Table 6. — Fishing power correction factors for the 

 vessels Anna Marie. Pa: San Marie, and Oregon 

 relative to Miller Freeman. 





Anna 



Pal San 





Species 



Marie 



Marie 



Oregon 



Walleye 









pollock 









(<20 cm) 



0.52 



0.34 



0.21 



Walleye 









pollock 









(a 20 cm) 



0.79 



0.61 



0.63 



All pollock 



0.75 



0.57 



0.35 



Rock sole 



0.65 



0.76 



1.21 



Snow 









(Tanner) crab 



0.66 



0.75 



2.53 



King Crab 



0.70 



1.05 



1.63 



'For all other species, the fishing power correc- 

 tion factor for each vessel was assumed equal to 1 .00. 



£ (CPU% - CPUE /lt ) 2 



VAR CPUE,* = 



"iff, - 1) 



(3) 



The overall mean CPUE for the entire survey area (CPUE rA .) was 

 determined as the sum of the weighted mean CPUE values of the 

 individual subareas: 



CPUE™ = 



(CPUE,* • Ai : 



A-, 



(4) 



and pis the average effective path width swept. The ratio p/A. 

 relates the size of individual sampling units (1.0 km in length) to 

 the size of each survey subarea. 



The coefficient of vulnerability (Q can be considered to consist 

 of two components: 1) C h , the efficiency of the gear to capture 

 fish within the path of the trawl's cross-sectional mouth area; and 

 2) C u , the proportion of the target fish population distributed in 

 the water column within the trawl's vertical sampling path. 

 Although the specific vulnerabilities of Bering Sea demersal fish 

 to the sampling gear and methods used in the present study have 

 not been well evaluated, all analyses of this study have assumed 

 100% capture efficiency (C = 1.00). 



The average effective horizontal spread (p) of the modified 

 eastern trawl used by the reference vessel Miller Freeman was 

 estimated by diving observations to be 0.017 km (Pereyra et al. see 

 footnote 2). Other studies of eastern trawl performance have esti- 

 mated the vertical opening to average approximately 2.3 m, with a 

 range of 1.9-2.7 m (Wathne 1977). 



The biomass of species k within subarea ;' was then estimated by 

 the expansion: 



«* 



Ai 



CPUE, 



(8) 



having a variance of 



VAR £,* = (4i\ ' VAR CPUE,* . 



Confidence intervals for biomass estimates were computed as 



where A- t is the area of the rth subarea and A T is the total area of 

 all subareas combined. 



The variance of this estimate was determined as the weighted 

 sum of the individual variances of each subarea: 



%± 'k»J VVARS, , 



(10) 



where the number of effective degrees of freedom (n e ) was 

 determined according to Cochran (1977), equation 5.16: 



VAR CPUEt* 



~* 



• VAR CPUE 



(5) 



Standing stock estimates. — Estimates of population weight 

 (biomass) were made using the methods described by Alverson 

 and Pereyra (1969) that relate CPUE and stock density within an 

 area surveyed as 



i 



£ _/}-VARCPU%. 



L 



where 



ft • (VAR CPU%) 2 



(ID 



i = i n-l 



B ik = CPUE ik /q k , 



(6) 



f 



"pr, - «,) 



(12) 



where B ik is the estimated standing stock weight (kg) of the kth 

 species in the ith subarea, and q k is a coefficient of catchability. 

 This coefficient, relating the capture efficiency of the sampling 

 gear and sampling unit size, is defined: 



Q k = C.ip/Ai), 



(7) 



Nj equals the total number of sampling units in the ith subarea [ = 

 A/ ( p • 1 .0 km)] , and n j equals the number of stations in subarea 

 i. The biomass estimate for a given species or taxonomic group 

 and its variance for the total survey area were obtained by sum- 

 ming the subarea biomasses and variances, respectively: 



where C k is a proportionality coefficient describing the 

 vulnerability of individuals of species k to be caught and retained, 



OTk = w B ik 



i*=l 



(13) 



