Diamond: Estimation of shrimp trawl bycatch 



493 



spot, the grand F:S ratio estimate (ratio of the means) was 

 intermediate between the basic F:S ratio estimate and the 

 CPUE-mean-per-unit estimate in the northern region, but 

 was lower than either of the other estimates in the south- 

 em region. The grand F:S ratio estimate for weakfish was 

 larger by both weight and number than either of the other 

 two estimates in the northern region and was the smallest 

 in the southern region. The Hartley-Ross bias-corrected 

 F:S ratio estimator gave estimates for the northern region 

 that fell between the basic and grand F:S methods but gave 

 negative estimates for the southern region (Table 4). CVs of 

 the catch rates were generally larger for the basic F:S ratio 

 estimator method than for CPUE-mean-per-unit estima- 



tor for spot in the northern region and Atlantic croaker in 

 both regions, and smaller for spot in the southern region 

 and weakfish in both regions. CVs estimated by the grand 

 F:S ratio method were much smaller than those for either 

 of the other methods for the northern region and much 

 larger than the others for the southern region (Table 5). 

 The variance of the catch rates with both methods was 

 usually much larger than the mean (sometimes by an or- 

 der of magnitude), indicating that catches were aggregated 

 (Table 5). The confidence intervals around the bycatch esti- 

 mates were huge regardless of method because of the small 

 number of observed fishing days and the large variability 

 in catch rates (Table 4). 



