NEFSC Ref Doc 06-17

This assessment of the summer flounder (Paralichthys dentatus) stock along the Atlantic coast (Maine to North Carolina) is an update through 2005/2006 of commercial and recreational fishery catch data, research survey indices of abundance, and the analyses of the data. For 2005, commercial and recreational fishery final quotas were 8,246 mt and 5,434 mt, respectively, for a total of 13,553 mt. The reported commercial landings used in this assessment for 2005 were 7,765 mt, while estimated recreational landings were 4,550 mt, for a 2005 landings total of 12,315 mt. An analytical assessment (virtual population analysis, VPA) of commercial and recreational total catch at age (landings plus discards) was conducted. Indices of recruitment and stock abundance were developed from Northeast Fisheries Science Center winter, spring and autumn, Massachusetts spring and autumn, Rhode Island annual, Connecticut spring and autumn, New Jersey annual, and Delaware annual trawl survey data. Recruitment indices were also developed from young-of-year surveys conducted by the states of North Carolina, Virginia, and Maryland.

The stock assessment indicates the summer flounder stock is not overfished, but overfishing is occurring relative to the biological reference points. The fishing mortality rate declined from 1.32 in 1994 to 0.46 during 2003-2004, before increasing to 0.53 in 2005, well above the overfishing definition reference point (F_threshold= F_target= F_max= 0.276). There is an 80% chance that the 2005 F was between 0.42 and 0.75. The estimate of F for 2005 may understate the actual fishing mortality; retrospective analysis shows that the current assessment method tends to underestimate recent fishing mortality rates. Over the last 5 years, the annual retrospective increase in fishing mortality has averaged 33%. Total stock biomass increased substantially during the 1990s and through 2004, but has decreased slightly since 2004 and was estimated to be 47,800 mt on January 1, 2006 -- just above the biomass threshold (46,323 mt). There is an 80% chance that total stock biomass in 2006 was between 41,600 and 56,900 mt. Spawning stock biomass (SSB; Age 0+) declined 72% from 1983 to 1989 (18,800 mt to 5,200 mt), but with improved recruitment and decreased fishing mortality had increased to 32,600 mt by 2004, before decreasing to 30,600 mt in 2005. Retrospective analysis shows a tendency to overestimate the SSB in the most recent years. Over the last 5 years, the annual retrospective decrease in SSB has averaged 17%.The age structure of the spawning stock has expanded, with 74% at ages 2 and older, and 23% at ages 5 and older. Under equilibrium conditions at F_max, about 85% of the spawning stock biomass would be expected to be ages 2 and older, with 50% at ages 5 and older. The arithmetic average recruitment from 1982 to 2005 is 35 million fish at age 0, with a median of 33 million fish. The 2005 year class is currently estimated to be well below the median recruitment level. Retrospective analysis shows that the current assessment method tends to overestimate the abundance of age 0 fish in the most recent years. Over the last 5 years, the annual retrospective decrease in recruitment has averaged 10%.

Stochastic forecasts only incorporate uncertainty in 2006 stock sizes due to survey variability and assume current discard to landings proportions. If landings in 2006 are 10,700 mt (23.6 million lbs) and discards are 800 mt (1.8 million lbs), the forecast estimates a median (50% probability) F in 2006 = 0.35 and a median total stock biomass on January 1, 2007 of 51,200 mt, above the biomass threshold of½B_MSY= 46,323 mt. A subsequent reduction in fishing mortality in 2007 to F = 0.276 is forecast to yield landings of 9,026 mt (19.899 million lbs) and a median total stock biomass level on January 1, 2008 of 58,100 mt.

The Stock Assessment Workshop (SAW) Southern Demersal Working Group (SDWG) met on June 20, 2006 to update the assessment of summer flounder through 2005/2006. The following scientists and managers participated in the meeting:

Although they were unable to attend the meeting, Laura Lee of the RIDFW, David Simpson of the Connecticut Department of Environmental Protection (CTDEP), Anne Mooney of the New York Department of Environmental Conservation (NYDEC), Don Byrne and Jeff Brust of the New Jersey Department of Fish and Wildlife (NJDFW), Stew Michels of the DEDFW, Steve Doctor of the Maryland Department of Natural Resources (MDDNR), Chris Bonzak of the Virginia Institute of Marine Science (VIMS), Rob O’Reilly of the Virginia Marine Resources Commission (VMRC), and Chris Batsavage of the North Carolina Division of Marine Fisheries (NCDMF) provided research survey and/or fisheries catch data that were used in the assessment.

For assessment purposes, the previous definition of Wilk et al. (1980) of a unit stock extending from Cape Hatteras north to New England has been accepted in this and previous assessments (e.g., NEFSC 2002). The joint Mid-Atlantic Fishery Management Council (MAFMC) Atlantic States Marine Fisheries Commission (ASMFC) Fishery Management Plan (FMP) for summer flounder has as a management unit all summer flounder from the southern border of North Carolina, northeast to the U.S.-Canadian border. A recent summer flounder genetics study, which revealed no population subdivision at Cape Hatteras (Jones and Quattro, 1999), is consistent with the definition of the management unit. A recent consideration of summer flounder stock structure incorporating new tagging data concluded that evidence supported the existence of stocks north and south of Cape Hatteras, with the stock north of Cape Hatteras possibly composed of two distinct spawning aggregations, off New Jersey and Virginia-North Carolina (Kraus and Musick, 2003). The conclusions of Kraus and Musick (2003) are consistent with the current assessment stock unit.

The most recent benchmark stock assessment was completed in 2002 (SARC 35; NEFSC 2002) and found that the summer flounder stock was overfished and overfishing was occurring relative to the current biological reference points. The fishing mortality rate had declined from 1.32 in 1994 to 0.27 in 2001, marginally above the overfishing definition reference point (F_threshold= F_target= F_max= 0.26). Total stock biomass in 2001 was estimated to be 42,900 mt, 19% below the biomass threshold (53,200 mt). In the review of the 2002 stock assessment, SARC 35 had concluded that updating the biological reference points was not warranted at that time (NEFSC 2002). Updates to the stock assessment were completed in 2003 (Terceiro 2003), 2004 (SDWG 2004), and 2005 (SAW 41; NEFSC 2005). While the 2003 assessment found the stock not overfished and no overfishing occurring, the 2004 and 2005 assessments found the stock again experiencing overfishing. The 2005 SAW 41 assessment also provided updated values for the fishing mortality and stock biomass reference points. This 2006 assessment updates uses the same data compilation methods and analytical configurations as the 2005 SAW 41 assessment (NEFSC 2005).

Total U.S. commercial landings of summer flounder from Maine to North Carolina peaked in 1979 at nearly 18,000 mt (40 million lbs, Table 1). The reported landings in 2005 of 7,765 mt (about 17.1 million lbs) were about 6% under the final 2005 quota of 8,246 mt (18.2 million lbs). Since 1980, 70% of the commercial landings of summer flounder have come from the Exclusive Economic Zone (EEZ; greater than 3 miles from shore). The percentage of landings attributable to the EEZ was lowest in 1983 and 1990 at 63% and was highest in 1989 at 77%. Large variability in summer flounder landings exist among the states, over time, and the percent of total summer flounder landings taken from the EEZ has varied widely among the states.

Annual commercial landings data for summer flounder in years prior to 1994 were obtained from trip-level detailed landings records contained in master data files maintained by the NEFSC (the weighout system; 1963-1993) and from summary reports of the Bureau of Commercial Fisheries and its predecessor the U.S. Fish Commission (1940-1962). Beginning in 1994, landings estimates were derived from mandatory dealer reports under the current NMFS Northeast Region (NER) summer flounder quota monitoring system.

Prior to 1994, summer flounder commercial landings were allocated to NEFSC 3-digit statistical area according to interview data (Burns et al.1983). During 1994-2002, dealer landings were allocated to statistical area using fishing Vessel Trip Reports (VTR data) according to the general procedures developed by Wigley et al. (1997), in which a matched set of dealer and VTR data is used as a sample to characterize the statistical area distribution of monthly state landings. A comparison of the distribution of landings by state and month as indicated by the dealer, VTR, and matched set data for 2003-2005 is presented in Table 2, Table 3, and Table 4 (see Terceiro 2003 for years prior to 2003). Since the implementation of the annual commercial landings quota in 1993, the commercial landings have become concentrated during the first calendar quarter of the year, with 52% of the landings taken during the first quarter in 2005 (Table 4).

The distribution of Northeast Region (ME to VA) 1992-2005 landings by three-digit statistical area is presented in Table 5. Areas 537-539 (Southern New England), areas 611-616 (New York Bight), areas 621, 622, 625, and 626 (Delmarva region), and areas 631 and 632 (Norfolk Canyon area) have generally accounted for over 80% of the NER commercial landings. A summary of length and age sampling of summer flounder landings collected by the NEFSC commercial fishery port agent system in the NER is presented in Table 6. For comparability with the manner in which length frequency sampling in the recreational fishery has been evaluated, sampling intensity is expressed in terms of metric tons of landings (mt) per 100 fish lengths measured. The sampling is proportionally stratified by market category (jumbo, large, medium, small, and unclassified), with the sampling distribution generally reflecting the distribution of commercial landings by market category. Overall sampling intensity has improved markedly since 1995, from 165 mt per 100 lengths to 30-60 mt per 100 lengths, and temporal and geographic coverage has generally improved as well (e.g., for 2003-2005 in Table 7, Table 8, and Table 9).

The age composition of the NER commercial landings for 1982-1999 was generally estimated semiannually by market category and (usually) 1-digit statistical area (e.g., area 5 or area 6), using standard NEFSC procedures (market category length frequency samples converted to mean weights by length-weight relationships; mean weights in turn divided into landings to calculate numbers landed by market category; market category numbers at length apportioned to age by application of age-length keys, on semiannual area basis). For 2000-2002, sampling was generally sufficient to make quarterly estimates of the age composition in area 6 (in some cases, by division) for the large and medium market categories. For 2003-2005, sampling was generally sufficient to make quarterly estimates of the age composition in areas 5 and 6 for the jumbo, large, and medium market categories.

The distribution of 2003-2005 length frequency samples by market category, 1- and 2-digit statistical area (division), and calendar quarter is presented in Table 7, Table 8, and Table 9 (see Terceiro 2003 for years prior to 2003). NER landed numbers at age were raised to total NER (general canvas) commercial landings when necessary by assuming that landings not accounted for in the weighout/mandatory reporting system had the same age composition as that sampled, as follows: calculate proportion at age by weight; apply proportions at age by weight to total NER commercial landings to derive total NER commercial catch at age by weight; divide by mean weights at age to derive total NER commercial landed numbers at age. The proportion of large and jumbo market category fish in the NER landings has increased since 1996, while the proportion of small market category landings has become very low. The mean size of fish landed in the NER commercial fishery has been increasing since 1993, and was 0.9-1.0 kg (2.0-2.2 lbs) during 2000-2005, typical of an age 3 summer flounder (Table 10 and Table 11).

The North Carolina winter trawl fishery accounts for about 99% of summer flounder commercial landings in North Carolina. A separate landings at age matrix for this component of the commercial fishery was developed from North Carolina Division of Marine Fisheries (NCDMF) length and age frequency sampling data. The NCDMF program samples about 10% of the winter trawl fishery landings annually, most recently (2003-2005) at a mean rate of 10 mt of landings per 100 lengths measured (Table 12). All length frequency data used in construction of the North Carolina winter trawl fishery landings at age matrix were collected in the NCDMF program; age-length keys from NEFSC commercial data and NEFSC spring survey data (1982-1987) and NCDMF commercial fishery data (1988-2005) were combined by appropriate statistical area and semiannual period to resolve lengths to age. Fishery regulations in North Carolina also changed between 1987 and 1988, with increases in both the minimum mesh size of the codend and minimum landed fish size taking effect. It is not clear whether the change in regulations or the change in keys, or some combination, is responsible for the decreases in the numbers of age-0 and age-1 fish estimated in the North Carolina commercial fishery landings since 1987. Landed numbers at age and mean weights at age from this fishery are shown in Table 13 and Table 14.

Analysis of variance of fishery observer data for summer flounder was used to identify stratification variables for an expansion procedure to estimate total landings and discards from the observer data kept and discard rates (weight per day fished) in the commercial fishery. Initial models included year, quarter, fisheries statistical division (2-digit area), area (divisions north and south of Delaware Bay), and tonnage class as main effects. Quarter and division consistently emerged as significant main effects without significant interaction with the year (NEFSC 1993). The estimation procedure expands transformation bias-corrected geometric mean catch (landings and discards) rates in year, quarter, and division strata by total days fished (days fished on trips landing any summer flounder by any mobile gear, including fish trawls and scallop dredges) to derive fishery landings and discards. The days fished metric correlates better with the observed summer flounder discards on a per trip basis than other potential expansion factors such as total summer flounder landings or total trip landings of all species. The use of fishery effort as the multiplier (raising factor) also allows estimation of landings from the fishery observer data for comparison with dealer reported landings, to help judge the potential accuracy of the procedure and/or sample data.

For strata with no fishery observer sampling, catch rates from adjacent or comparable strata were substituted as appropriate (except for Division 51, which generally has very low catch rates and negligible catch). Estimates of discard were stratified by 2 gear types (scallop dredges; trawls) for years when data were adequate (1992 and later years). Estimates at length and age were stratified by gear for 1994-2000 and 2002-2005, again due to sample size considerations. Only 11 fish were sampled from the sea scallop dredge fishery 2001, and so the scallop dredge discards were assumed to have the same length and age composition as the trawl fishery discards in 2001.

While estimates of catch rates from the NER fishery observer data were used in this assessment to estimate total discards, catch rate information is also reported in the VTR data. A comparison of discard to total catch ratios for the fishery observer and VTR data sets for trawl and scallop dredge gear indicates similar discard rates from the two data sources. Overall fishery observer and VTR discard to total catch ratios for 1994-2005 were generally within 10-15% of each other; 2001 was an exception, with an overall discard to total catch ratio of 49% in the fishery observer data and 29% in the VTR data. Discard rates of summer flounder in the scallop dredge fishery were much higher than in the trawl fishery (Table 15 and Table 16).

The change in mid-1994 from the interview/weighout data reporting system to the VTR/mandatory dealer report system required a change in the estimation of effort (days fished) to estimate total discards. An initial examination of days fished and catch per unit effort (CPUE; landings per day fished) for cod conducted at SAW 24 (NEFSC 1997a) compared these quantities as reported in the full weighout and VTR data sets (DeLong et al., 1997). This comparison indicated a shift to a higher frequency of short trips (trips with one or two days fished reported), and to a mode at a lower rate of CPUE. It was not clear at SAW 24 if these changes were due to the change in reporting system (units reported not comparable), or real changes in the fishery, and so effort data reported by the VTR system were not used quantitatively in the SAW 24 assessments. In the SAW 25 assessment for summer flounder (NEFSC 1997b), a slightly different comparison was made. The port agent interview data for 1991-93 and merged dealer/VTR data for 1994-1996 (the matched set data), which under each system serve as the“sample”to characterize the total commercial landings, were compared in relative terms (percent frequency). For summer flounder, the percent frequency of short trips (lower number of days fished per trip) increased during 1991-1996, but not to the degree observed for cod, and the mode of CPUE rates for summer flounder increased in spite of lower effort per trip. For the summer flounder fishery, these may reflect actual changes in the fishery, due to increased restrictions on allowable landings per trip (trip landings limits might lead to shorter trips) and stock size increases (higher CPUE). As for cod, however, the influence of each of these changes (reporting system, management changes, stock size changes) has not been quantified. Total days fished in the summer flounder fishery were comparable between 1989-1993 period and 1994. Since 1994, total days fished have ranged from 20,700 days in 1999 to 9,300 days in 2004, with a mean of about 12,000 days, a substantial decline relative to the 1989-1993 mean of 22,000 days. Because the effort measure is critical to the estimation of discards for summer flounder, the VTR data were used as the best data source to estimate summer flounder fishery days fished for 1994-2005.

Two adjustments were made to the dealer/VTR matched data subset days fished estimates to fully account for summer flounder fishery effort during 1994-2005. First, the landings to days fished relationship in the matched set was assumed to be the same for unmatched trips, and so the days fished total in each discard estimation stratum (2-digit area and quarter) was raised by the dealer to matched set landings ratio. This step in the estimation accounted for days fished associated with trips landing summer flounder, and provided an estimate of discard for trips landing summer flounder.

Given the restrictions on the fishery however, there is fishing activity which results in summer flounder discard, but no landings, especially in the scallop dredge fishery. The days fished associated with these trips was accounted for by raising strata discard estimates by the ratio of the total days fished on trips catching any summer flounder (trips with landings and discard, plus trips with discard only) to the days fished on trips landing summer flounder (trips with landings and discard), for VTR trips reporting discard of any species (DeLong et al.1997). For this step , it is necessary to assume that the discard rate (as indicated by the fishery observer data, which includes trips with discard but no landings, and which is used in previous estimation procedure steps) is the same for trips with only discards as for trips with both landings and discards.

Discard estimates for 1989-2005 are summarized in Table 17. Discards as a proportion of the fishery observer data estimated landings were highest in 2001 (53%), and lowest in 1995 and 1996 (5 and 7%). Estimates of landings from observer data ranged from +53% (1999) to -70% (2001) of the reported landings in the fisheries (Table 18), with discards ranging from 41% (1990) to 6% (1995) of the reported landings. Total discards estimated for 2003, 2004, and 2005 were 10%, 4%, and 4% of the reported landings. Scallop dredge fishery discard to landed ratios are much higher than trawl fishery ratios, purportedly because of closures and trip limits. Although the scallop dredge landings of summer flounder are less than 5% of the total, the discards of summer flounder are of the same order of magnitude as in the trawl fishery.

These discard estimates were based only on the days fished data for ports in the NER during 1989-1996, and so it was necessary to raise the discard estimate to account for discarding occurring outside the NER reporting system (i.e., NER state reporting systems such as Connecticut and Virginia, and North Carolina). To determine the proper raising factor, landings accounted for by the NER reporting system (which result from the fishing effort on which the fishery observer discard estimate is based) were compared with total NER landings, plus that portion of North Carolina landings from the EEZ (it is assumed that only the North Carolina fishery in the EEZ would experience significant discard, as mesh regulations in state waters have resulted in very low discards in state waters since implementation of the regulation in 1989; R. Monaghan, NCDMF; personal communication, June 30, 1997). As a result of this exercise, the total discard estimates were raised by 11 to 38% for the 1989-1996 period. Since 1996, all states’landings and are included in the NER dealer reporting system, so no raising is necessary to account for missing landings. As recommended by SAW 16 (NEFSC 1993), a commercial fishery discard mortality rate of 80% was assumed to develop the final estimate of discard mortality.

Existing fishery observer data were used to develop estimates of commercial fishery discard for 1989-2005. However, adequate data (e.g., interviewed trip data, survey data) are not available to develop summer flounder discard estimates for 1982-1988. Discard numbers were assumed to be very small relative to landings during 1982-1988 (because of the lack of a minimum size limit in the EEZ), but to have increased since 1989 with the implementation of fishery regulations under the FMP. It is recognized that not accounting directly for commercial fishery discards in 1982-1988 likely results in an underestimation of fishing mortality and population sizes in these years.

NEFSC fishery observer length frequency samples were converted to sample numbers at age and sample weight at age frequencies by application of NEFSC survey length-weight relationships and fishery observer, commercial fishery, and survey age-length keys. Sample weight proportions at age were next applied to the raised fishery discard estimates to derive fishery total discard weight at age. Fishery discard weights at age were then divided by fishery observer mean weights at age to derive fishery discard numbers at age. Classification to age for 1989-1993 was done by semiannual (quarters 1 and 2 pooled, quarters 3 and 4 pooled) periods using NEFSC fishery observer age-length keys, except for 1989, when first period lengths were aged using combined commercial landings (quarters 1 and 2) and NEFSC spring survey age-length keys. For 1994-2005, only NEFSC winter, spring, and fall survey age-length keys were used, since fishery observer age-length keys were not yet available and commercial landings age-length keys contained an insufficient number of small summer flounder (<40 cm = 16 inches) that comprise most of the discards. Fishery observer sampling intensity is summarized in Table 17. Estimates of discarded numbers at age, mean length and mean weight at age are summarized in Table 19, Table 20, and Table 21.

The reason for discarding in the trawl and scallop dredge fisheries has been changing over time. During 1989 to 1995, the minimum size regulation was recorded as the reason for discarding summer flounder in over 90% of the observed trawl and scallop dredge tows. In 1999, the minimum size regulation was provided as the reason for discarding in 61% of the observed trawl tows, with quota or trip limits given as the discard reason in 26% of the observed tows, and high-grading in 11% of the observed tows. In the scallop fishery in 1999, quota or trip limits was given as the discard reason in over 90% of the observed tows. During 2000-2005, minimum size regulations were identified as the discard reason in 40-45% of the observed trawl tows, quota or trip limits in 25-30% of the tows, and high grading in 3-8%. In the scallop fishery during 2000-2005, quota or trip limits was given as the discard reason for over 99% of the observed tows. As a result of the increasing impact of trip limits, fishery closures, and high grading as reasons for discarding, the age structure of the summer flounder discards has also changed, with a higher proportion of older fish being discarded (Table 19).

Landings statistics for the summer flounder recreational fishery (catch type A+B1) are estimated by the National Marine Fisheries Service (NMFS) Marine Recreational Fishery Statistics Survey (MRFSS; Table 22 and Table 23). Recreational fishery landings decreased 13% by number and 6% by weight from 2004 to 2005, as the fishery landed 84% (4,550 mt, 10.0 million lbs) of the 5,434 mt (11.98 million lbs) harvest limit established for 2005. The Proportional Standard Error (PSE) of MRFSS landings estimates by number and weight averaged 6% over the 1982-2005 period, ranging from 26% in 1982 to 3% in 1996 and 2000 (Table 22 and Table 23).

The commercial fishery VTR system provides an alternative set of reported recreational landings by the party/charter boat sector. A comparison of VTR reports and MRFSS estimates indicates that MRFSS estimates are higher by an average factor of 2.30 for the 1995-2005 period, ranging from a factor of 1.02 in 1998 to 4.35 in 2005 (Table 24). It is not clear if this is due mainly to under-reporting of party/charter boat recreational landings in the VTR system, or a systematic positive bias of MRFSS landings estimates for the party/charter boat sector.

Length frequency sampling intensity for the recreational fishery for summer flounder was calculated by MRFSS subregions (North - Maine to Connecticut; Mid - New York to Virginia; South - North Carolina) on a metric tons of landings per hundred lengths measured basis (Burns et al. In Doubleday and Rivard, 1983). For 2005, aggregate sampling intensity averaged 162 mt of landings per 100 fish measured (Table 25).

MRFSS sample length frequency data, NEFSC commercial age-length data, and NEFSC survey age-length data were examined in terms of number of fish measured/aged on various temporal and geographical bases. Correspondences were made between MRFSS intercept date (quarter), commercial quarter, and survey season (spring and summer/fall), and between MRFSS subregion, commercial statistical areas, and survey depth strata to integrate data from the different sources. Based on the number, size range, and distribution of lengths and ages, a semiannual (quarters 1 and 2; quarters 3 and 4), subregional basis of aggregation was adopted for matching of commercial and survey age-length keys with recreational length frequency distributions to convert lengths to ages.

The recreational landings historically were dominated by relatively young fish. Over the 1982-1996 period, age 1 fish accounted for over 50% of the landings by number; summer flounder of ages 0 to 4 accounted for over 99% of landings by number. No fish from the recreational landings were determined to be older than age 7. With increases in the minimum size since 1996 (to 14.5 in [37 cm] in 1997, 15 in [38 cm] in 1998-1999, generally 15.5 in [39 cm] in 2000, and various state minimum sizes from 15.5 [38 cm] to 17.5 in [44 cm] in 2001-2005) and a trend to lower fishing mortality rates, the age composition of the recreational landings now includes mainly fish at ages 2 and 3. The number of summer flounder of ages 4 and older landed by the recreational fishery in 2005 (28% of the landings by number) was the highest in the time series (Table 26). Limited MRFSS length sampling for larger fish resulted in a high degree of variability in mean length for older fish, especially at ages 5 and older during the first decade of the time series. Attempts to estimate length-weight relationships from the MRFSS biological sampling data provided unsatisfactory results. As a result, quarterly length (mm) to weight (g) relationships from Lux and Porter (1966) were used to calculate annual mean weights at age from the estimated age-length frequency distribution of the landings.

The MRFSS estimates of live discard (catch type B2) to total catch (catch types A+B1+B2) in the recreational fishery for summer flounder has varied from about 18% (1985) to about 85% (2005; Table 27). The Proportional Standard Error (PSE) of MRFSS live discards estimates by number averaged 8% over the 1982-2005 period, ranging from 59% in 1982 to 3% in 2001 and 2002 (Table 27).

To account for all removals from the summer flounder stock by the recreational fishery, some assumptions about the biological characteristics and hooking mortality rate of the recreational live discard needed to be made, because biological samples are not routinely taken of MRFSS catch type B2 fish. In previous assessments, data available from New York Department of Environmental Conservation (NYDEC) surveys (1988-92) of New York party boats suggested the following: 1) nearly all (>95%) of the fish released alive from boats were below the minimum regulated size (during 1988-92, 14 in [36 cm] in New York state waters); 2) nearly all of these fish were age 0 and age 1 summer flounder; and 3) age 0 and 1 summer flounder occurred in approximately the same proportions in the live discard as in the landings. It was therefore assumed that all B2 catch would be of lengths below regulated size limits, and be either age 0 or age 1 in all three subregions during 1982-1996. Catch type B2 was allocated on a semi-annual, subregional basis in the same ratio as the annual age 0 to age 1 proportion observed in the landings during 1982-1996. Mean weights at age were assumed to be the same as in the landings during 1982-1996.

The minimum landed size in federal and most state waters increased to 14.5 in (37 cm) in 1997, to 15.0 in (38 cm) in 1998-1999, and to 15.5 in (39 cm) in 2000. Applying the same logic used to allocate the 1982-1996 recreational released catch to size and age categories during 1997-2000 implied that the recreational fishery released catch included fish of ages 2 and 3. Investigation of data from the CTDEP Volunteer Angler Survey (VAS) for 1997-1999 and from the American Littoral Society (ALS) for 1999, and comparing the length frequency of released fish in these programs with the MRFSS data on the length frequency of landed fish below the minimum size, indicated this assumption was valid for 1997-1999 (MAFMC 2001). The CTDEP VAS and ALS data, along with data from the NYDEC Party Boat Survey (PBS) was used to validate this assumption for 2000. For 1997-2000 all B2 catch was assumed to be of lengths below regulated size limits, and therefore comprised of ages 0 to 3. Catch type B2 was allocated on a sub-regional basis in the same ratio as the annual age 0 to age 3 proportions observed in the landings at lengths less than 37 cm in 1997, 38 cm in 1998-1999, and 39 cm in 2000.

In 2001, many states adopted different combinations of minimum size and possession limits to meet management requirements. As a result, minimum sizes for summer flounder ranged from 15.5 in (39 cm) in Federal, VA, and NC waters, 16 in (41 cm) in NJ, 16.5 in (42 cm) in MA, 17 in (43 cm) in MD and NY, to 17.5 in (44 cm) in CT, RI, and DE. Examination of data provided by MD sport fishing clubs, the CTDEP VAS, the ALS, and the NYDEC PBS indicated that the assumption that fish released are those smaller than the minimum size remained valid for 2001, and so catch type B2 was characterized by the same proportion at length as the landed catch less than the minimum size in the respective states. The differential minimum sizes by state has continued since 2001. For 2002-2005, increased samples of the recreational fishery discards by the CT VAS, NYDEC PBS, and, for 2005, the MRFSS For Hire Survey (FHS), has allowed direct characterization the length frequencies of the discards from sample data (Table 28).

Studies conducted to estimate hooking mortality for striped bass and black sea bass indicated a hooking mortality rate of 8% for striped bass (Diodati and Richards 1996) and 5% for black sea bass (Bugley and Shepherd, 1991). Work by the states of Washington and Oregon with Pacific halibut (a potentially much larger flatfish species, but otherwise morphologically similar to summer flounder) found "average hooking mortality...between eight and 24 percent" (IPHC, 1988). An unpublished tagging study by the NYDEC (Weber MS 1984) on survival of released sublegal summer flounder caught by hook-and-line suggested a total, non-fishing mortality rate of 53%, which included hooking plus tagging mortality as well as deaths by natural causes (i.e., predation, disease, senescence). Assuming deaths by natural causes to be about 18%, (an instantaneous rate of 0.20), an annual hooking plus tagging mortality rate of about 35% can be derived from the NYDEC results. In the SARC 25 (NEFSC 1997b) and earlier assessments of summer flounder, a 25% hooking mortality rate was assumed for summer flounder released alive by anglers.

However, three subsequent investigations of summer flounder recreational fishery release mortality suggested that a lower release mortality rate was more appropriate. Lucy and Holton (1998) used field trials and tank experiments to investigate the release mortality rate for summer flounder in Virginia, and found rates ranging from 6% (field trials) to 11% (tank experiments). Malchoff and Lucy (1998) used field cages to hold fish angled in New York and Virginia during 1997 and 1998, and found a mean short term mortality rate of 14% across all trials. Gearhart (1999) used field cages to hold summer flounder angled in North Carolina for 72 hours after a capture, and found a short-term mortality rate of 7%. Given the results of these release mortality studies conducted specifically for summer flounder, a 10% release mortality rate was adopted in the Terceiro (1999) stock assessment and has been retained in all subsequent assessments. Ten percent of the total B2 catch at age is therefore the basis of estimates of summer flounder recreational fishery discard at age (Table 29). In 2005, the number of fish discarded and assumed dead in the recreational fishery (2.2 million fish, 899 mt) was 57% by number and 20% by weight of the total landed (3.8 million fish, 4,550 mt) in the recreational fishery. The total catch and mean weight at age for the recreational fishery are provided in Table 30 and Table 31.

NER commercial fishery landings and discards at age, North Carolina winter trawl fishery landings and discards at age, and MRFSS recreational fishery landings and discards at age totals were summed to provide a total fishery catch at age matrix for 1982-2005 (Table 32). The percentage of age-3 and older fish in the total catch in numbers has increased during the last decade from only 4% in 1993 to over 40% in 2002 and later years. Overall mean lengths and weights at age in the total catch were calculated as weighted means (by number in the catch at age) of the respective mean values at age from the NER commercial (Maine to Virginia), North Carolina commercial, and recreational (Maine to North Carolina) fisheries (Table 33 and Table 34). The respective components of the total summer flounder catch in weight are summarized in Table 35.

Long-term trends in summer flounder abundance were derived from a stratified random bottom trawl survey conducted in spring by NEFSC between Cape Hatteras and Nova Scotia since 1968 (Clark 1979). NEFSC spring survey indices suggest that total stock biomass last peaked in 2003 (2.42 kg/tow; a new historical high), about 20% above the previous peak value in 1976 of 2.00 kg/tow (Table 36, Figure 1). Since 2003, the spring biomass index has declined by about 45%. Age composition data from the NEFSC spring surveys indicate a substantial reduction in the number of ages in the stock between 1976-1990 (Table 37). Between 1976-1981, fish of ages 5-8 were captured regularly in the survey, with the oldest individuals aged 8-10 years. Between 1982-1986, fish aged 5 and older were only occasionally observed in the survey, and by 1986, the oldest fish observed in the survey were age 5. In 1990 and 1991, only three age groups were observed in the survey catch, and there was an indication that the 1988 year class was very weak. Since 1991, the survey age composition has expanded significantly (Figure 2). Mean lengths at age in the NEFSC spring survey are presented in Table 38.

Summer flounder are frequently caught in the NEFSC autumn survey at stations in inshore strata (<27 meters = 15 fathoms = 90 feet) and at offshore stations in the 27-55 meter depth zone (15-30 fathoms, 90-180 feet) at about the same level as in the spring survey (Table 36, Figure 1). Furthermore, the autumn survey catches age-0 summer flounder in abundance, providing an index of summer flounder recruitment (Table 39, Figure 3). Autumn survey indices suggest improved recruitment since the late 1980s, and an increase in abundance of age-2 and older fish since 1995. Mean lengths at age in the NEFSC autumn survey are presented in Table 40.

A new series of NEFSC winter trawl surveys was initiated in February 1992 to provide improved abundance indices for flatfish, including summer flounder. The surveys target flatfish when they are concentrated offshore during the winter. A modified 36 Yankee trawl is used that differs from the standard trawl employed during the spring and autumn surveys in that long trawl sweeps (wires) are added before the trawl doors to better herd fish to the mouth of the net, and the large rollers used on the standard gear are absent with only a chain"tickler"and small spacing"cookies"present on the footrope.

The design and conduct of the winter survey (timing, strata sampled, and the use of the modified 36 Yankee trawl gear) has resulted in greater catchability of summer flounder compared to the other surveys. Most fish area captured in survey strata 61-76 (27-110 meters; 15-60 fathoms) off the Delmarva and North Carolina coasts . Other concentrations of fish are found in strata 1-12, south of the New York and Rhode Island coasts, in slightly deeper waters. Significant numbers of large summer flounder are often taken along the southern flank of Georges Bank (strata 13-18).

Indices of summer flounder abundance from the winter survey indicate stable stock size during 1992-1995, with catch per tow values ranging from 10.9 in 1995 to 13.6 in 1993 (Table 36). For 1996, the winter survey index increased by 290% over 1995, from 10.9 to 31.2 fish per tow. The largest increases in 1996 occurred in the Mid-Atlantic Bight region (offshore strata 61-76), where increases up to an order of magnitude occurred in several strata, with the largest increases in strata 61, 62, and 63 off the northern coast of North Carolina. Most of the increased catch in 1996 consisted of age-1 summer flounder from the 1995 year class. In 1997, the index dropped to 10.3 fish per tow, due to the lower numbers of age-1 (1996 year class) fish caught. Since 1998, the Winter trawl survey indices have increased, with the Winter 2003 survey number and weight per tow indices the highest in the time series. Since 2003, the Winter survey abundance and biomass indices have declined by about 40%. As with the other two NEFSC surveys, there is strong evidence in recent winter surveys of increased abundance of age-3 and older fish relative to earlier years in the time series (Table 41, Figure 1). Mean lengths at age in the NEFSC winter survey are presented in Table 42.

Spring and fall bottom trawl surveys conducted by the Massachusetts Division of Marine Fisheries (MADMF) show a decline in abundance in numbers of summer flounder from high levels in 1986 to record lows in 1990 (MADMF fall survey), and 1991 (MADMF spring survey). In 1994, the MADMF survey indices increased to values last observed during 1982-1986, but then declined substantially in 1995, although the indices remain higher than the levels observed in the late 1980s. Since 1996, both the MADMF spring and fall indices have increased to record high levels (Table 43 and Table 44; Figure 4). The spring index peaked in 2000, and has since declined by about 40%. The fall survey index, in contrast, has continued to increase and reached an historical peak in 2005. The MADMF also captures a small number of age-0 summer flounder in a seine survey of estuaries, and these data constitute an index of recruitment (Table 45, Figure 5).

Spring and fall bottom trawl surveys are conducted by the Connecticut Department of Environmental Protection (CTDEP). The CTDEP surveys show a decline in abundance in numbers of summer flounder from high levels in 1986 to record lows in 1989. The CTDEP surveys indicate recovery since 1989, and evidence of increased abundance at ages 2 and older since 1995. The 2002 spring and autumn indices were the highest in the respective time series, and both series have since declined by over 60% (Table 46 and Table 47, Figure 6). An index of recruitment from the autumn series is available (Table 47, Figure 3).

Standardized bottom trawl surveys have been conducted since 1979 during the spring and autumn months in Narragansett Bay and state waters of Rhode Island Sound by the Rhode Island Department of Fish and Wildlife (RIDFW). Indices of abundance at age for summer flounder have been developed from the autumn survey data using NEFSC autumn survey age-length keys. Survey indices show that the 1984-1987, 1999, 2000, and 2002 year classes are all strong. The autumn survey reached a time series high in 2003, and has since declined by 40% (Table 48, Figure 4). An abundance index has also been developed from a set of fixed stations sampled monthly during 1990-2005. Indices of abundance at age from this series indicate that strong year classes recruited to the stock in 1996, 1999, 2000, and 2002, with age 2+ abundance peaking in 2003, and declining by about 35% since then (Table 49). Recruitment indices are available from both the autumn (Figure 5) and monthly fixed station surveys.

The New Jersey Bureau of Marine Fisheries (NJBMF) has conducted a standardized bottom trawl survey since 1988. Indices of abundance for summer flounder incorporate data collected from April through October. The NJBMF survey mean number per tow indices and frequency distributions were converted to age using the corresponding annual NEFSC combined spring and fall survey age-length keys. Indices of the 1995 year class at age-0 and at older ages in subsequent years indicate that this cohort is the strongest in the time series. Indices of the 1996-2001 year classes are below average, while the 2002 year class is average. The NJBMF survey indices reached a peak in 2002, and have declined by about 35% since then (Table 50, Figure 7). Age 0 recruitment indices are available from the NJBMF survey (Figure 3).

The Delaware Division of Fish and Wildlife (DEDFW) has conducted a standardized bottom trawl survey with a 16 foot headrope trawl since 1980, and with a 30 foot headrope trawl since 1991. Recruitment indices (age 0 fish; one index from the Delaware estuary proper for 1980 and later, one from the inland bays for 1986 and later) have been developed from the 16 foot trawl survey data. Indices for age-0 to age-4 and older summer flounder have been compiled from the 30 foot headrope survey. The indices use data collected from June through October (arithmetic mean number per tow), with age 0 summer flounder separated from older fish by visual inspection of the length frequency. The 16 foot headrope survey indices suggest poor recruitment in 1988 and 1993, improved recruitment in 1994-1995, above average recruitment in 2000 and 2004, and poor recruitment in 2005 (Table 51, Figure 5). The 30 foot headrope survey indices suggest stable stock sizes over the 1991-2001 time series, with strong recruitment in 1991, 1994, 1995, and 2000. The index from the 30 foot survey was near the time series peak in 2001, and has declined over 80% since then (Table 52, Figure 7).

The Maryland Department of Natural Resources (MDDNR) has conducted a standardized trawl survey in the seaside bays and estuaries around Ocean City, MD since 1972. Samples collected during May to October with a 16 foot bottom trawl have been used to develop a recruitment index for summer flounder for the period 1972-2002. This index suggests that weakest year class in the time series recruited to the stock in 1988, and the strongest in 1972, 1983, 1986, and 1994, and 2001. The 2005 index was the lowest since 1988 (Table 53, Figure 8).

The Virginia Institute of Marine Science (VIMS) conducts a juvenile fish survey using trawl gear in Virginia rivers and the mainstem of Chesapeake Bay. The time series for the rivers began in 1979. With the Bay included, the series is available only since 1988, but many more stations are included. Trends in the two time series are very similar. An index of recruitment developed from the rivers only series suggests weak year classes recruited to the stock in 1987 and 1999, with strong year classes recruiting during 1980-1984, and 1990. Recruitment indices since 1990 have been below average, and the 2005 index was the lowest since 1975 (Table 54, Figure 8).

The NCDMF has conducted a stratified random trawl survey using two 30 foot headrope nets with 3/4" mesh codend in Pamlico Sound since 1987. An index of recruitment developed from these data suggests weak year classes recruited to the stock in 1988 and 2000, with strong year classes in 1987, 1992, and 1996, 2001, and 2002 (Table 55, Figure 8). The survey normally takes place in mid-June, but in 1999 was delayed until mid-July. The 1999 index is therefore inconsistent with the other indices in the time series, and the 1999 value was excluded from the VPA calibration in the SARC 31 (NEFSC 2000) and subsequent assessments.

ESTIMATES OF MORTALITY AND STOCK SIZE

Virtual population analysis

Fishing mortality rates in 2005 and stock sizes in 2006 were estimated using the ADAPT method for calibration of the VPA (Parrack 1986, Gavaris 1988, Conser and Powers 1990) as implemented in the NOAA Fisheries Toolbox (NFT) version 2.51 VPA. As recommended by the MAFMC S&S Committee during the review of the Terceiro (1999) assessment and by the National Research Council review of the summer flounder assessment (NRC 2000), ages 0-6 were included in the analysis as true ages, with ages 7 and older combined as a plus group. An instantaneous natural mortality rate of M = 0.2 was assumed for all ages in all years. Alternative estimates of M were considered in the SAW 20 assessment (NEFSC 1996). In the SAW 20 work, estimates were derived with the methods described by: 1) Pauly (1980) using growth parameters derived from NCDMF age-length data and a mean annual bottom temperature (17.5°C) from NC coastal waters; 2) Hoenig (1983) using a maximum age for summer flounder of 15 years; and 3) consideration of age structure expected in unexploited populations (5% rule, 3/M rule, e.g., Anthony 1982). SAW 20 (NEFSC 1996) concluded that M = 0.2 was a reasonable value given the mean (0.23) and range (0.15-0.28) obtained from the various analyses, and this value for M has been used in all subsequent assessments. Maturities at age for all years were 38% for age-0, 72% for age-1, 90% for age-2, and 100% for ages 3 and older. Stock sizes in 2006 were directly estimated for ages 1-6, while the age 7+ group was calculated from Fs estimated in 2005. Fishing mortality on the oldest true age (6) in the years prior to the terminal year was estimated from back-calculated stock sizes for ages 3-6. Fishing mortality on the age 7+ group was assumed equal to the fishing mortality for age 6. Winter, spring, and mid-year (e.g., RIDFW monthly fixed station, DEDFW, and NJBMF) survey indices and all survey recruitment (age-0) indices were compared to population numbers of the same age at the beginning of the same year. Fall survey indices were compared to population numbers one year older at the beginning of the next year. Tuning indices were unweighted.

The final VPA run (run F06_1) included the same set of indices (n=41) in terms of source and age range as used in the 2002-2005 assessments (NEFSC 2002, Terceiro 2003, SDWG 2004, NEFSC 2005). In addition to a run including all available indices (F06_ALL) and the run chosen as final (F06_1), the results from two other runs were also considered. The NEFSC survey indices generally had the lowest partial variances within the VPA, but sometimes indicated patterns in stock size dissimilar to those in the state surveys. Therefore runs were also examined that contrasted the VPA solutions provided by NEFSC (F06_NEFSC) versus state survey (F06_STATE) series. The sensitivity of the summer flounder VPA estimates of F and SSB is illustrated in Figure 9. The output for the final 2006 assessment VPA (run F06_1) is presented in Table 56.

The annual partial recruitment of age-1 fish decreased from near 0.50 during the first half of the VPA time series to less than 0.30 since 1994, and to about 0.20 during 2000-2002 and less than 0.20 during 2003-2005. The partial recruitment of age-2 fish has decreased from 1.00 in 1993 to about 0.80 during 2000-2002 and to about 0.70 during 2003-2005. These decreases in partial recruitment at age are in line with expectations given recent changes in commercial and recreational fishery regulations. For these reasons, summer flounder are currently considered to be fully recruited to the fisheries at age 3, and fully recruited fishing mortality is expressed as the unweighted average of fishing mortality at age for ages 3 to 5.

Fishing mortality calculated from the average of the currently fully recruited ages (3-5) was very high, varying between 0.9 and 2.2 during 1982-1997 (55%-83% exploitation), far in excess of the revised FMP Amendment 12 (MAFMC 1999) overfishing definition, F_threshold= F_target= F_max= 0.276. The fishing mortality rate has declined since 1997 and was estimated to be 0.46 during 2003-2004, rising to 0.53 (37% exploitation) in 2005 (Figure 10). The estimate of F for 2005 may understate the actual fishing mortality; retrospective analysis shows that the current assessment method tends to underestimate recent fishing mortality rates, continuing the pattern observed in the last six assessments (NEFSC 2000, MAFMC 2001, NEFSC 2002, Terceiro 2003, SDWG 2004, NEFSC 2005). Over the last 5 years, the annual retrospective increase in fishing mortality has averaged 33% (Figure 11).

Total stock biomass increased substantially during the 1990s and through 2004, but has decreased slightly since 2004 and was estimated to be 47,800 mt on January 1, 2006, just above the biomass threshold (46,323 mt; Figure 12 and 14). Spawning stock biomass (SSB; Age 0+) declined 72% from 1983 to 1989 (18,800 mt to 5,200 mt), but with improved recruitment and decreased fishing mortality had increased to 32,600 mt by 2004, before decreasing to 30,600 mt in 2005 (Figure 12 and Figure 13). Retrospective analysis shows a tendency to overestimate the SSB in the most recent years, continuing the pattern observed in the last six assessments (NEFSC 2000, MAFMC 2001, NEFSC 2002, Terceiro 2003, SDWG 2004, NEFSC 2005). Over the last 5 years, the annual retrospective decrease in SSB has averaged 17% (Figure 11). The age structure of the spawning stock has expanded, with 74% at ages 2 and older, and 23% at ages 5 and older. Under equilibrium conditions at F_max, about 85% of the spawning stock biomass would be expected to be ages 2 and older, with 50% at ages 5 and older (Figure 15).

Summer flounder spawn in the late autumn and early winter (peak spawning on November 1), and age 0 fish recruit to the fishery during the autumn after they are spawned. For example, summer flounder spawned in autumn 1987 (from the November 1, 1987 spawning stock biomass) recruit to the fishery in autumn 1988, and appear in VPA tables as age 0 fish in 1988. This assessment indicates that the 1982 and 1983 year classes were the largest of the VPA series, at 74 and 80 million fish, respectively. The 1988 year class was the smallest of the series, at only 13 million fish. The arithmetic average recruitment from 1982 to 2005 is 35 million fish at age 0, with a median of 33 million fish. The 2005 year class is estimated to be the smallest since 1988, at about 15 million fish (Table 56, Figure 12, Figure 13, and Figure 14). Retrospective analysis shows that the current assessment method tends to overestimate the abundance of age 0 fish in the most recent years. Over the last 5 years, the annual retrospective decrease in recruitment has averaged 10% (Figure 11).

The precision and bias of the 2005 fishing mortality rates, 1 January 2006 stock sizes, and stock biomass estimates are presented in Table 57. Bias was less than 10% for all parameters estimated. The bootstrap estimate of the 2006 total stock biomass was relatively precise, with a corrected CV of 10%. The bootstrap mean (48,616 mt) was slightly higher than the VPA point estimate (47,789 mt). There is an 80% chance that total stock biomass in 2006 was between 41,600 and 56,900 mt. The bootstrap mean F for 2005 (0. 5611) was slightly higher (6%) than the point estimate from the VPA (0.5279). There is a 80% chance that F in 2005 was between about 0.42 and 0.75, given variability in survey observations.

BIOLOGICAL REFERENCE POINTS

The calculation of biological reference points based on yield per recruit for summer flounder using the Thompson and Bell (1934) model was detailed in the 1990 SAW 11 assessment (NEFC 1990). The 1990 analysis estimated F_max = 0.23. In the 1997 SAW 25 assessment (NEFSC 1997b), an updated yield per recruit analysis reflecting the partial recruitment pattern and mean weights at age for 1995-1996 estimated that F_max = 0.24. The analysis in the Terceiro (1999) assessment, reflecting partial recruitment and mean weights at age for 1997-1998, estimated that Fmax = 0.263.

The Overfishing Definition Review Panel (Applegate et al. 1998) recommended that the MAFMC base MSY proxy reference points on yield per recruit analysis, and this recommendation was adopted in formulating the FMP Amendment 12 reference points (see Introduction), based on the 1999 assessment (Terceiro 1999). The 1999 assessment yield per recruit analysis indicated that F_threshold = F_target= F_max = 0.263, yield per recruit (YPR) at F_max was 0.55219 kg/recruit, and January 1 biomass per recruit (BPR) at Fmax was 2.8127 kg/recruit. The median number of summer flounder recruits estimated from the 1999 VPA for the 1982-1998 period was 37.844 million fish. Based on this recruitment, maximum sustainable yield (MSY) was estimated to be 20,897 mt (46 million lbs) at a biomass (B_MSY) of 106,444 mt (235 million lbs). The biomass threshold, one-half B_MSY, was therefore estimated to be 53,222 mt (118 million lbs). The Terceiro (1999) reference points were retained in the 2000 and 2001 stock assessments (NEFSC 2000, MAFMC 2001) because of the stability of the input data. In the review of the 2002 stock assessment, SARC 35 concluded that updating the reference points was not warranted (NEFSC 2002).

The biological reference points for summer flounder were reviewed and updated for the 2005 SAW 41 assessment (NEFSC 2005). The yield per recruit analysis conducted for the 2005 assessment indicated that F_max = 0.276, which was used as a proxy for F_target and F_threshold (Figure 14). Updated FMP Amendment 12 (MAFMC 1999) stock biomass reference points were estimated as the product of yield per recruit (0.576 kg per recruit) and total stock biomass per recruit (2.798 kg per recruit) at F_max = 0.276, and median recruitment of 33.111 million fish per year (1982-2004; from NEFSC (2005)). Yield at F_max, used as a proxy for MSY, was estimated to be 19,072 mt (42.0 million lbs), and the corresponding stock biomass, used as a proxy for B_MSY, was estimated to be 92,645 mt (204.2 million lbs; Figure 14). The biomass threshold of 0.5* B_MSY = 46,323 mt (102.1 million lbs).

PROJECTIONS

Stochastic projections were made to provide forecasts of stock size and catches in 2006-2007 consistent with target reference points established in the FMP. The projections assume that recent patterns of discarding will continue over the time span of the projections. Different patterns that could develop in the future due to additional trip and bag limits and fishery closures have not been evaluated. The partial recruitment pattern (including discards) used in the projections was estimated as the geometric mean of F at age for 2003-2005, reflecting recent conditions in the fisheries. Mean weights at age were estimated as the geometric means of 2003-2005 values. Separate mean weight at age vectors were developed for the January 1 biomass, landings, and discards.

One hundred projections were made for each of the 1000 bootstrapped realizations of 2006 stock sizes from the final VPA, using algorithms and software described by Brodziak and Rago (MS 1994) as implemented in the NFT AGEPRO version 3.20. Future recruitment at age 0 was generated randomly from a cumulative density function of the VPA recruitment series for 1982-2005 (median recruitment = 32.7 million fish).

If landings in 2006 are 10,700 mt (23.6 million lbs) and discards are 800 mt (1.8 million lbs), the forecast estimates a median (50% probability) F in 2006 = 0.35 and a median total stock biomass on January 1, 2007 of 51,200 mt, above the biomass threshold of ½B_MSY= 46,323 mt. A subsequent reduction in fishing mortality in 2007 to F = 0.276 is forecast to yield landings of 9,026 mt (19.899 million lbs) and a median total stock biomass level on January 1, 2008 of 58,100 mt (Table 58).

CONCLUSIONS

Assessment results

The summer flounder stock is not overfished but overfishing is occurring relative to the biological reference points. The fishing mortality rate declined from 1.32 in 1994 to 0.46 during 2003-2004, before increasing to 0.53 in 2005, well above the overfishing definition reference point (F_threshold= F_target= F_max= 0.276). There is an 80% chance that the 2005 F was between 0.42 and 0.75. The estimate of F for 2005 may understate the actual fishing mortality; retrospective analysis shows that the current assessment method tends to underestimate recent fishing mortality rates. Over the last 5 years, the annual retrospective increase in fishing mortality has averaged 33%.

Total stock biomass increased substantially during the 1990s and through 2004, but has decreased slightly since 2004 and was estimated to be 47,800 mt on January 1, 2006, just above the biomass threshold (46,323 mt). There is an 80% chance that total stock biomass in 2006 was between 41,600 and 56,900 mt. Spawning stock biomass (SSB; Age 0+) declined 72% from 1983 to 1989 (18,800 mt to 5,200 mt), but with improved recruitment and decreased fishing mortality had increased to 32,600 mt by 2004, before decreasing to 30,600 mt in 2005. Retrospective analysis shows a tendency to overestimate the SSB in the most recent years. Over the last 5 years, the annual retrospective decrease in SSB has averaged 17%.The age structure of the spawning stock has expanded, with 74% at ages 2 and older, and 23% at ages 5 and older. Under equilibrium conditions at F_max, about 85% of the spawning stock biomass would be expected to be ages 2 and older, with 50% at ages 5 and older. The arithmetic average recruitment from 1982 to 2005 is 35 million fish at age 0, with a median of 33 million fish. The 2005 year class is currently estimated to be well below the median recruitment level. Retrospective analysis shows that the current assessment method tends to overestimate the abundance of age 0 fish in the most recent years. Over the last 5 years, the annual retrospective decrease in recruitment has averaged 10%.

Special Comments

The landings from the commercial fisheries used in this assessment assume no under reporting of summer flounder landings. Therefore, reported landings from the commercial fisheries should be considered minimum estimates.
The recreational fishery landings and discards used in the assessment are estimates developed from the Marine Recreational Fishery Statistics Survey (MRFSS). While the estimates of summer flounder catch are among the most precise produced by the MRFSS, they are subject to possible error and bias. A sensitivity analysis to examine to determine the impact of uncertainty in the recreational data on the assessment results revealed that the Proportional Standard Errors (PSEs) of MRFSS estimated landed numbers of fish (1982-2005 average = 6%) and discarded numbers of fish (1982-2005 average = 8%) are relatively small. Therefore, the impact of potential bias within the range of +/- 2 PSEs is also small. Changes in biomass estimates are directly proportional to the input changes in catch (e.g., increasing/decreasing total catch by 5% increases/decreases total biomass by 5%). Changes in fishing mortality are relatively smaller, because of restriction to fully recruited ages 3-5.
Given the persistent retrospective underestimation of fishing mortality in the assessment, managers should consider adopting a lower TAL for 2007 than indicated by the median projection results to reduce the risk that overfishing will occur. Managers should note that over the last 5 years, the annual retrospective increase in fishing mortality has averaged 33%, the annual retrospective decrease in SSB has averaged 17%, and the annual retrospective decrease in recruitment has averaged 10%.
The SAW 2006 Southern Demersal Working Group reviewed the in-progress and proposed research recommendations from the SAW41 assessment, and developed a prioritized list of research recommendations for future consideration. Research recommendations are grouped into high, medium, and low priority categories:

High

Conduct further research to better determine the discard mortality rate of commercial fishery summer flounder discards.
Continue ongoing age structure exchanges between the NEFSC and all interested state agencies and academic institutions.
Complete the NEFSC comparison study between scales and otoliths as aging structures for summer flounder.
Develop a long term protocol to sample otoliths from summer flounder caught in the recreational and commercial fisheries (e.g., purchase samples; as a component of Research Set-Aside projects; as Cooperative Research with industry).
As an alternative to"production sampling"of ototliths, develop a long term protocol to correct summer flounder scale ages using a more limited sample of otolith ages.

Medium

Conduct further research to better determine the discard mortality rate of recreational fishery summer flounder discards.
Evaluate use of a forward calculating age-structured model for comparison with VPA. Forward models would facilitate use of expanding age/sex structure and allow inclusion of historical data. If sex-specific assessments are explored, the implications on YPR should also be investigated.
Evaluate trends in the regional components of the NEFSC surveys and contrast with the state surveys that potentially index components of the stock. Explore statistical methods to develop“combined”survey abundance indices (by age if possible) from state agency survey data, for use in calibration of analytical assessment models.

Low

Explore the sensitivity of the VPA results to separating the summer flounder stock into multiple components.
Develop a program to annually sample the length and age frequency of summer flounder discards from all sectors of the recreational fishery. The NMFS For-Hire Survey (FHS) has been implemented for 2005 and subsequent years to collect data from the party/charter boat discards, but synoptic data for private boat and shore discards continue to be needed.

ACKNOWLEDGMENTS

Special thanks to Jay Burnett and the staff of the NOAA Fisheries NEFSC Population Biology Branch for their timely preparation of the summer flounder ages used in this assessment.

LITERATURE CITED

Anthony V. 1982. The calculation of F 0.1: a plea for standardization. Halifax NS: Northwest Atlantic Fisheries Organization, Serial Document SCR 82/VI/64.

Applegate A, Cadrin S, Hoenig J, Moore C, Murawski S, Pikitch E. 1998. Evaluation of existing overfishing definitions and recommendations for new overfishing definitions to comply with the Sustainable Fisheries Act. Overfishing Definition Review Panel Final Report; 179 p.

Brodziak J, Rago P. MS 1994. A general approach for short-term stochastic projections in age-structured fisheries assessment methods. Woods Hole MA: Population Dynamics Branch, Northeast Fisheries Science Center.

Bugley K, Shepherd G. 1991. Effect of catch-and-release angling on the survival of black sea bass. N. Amer. J. Fish. Mgmt. 11:468-471.

Burns TS, Schultz R, Brown BE. 1983. The commercial catch sampling program in the northeastern United States. In Doubleday WG, Rivard D, eds. Sampling commercial catches of marine fish and invertebrates. Can. Spec. Publ. Fish. Aquat. Sci. 66: 290 p.

Clark SH. 1979. Application of bottom-trawl survey data to fish stock assessments. Fisheries 4: 9-15.

Conser RJ, Powers JE. 1990. Extension 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: 461-47.

DeLong A, Sosebee K, Cadrin S. 1997. Evaluation of vessel logbook data for discard and CPUE estimates. Stock Assessment Workshop (SAW) 24 Stock Assessment Review Committee (SARC) Working Paper Gen 5. 33 p.

Diodati PJ, Richards RA. 1996. Mortality of striped bass hooked and released in salt water. Trans. Am. Fish. Soc. 125(2):300-307.

Gavaris S. 1988. An adaptive framework for the estimation of population size. Canadian Atl. Fish. Sci. Adv. Comm. (CAFSAC) Res. Doc. 88/29. 12 p.

Gearhart J. 1999. Short-term hooking mortality of summer flounder. Semi-annual report for NCDMF ACFCMA Project 2-ACA-003, 10 p.

Hoenig JM. 1983. Empirical use of longevity data to estimate mortality rates. Fish. Bull. 81: 898-902.

International Pacific Halibut Commission (IPHC). 1988. Annual Report, 1987. Seattle, Washington. 51 p.

Jones WJ, Quattro JM. 1999. Genetic structure of summer flounder (Paralichthys dentatus) populations north and south of Cape Hatteras. Marine Biology 133: 129-135.

Kraus RT, Musick JA. 2003. A brief interpretation of summer flounder, Paralichthys dentatus, movements and stock structure with new tagging data on juveniles. Mar. Fish. Rev. 63(3): 1-6.

Lucy JA, Holton TD. 1998. Release mortality in Virginia’s recreational fishery for summer flounder, Paralichthys dentatus. Virginia Marine Resource Report Number 97-8, 48 p.

Lux FE, Porter LR. 1966. Length-weight relation of the summer flounder (Paralichthys dentatus (Linneaus). U.S. Bureau Comm. Fish., Spec. Sci. Rept. Fish. No. 531; 5 p.

Malchoff MH, Lucy J. 1998. Short-term hooking mortality of summer flounder in New York and Virginia. Interim report for Cornell Univ/DEC Project MOU 000024, 6 p.

Mid-Atlantic Fishery Management Council (MAFMC). 2001. SAW Southern Demersal Working Group 2001 Advisory Report: Summer Flounder; 12 p.

National Research Council (NRC). 2000. Improving the collection, management, and use of marine fisheries data. National Academy Press, Washington, DC. 222 p.

Northeast Fisheries Center (NEFC). 1990. Report of the Eleventh NEFC Stock Assessment Workshop Fall 1990. NEFC Ref. Doc. No. 90-09; 121 p.

Northeast Fisheries Science Center (NEFSC). 1993. Report of the 16th Northeast Regional Stock Assessment Workshop (16th SAW). NEFSC Ref. Doc. No. 93-18; 116 p.

NEFSC. 1996. Report of the 20th Northeast Regional Stock Assessment Workshop (20th SAW): Stock Assessment Review Committee (SARC) Consensus Summary of Assessments. NEFSC Ref. Doc. No. 95-18; 211 p.

NEFSC. 1997a. Report of the 24th Northeast Regional Stock Assessment Workshop (24th SAW): Stock Assessment Review Committee (SARC) Consensus Summary of Assessments. NEFSC Ref. Doc. No. 97-12; 291 p.

NEFSC. 1997b. Report of the 25th Northeast Regional Stock Assessment Workshop (25th SAW): Stock Assessment Review Committee (SARC) Consensus Summary of Assessments. NEFSC Ref. Doc. No. 97-14; 143 p.

NEFSC. 2000. Report of the 31st Northeast Regional Stock Assessment Workshop (31st SAW): Stock Assessment Review Committee (SARC) Consensus Summary of Assessments. NEFSC Ref. Doc. No. 00-15; 400 p.

NEFSC. 2002. Report of the 35th Northeast Regional Stock Assessment Workshop (35th SAW): SARC Consensus Summary of Assessments. NEFSC Ref. Doc. 02-14; 259 p.

NEFSC. 2005. Report of the 41st Northeast Regional Stock Assessment Workshop (41st SAW): 41 st SAW Assessment Summary Report. NEFSC Ref. Doc. 05-10; 36 p.

Parrack ML. 1986. A method of analyzing catches and abundance indices from a fishery. Int. Comm. Conserv. Atlantic Tunas, Coll. Vol. Sci. Pap. 24: 209-221.

Pauly D. 1980. On the interrelationship between natural mortality, growth parameters, and mean environmental temperature in 175 fish stocks. J. Cons. int. Explor. Mer 42:116-124.

Stock Assessment Workshop Southern Demersal Working Group (SDWG). 2004. Summer flounder assessment summary for 2004; 9 p.

Terceiro M. 1999. Stock assessment of summer flounder for 1999. NEFSC Ref. Doc. 99-19; 178 p.

Terceiro M. 2003. Stock assessment of summer flounder for 2003. NEFSC Ref. Doc. 03-09; 179 p.

Thompson WF, Bell FH. 1934. Biological statistics of the Pacific halibut fishery. 2. Effect of changes in intensity upon total yield and yield per unit of gear. Rep. Int. Fish. (Pacific halibut) Comm. 8; 49 p.

Weber AM. MS 1984. Summer flounder in Great South Bay: survival of sublegals caught by hook-and-line and released. Stony Brook, NY: New York State Department of Environmental Conservation, Division of Marine Resources; 27 p.

Wigley S, Terceiro M, DeLong A, Sosebee K. 1997. Proration of 1994-96 commercial landings of cod, haddock, and yellowtail flounder. SAW 24 SARC Working Paper Gen 4; 32 p.