Diamond: Estimation of shrimp trawl bycatch 



487 



To sample the catch, I used the NMFS bycatch samphng 

 protocol as described below. If the boat carried two nets and 

 no try net (the small net towed in front of the main nets 

 which is used to survey the catch at short time intervals), 

 I randomly picked one net (the "selected net") by flipping 

 a coin. If the boat had a try net, I picked the opposite net. 

 I weighed the total catch of the selected net on a flat agri- 

 cultural scale by emptying the net into a plastic tub placed 

 on a scale. After having been weighed, the catch of the se- 

 lected net was dumped onto the deck or into a culling tray 

 that was divided so that the catch of the selected net was 

 separated from the catch of the unselected net. Following 

 the NMFS protocol, I mixed the selected net contents thor- 

 oughly with a shovel, then took a random sample and set it 

 aside until after the rest of the net contents had been sort- 

 ed. To sort the net contents, marketable shrimp, which are 

 pink shrimp (Faifantepenaeus duorarum), brown shrimp 

 {Farfantepenaeus aztecus), and white shrimp (Litopenaeus 

 setiferus) larger than would comprise about a 70-80 count 

 (i.e. 70-80 shrimp per pound) were separated from the rest 

 of the contents of the selected net, weighed, and then re- 

 turned to the fisherman. The unsampled bycatch from the 

 selected net was discarded overboard. The random sample 

 taken from the selected net was then weighed. Market 

 shrimp in the sample were taken out, weighed and counted 

 by species, and returned to the fisherman. The bycatch por- 

 tion of the sample, including undersized market shrimp, 

 mantis shrimp, and all other fish and invertebrates, was 

 packaged in plastic bags and placed on ice for the remain- 

 der of the trip. Bycatch samples were brought back to the 

 laboratory and frozen. Samples, including market shrimp, 

 averaged 12% by weight of the total catch of the selected 

 net. and ranged from 5% to 37% by weight. 



Expansion of observed bycatch to the entire tow In the 



laboratory, I thawed and rehydrated the bycatch sample in 

 water. I sorted each sample by species and weighed each 

 species as a group. All individuals of each species were then 

 weighed and measured separately. To account for differ- 

 ences between the scales used on the boat and those used 

 in the laboratory, and for weight loss due to freezing, I cor- 

 rected the weight of the total catch of each net measured on 

 the boat by the ratio of the sample weight from the labora- 

 tory to the sample weight from the boat as follows: 



Corrected total weight ^ = 



lab sample weight 



boat total weighty x 



(1) 



boat sample weighty 



where corrected total weighty = the corrected weight of the 



f-^ selected net; 

 boat total iveightj = the weight of the entire 

 catch of the/^" selected net 

 measured on the boat; 

 lab sample iveightj = the weight of the bycatch 

 sample of the/'' net mea- 

 sured in the laboratory 

 plus the shrimp sample 

 weight from the boat; and 



boat sample iveightj = the weight of the entire 

 sample (including shrimp) 

 from the 7^'' net weighed 

 on the boat. 



This correction averaged less than 5% across all selected 

 nets. To expand the catch in weight of each bycatch spe- 

 cies from the sample to the entire selected net (called the 

 "species net weight"), the total corrected weight of each 

 selected net was multiplied by the fraction of the sample 

 from the selected net that consisted of the bycatch species, 

 as follows: 



corrected total weight, x 



Species net weighty ^ = 



species sample weighty ^ 



total sample weight j 



(2) 



where species net weighty : = the estimated catch in 



weight of the i^^ species in 

 the/'' net; 

 corrected total weight = the corrected weight of the 

 total catch of the/'' net from 

 Equation 1; 

 species sample iveightj , = the weight of the j"' species 



in the sample from the /'' 

 net; and 

 total sample iveightj = the weight of the bycatch 

 sample from the/'' net mea- 

 sured in the laboratory plus 

 the weight of the market 

 shrimp in that sample mea- 

 sured on the boat. 



Because the net contents were thoroughly mixed before 

 sampling, I assumed, following the NMFS protocol, that 

 there would be minimal variance among samples if more 

 than one were taken. 



Expanding the catch in numbers of each bycatch spe- 

 cies from the sample to the entire selected net (called the 

 "species net number") could not be done in the same way 

 as the expansion for the species net weight because there 

 were often organisms like sea lettuce or pieces of fish or 

 crabs that were weighed but that could not be counted. The 

 species net number was therefore calculated by dividing 

 the species net weight by the average weight per whole 

 individual: 



species net weight, ^ 



Species net number^ - = 



species sample weight, ^ 



species sample number,^ 



(3) 



where species net number 



species net weight, 



the estimated number of 

 individuals of the i''' spe- 

 cies in the/'' net; 

 the estimated total weight 

 of the i*^ species in the /•■ 

 net from Equation 2; 



