Trawling Efficiency Device: A New Concept for Selective Shrimp Trawling Gear

JOHN W. WATSON, JOHN F. MITCHELL, and ARVIND K. SHAH

Introduction cussed by Seidel (1975). Juhl et al. 1 Selective trawls also were used in Bel­ reported incidental finfish catch rates of gium, Norway, Iceland, and in the The incidental catch of finfish by the 3-20 pounds of fish for each pound of northwestern United States on crangonid penaeid shrimp fishery in the southeast­ shrimp caught. Other investigations and pandalid shrimp in the 1960's. A ern United States is an increasing prob­ have reported average fish-to-shrimp summary of this work was presented at lem in the management and use of weight ratios ranging between 3:1 and an FAO (Food and Agriculture Organi­ fishery resources. The need for selec­ 6:1 (Blomo and Nichols, 1974; Bryan, zation of the United Nations) sponsored tive shrimp trawling gear was first dis- 1980; Chittenden and McEachran, conference in 1973 (Anonymous, 1973). 1976). Pellegrin et aI.2 recently esti­ Efforts to develop selective shrimp mated fish bycatch for the northern Gulf trawls have continued in Canada (yVay The authors are with the Mississippi Laboratories, of Mexico penaeid shrimp fleet at over and Hickey, 1978), Norway (Karlsen, Southeast Fisheries Center, National Marine Fish­ 510,000 metric tons annually, based on 1976; and Isaksen3), and Great Britain eries Service, NOAA, Pascagoula, MS 39568­ 12m. Arvind K. Shah is also associate professor data collected by onboard observers and (Main and Sangster, 1982). These sep­ of statistics at the University of South Alabama, computed by geographical areas, sea­ arator designs all use panels of webbing Mobile. sons, and depths. The studies showed a placed in the mouth, throat, or along the high mortality of finfish associated with wings of the trawl to lead fish toward the penaeid fishery. escape openings, allowing shrimp to A separator trawl that could reduce pass through relatively large panel finfish bycatch mortality would benefit meshes into the cod ends. Other designs ABSTRACT-The Trawling Efficiency both the bottomfish fishery (by pro­ divide the trawl into upper and lower Device (TED) is inserted between the body viding greater resource abundance) and halves or use a "trawl within a trawl" and cod end ofa shrimp trawl. A steel grid the shrimp fishery (by reducing trawl design concept (Anonymous, 1973). and trap door ejects unwanted shrimp by­ catch such as turtles, sharks, rays, jellyfish, drag and fuel consumption). An effec­ Mechanical separation of fish and crabs, sponge, etc. The TED also can be tive separator trawl also would signifi­ shrimp with webbing panels has been used to reduce finfish bycatch by employing cantly reduce labor needed for sorting successful in fisheries where the differ­ afinfish deflector grid, leading panels, and shrimp from the finfish and improve ence between sizes of shrimp and fish exit openings. The device eliminates finfish shrimp quality by reducing damage to by taking advantage of the difference in is significant. Panel-type separator swimming ability and behavior between fin­ shrimp in the trawl. trawls of various designs tested in the fish and shrimp. Shrimp are carried into the Selective shrimp trawls have been Gulfof Mexico, however, have not been cod end by accelerating waterflow through used with varying success in several very successful (yVatson and McVea, the device with a webbing funnel. Finfish are shrimp fisheries around the world. The 1977). The principal difficulty was the stimulated into an escape reaction by a fin­ fish deflector grid and are guided to exit first separator trawls were developed in diversity of sizes and types of fish openings by leading panels. Finfish separa­ France and the Netherlands in 1964. associated with the shrimp fishery, re­ tion rates averaging 78 percent and 53 per­ sulting in panels becoming clogged with cent were achieved during day trawling and fish, and affecting separation and caus­ night trawling, respectively, with no signifi­ 'JuW, R., S. B. Drummond, E. 1. Gutherz, C. M. ing unacceptable losses in shrimp cant difference in shrimp catch rates. Fin­ Roithmayr,1. A. Beningo, and 1. A. Butler, 1976. fish separation rates varied by species and Oceanic resource surveys and assessment task production. Additionally, some panel­ total separation varied as a function ofcatch status report. NMFS Mississippi Laboratories, type separator trawls were too complex composition. The TED is being introduced P.O. Drawer 12m, Pascagoula, MS 39568-12m. into the shrimp fishery in the southeastern Unpubl. rep., 50 p. 2pellegrin, G., Jr., S. B. Drummond, and R. S. United States to reduce incidental turtle cap­ Ford, Jr. 1985. The incidental catch of fish by the 'Isaksen, B. 1982. Forsok moo vertikalt-stilte tures, conserve finfish resources discarded Northern Gulf of Mexico shrimp fleet. NMFS sidesorteringsnett i reketral. Fiskeriteknologisk by the shrimp fleet, and increase trawling Mississippi Laboratories, P.O. Drawer 12m, Pas­ Forskningsinstitutt Arbeidsnotat (working note), efficiency. cagoula, MS 39568-12m. Unpubl. rep. Bergen, Norway, 16 p.

48(1), 1986 1 N fUNNEL

~.------_.,

DEFLECTOR GR\D F\NF\SH DEFLECTOR

F\NF\SH QPEN\NG

"go« \.-coo,"oo,"~ 0' tho ,,",whog Hr."oo" D""co (TED) ~ p... 'S' '"';!\ '";:so... '";;" '" ~ ~ or too fragile for production fishing. angled bars and into the cod end of the tained from a commercial shrimp Another drawback is the complexity of trawl. Deflector bar spacing can be ad­ trawler operating off Louisiana in May fitting panels to the many diverse trawl justed between 3 and 6 inches to exclude 1985. types used in the shrimp fisheries. bycatch, such as cannonball jellyfish, Results In 1980, a unique separator trawl horseshoe crabs, and sponges. Finfish design was introduced by the National separation is accomplished by employ­ Data from nighttime tows for TED­ Marine Fisheries Service (NMFS). The ing a smaller finfish deflector grid and equipped and standard trawls are pre­ separator, initially called the Turtle Ex­ openings with leading panels to guide sented in Tables 1-6. The TED was cluder Device (TED), was developed in fish out of the trawl. The TED is in­ tested at night because previous data response to a critical conservation prob­ stalled between the trawl body and the have shown that the TED was effective lem involving the incidental capture and beginning of the cod end in the exten­ only during the daytime at separating mortality of endangered sea turtles. The sion of the trawl. A detailed description finfish (Watson4). Shown in Tables I, TED was designed to allow turtles to of the TED and installation instructions 3, and 5 are mean catch rates for stand­ escape from shrimp trawling gear are presented by Taylor et al. (1985). ard trawls and the trawls equipped with through a trapdoor positioned in the Three TED designs were evaluated the device. All data were normalized to throat of the trawl (Watson and Seidel, during this study: A collapsible TED I-hour tows. Data on finfish catch com­ 1980). During development ofthe TED, constructed from Ys-inch steel pipe, a position and reduction rates for each scuba-diver observations of fish and collapsible TED constructed from %­ TED design tested are presented in shrimp in the experimental devices in­ inch fiberglass rod, and a rigid-frame Tables 2,4, and 6. Species representing dicated a marked behavioral difference TED constructed from %-inch fiberglass less than 1 percent of the catch were not in fish and shrimp responses that could rod. Collapsible TED designs are included. The catch reduction rate for allow effective separation. Design hinged between the front and rear each species was estimated based on modifications proved effective at elim­ frames and the slanted bars, enabling measured differences between standard inating finfish, jellyfish, sharks, rays, the device to fold to a flat position when and TED-equipped trawls. sponge, and other bycatch (Watson4,5). not in use (Fig. 2). When fishing, ten­ A multivariate paired t-test was per­ Because of operational benefits, the sion in the trawl opens the TED to its formed to test the hypothesis of equal name of the device was changed to working configuration (Fig. 3). Advan­ catch rates for shrimp, finfish, and total "trawling efficiency device (TED)". It tages of the collapsible design are the catch simultaneously for standard and is presently being introduced through a light weight and easier handling and TED-equipped trawls. The null hypoth­ technology transfer program into the storage. All three designs are equipped esis was: shrimp fishery of the southeastern with finfish separator modifications, in­ United States. cluding finfish exits and a finfish deflec­ (j.lshrimp, j.lfinfish, j.ltotal) Std. tor grid constructed of Ys-inch fiberglass = (j.lshrimp, j.lfinfish, j.ltotal) TED. Materials and Methods tubing and Yt6-inch stainless steel strand The TED consists of a 36 X 42 X cable. 30-inch frame constructed of Ys-inch The TED designs were evaluated in Data collected for comparisons of galvanized pipe, or %-inch fiberglass 65-foot flat trawls spread by 9-foot X standard and collapsible steel TED­ rod (Fig. 1). Inside the frame are deflec­ 40-inch otter doors (Watson et aI., equipped trawls provided strong evi­ tors bars angled at about 45° and spaced 1984). Comparative 1- and 2-hour tows dence to refute null hypothesis of equal­ 3-6 inches apart with a 30- X 30-inch were conducted from a double-rigged ity in the two mean vectors (F = 8.53165 door at the top ofthe deflector bars. Ob­ 68-foot chartered commercial shrimp with 3 and 13 d.f.). The P value jects that cannot pass through the deflec­ vessel. A trawl with TED installed was (chances of observing what we have ob­ tor bars are forced through the door. towed simultaneously with an identical­ served in the sample or even more ex­ The door opens on hinges, allowing ob­ ly rigged flat trawl without a TED. The treme if the above null hypothesis is jects to pass out of the trawl, and then trawls were calibrated and tuned after true) was only 0.0027. In other words, closes as the object is released. Smaller Watson and Seidel (1985). Catches from the chance of yielding F = 8.53165 or objects (fish, shrimp, etc.) pass through each trawl were weighed and bycatch larger under the above null hypothesis a webbing funnel, which accelerates the samples taken. Data were collected on was only 0.0027. water flow and carries them through the shrimp catch rates, finfish catch rates, However, rejection of the null hypoth­ and total catch rates. Samples were esis does not indicate which of the three sorted and identified to species and variable differences have caused rejec­ ·Watson, 1. W. 1983. FYI982 sea turtle excluder mean weights determined. The gear was tion of that hypothesis. The use of trawl project report. NMFS Mississippi Labor­ atories, p.o. Drawer 1207, Pascagoula, MS 39568­ evaluated at night on shrimping grounds several univariate t-tests for this purpose 1207. Unpubl. rep., 22 p. off Alabama and Mississippi between is not appropriate as they will produce 'Watson 1. W. 1983. Cruise report, FRS OREGON lat. 30°06', long. 88°18' and lat. 30°08', an attained level (i .e., the experimental II Cruise 137. NMFS Mississippi Laboratories, P.O. Drawer 1207, Pascagoula, MS 39568-1207. long. 89°01' (14 August - 25 Septem­ error rate which is the chance ofdeclar­ Unpubl. rep., 16 p. ber 1984). Additional data were ob- ing at least one significant difference

48(1), 1986 3 Figure 2.-Collapsible TED's in the folded configuration.

Table 1.-Comparative catch results (pounds per Table 2.-Finfish catch reduction rates and catch composition by weight for com­ hour) between a standard rigged 65-foot flat trawl parative tows between a standard 65-foot flat trawl and a 65-foot flat trawl equipped and a 65-foot flat trawl equipped with a collapsible with a collapsible steel TED. steel TED. Catch Reduction Catch (pounds/hour) Species composition (0/0) rate ('!o)

Trawl No. Shrimp Finfish Total Atlantic croaker, Micropogon undu/atus 54 56 Spot. Leiostomus xBnthurus 10 72 Standard 16 7.9 377 469 Atlantic threadfin, Polydaclylus octonemus 3 49 TED 16 7.7 185 263 Seatrout, eynoscion sp. 3 40 Atlantic cutlassfish, Trichiurus lepturus 2 70 Percent Gulf butlerfish, Peprilus burti 1 56 difference 2 '51 '44 Atlantic bumper, Ch/oroscombrus chrysurus 1 67

'Significant difference at 95 percent level.

when in fact none exist) much higher was used here to test for shrimp, finfish, Collapsible Steel TED than the chosen or prespecified level. and total catch differences between the The approach discussed by Morrison standard and TED-equipped trawls The mean shrimp catch rate for the (1976) to control experimental error rate individually. collapsible steel TED was 7.7 pounds/

4 Marine Fisheries Review Figure 3.-Collapsible TED's in the working configuration.

Table 3.-Compamtive catch results between a standard Table 4.-Flnfish catch reduction rates and catch composition by weight lor com­ rigged 65-loot Ilat tmwl and a 65-loot IIat tmwl equipped parative tows between a standard 65-loot IIat trawl and a 65 loot-llat trawl equipped with a solid liberglass TED. with a solid fiberglass TED.

Catch (pounds/hour) Catch Reduction Species composition (oAl) rate (oro) Trawl No. Shrimp Finlish Total Atlantic croaker, Micropogon undulatus 63 58 Standard 14 6.79 422 479 Spot. Leiostomus xanthurus 5 72 TED 14 6.43 198 259 Gull butlerfish, Peprilus burti 3 80 Hardhead catfish. Arius telis 3 88 Percent Seatrout. Cynoscion sp. 2 11 difference 5 '53 '46 Atlantic threadfin. Polydactylus octonemus 1 12

'Denotes a significant diflerent at 95 percent level.

hour and 7.9 pounds/hour for the stand­ = 0.2335 with 3 and 13 d.f., P = hour and J77 pounds/hour for the stand­ ard trawl. There was no statistical evi­ 0.8714). ard trawl. These differences were highly dence to indicate differences between The mean finfish catch rate for the significant (F = 6.3935 with 3 and 13 average catch rates for the two trawls (F collapsible steel TED was 185 pounds/ d. f., P = 0.0068). The total catch rates

48(1), 1986 5 were 263 pounds/hour for the collaps­ was 422 pounds/hour for the standard trawl and a traWl equipped with a col­ ible steel TED and 469 pounds/hour for trawl and 198 pounds/hour for the solid lapsible fiberglass TED. This multi­ the standard trawl. Again, the difference fiberglass TED. This difference was variate hypothesis was rejected with in the average total catch was highly judged highly significant (F = 9.3430 strong evidence (F = 5.9612 with 3 and significant (F = 7.7525 with 3 and 13 with 3 and 11 d.f., P = 0.0023). The 17 d.f., P = 0.0057). d.f., P = 0.(032). The collapsible steel mean total catch was 479 pounds/hour Individual tests for each variable also TED had a 51 percent reduction in fin­ for the standard trawl and 259 pounds/ were performed. The mean shrimp fish catch and a 44 percent reduction in hour for the solid fiberglass TED. The catch rate was 11.4 pounds/hour for the total catch (Table I). difference in these catch rates also was standard trawl and 11.7 pounds/hour for The predominant finfish species in the judged to be highly significant (F = the trawl with the collapsible fiberglass standard catch were Atlantic croaker (54 8.8515 with 3 and 11 d.f., P = 0.0029). TED. The difference in mean shrimp percent) and spot (10 percent) (Table 2). The trawl equipped with the solid fiber­ catches was not judged significant (F = The Atlantic threadfin and seatrout each glass TED showed a 53 percent reduc­ 0.1589 with 3 and 17 d.f., P = 0.9225). comprised 3 percent of the catch. Atlan­ tion in finfish catch and 46 percent Mean finfish catch rates were 254 tic cutlassfish represented 2 percent of reduction in total catch when compared pounds/hour for the standard trawl and the finfish catch. Gulf butterfish and with the standard trawl (Table 3). 122 pounds/hour for the trawl with the Atlantic bumper each represented 1 per­ The species composition of the fin­ collapsible fiberglass TED. The differ­ cent of the catch. Finfish separation fish catch in the standard trawl is shown ence in mean finfish catch was judged rates varied between 40 percent and 72 in Table 4. The predominant species was significant (F = 3.6211 with 3 and 17 percent for the individual species. The Atlantic croaker, which made up 63 per­ d.f., P = 0.0346). The mean total catch best separation rates were for spot (72 cent of the catch. Five other species was 2fJ7 pounds/hour for the standard percent), Atlantic cutlassfish (70 per­ (spot, gulf butterfish, hardhead catfish, trawl and 159 pounds/hour for the trawl cent), and Atlantic bumper (67 percent). seatrout, and Atlantic threadfin) repre­ with the collapsible fiberglass TED. The There was a 56 percent reduction for sented 1-5 percent of the finfish catch. difference in these catch rates was Atlantic croaker and butterfish. Calculated finfish reduction rates for the judged significant (F = 3.5423 with 3 solid fiberglass TED were 58 percent and 17 d. f., P = 0.0370). The collaps­ Solid Fiberglass TED for Atlantic croaker, 72 percent for spot, ible fiberglass TED had a 52 percent A multivariate paired t-test was per­ 80 percent for gulf butterfish, 88 per­ reduction in finfish catch and a 46 per­ formed to test the hypothesis of equal cent for hardhead catfish, 11 percent for cent reduction in total catch compared catch rates for shrimp, finfish, and total seatrout, and 12 percent for Atlantic with the standard trawl (Table 5). catch simultaneously for the standard threadfin. Evaluation of the collapsible fiber­ trawl and the trawl with a solid fiber­ glass TED was conducted in two dif­ Collapsible Fiberglass TED glass TED. This multivariate hypothesis ferent geographical areas and the finfish was rejected (F = 18.171 with 3 and 11 A multivariate paired t-test was per­ catch composition varied between them d. f., P = 0.0001). formed to test the hypothesis of equal (Table 6). In Area I the predominant Individual tests for each variable were catch rates for shrimp, finfish, and total finfish species were gulf butterfish (19 also performed. Mean shrimp catch catch simultaneously for the standard percent), gulf menhaden (19 percent), rates for the solid fiberglass TED and standard trawl were 6.43 pounds/hour and 6.79 pounds/hour, respectively. The difference was judged nonsignificant (F Table 6.-Finlish catch reduction rales and catch composition by weight lor com­ = 0.7106 with 3 and 11 d.f., P = parative tows belween a standard 65-1001 flat trawl and a 65-loot flat trawl equipped 0.5657). The mean finfish catch rate wilh a collapsible liberglass TED.

Catch Reduction Species composition (%) rate (%)

Area 1 Gulf butterfish. Peprifus burti 19 25 Gulf menhaden. Brevoortia patronus 19 69 Table 5.-Comparative catch results between a Spot. Leiostomus xanthurus 14 69 65-loot Ilat trawl with 9-loot x 40-inch doors and Atlantic croaker, Micropogon undulatus 12 64 an indentical trawl rigged with a collapsible liber­ Spanish mackerel, Scomberomorus maculatus 9 100 glass TED. Seatrout. Cynoscion sp. 7 73 Hardhead catfish. Arius felis 7 64 Catch (poundslhour)

Trawl No. Shrimp Finfish Total Area 2 Atlantic croaker. Micropogon undulatus 54 62 Standard 20 11.4 254 297 Atlantic cutlassfish, Trichiurus lepturus 10 56 TED 20 11.7 122 159 Spot, Leiostomus xanthurus 8 75 Atlantic bumper, Ghloroscombrus chrysurus 8 100 Percent Scaled sardine, Harengula jaguana 6 100 difference -3 '52 '46 Seatrout, Gynoscion sp. 4 100 Hardhead catfish, Arius felis 4 49 'Denotes a significant difference at the 95 percent level.

6 Marine Fisheries Review spot (14 percent), and Atlantic croaker one side of the vessel and two identical using webbing panels have been effec­ (12 percent). Other species representing trawls without TED's were used on the tive under some conditions, but have >1 percent of the catch included Span­ other side. Data collected by the cap­ limitations and have not been successful ish mackerel (9 percent), seatrout (7 tain of the Miss Santrina included total in the penaeid shrimp fisheries. The percent), and hardhead catfish (7 per­ shrimp catch and total bycatch. The TED which evolved from the turtle ex­ cent). Finfish reduction rates calculated catches were kept separate, shrimp were cluder device uses a different mechan­ for the collapsible fiberglass TED for sorted and weighed, and total bycatch ism to separate finfish and other bycatch Area 1 ranged from 25 to 100 percent was estimated by filling shrimp baskets from the shrimp catch. It uses a rigid for individual species. The reduction and counting the number of baskets for frame placed in the zone of the trawl rates were 100 percent for Spanish each side. The number of baskets was where the wings and body taper into the mackerel, 60-70 percent for gulf men­ then multiplied by the average weight cod end. Finfish gilling is common in haden, spot, Atlantic croaker, and hard­ of a full basket. The Miss Santrina this section ofa standard shrimp trawl, head catfish, 73 percent for seatrout, shrimped both day and night, and the indicating that finfish escape reactions and 25 percent for gulf butterfish. data were analyzed separately and in occur in this zone. The catch composition in Area 2 was combination for daytime and nighttime The TED utilizes differences in the predominantly Atlantic croaker (54 per­ periods. behavioral reaction of finfish and cent). Atlantic cutlassfish made up 10 Three multivariate paired t-tests were shrimp and the better swimming abil­ percent of the catch, followed by spot performed on the commercial vessel ity of the fish to separate and exclude and Atlantic bumper (8 percent) and data (day, night, combined) to test the fish from the catch. A funnel of web­ scaled sardine (6 percent). Seatrout and hypothesis of equal catch rates for bing in the TED accelerates water flow hardhead catfish, each made up 4 per­ shrimp and bycatch simultaneously for entering the cod end of the trawl. Water cent of the catch. Finfish reduction rates both trawls (standard trawls and trawls entering the cod end is accelerated by in Area 2 were 100 percent for Atlantic with collapsible steel TED's). This the funnel carrying shrimp, which are bumper, scaled sardine, and seatrout, 75 multivariate hypothesis was rejected weak swimmers, past the deflector grid percent for spot, 62 percent for Atlan­ with strong evidence (F = 14.0463 with into the cod end (Watson4). Finfish ac­ tic croaker, 56 percent for Atlantic 2 and 34 d. f., P = 0.0000 for the day­ tively swimming in the trawl also pass cutlassfish, and 49 percent for hardhead time data; F = 5.49479 with 2 and 21 through the funnel in the accelerated catfish. d.f., P = 0.0120 for the nighttime data; water flow, but are stimulated by the and F = 17.8719 with 2 and 57 d. f., P closeness of the webbing to escape the Commercial Shrimp Trawls = 0.0000 for the combined data). trawl. The commercial shrimp vessel Miss Individual tests for each variable were As the fish pass through the funnel, Santrina, fishing out of Lafitte, La., col­ also performed on each of the three data they either strike a finfish deflector or lected data on the effectiveness of the sets (daytime, nighttime, combined). enter an area of less water flow to the TED during commercial operations in Average catch rates for the trawls and side of the main flow exiting the funnel May 1985. The Miss Santrina was catch reduction rates due to the TED are (Fig. 4). There, fish are guided by web­ rigged with four trawls (twin rigs). Two presented in Table 7 for each data set. bing panels and can exit the trawl trawls were equipped with TED's on Shrimp catch differences were not sig­ through the side exits. Shrimp do not nificant (F = 0.2619 with 2 and 34 d.f., have the swimming ability or behavior P = 0.7711 for the day data; F = 0.0398 necessary to reach the exit openings and with 2 and 21 d.f., P = 0.9610 for the are carried on into the cod end. Larger nighttime data; and F = 0.3013 with 2 objects or organisms that cannot pass Table 7.-Comparative catch results between TED­ and 57 d.f., P = 0.7410 for the com­ through the 3- to 6-inch openings of the rigged twin trawls and standard twin trawls from the bined data), while bycatch rates were commercial shrimp vessel Miss Santrins. main deflector grid are ejected through significant (F = 14.0019 with 2 and 34 the hinged door at the top of the TED. Catch (Ib/h) d.f., P = 0.0000 for the daytime data; The potential of the TED as a finfish F 4.5697 with 2 and 21 d.f., Trawl No. Shrimp Bycatch = P = separator was discovered by scuba 0.0225 for the nighttime data; and F = divers observing turtles passing through Daytime Standard 36 19.22 1,828 17.5585 with 2 and 57 d.f., P = 0.0000 the TED during tests ofthe device. They TED 36 19.00 831 for the combined data). noticed that fish had a tendency to turn Percent ditterence 1 '55 and swim out ofthe accelerated flow and Nighttime Discussion Standard 23 11.87 1,385 then swim forward inside the TED in TED 23 11.78 954 The trawling efficiency device repre­ the zone of relatively slack water around Percent difference 1 '31 sents a new concept in selective fishing the funnel. Other fish carried further in­ Combined Day and Night gear. Historically, separator trawl to the cod end tended to swim forward Standard 59 16.36 1,655 TED 59 16.19 879 designs have relied on webbing panels along the bottom and sides of the cod Percent difference 1 '47 to sort bycatch by species or size from end until they reached the wall of web­ 'Denotes a significant difference at the 95 percent level. the rest of the catch. Separator trawls bing at the front frame of the TED. The

48(1), 1986 7 finfish escape openings were designed to take advantage of the reaction of the fish within the trawl to allow them to escape without loss in shrimp produc­ tion. The TED introduced in 1980 was used by shrimp fishermen in some areas because it was effective in reducing the incidental catch of cannonball jellyfish, Stemolophus sp.; sponges, and horse­ shoe crabs, Limulus sp., which at times causes serious problems for shrimp fishermen. ---I---TRAWL BODY In 1982, the finfish separator modifi­ cations were introduced with limited success. Finfish separation rates up to 53 percent were achieved during the day compared with standard trawling gear, but only 10 percent reduction was achieved during nighttime trawling. Fish appeared to show different nocturnal and diurnal behavior and we tested several design modifications to achieve better nocturnal finfish separation. These included lights, luminescent materials, and various types of deflec­ --.~-f---WEBBING tors. The most successful ofthe modifi­ ACCELERATOR cations was a small deflector grid placed FUNNEL behind the main deflector frame. The ,J' finfish deflector grid was introduced in 1983 and resulted in improved finfish LFINFISH EXIT OPENING separation rates averaging 78 percent __--;;---LEADING PANEL during the day and 48 percent at night (Watson5). The finfish deflector grid acts as a mechanical stimulus and gen­ erates sound vibrations. The data presented in this paper represent the best separation rates achieved during the night. Chemical light sticks were tested to illuminate fin­ fish deflectors and the side openings. Preliminary results indicated that fmfish separation was improved but the data Figure 4.-TED finfish separation technique. were limited and not statistically signifi­ cant. Other modifications that may lead to further improvements in separation include placement and spacing of the ac­ species composition of the catch and results indicate that 61 percent of had­ celerator funnel, finfish opening, and may also be related to the size ofthe in­ dock <39 cm long and 30 percent ofcod the finfish deflector grid, and the spac­ dividual species as it relates to their <42 cm long escaped from shrimp ing and tension in the wires of the swimming ability. trawls equipped with the RES. The deflector grid. The separation concept used in the FTFI design does not use a rigid frame Effectiveness of the TED in separ­ TED has been tried in Norway by because trawls are handled using net ating finfish varies with individual Fiskeriteknologisk Forskningsinstitatt reels. The RES was constructed using species and appears to be related to the (FTFI) in a separator design called the two webbing funnels and large mesh swimming ability of the individual "Radial Escape Section" (RES) to sep­ webbing around the funnels to allow fish species and their behavior patterns. arate haddock and cod from shrimp escapement. Total separation rates thus vary with the trawls (West et aI., 1984). Preliminary The rigid frame structure of the TED

8 Marine Fisheries Review has several operational advantages over TED by the shrimp industry has in­ Chittenden, M. E., Jr., and 1. D. McEachran. 1976. Composition, ecology and dynamics of the webbing type structures of the RES, creased. TED offers long-term potential demersal fish communities on the northwestern although it is more cumbersome to as a tool for the conservation of finfish Gulf of Mexico continental shelf, with a similar handle. The rigid frame physically holds resources and improved efficiency of the synopsis for the entire Gulf. Tex. A&M Univ., Sea Grant Pub!. TAMU-SG-76-208, 104 p. the webbing in the desired shape, keep­ shrimping fleet. Karlsen, L. 1976. Experiments with selective ing the funnel and finfish exits open. prawn trawls in Norway. I.e.E.S. C.M. 1976/ B:28. Flexible designs of webbing or rope are Main, 1., and G. I. Sangster. 1982. A study of more difficult to design and maintain Acknowledgments separating fish from Nephrops Norregicus L. properly, and may not produce consis­ in a bottom trawl. Scotl. Fish. Res. Rep. 24, We recognize Andrew 1. Kemmerer 8 p. tent operational results because the and Wilber R. Seidel for their leader­ Morrison, D. F. 1976. Multivariate statistical structure will change shape when forces ship in the successful development ofthe methods. McGraw-Hill Pub!. N.Y., 415 p. are exerted on the flexible structure by Seidel, W R. 1975. A shrimp separator trawl for TED, and Charles W. Taylor and An­ southwest fisheries. Proc. Gulf Caribb. Fish. changing configurations of the trawl thony F. Serra for their critical technical Inst., 27th Annu. Sess., p. 66-76. under various trawling conditions. contributions and dedicated efforts. Ian Taylor, e. W, A. F. Serra, 1. F. Mitchell, and 1. W Watson. 1985. Construction and installation FTFI, however, has shown that a flex­ K. Workman, Jane P. Corliss, Noel H. instructions for the trawling efficiency device. ible design does have potential and more Watts, Charles McVea, and Alan R. NOAA Tech. Memo. NMFS-SEFC-71, 31 p. research may achieve better and more Bunn are acknowledged for their field Watson, 1. W, and e. McVea, Jr. 1977. Develop­ ment ofa selective shrimp trawl for the south­ efficient separation. work in testing the TED, Rodney C. eastern United States penaeid shrimp fisheries. The TED is presently being intro­ Sawyer for his contribution in TED Mar. Fish. Rev. 39(10):18-24. duced into the shrimp industry in the ____ , and W R. Seidel. 1980. Evaluation technology transfer, and Sally Glynn for oftechniques to decrease sea turtle mortalities southeastern United States through a her many dedicated hours for the TED in the southeastern United States shrimp fish­ technology transfer effort. The original project. ery. I.e.E.S. e.M. 1976/B:28. rigid-frame TED weighed CJ7 pounds _~,------: ' and . 1985. Techniques and methodology for "calibrating" shrimp and and required considerable deck space bottomfish sampling gear. Proc. SEAMAP for storage. Shrimp fishermen objected Literature Cited Gear Workshop, Gulf States Mar. Fish. Comm. Pub!. 12, II p. to this during initial efforts to introduce Anonymous. 1973. Report of the expert consulta­ ____ , I. K. Workman, e. W Taylor, and the gear. In 1984, the collapsible frame tion on selective shrimp trawls. FAO Fish. Rep. A. F. Serra. 1984. Configurations and relative was introduced, which solved handling 139, 71 p. efficiencies of shrimp trawls employed in south­ Blomo, V. 1., and 1. P. Nichols. 1974. Utilization eastern United States waters. NOAA Tech. Rep. and storage problems. of finfishes caught incidental to shrimp trawl­ NMFS 3, 12 p. The collapsible TED weighs 37 ing in the western Gulf of Mexico, Part I: Way, E., and W Hickey. 1978. Shrimp sorting pounds and folds flat for storage. It is Evaluation of markets. Tex. A&M Univ., Sea trawl. Environ. Can., Fish Mar. Ser., Indust. Grant Publ. TAMU-SG-74-212, 85 p. Rev. Branch, Newfoundland Reg., P.o. Box just as effective as the original TED Bryan, e. E. 1980. Organisms captured by the 5667, St. Johns, Rep. 081-7-7000, 28 p. design. Since the introduction ofthe col­ commercial shrimp fleet on the Texas brown West, e. W, 1. W Valdermarsen, and B. Isaksen. shrimp (Penaeus aztecus Ives) grounds. Masters 1984. Preliminary tests of a shrimp-fish separ­ lapsible design and the improvement of thesis, Corpus Christi State Univ., Corpus ator section for use in shrimp trawls. I.e.E.S. night finfish separation rates, the use of Christi, Tex., 44 p. e. M. 1984/B:12.

48(1), 1986 9