Interactions between sea turtles and dredge gear in the U.S. sea scallop (Placopecten magellanicus) fishery, 2001–2008

Since 2006, the National Marine Fisheries Service (NMFS) has mandated gear modifications (“chain mats”) and fishing effort reductions in the U.S. Mid-Atlantic sea scallop dredge fishery to alleviate or minimize interactions with sea turtles. Turtle interactions with gear can be defined as those that are “observable” based on standard fishery observer protocols, plus unobserved interactions, which include both quantifiable and unquantifiable interactions. Once a gear modification is in place, a turtle interaction that was once observable may become unobservable, because the gear modification successfully prevented the turtle from being captured. This paper describes turtle interactions in scallop dredge gear from 2001 to 2008, identifies gear and environmental correlates with observable interaction rates, and reports the average annual number of interactions and adult-equivalent interactions before and after chain mats were mandated in the fishery. Fisheries observer data were used to develop a Generalized Additive Model (GAM) to estimate rates of observable interactions of hard-shelled turtles. These rates were applied to commercial dredge fishing effort to estimate the total number of observable interactions, and to infer the number of unobservable, yet quantifiable interactions after chain mats were implemented. Interaction rates of hard-shelled turtles were correlated with sea surface temperature, depth, and use of a chain mat. The average number of annual observable interactions of hard-shelled turtles in the Mid-Atlantic scallop dredge fishery prior to the implementation of chain mats (1 January 2001 through 25 September 2006) was estimated to be 288 turtles (CV = 0.14, 95% CI: 209–363), which is equivalent to 49 adults. After implementation of chain mats, the average annual number of observable interactions was estimated to be 20 turtles (CV = 0.48, 95% CI: 3–42), equivalent to 4 adults. If the rate of observable interactions from dredges without chain mats had been applied to trips with chain mats, the estimated number of observable and inferred interactions of hard-shelled species after chain mats were implemented would have been 125 turtles per year (CV = 0.15, 95% CI: 88–163). Results from this analysis suggest that chain mats and fishing effort reductions contributed to the decline in estimated turtle interactions after 2006.