Vaughan and Prager: Decline in abundance of Pagrus pagnK off \he southieastem United States 



357 



tic) model (Schaefer, 1954, 1957; Pella, 1967; Schnute, 

 1977; Prager, 1994). The approach is an ohservation- 

 error estimator assuming proportional error variance 

 and conditioned on observed landings. We attempted 

 to fit the generalized production model (Fella and 

 Tomlinson, 1969; Fletcher, 1978) by similar methods. 

 Data were total landings in weight. 1972-97, and CFE 

 indices derived separately from fishery-independent 

 hook-and-line and trap data, 1979-97. 



Initial difficulty in obtaining estimates was traced 

 to two years of zero CPE in the fishery-independent 

 hook-and-line index. Because the fitting was conduct- 

 ed in logarithmic transform, zeroes could not be ac- 

 commodated; our initial, unsuccessful approach was 

 to substitute very small values. We obtained estimates 

 through the alternate procedure of considering those 

 values missing, which is statistically equivalent to 

 setting the observation weights to zero. 



Because some estimates were moderately sensitive 

 to assumptions about the starting stock biomass, Bj in 

 relation to carrying capacity. A', we obtained estimates 

 at a range of values, namely B^IK = (0.55, 0.65, 0.75, 

 0.85, 0.951 , a range chosen to reflect the stocks lightly 

 fished status before the first year (1972). Fixing the 

 starting value of relative biomass can reduce variance 

 of estimates considerably (Punt, 1990). Bootstrapping 

 with 700 repetitions was used to estimate 80% bias- 

 corrected confidence intervals (Efron and Tibshirani, 

 1986) on the central set of estimates (Bj/A'=0.75). 



Results 



S 750 



-■ 250 



B Headboat D Recreational B Commercial 



1250 

 1000 



S Headboat D Recreational Commercial I 



Figure 1 



Annual landings of red porgy off the southeastern United 

 States by fishery (headboat, recreational, and commercial) in 

 (A) weight, and (Bi numbers. 



In this section, all results are reported for the base 

 value of M = 0.28/yr unless otherwise stated. Results 

 reflect analysis of the primary catch matrix unless use of 

 the alternate matrix is specifically mentioned. 



Summary of landings and CPE 



During the study period (1972-97), total landings in 

 weight rose from about 325 t in 1972 to a maximum of 

 880 t in 1982 (Fig. lA). Landings have been at or below 

 300 t since 1992, with 234 t in 1997. The headboat fishery 

 accounted for about 52^-80% of total landings through 

 1977; commercial landings have been most prevalent since 

 (Fig. IB, Table 3). Maximum landings in numbers were 

 about 1.13 million fish in 1982 (Table 4). 



Commercial landings in weight rose from 45 t in 1972 

 to 678 t in 1982 and declined to 200 t/yr or less since 

 1992 (Fig. lA); landings in numbers had a similar pat- 

 tern (Fig. IB). Commercial trawl landings averaged about 

 14*^^ of commercial landings. 1972-84 but were quite small 

 by 1985, and trawl gear was prohibited as of 12 January 

 1989 by amendment 1 to the Fishery Management Plan 

 (amendment 12 [SAFMC^]). Hook-and-line landings have 

 been about 95'7f of commercial landings by weight since 

 1985. 



Recreational landings in numbers have shown no con- 

 sistent trend, averaging about 70,000 fish per year, with 



some outstanding years of more than 100,000 fish between 

 1985 and 1990 (F'ig. IB). Catch by mode of fishing has 

 been highly variable, but recreational landings have main- 

 ly been taken by charter boat (43*? by number during the 

 1990s) and private boat (55*7^). Although remaining land- 

 ings (<3%) are attributed to fishing for red porgy from 

 shore, this amount is unlikely to be correct. These records 

 may result from errors in recorded fishing mode or from 

 erroneous species identification, but in either case would 

 have negligible effect on our analyses. 



Headboat landings have declined from 200,000 fish/yr 

 during the 1970s to around 100,000-200,000 fish/yr in the 

 1980s and fewer than 100,000 fisli/yr since (Fig. IB). Most 

 landings have come from NC and SC. 



Mean weight in the commercial hook-and-line and head- 

 boat fisheries in the Carolinas has decreased, especially 

 since about 1978 (Fig. 2). Mean weight in all fisheries com- 

 bined has shown a general decline from 1972 to 1991 (Fig. 

 2) but has since increased somewhat. 



All data on catch per effort (CPE) exhibited long-term 

 decline. Declining headboat CPE was apparent from both 

 NC and SC (Fig. 3). The fishery-independent abundance 

 indices, also based on CPE, showed corresponding declines 

 (Fig. 3); the extended chevron-trap index decline was pre- 

 cipitous; that in the hook-and-line index was similar, but 

 somewhat less dramatic. (The CPE series derived from 



