486 



Fishery Bulletin 101(3) 



Sound 



36°00' 



35°00' 



Figure 1 



Map of North Carolina waters. Shrimping operations were observed in northern Pamlico 

 Sound (between the mouth of the Pamlico River and southern Roanoke Sound) and the lower 

 third of the Cape Fear River. For total bycatch, fleet shrimp landings and fleet shrimp effort, 

 the northern region includes Pamlico Sound and its tributaries, and the southern region 

 includes from the Cape Fear River to the New River. 



and other fisheries, fishery biologists need clear guidance 

 on which method to use to estimate bycatch and they need 

 a definitive knowledge of which methods are best under the 

 varying conditions that might be found in a field observer 

 study. 



In this article, I use both field data and computer simula- 

 tions to compare the methods of bycatch estimation used 

 in past studies. First, using field observations of Atlantic 

 croaker {Micropogonias undulatus), spot (Leiostomas xan- 

 thurus), and weakfish bycatch from shrimp trawlers in 

 North Carolina, I compare bycatch estimates generated 

 by the CPUE-mean-per-unit estimator with two different 

 forms of the F:S ratio estimator, the mean of the individual 

 fish to shrimp ratios and the ratio of the mean catch offish 

 to the mean catch of shrimp. I then simulate fishing fleets 

 with different catches of fish and shrimp, and estimate 

 bycatch using the following five different estimators, a 

 mean-per-unit estimator and four forms of the ratio esti- 

 mator, respectively: 1 ) the mean fish catch per unit of effort, 

 where unit effort is a proxy for sample size, 2) the mean 

 of the individual fish to shrimp ratios, 3) the ratio of mean 

 fish catch to mean shrimp catch, 4) the mean of the ratios 

 offish catch per time fished (a variable measure of effort), 

 and 5) the ratio of mean fish catch per mean time fished. 

 The simulations employ different mean catches offish and 

 shrimp, different levels of variability around the catches 

 offish and shrimp and around the variable measure of ef- 

 fort in the ratio estimator, and different levels of observer 

 coverage, or the number of observations. I also investigate 



the effects on the bycatch estimates of different underlying 

 distributions offish and shrimp, including normal distribu- 

 tions offish and shrimp with different levels of correlation 

 between the catches of fish and shrimp, and delta lognor- 

 mal distributions of both fish and shrimp, with differing 

 probabilities of catching fish or shrimp. 



Materials and methods 



Field sampling 



To compare the methods described in the literature using 

 field data, I observed shrimping operations aboard com- 

 mercial shrimp boats from July through October 1995 in 

 Pamlico Sound, North Carolina, and from August through 

 October 1995, in the Cape Fear River, North Carolina 

 (Fig. 1). These two areas have different levels of fishing 

 effort, different fish-to-shrimp ratios, and different prob- 

 abilities of catching fish and shrimp. All fishermen coopera- 

 tors were unpaid volunteers, and I did not direct them in 

 any way regarding where or how to fish. Although sampled 

 boats were not randomly chosen, the fishermen appeared 

 to use gear and fishing methods similar to those of other 

 shrimpers, and other shrimpers were often seen fishing in 

 the area near the sampled boats. 



Sampled shrimp boats towed one or two nets, and all 

 nets contained some form of turtle excluder device (TED) 

 and bycatch reduction device (BRD) required by regulation. 



