154 



Fishery Bulletin 89(1), 1991 



- Atlantic croaker- 

 silver/sand seatrout 



' Spotted \ 

 seatrout — * 



Apr. May Jun. Jul. Aug. Sep. Oct. Nov. Dec. Jan. Feb. 

 1987 ,o 88 



Figure 1 



Calculated values of Atlantic croaker-silver/sand seatrout, 

 bluefish-red drum, and spotted seatrout abundance on ortho- 

 gonal of month from multiple regressions. 



depth x submerged surface area (Table 5). Platform 

 age did not affect the abundance of reef fish as evi- 

 denced by the negative regression coefficient for quad- 

 ratic age of the platform (Table 5). Angler character- 

 istics were not good predictors of reef fish catches, with 

 only the presence of graph recorders and fishing 

 method being significant (Table 5). Since the regres- 

 sion coefficient of fishing method was negative, reef 

 fish catches were highest while bottom fishing. 



Catches of pelagic fish were higher while trolling in 

 relatively small, well-equipped vessels near large un- 

 manned structures in intermediate water depths (Fig. 

 3). Retention of 10 significant predictor variables 

 accounted for 31.5% of variance in the pelagic fish 

 factor (Table 5). Regression coefficients for fishing 

 method, LORAN presence, linear date, submerged sur- 

 face area, water depth, and the interaction of sub- 

 merged surface area and volume of water enclosed 

 were positive from the MRA, while negative regres- 

 sion coefficients were found for the interactions of 

 structure manned x structure in production, water 

 depth x submerged surface area, and water depth x 

 volume of water enclosed and boat length (Table 5). 



Highest abundances of Atlantic croaker and silver/ 

 sand seatrout were found near small, manned plat- 

 forms in deep water. Angler characteristic variables 



Water Depth 

 (m) 



Volume Enclosed 

 (•100,000 m 3 ) 



Figure 2 



Response surface plot of water depth, volume of water en- 

 closed by oil and gas platforms, and reef fish abundance from 

 the multiple regression of principal component scores for catch 

 data from logbook participants. Note: stippled area represents 

 realistic values of volume of water enclosed at various depths 

 for oil and gas platforms. 



had little influence on their catch rates. Eight signifi- 

 cant predictor variables explained 18.4% of the Atlan- 

 tic croaker-silver/sand seatrout factor (Table 5). Tem- 

 poral variables (quadratic date and linear date) had the 

 highest partial r 2 values (Table 5) and indicated Atlan- 

 tic croaker-silver/sand seatrout were most abundant 

 in the early spring (Fig. 1). Positive regression coeffi- 

 cients were found for water depth and the interaction 

 of structure manned x structure in production, and 

 negative regression coefficients were found for volume 

 of water enclosed and structure age x submerged sur- 

 face area. LORAN presence and fishing method were 

 the only significant angler characteristic variables re- 

 tained in MRA (Table 5). 



Based on MRA results, highest catches of bluefish 

 and red drum occurred while trolling from late winter 

 to early spring near platforms with complex construc- 

 tion. Retention of five significant predictor variables 

 explained 14.5% of the variance in the bluefish-red 

 drum factor (Table 5). The positive regression coeffi- 

 cient and the high partial r 2 (Table 5) for the interac- 



