HIGHTOWER and GROSSMAN: TILEFISH FISHERY 



Table 5.— Estimates of virgin levels of recruitment, adult population density, and biomass; recommended levels of fishing mortality, biomass, 

 and yield; and 1987 biomass. Estimates were obtained from research CPUE data, using two sets of selectivity (LSOg^,) parameters, four 

 levels of natural mortality (M). and two assumptions about the stock-recruitment relationship (A = 0.889 - recruitment dependent on 

 spawning stock, A = 1.000 - recruitment constant). 



Female/male selectivity parameter LSOj^, 



M^ 



Virgin recruitment level (thousands) 

 Virgin population density (#/km^) 

 Virgin biomass (t) 

 Recommended F (A = 0.889) 

 Recommended F (A = 1 .000) 

 Recommended biomass (t) (A = 0.889) 

 Recommended biomass (t) (A = 1 .000) 

 Recommended yield (t) (A = 0.889) 

 Recommended yield (t) (A = 1 .000) 

 Estimated 1987 biomass (t) 



The two sets of CPUE estimates resulted in differ- 

 ent conclusions about the current status of the stock. 

 Based on commercial CPUE data, 1987 stock bio- 

 mass was 22-45% of virgin biomass and 51-105% 

 of the recommended level (Table 4). Estimated fish- 

 ing mortality rates increased from about 0.1 in 1981 

 to a range of 0.3-1.4 in 1986 (Fig. 3). Based on 

 research CPUE data, 1987 stock biomass was 55- 

 70% of virgin biomass and 132-145% of the recom- 

 mended level (Table 5). Estimated fishing mortality 

 rates were much lower than from commercial CPUE 

 data, increasing from about 0.08 in 1981 to 0.2 in 

 1986 (Fig. 3). 



We believe that the results obtained from com- 

 mercial CPUE data are far more likely, given recent 

 reported declines in directed fishing. The decrease 

 in landings observed in 1985 was attributed in part 

 to reduced fishing pressure. A large group of boats 

 from the Port Canaveral, FL area left the fishery, 

 and a number of Georgia vessels began fishing fur- 

 ther north (M. V. Rawson^). As of April 1988, most 

 South Carolina longline vessels had switched to 

 other fisheries and little directed tilefish fishing was 

 occurring (R. LoW). 



The difference in results for commercial and re- 

 search CPUE estimates may be due to differences 

 in areas fished. The RV Georgia Bulldog cruises 

 were exploratory in nature, and catches were ob- 

 tained primarily in the southern section of tilefish 

 habitat off the Georgia coast (Harris and Grossman 



'M. V. Rawson. University of Georgia Marine Extension Ser- 

 vice, Brunswick, GA 31523, pers. commun. March 1987. 



'R. Low, South Carolina Wildlife and Marine Resources Depart- 

 ment, Charleston, SC 29412, pers. commun. April 1988. 



1985). Early commercial effort was concentrated in 

 the more northerly part of the tilefish habitat (Low 

 et al. 1983). We recognize the commercial catch data 

 provide a biased measure of abundance. Neverthe- 

 less, because South Carolina landings predominated 

 in the developing fishery, we believe that commer- 

 cial catch data from the primary fishing grounds will 

 be a better overall measure of changes in abundance. 

 For that reason, we restrict our remaining com- 

 ments to results from the commercial CPUE data. 

 Declines in commercial CPUE may underrepresent 

 actual dechnes in abundance because fishermen 

 presumably would change tactics over time in order 

 to maintain high catch rates. If so, our use of com- 

 mercial catch data may result in an optimistic esti- 

 mate of current abundance. 



We obtained equivalent estimates of 1987 biomass 

 for the constant recruitment and density-dependent 

 cases because of the short length of the data series, 

 relative to the 6-yr lag between a reduction in 

 spawning stock and subsequent lower recruitment 

 to the fishery. Equilibrium yield curves differed 

 substantially for the two recruitment assumptions 

 (Fig. 4). Despite differences in equilibrium yield, 

 recommended Fs, were very similar for the two cases 

 because of the different criteria used to develop F 

 and yield recommendations (Table 4). Recommended 

 yield was moderately higher under the optimistic 

 constant recruitment assumption (Table 4). Recom- 

 mended F was higher when LSO^,,, was increased; 

 however, differences in recommended yield were 

 negligible. 



Estimated Fs differed substantially for the four 

 levels of natural mortality (upper panel, Fig. 3). The 

 large differences in 1984-86 Fs probably were due 



183 



