Effects of Turtle Excluder Devices (TEDs) on the Bycatch of Three Small Coastal Sharks in the Gulf of Mexico Penaeid Shrimp Fishery

The stock of blacknose sharks Carcharhinus acronotus in the U.S. South Atlantic and the Gulf of Mexico is overfished, and according to the 2007 stock assessment conducted by the National Marine Fisheries Service overfishing continues to occur. Penaeid shrimp trawl bycatch rates in the Gulf of Mexico were modeled for this species as well as for the Atlantic sharpnose shark Rhizoprionodon terraenovae and bonnethead shark Sphyrna tiburo using a combination of research trawl and observer data. Research trawls have never used turtle excluder devices (TEDs), which are expected to exclude larger specimens of blacknose sharks. Most of the observer data that contain blacknose shark occurrences were collected during the pre-TED era when the two data sets tracked one another. Minimum observer data were available for the post-TED period (1990–present). As a consequence, the pre-TED (1972–1989) relationship between observer and research trawl catch per unit effort (CPUE) is driving the observer CPUE estimates from 1990 to the present, a period characterized by increased blacknose shark abundance. We suspected that the increase in predicted observer CPUE in the post-TED era is an artifact of application of the pre-TED observer and research trawl relationship to the post-TED era. This suspicion led us to question whether the bycatch of these species was altered due to the use of TEDs. We used negative binomial regression in a before-after-control-impact setting to test the effects of TEDs on the bycatch rates of these small coastal sharks. The TED effect was found to substantially reduce the bycatch of blacknose sharks (by 94%) and to do so moderately for bonnethead sharks (31%); the results were inconclusive for Atlantic sharpnose sharks. The management implication of our findings is that the existing small coastal shark–penaeid shrimp fishery bycatch model needs to be modified or replaced with a model that explicitly incorporates the potential for a TED effect.