SIGLER and FUJIOKA: VARIABILITY IN SABLEFISH 



catch data to a stratum was based on the recorded 

 depth of every fifth hachi or the interpolated depth 

 of each of the intervening four hachis. The number 

 of strata totaled nine, with the first stratum repre- 

 senting 101-200 m, the second representing 201- 

 300 m, and so on. A catch per hachi value was 

 calculated for each stratum of a station (Fig. 2), and 

 the resultant values within each statistical area were 

 averaged. 



Not all strata were sampled at each station and 

 as a result, sometimes no stations were sampled 

 within certain strata (Table 1). Although not sam- 

 pled one year, these strata generally were sampled 

 the previous year, because of slight annual differ- 

 ences in the depths sampled at each station. Catch 

 per hachi values CPUE,j were estimated for sta- 

 tions within the unsampled strata from the catch per 

 hachi values for sampled stations from the previous 

 year and the annual change in the catch per hachi 

 values in the adjacent stratum. For each station: 



CPUE, J = CPUE,_i J * CPUE ,,j /CPUEi_i__,. 



where i = yed.r,j = stratum, a.ndj' = the adjacent 

 stratum. Catch per hachi values were estimated only 

 for stations within unsampled strata. If one or more 

 stations within a stratum were sampled, then no 

 catch per hachi values were estimated for the un- 

 sampled stations. The effect of these missing values 

 on the survey results vdll be discussed later. 



Relative population numbers were calculated from 

 the catch per hachi values to index annual changes 

 in sablefish abundance. As in Sasaki (1985), the aver- 

 age catch per hachi was multiplied by the areal size 

 of each stratum and statistical area to calculate an 

 RPN for each stratum and statistical area. The areal 

 sizes used to calculate the RPN's are for the con- 

 tinental slope only (Table 2), which corresponds to 

 the area covered by the survey. The resultant RPN's 

 were summed across strata to calculate an RPN for 

 each statistical area, and these RPN's were summed 



Table 1.— Number of stations sampled by year, INPFC area, and depth during the Japan-U.S. cooperative longline survey, 1979-86. 



447 



