Georges Bank cod opens in June, and the majority of the landings are taken in the third calendar quarter (July-September).
USA landings were taken primarily by otter trawl gear (69%) and gillnet gear (16%) during 2000-2003. In 2004, otter trawl gear accounted for the majority (80%) of the USA landings (Table 3). Canadian landings were taken primarily by the longline (53%) and otter trawl (34%) fisheries during 2000-2003. In 2004, otter trawl gear accounted for 55% and longline gear accounted for 30% of the Canadian landings (Hunt et al. 2006).
Cod landings from Georges Bank, categorized by size as 'scrod' (small) , 'market' (medium), and 'large', continued to be dominated by 'market' cod in both weight (63%) and number (69%) in 2000 (Table 4). Historically, 'market' cod have accounted for 37-67% of the landings by weight.
Preliminary estimates of the weight of fish discarded on otter trawl and gillnet trips were derived for 1989-2004 using the Sea Sampling Database. Discard ratios were estimated as the amount of cod discarded to the amount kept for catch taken for all of Georges Bank (SA 521-522, 525-526, 537-539, 561-562). Discard to kept ratios ranged from 0.001 to 0.12 in the otter trawl fishery and from 0.001 to 0.193 in the gillnet fishery, but were predominantly less than 0.10 in both fisheries (Appendix 2).
Discard estimates were not included in the assessment due primarily to the lack of data for 1978-1988. In addition, the available sea sampling data from 1989-2000 are limited by both inadequate coverage of trips and few biological samples.
Recreational cod landings during 1981-2000 ranged between 300 and 8,600 mt, accounting for 1-19% of the total landings (Table 5). Recreational landings were 346 mt in 2004 and accounted for 2.6% of the total (total commercial + recreational) landings.
A previous assessment that incorporated recreational landings in the catch at age resulted in slightly elevated stock sizes with little change in fishing mortality or spawning stock biomass estimates from the VPA (O'Brien 1999) . The 24th SARC recommended that recreational catches not be included in the assessment analysis at that time because (1) the recreational catch at age is based on very few length samples and may not fully characterize the recreational landings, (2) including the recreational catch at age would require excluding the first three years of the time series given the lack of recreational landings data for 1978-1980, and (3) there was a minimal difference in estimates of fishing mortality and spawning stock biomass observed in the terminal year from comparable ADAPT formulations that had commercial catch at age only vs. commercial plus recreational catch at age (Northeast Fisheries Science Center 1997).
The numbers of samples taken to characterize the length and age composition of the USA and Canadian commercial cod landings from Georges Bank are summarized in Table 6 and Table 7. Sampling intensity was very high in 2004 with 1 sample per 27 mt for the USA (Table 7) and 1 sample per 9 mt for the Canadian fishery. The average number in each length sample was 70 fish for the USA and 141 fish for Canada during 2004. Although overall sampling intensity was high, the spatial and temporal pattern of sampling for USA landings resulted in semi-annual pooling of quarterly samples for the 'scrod' market category, and the third and fourth quarter for the ‘medium’ and ‘large’ market category.
During 1981-2003, the number of fish sampled ranged from 0.01 to 0.06% of the total number landed. In 2004, 0.11% of the fish landed were sampled (a total of 190 fish).
Commercial Landings at Age
The age composition of the 1978-1993 USA landings was estimated, by market category, from length frequency and age samples pooled by calendar quarter. Landed mean weights were estimated by applying the length-weight equation:
ln Weight (kg,live) = -11.7231 + 3.0521 ln Length (cm)
to the quarterly length frequency samples, by market category. Numbers landed, by quarter, were estimated by dividing the mean weight into the quarterly landings, by market category, and prorating the total numbers by the corresponding market category sample length frequency. Quarterly age-length keys were then applied to the numbers-at-length to estimate numbers landed at age. Annual estimates of landings at age were obtained by summing values over market category and quarter (Table 8). Derivation of landings by quarter, rather than by month, was performed since not all months had at least two length frequency samples per market category (i.e., minimum desired for monthly catch estimates).
The age composition of the 1994-1996 USA landings was also estimated, by market category, from quarterly length frequency and age samples, but in some years samples were pooled semi-annually due to an insufficient number of samples within a quarter. The landings were disaggregated into eastern (SA 561-562) and western Georges Bank (SA 521-522, 525-526, 537-539). The age composition of the USA landings from eastern Georges Bank was estimated by applying USA length frequencies and combined USA and Canadian age samples, while the age composition of the USA landings from western Georges Bank was estimated by applying USA length frequencies and age samples.
The age composition of the 1997-2004 USA landings was estimated in a similar manner, however; due to the lack of length samples from eastern Georges Bank, combined length frequencies were applied. The assumption was made that length frequencies from eastern and western Georges Bank would be similar, therefore all length frequencies were combined to characterize the eastern component of landings. In addition, for 2000 only, the otter trawl and gillnet samples were pooled, and the handline and longline samples were pooled and then applied to the corresponding combined landings. The 1994-2004 landings-at-age was then derived as described above for the 1978-1993 landings-at-age. The eastern and western Georges Bank landings-at-age were combined to obtain the landings-at-age matrix for USA Georges Bank cod landings for 2000 (Table 8). The USA eastern Georges Bank landings-at-age was included in the Canadian assessment of cod in area 5Zj,m (Hunt et al. 2006).
Canadian landings-at-age data (Table 9) from the Northeast Peak of Georges Bank (SA 551-552) were provided by J. Hunt (DFO, St. Andrews, NB, pers. comm) for 2004. Canadian and USA data were combined to produce a total landings-at-age matrix for 1978-2004 (Table 10). The USA fishery accounted for 75% and 76% of the total landings by number and weight, respectively in 2004.
Total commercial landings and USA landings in 2004 were dominated in numbers by age 3 fish from the 2001 year class and in weight by age 6 fish from the 1998 year class (Table 11, Figure 4). In the Canadian fishery the landings were dominated by the 1998 year class in both weight and numbers of fish (Table 11, Figure 4).
Commercial Mean Weights at Age
Mean lengths and weights at age for ages 1-10+ are summarized for USA, Canadian, and total landings in Tables 8-10. There does not appear to be a consistent trend in mean weight by age during the 23-year time series. The mean weight for age 2 fish in 2003 and 2004 are the highest in the time series and may be due to the increase in mesh size in 2002 and the high sampling intensity. Variability in mean weight of the older fish in recent years may be due to fewer older fish (age 10+) being landed. Beginning year stock mean weights at age, derived from catch mean weights at age (Rivard 1980) are presented in Table 12 and Figure 5.STOCK ABUNDANCE AND BIOMASS INDICES
Commercial Catch Rates
A general linear model (GLM) was applied to all USA interviewed otter trawl trips landing cod from Georges Bank and South during 1978-1993, to derive standardized fishing effort and commercial landings-per-unit-effort (LPUE) (O'Brien 1999, Mayo et al. 1994) . Standardized fishing effort and LPUE during 1994-2000 were estimated by applying the re-transformed GLM coefficients (area, quarter, tonnage class, and depth) to the effort estimate of all trips reporting cod landings in the Vessel Trip Reporting (VTR) database. Total standardized or "raised" effort was calculated by dividing total USA landings by the standardized LPUE (O'Brien and Munroe 2001).
Under the current management restrictions of days at sea (DAS), greater mesh sizes, closed areas since December of 1994, mandatory logbooks for collection of effort data implemented in May 1994, and other management measures, effort data collected since 1994 may no longer be equivalent to the historic 1978-1993 effort series. Additionally, the effort estimated since 1994 has been derived from provisional data. Given these uncertainties the effort data was not analyzed for 2001-2004. The LPUE series has not been used as an index of abundance in the calibration of the VPA since 1993.
Research Vessel Survey Indices
NEFSC spring and autumn research vessel bottom trawl surveys have been conducted off the Northeast coast of the USA since 1968 and 1963, respectively (Azarovitz 1981) . Indices of abundance (stratified mean number per tow) and biomass (stratified mean weight per tow [kg]) were estimated from both the spring and autumn surveys for Georges Bank cod (strata 13-25) during 1963-2004 (Table 13). All surveys were conducted with a Yankee #36 trawl except for spring surveys during 1973-1981, when a Yankee #41 trawl was employed. No adjustments were made for gear changes; however, the indices were adjusted for differences in fishing power of the Albatross IV and the Delaware II, and for differences between catchability of BMV and polyvalent doors introduced in 1985. The fishing power coefficients of 0.79 and 0.67 and the door conversion coefficients of 1.56 and 1.62 were applied to abundance and biomass indices, respectively (Northeast Fisheries Science Center 1991) . The entire time series for both spring and autumn was re-estimated in 2001 (O'Brien and Munroe 2001) to include any large tows that had previously been excluded, and these indices were used in the calibration of the VPA. Standardized catch per tow at age in number for NEFSC spring and autumn surveys and the catch per tow at age for Canadian spring surveys are presented in Appendix 3, Table 1.
NEFSC spring and autumn catch-per-tow biomass and abundance indices show similar trends throughout the time series (Table 13, Figure 6 and Figure 7). Survey biomass indices were relatively stable between 1963 and 1971, then increased to a record high in 1973. Biomass indices generally declined over the next two decades, reaching record low levels between 1991 and 1994. The indices have improved slightly since 1994 and have fluctuated with no trend. Both the spring and autumn biomass and abundance indices remained below average in 2004. Autumn survey abundance indices for both ages 1 and 2 indicate above-average recruitment of the 1965, 1966, 1971, 1975, 1977, 1980, 1985, and 1988 year classes (Appendix 3, Table 2; Figure 8 and Figure 9). The 2003 year class at age 1 is the strongest year class since the 1992 year class (Figure 10). The magnitude of an above-average year class has been declining over time, particularly noticeable in the recruits at age 1 (Figure 10).
Canadian research vessel bottom trawl surveys have been conducted during February on Georges Bank since 1986. Survey abundance indices for age 1 indicate a strong year class in 1996 and above average year classes in 1998 and 2003. Age 2 indices have fluctuated and generally declined during 1990-2003 (Appendix 3, Table 2 ; Figure 11). Indices for 2005 were not available due to vessel problems encountered during the survey (J. Hunt. pers. comm.). Abundance indices for ages 1 and 2 indicate above-average recruitment of the 1985, 1988, and 1990 year classes (Figure 11). In 1993, 1994, and 2005, the Canadian survey did not sample the western part of Georges Bank (Canadian strata 5Z5 - 5Z7), therefore the indices of stratified mean number per tow at age in those years were not used in the calibration of the VPA.MORTALITY
Instantaneous natural mortality (M) of Georges Bank cod is assumed to be 0.2 , the conventional value of M used for all Northwest Atlantic cod stocks (Paloheimo and Koehler 1968 , Pinhorn 1975 , Minet 1978).
Estimates of instantaneous total mortality (Z) were derived from both NEFSC spring and autumn survey catch-per-tow indices (Appendix 3, Table 1 and Appendix 3, Table 2). Total mortality was estimated using Heincke=s method (Ricker 1975) with spring data as:
ln (Σ age 4+ for years i to j/ Σ age 5+ for years i+1 to j+1) ,
and with autumn data as:
ln (Σ age 3+ for years i-1 to j-1/ Σ age 4+ for years i to j).
A three-year moving average was fit to each survey series (Figure 12a-12b) and also to the sequential spring and autumn mortality estimates (Figure 12c). The estimates are highly variable throughout the time series, although there appears to be a trend of increasing Z from the mid-1970s to the mid-1990s.ESTIMATES OF STOCK SIZE AND FISHING MORTALITY
The ADAPT calibration method (Parrack 1986 , Gavaris 1988 , Conser and Powers 1990) was used in a Virtual Population Analysis (VPA) to derive estimates of instantaneous fishing mortality (F) in 2004 and beginning-year stock sizes in 2005. The landings at age data used in the VPA consisted of combined USA and Canadian commercial landings from 1978-2004 for ages 1-9 with a 10+ age group (Table 10). The indices of abundance used to calibrate the VPA included the NEFSC 1978-2005 spring survey indices for ages 1-8, the Canadian 1986-1992 and 1995-2004 spring survey indices for ages 1-8, and the NEFSC 1977-2004 autumn survey indices for ages 0-6 (Appendix 3). The NEFSC spring survey was disaggregated into two series based on the use of the Yankee #36 or #41 trawls. The NEFSC employed the #41 trawl during 1973 to 1981. The spring indices were split into a index series for 1978-1981 for the #41 trawl and a series for 1982-2005 for the #36 trawl. The autumn survey indices were shifted forward one age and one year to match cohorts in the spring survey in the subsequent year. The transformed (ln) observed survey indices, standardized to the mean, generally show similar trends between surveys (Figure 13).
Comparative VPA Calibrations
Since the last assessment in 2002 (O’Brien et al. 2002) , modifications have been made to the VPA software. In addition, revisions occurred in the USA 2000 and 2001 age samples and the Canadian commercial catch at age for 2001. The VPA from the 2002 assessment (O’Brien et al.2002) was recalibrated with each of these modifications to determine any effects on the estimates of F and stock size in 2001. The VPA was first run using the new NFT software, then each change was sequentially added to the next VPA calibration (Table 13). Comparison of the new NFT software results vs. FACT software results did not show any substantial change in the original estimates, and neither did the upgraded age data, or the revised Canadian catch-at-age data (Table 14). Based on these results, the revised NFT software and the updated data were used for the final VPA calibration for terminal year 2004.
The base ADAPT formulation provided stock size estimates for ages 1-8 in 2005 and corresponding F estimates for ages 1-7 in 2004. Assuming full recruitment at age 4, the F on ages 8 and 9 in the terminal year was estimated as the average of the F on ages 4-7. The F on age 9 in all years prior to the terminal year was derived from weighted estimates of Z for ages 4-8. For all years, the F on age 9 was applied to the 10+ age group. Spawning stock size estimates were derived by applying pooled maturity ogives for 1978-1981, 1982-1985, 1986-1989, 1990-1994, 1995-1997, 1998-2000, 2001-2002, and 2003-2004 (Table 15) derived from NEFSC spring research survey data using methodology described in O'Brien (1990) . Due to insufficient numbers of fish at age in annual samples, data for adjacent years that had similar annual median maturity at length and age were pooled to derive a more representative ogive.
The final ADAPT calibration results are presented in Appendix 4 for estimates of F, stock size, and SSB at age, and are summarized in Table 15. Estimates of stock size were more precise for ages 2-8, with CVs ranging from 0.29 to 0.46, than for age 1 (CV=0.57). The residual patterns of the indices did not show any strong trends for the four surveys (Figure 14).
Average fully-recruited fishing mortality (ages 4-8) in 2004 was estimated at 0.24 (19% exploitation), a decrease from the 2003 estimate of 0.39 (30% exploitation; Table 15, Figure 15). In the terminal year, fishing mortality estimates from the VPA do not indicate a pronounced domed partial recruitment as observed in previous assessments (O’Brien et al. 2002 , O'Brien and Munroe 2001) . The pattern of domed partial recruitment that had started in 1994 in previous assessments is not as apparent in the present assessment. This may be due to the improved sampling of the ‘large’ market category during 2001-2004 compared to the minimal samples taken during 1994-1999 (Table 7). The lack of samples during 1994-1999 may have caused a biased characterization of the landings to the smaller "market" and "scrod" fish in the catch at age.
Spawning stock biomass declined from about 93,000 mt in 1980 to a record low of 17,340 mt in 1995. SSB slowly increased to 30,000 mt in 2001 primarily due to the growth of the 1996 and 1998 year classes.
The 2004 estimate of SSB was 22,564 mt, a decrease of 25% from 2001 -- the last year the 1996 year class dominated the SSB. The 2004 estimate is an increase, however, of 30% from the record low 1995 estimate (Table 15, Figure 16).
Since 1978, recruitment at age 1 has ranged from 2 million (2002 year class) to 43 million (1985 year class). The 2004 year class is estimated to be about 10.4 million fish at age 1, below the long term (1978-2004) average of 14.7 million age 1 fish. The 2003 year class, however, is estimated to be 21.2 million age 1 fish, well above the long term average. The previous above average year class occurred in 1990 (17.8 million age 1 fish). The 1998 year class (12.8 million age 1 fish), although below average, was the strongest since 1990. The 1994, 2001, and 2002 year classes are the poorest of the 28-year time series (Table 15, Figure 16). The survival ratio of recruits to spawning stock biomass initially indicates a trend similar to the estimates of recruits at age 1; however, after 1991 the relationship is not as apparent (Figure 17). In the last decade, low recruitment at age 1 has generally been associated with low SSB in contrast to the earlier time period (1978-1990) when higher recruitment was realized from higher SSB (Figure 18).
Precision of F and Stock Biomass Estimates
A conditional non-parametric bootstrap procedure (Efron 1983) was used to evaluate the uncertainty associated with the estimates of fishing mortality and spawning stock biomass from the final VPA. One thousand bootstrap iterations were performed to estimate standard errors, coefficients of variation (CVs) and bias for age 1-8 stock size estimates at the start of 2005, the catchability estimates (q) for each index of abundance used in calibrating the VPA, and the F at ages 1-7 in 2004 (Appendix 5).
The bootstrap results indicate that stock sizes were well estimated for ages 2-8 with coefficients of variation (CVs) varying between 0.27 and 0.45. Stock size for age 1 was not well estimated with a CV of 0.90. The CVs for the catchability coefficients for all indices ranged between 0.09 and 0.83. The fully recruited F for ages 4+ was well estimated with a CV=0.18. The distribution of the 2004 F estimates, derived from 1,000 bootstrap iterations, ranged from 0.13 to 0.35 (Figure 19). The spawning stock biomass was reasonably well estimated (CV=0.13) and slightly higher than the NLLS estimate of 22,564 mt (Appendix 5). The distribution of the 2004 spawning stock biomass estimates, derived from the 1000 bootstrap iterations, ranged from 15,000 mt to 33,000 mt (Figure 19).
A retrospective analysis was performed to evaluate how well the current ADAPT calibration would estimate recruits at age 1, spawning stock biomass, and fishing mortality for the five years prior to the current assessment, 1999-2003. Convergence of the estimates generally occurs after about six years (Figures 20a-c). The retrospective analysis indicates a pattern of overestimating the recruits at age 1 in the recent years, and then underestimating the recruits in the earlier years (Figure 20a). There is no retrospective pattern for SSB estimates in the recent years, but there was a pattern of overestimation for 1996-1999 (Figure 20b). Estimates of fishing mortality (F) are consistently underestimated (Figure 20c). Factors influencing the retrospective pattern may include mis-reporting of catch, immigration or emigration; an unrepresentative estimate of natural mortality; mis-specification of the model; and inadequate sampling that does not properly characterize the landings.BIOLOGICAL REFERENCE POINTS
Yield and Spawning Stock Biomass per Recruit Biological Reference Points
Yield and spawning stock biomass biological reference points (BRPs) were not re-estimated during this assessment. The Working Group on Re-Evaluation of Biological Reference Points did, however, re-estimate the BRPs with input data (Table 16) derived in the 2001 assessment using the methods of Thompson and Bell (1934) (NEFSC 2002). Input data for catch weights (ages 1-10+) and stock weights (ages 1-9) were derived from the long term average weight during 1978-2000 (O’Brien and Munroe 2001). Stock mean weights for ages 10+ were derived from an expanded age structure out to age 18 (oldest age observed in survey) at F = F40% = 0.167 and M= 0.2. The mean weights for ages 10 to 18 were estimated from the length- weight equation (O’Brien and Munroe 2001): ln Weight (kg, live) = -11.7231 + 3.0521 ln Length (cm). The mean length at ages 10-18 were derived from the linear regression of length vs. ln (age) using the 1978-1997 commercial length sample data. The partial recruitment (PR) was based on a normalized geometric mean of 1996-1999 fishing mortality and the maturity ogive was from the 2001 assessment (O'Brien and Munroe 2001) .
The newly estimated biological reference points for F0.1=0.169, Fmax = 0.331, and F40% = 0.167 (Table 16, Figure 21) are slightly lower than those reported in O’Brien and Munroe (2001). These BRPs were applied in the 2002 GARM assessment (O’Brien et al. 2002) , and in the current assessment.
Age-structured Production Model Reference Points
Maximum sustainable yield (MSY) reference points were derived from an age-structured production model (Sissenwine and Shepherd 1987) using data and results from the 2001 assessment (O'Brien and Munroe 2001) . The Working Group on Re-Evaluation of Biological Reference Points developed the reference points using a Beverton-Holt stock recruit relationship with an assumed prior for the unfished recruitment and yield- and spawning stock-per-recruit information (NEFSC 2002.) . The model provided the following reference points:
FMSY = 0.175,
MSY = 35,200 mt and
SSBMSY = 217,000 mt.
The MSY includes commercial landings only and does not include recreational landings or discards.
Short term projections will not be presented in this assessment. These analyses will be conducted by Plan Development Team (PDT) of the New England Fisheries Management Council. A comparison of the projections conducted for Amendment 13 and the current bootstrap estimates of F and SSB are presented in Figure 22.CONCLUSIONS
Georges Bank Atlantic cod are overfished, and overfishing is occurring. Biomass indices derived from research surveys indicate that the stock remains below the long-term average of the 43-year time series. Fishing mortality has been steadily declining since 1997, except for a slight increase in 2001, and is currently at the lowest exploitation in the time series. Spawning stock biomass reached a record low in 1995 and slowly increased, due to the growth of the 1996 and 1998 year classes; however, since 2001 SSB has been declining. The 2002-2004 F trajectory is less than that projected for A13 and the SSB is slightly higher than the A13 projection. Catch during 2002-2004 was also less than the A13 projection. Retrospective analysis indicates a pattern in which estimates of F in the last year of the VPA are less than the converged estimates of F.
The 1999 and 1998 year classes account for the majority of the USA catch and the 1998 year class accounts for the majority of the Canadian catch in 2004. The 1998 (12.8 million age 1 fish) year class, while below the long term average (14.7 million age 1 fish), represents the strongest year class since the last above-average year class that occurred in 1990 (17.8 million age 1 fish). The 1994, 2000, 2001, and 2002 year classes are the lowest in the time series. The 2003 (21.2 million age 1 fish) year class is the first above average year class since 1990, and recruited to the fishery during 2005.
The NEFSC and DFO survey biomass and abundance indices fluctuated during 2002 to 2005; however, all the indices continue to remain below the long term average. The most recent NEFSC surveys indicate that the 2003 year class may be similar in size to the 1998 year class, and the DFO spring survey indicates that the year class is above average.
The lack of strong recruitment in the last decade suggests that recovery of this stock will be largely dependent on maintaining reduced fishing mortality in the near term and husbanding the strong 2003 year class -- and potentially the 2004 year class -- to increase SSB.ACKNOWLEDGMENTS
We appreciate the constructive review of all the members and participants of the GARM. We thank J.J. Hunt and B. Hatt for providing data for the Canadian fishery, and Tom Nies for the most recent NEFMC management measures.LITERATURE CITED
Azarovitz, T.R. 1981. A brief historical review of the Woods Hole Laboratory trawl survey time series. Can. Spec. Publ. Fish.. Aq. Sci. 58: 62-67.
Conser, R.J. and J.E. Powers. 1990. Extensions of the ADAPT VPA tuning method designed to facilitate assessment work on tuna and swordfish stocks. Int. Comm. Conserv. Atlantic Tunas. Coll. Vol .Sci. Pap. 32: 461-467.
Efron, B. 1983. The jackknife, the bootstrap and other resampling plans. Bristol, England: JW Arrowsmith; 92 p.
Gavaris, S. 1988. An adaptive framework for the estimation of population size. CAFSAC Res. Doc. 88/29; 12 p.
Hunt, J.J., L. O'Brien, and B. Hatt. 2006. Population status of eastern Georges Bank cod (Unit areas 5Zj,m) for 1978-2006. Transboundary Resource Assessment Committee 2006/xx; 47 p.
Mayo, R.K., T.E. Helser, L. O'Brien, K.A. Sosebee, B.F. Figuerido, and D. Hayes. 1994. Estimation of standardized otter trawl effort, landings per unit effort, and landings at age for Gulf of Maine and Georges Bank cod. NEFSC Ref. Doc. 94-12; 17 p.
Minet, J.P. 1978. Dynamics and yield assessment of the northeastern Gulf of St. Lawrence cod stock. Int. Comm. Northw. Atlant. Fish. Selected Papers. 3:7-16.
Northeast Fisheries Science Center. 1991. Report of the 12th NE Regional Stock Assessment Workshop (12th SAW). Spring 1991. NEFSC Ref. Doc. 91-03 187 p.
Northeast Fisheries Science Center 1997. 24th Northeast Regional Stock Assesssment Workshop (24th SAW) Stock Assessmemt Review Committee (SARC) consensus summary of assessments. NEFSC Ref.Doc. 97-12 291 p.
Northeast Fisheries Science Center. 2002. Working group on re-evaluation of biological reference points for New England groundfish. NEFSC Ref. Doc. 02-04; 254 p.
O'Brien, L. 1990. Effects of fluctuations in stock abundance upon life history parameters of Atlantic cod Gadus morhua L., for the 1970-1987 year classses from Georges Bank and the Gulf of Maine. Masters of Science Thesis, University of Washington; 95 p.
O'Brien, L. 1999. Assessment of the Georges Bank cod stock for 1997. NEFSC Ref. Doc. 99-02; 122 p.
O'Brien, L. and N.J. Munroe. 2001. Assessment of the Georges Bank Atlantic cod stock for 2001. NEFSC Ref. Doc. 01-10; 126 p.
O’Brien, L., N.J. Munroe, and L. Col. 2002. Georges Bank Atlantic Cod in: Assessment of 20 Northeast groundfish stocks through 2001. A report of the groundfish assessment review meeting (GARM), Northeast Fisheries Science Center, Woods Hole, Massachusetts, October 8-11, 2002. NEFSC Ref. Doc. 02-16; 522 p.
Paloheimo, J.E. and A.C. Koehler. 1968. Analysis of the southern Gulf of St. Lawrence cod populations. J. Fish. Res. Board Can. 25: 555-578.
Parrack, M.L. 1986. A method of analyzing catches and abundance indices from a fishery. Int. Comm. Conserv. Atlantic Tunas. Coll. Vol. Sci. Pap. 24: 209-221.
Pinhorn, A.T. 1975. Estimates of natural mortality of the cod stock complex in ICNAF Division 2J, 3K and 3L. Int. Comm. Northw. Atlant. Fish. Res. Bull. 11: 31-36.
Power, G., K. Wilhelm, K. McGrath, and T. Theriault. 1997. Commercial fisheries dependent data collection in the Northeastern United States. SAW-24 Working Paper Gen.; 3 p.
Ricker, W.E. 1975. Computation and interpretation of biological statistics of fish populations. Bull. Fish. Res. Board Canada; 382 p.
Rivard, D. 1980. APL programs for stock assessment. Can. Tech. Rep. Fish. Aquat. Sci. 953; 103 p.
Sissenwine, M.P., and J.G. Shepherd. 1987. An alternative perspective on recruitment overfishing and bilogical reference points. Can. J. Fish. Aquat. Sci. 44: 913-918.
Thompson, W.F. and F.H. Bell. 1934. Biological statistics of the Pacific halibut fishery. (2) effect of changes in intensity upon total yield and yield per unit of gear. Rep. Inter. Fish. Comm. No. 8; 49 p.
Wigley, S.E., M.Terceiro, A.DeLong, and K.Sosebee. 1998. Proration of 1994-1996 USA commercial landings of Atlantic cod, haddock, and yellowtail flounder to unit stock areas. NEFSC Ref. Doc. 98-02; 32 p.