American Samoa Study Shows Effects of Removing Longline Hooks Close to Floats

One of the ways to reduce incidental capture of sea turtles in pelagic longline gear is to remove hooks adjacent to floats, which fish in the shallower depths were turtle are most likely to occur. A side effect of removing the shallower hooks could be a change in the catch rates (CPUE) of target fish species. To study this problem, PIFSC scientist Keith Bigelow his colleague Eric Fletcher of the University of Hawaii Joint Institute for Marine and Atmospheric Research (JIMAR) conducted a field experiment in the American Samoa longline fishery. They used time-depth recorders (TDRs) to measure hook depths in the fishery and determined the frequency distribution of "hook-at-capture" - the tendency of fish to be caught on a given hook in relation to the hook’s proximity to the nearest longline float. The study involved TDRs deployed by vessels carrying scientific observers. It was designed to provide data on the efficacy of removing hooks adjacent to floats as a way to reduce turtle interactions and to estimate corresponding changes in CPUE of target fish species as well as incidental and bycatch species. The full results of the study are contained in a PIFSC Internal Report IR-09-008 issued in March 2009. The work was funded in part by the Fisheries Disaster Relief Program and JIMAR's Pelagic Fisheries Research Program.

Frequency distribution of the settled longline depth (m), or depth of the deepest hook, as determined by time-depth recorders on 320 observed longline sets in the American Samoa longline fishery.

Thirty longline trips were observed from April 2006 to December 2008. TDR data were obtained on 320 sets from 12 trips. The TDR sample data provide an estimate of the depth of the deepest longline hook. The depth of shallower hooks was estimated by interpolation. The sample frequency distribution of the depth of the deepest settled hook was computed, and shows that the longline fishery in American Samoa typically fishes shallower (deepest hook at an average of 217 m) than the Hawaii-based tuna fishery (deepest hook at an average of 248 m). The American Samoa fishery targets albacore, which typically occupy shallower depths than the bigeye tuna targeted by the Hawaii fleet. Floatlines are longer (26 m) and branchlines shorter (10 m) in the American Samoa fishery than in the Hawaii-based fishery.

Based on bootstrap resampling of the data, 95% confidence intervals of mean depth of the first (i.e., closest), second (next closest) and third hooks adjacent to the longline float were 44−63 m, 55−85 m and 64−108 m, respectively. The percentages of hooks in various depth strata that would be expected to result from removing hook #1, hooks #1-2 and hooks #1-3 adjacent to the float were calculated. These indicated that removal of hook #1, i.e., hooks closest to the float, would effectively eliminate hooks from the upper 50 m of the water column. Removal of hooks #1-3 would effectively reduce effort from the surface down to 75 m depth to 0.3% of the total hooks deployed, although 5.6% of the total hooks would remain between the surface and 100 m.

A catch analysis was conducted using data for 488 longline sets from 16 observed trips by the American Samoa fleet which had been cleared through data quality control procedures by the NOAA Fisheries Pacific Islands Regional Office. During these trips, observers monitored 1,396,009 hooks and recorded hook position for each animal caught, with hooks numbered sequentially between each pair of adjacent floats. The resulting CPUE by hook number is illustrated for four fish species in the accompanying graphic (CPUE results for 19 species are illustrated in the PIFSC Internal Report). The shallowest hooks, those nearest the longline floats, had substantially higher CPUE for mahimahi, skipjack, wahoo, barracuda and shortbill spearfish than any deeper hooks. The deepest hooks had higher CPUE for bigeye tuna, sickle pomfret and longfin escolar.

Six scenarios based on catch rate by hook number were developed to portray the change in catches expected to occur by removing hooks adjacent to longline floats. Three scenarios considered the removal of hook #1, hooks #1-2 and hooks #1-3 adjacent to the float. Removal of hooks was modeled for both sides of each float; thus the scenario for removal of hooks #1-3 meant removal of six hooks. Three additional scenarios were explored in which the same sequence of hooks was removed, but the removed hooks were considered to be redeployed deeper in the water column by extending the total length of the gear (as if additional mainline and floats were deployed to carry these hook in the deeper pattern).

In the first three removal scenarios, the largest percentage catch reductions were for mahimahi, barracuda and shortbill spearfish. Target albacore catches declined 2%, 5% and 10%, respectively, by removing hook #1, hooks #1-2 or hooks #1-3 adjacent to the float. Total catches of other landed fish species declined by higher percentages (5%, 11% and 17%), mainly due to lower CPUE of wahoo and mahimahi. In the second set of scenarios, the redistribution of hook #1, hooks #1-2 and hooks #1-3 to greater depths increased target albacore catches by 4%, 8% and 11%, respectively.

Variation of catch rate (CPUE) with proximity of the longline hook to the nearest float (hook number) on 488 sets in the American Samoa longline fishery. Hook # 1 is closest to the float and hook # 18 is assumed to be the deepest hook. Wahoo and skipjack are more likely to be caught nearer the sea surface (closer to the float), whereas yellowfin tuna and albacore are more likely to be caught at greater depths in the water column (farther from the float).