132 



Fishery Bulletin 91 [1). 1993 



For each tow, shrimp CPUE (heads-off lb/h/100 ft of 

 headrope towed) from all TED-equipped nets were av- 

 eraged and compared against the average shrimp 

 CPUE of all standard nets, to provide one TED- 

 standard data pair per tow. Unless otherwise stated, 

 shrimp CPUE will refer to heads-off lb/h/100 ft of 

 headrope. The average CPUEs of two TED-equipped 

 and two standard nets were paired for each tow for 26 

 quad-rigged vessels and 1 twin-rigged vessel. However, 

 if one net was excluded from the analysis due to unac- 

 ceptable operation (refer to Gear Performance), then 

 the CPUE value from the remaining net was paired 

 with the average of CPUEs from the other two nets. If 

 both nets of a given gear type malfunctioned, all data 

 from that tow were deleted from the analysis. Stan- 

 dard and experimental nets were compared on twin- 

 rigged vessels and these data pooled with those from 

 quad-rigged vessels. 



Biological models Deterministic population models 

 were produced for brown shrimp Penaeus aztecus, white 

 shrimp P. setiferus, and pink shrimp P. duorarum by 

 linking a Ricker-type yield-per-recruit model to recruit- 

 ment estimates that were independent of parent stock 

 (Ricker 1975, Nichols 1984, Nance & Nichols 1988). 

 Recruitment level was set at the geometric mean for 

 the complete data set (1960-88). Estimates for 1986- 

 89 fishing mortality rates (F) were derived from vir- 

 tual population analysis, and the average was used as 

 the baseline for current conditions. Yield estimates were 

 made for all three species for a range of "F-multiplier" 

 values of 0-2 by 0.02 increments. Tables of these yield 

 estimates were used to determine effects of TED- 

 equipped nets on the shrimp yield in the Gulf of Mexico. 

 This was possible because yield estimates (Y t ) are a 

 direct result of fishing mortality rates (Royce 1972). 

 The yield model was 



Commercial shrimp catch Effort data for a given tem- 

 poral and spatial area were calculated by taking the 

 average trip CPUEs (heads-off lbs/24 h day/4 nets), ob- 

 tained by interviewing vessel captains, and extrapo- 

 lating to total effort by using the total-pounds value 

 from dealers' records. Fishing-effort data on the shrimp 

 fleet have been collected in this manner since 1960. 

 These data were compared with CPUEs (heads-off lbs/ 

 24 h day/4 nets) from our observer trips. The assump- 

 tion that shrimp CPUEs were equal, both for vessels 

 from this study and from the commercial fleet fishing 

 during the same seasons and in the same Statistical 

 Areas, was tested using a paired £-test with a prob- 

 ability level of 0.05. 



Y, = F t N t W, dt , 



where N, is the number of animals (R) in a cohort 

 subject to fishing (F) and natural (M) mortality at a 

 given time (t), using the formula 



N, = Re-' F+M,lt -V' 



F, = fishing mortality at a given time, 

 W t = average weight of an individual at time 

 t, estimated from growth equations. 



Fishing mortality rate (F) is the product of two sepa- 

 rate variables, a catchability coefficient (q) and directed 

 nominal fishing effort (f): 



Gear performance Each net was characterized by an 

 operation code based on its performance in the water. 

 Codes were used to describe successful tows or prob- 

 lems encountered, such as tangling of trawl doors, 

 gear fouling, twisted cables, bag choking, etc. Two 

 codes were occasionally required to describe trawl 

 performance. 



Data collected from the problematic tows not related 

 to TEDs, e.g., cod end coming untied, gear not fishing 

 properly, torn nets, and broken cables, were not in- 

 cluded in the analyses. Chi-square (P<0.05) analysis 

 was used to determine if the problematic tows were 

 independent of net type (e.g., TED-equipped nets or 

 standard nets) by area (Gulf of Mexico or Atlantic). 



F = qf. 



TED-equipped nets influence fishing mortality (F) by 

 affecting shrimp catchability (q), and not fishing effort 

 (f). Any percentage change in shrimp catchability 

 caused by TED-equipped nets was assumed to be di- 

 rectly reflected in an equal percentage change in fishing 

 mortality. This is based on an assumption of direct 

 proportionality between change in CPUE and change 

 in q. Thus, any change in CPUE as a result of TED 

 use is translated into a proportional change in q. 



Results 



Statistical analyses 



Paired t-tests Paired <-tests were performed to test the 

 hypothesis of equal CPUE of shrimp by standard and 

 TED-equipped trawls. Data were paired by tow. Confi- 

 dence intervals (95% ) on CPUE were also calculated. 



Descriptive data summary 



Paired data In the Gulf of Mexico, 589 data pairs 

 were collected using Georgia TEDs equipped with ac- 

 celerator funnels, 59 pairs from Georgia TEDs without 

 funnels, and 50 pairs from Super Shooter TEDs with 



