TRUMBLE ET AL.: STRAIT OF GEORGIA HERRING FISHERY 



Table 5.— Stock assessment summary, U.S. Strait of Georgia 

 sac-roe herring, 1976-79. 



10 15 20 25 30 



April 



10 15 20 25 30 4 

 May June 



Figure 5.— Biomass estimates of adult roe-herring in the 

 Strait of Georgia, 1979. 



recruitment. Average estimate of total adult 

 herring biomass was 8,950 tons. The fishery 

 harvested 21.8% of this estimated biomass in 

 1979. 



DISCUSSION AND CONCLUSIONS 



The combination of techniques applied in the 

 management of the sac-roe herring fishery 

 provides a timeliness and accuracy greater than 

 any single technique. The catch records are ob- 

 tained rapidly, but by themselves have little 

 management value. CPUE data are difficult to 

 evaluate in a timely manner and has question- 

 able application in a mixed gill net and purse 

 seine fishery on schooling fishes whose migration 

 patterns and timing vary annually; consequent- 

 ly, CPUE data are not used in the sac-roe herring 

 fishery. The spawning ground surveys provide 

 escapement data, but are not timely for in-season 

 management of the fishery. Conceivably the 

 excess biomass could be harvested after escape- 

 ment goals have been met, but this approach 

 forces the fishery to the end of the season when 

 fish are younger, smaller, and less valuable than 

 early in the season. 



The hydroacoustic-trawling data provide the 

 single most useful information for in-season 

 management. The estimates of potential spawn- 

 ing biomass are available for management 

 decisions by the end of the day following the 

 nighttime survey. The acoustic-trawl data have 

 provided good agreement with the other 

 measure of biomass. The average of the weekly 

 total run size estimates from the sum of all three 

 data sources have varied from 1% to 14% of the 

 final estimate from the catch and spawner 

 escapement estimate (Table 5). 



In all 4 yr the peak acoustic-trawling estimate 

 in conjunction with these data has provided a 



'This estimate consists of the average of acoustic-trawl plus cumulative 

 escapement plus cumulative catch point estimates from the time of full re- 

 cruitment through the final catch plus escapement estimate. 



reasonable and timely estimate of total run size. 

 However, the estimate of total run size obtained 

 using the peak acoustic-trawl estimate was 

 higher than the final estimate for all years ex- 

 cept 1979, and higher than all in-season point 

 estimates. Variability (random or unsystematic 

 factors) may contribute to the result, but in 

 general the high sampling power of acoustics 

 and the fairly uniform distribution of the 

 herring render this component inconsequential 

 (Saville 1977), and 95% confidence intervals 

 calculated from the acoustic data are typically 

 on the order of ±10%. Variability associated with 

 the trawling data for species composition is 

 probably more important, but is difficult to in- 

 corporate. Combined acoustic-trawling variance 

 estimation procedures have been developed for 

 other studies (Thomas 1979; Thomas et al. 1979); 

 however, they were not applied in this study 

 since we were more concerned with the sources 

 of potential error (systematic factors or bias). 

 Three sources of bias may contribute to the ob- 

 served differences between the peak acoustic- 

 trawling estimate and the final estimate. The 

 acoustic estimates may be biased high because of 

 the target strength assumption, but the acoustic 

 techniques may underestimate later in the run 

 when the fish move into shallow water just prior 

 to spawning. Studies by other investigators indi- 

 cate that a value of —32 dB/kg (which would 

 result in a 20% lower estimate) may be more 

 reasonable (Nakken and Olsen 1977; FAO 1978) 

 than the —33 dB/kg value used. Alternatively, 

 the estimates from spawning ground data could 

 be biased to the low side. 



Clearly more information on target strength is 

 needed to confidently establish the accuracy of 

 the acoustic technique as a measure of fish 

 biomass. The reasonable agreement with the 

 sum of the spawning escapement estimates and 

 catch is reassuring, but the spawning escape- 

 ment estimates are also subject to bias and un- 

 certainty, and the exploitation rate has been too 

 consistent to give much insight into the magni- 



387 



