Incidental capture of marine mammals in New Zealand trawl fisheries, 1995–96 to 2011–12

Citation

Thompson, F. N., Berkenbusch, K., & Abraham, E. R. (2016). Incidental capture of marine mammals in New Zealand trawl fisheries, 1995–96 to 2011–12. New Zealand Aquatic Environment and Biodiversity Report No. 167. Retrieved from http://mpi.govt.nz/document-vault/11947

Summary

A number of marine mammal species are incidentally captured in New Zealand commercial trawl fisheries each year, including common dolphin (Delphinus delphis), New Zealand fur seal (Arctocephalus forsteri), and New Zealand sea lion (Phocarctos hookeri). To monitor the extent of these incidental captures in New Zealand waters, government fisheries observers are placed on-board commercial fishing vessels to record the number and identity of protected species that are captured. As fisheries observers only cover a proportion of the overall fishing effort, assessments rely on estimation methods to derive the total number of incidental captures of protected species. These estimations include the development of statistical models that incorporate observer and fishing effort data. This report presents the most recent estimates of the total incidental captures of common dolphin, New Zealand fur seal and New Zealand sea lion in New Zealand trawl fisheries, including the 2011–12 fishing year.

Common dolphin is the most frequently captured cetacean species in New Zealand waters, with the majority of captures occurring in the mackerel trawl fishery on the North Island west coast. In the most recent fishing year included in this study, 2011–12, there were five observed common dolphin captures in this fishery, involving four separate capture events. Observer coverage was unusually high in this fishing year, with 79% of all tows being observed. In the 17-year reporting period between 1995–96 and 2011–12, there were 124 observed common dolphin captures.

Estimation of total common dolphin captures in the North Island mackerel trawl fishery involved a two-stage Bayesian hurdle model that also explored potential factors that may have contributed to the captures. Based on this model, the estimated total number of common dolphin captures in 2011–12 was 7 (95% c.i.: 5 to 14), which was a marked reduction from the previous estimate of 60 (95% c.i.: 24 to 113) common dolphin captures in 2010–11. The corresponding capture rates were an estimated 0.43 (95% c.i.: 0.30 to 0.85) common dolphins per 100 tows in 2011–12, compared with an estimated 3.86 (95% c.i.: 1.55 to 7.29) common dolphins per 100 tows in 2010–11. Of the covariates included in the model, headline depth (distance of the headline below the surface) best explained common dolphin captures. Observer data supported this finding, and the majority of observed captures occurred on trawl tows fished at headline depths less than 40 m. The model results suggest that increasing headline depth by 20 m would halve the capture event probability.

New Zealand fur seal are incidentally captured in a number of different trawl fisheries, involving a range of target species and fishing areas. In the 2011–12 fishing year, there were 82 observed fur seal captures in trawl fisheries in New Zealand waters. Over the entire 10-year reporting period from 2002–03 to 2011–12, 1008 incidental fur seal captures were recorded by fisheries observers. These observer data included inshore trawl fisheries (except for flatfish target fisheries) owing to recent increases in observer effort in these fisheries. Across the different trawl fisheries, observer coverage in 2011–12 was 10.8%, similar to that in the previous four years.

Capture estimates for 2011–12 were similar to estimates in the previous fishing year, with 442 (95% c.i.: 256 to 789) estimated fur seal captures and an estimated capture rate of 0.53 (95% c.i.: 0.30 to 0.94) fur seals per 100 tows, compared with 427 (95% c.i.: 246 to 743) fur seals and 0.50 (95% c.i.: 0.29 to 0.86) fur seals per 100 tows in 2010–11. Of the different target fisheries, hoki trawl fisheries contributed the greatest proportion of observed and estimated fur seal captures, with 200 (95% c.i.: 98 to 417) of the total 442 (95% c.i.: 256 to 789) estimated fur seal captures occurring in these fisheries. In addition to fishing areas and target fisheries, two other covariates were correlated with fur seal captures, these were distance from shore and day of year. The fur seal capture rate probability decreased with increasing distance from shore, whereas for day of year, it varied seasonally, with a distinct peak in August and September.

There was one observed New Zealand sea lion capture in trawl fisheries in the 2011–12 fishing year, on a tow targeting white warehou on the Stewart-Snares shelf. The sea lion was released alive. Observer effort in this fishing year was 42% across all trawl fisheries. The estimation of incidental sea lion captures across all trawl fisheries in 2011–12 resulted in a total of 13 (95% c.i.: 5 to 22) estimated sea lion captures, reflecting a decrease from the estimated 28 (95% c.i.: 17 to 41) sea lion captures in 2010–11.

The Auckland Islands squid fishery has been characterised by a significant number of New Zealand sea lion captures over time, with most observed captures between 1995–96 and 2011–12 occurring in this fishery. Nevertheless, there were no observed sea lion captures in this fishery in the two most recent fishing years, 2010–11 and 2011–12. Since 2004–05, this fishery has used a sea lion exclusion device (SLED) fitted to trawl nets to allow captured sea lion to exit the net. The use of SLEDs means that the number of sea lions that may have escaped the net and the post-escape survival of escapees are unknown. For this reason, the model used to estimate sea lion captures in the Auckland Islands squid fishery includes an estimate of the retention probability of sea lion on tows with SLEDs. The retention probability is subsequently used to estimate the number of interactions, which is the number of sea lions that would have been caught had no SLEDs been used. As the SLED design changed during the reporting period, the retention probability was estimated by combining two models that included either a single or a split SLED retention probability.

In the 2011–12 fishing year, there were an estimated 43 (95% c.i.: 2 to 206) sea lion interactions, with a corresponding strike rate of 3.3 (95% c.i.: 0.2 to 16.2) sea lion interactions per 100 tows. Both estimates were similar to those in 2010–11. At the same time, the large variation around the mean values highlights the uncertainty associated with these estimates. As the dataset becomes increasingly biased towards tows that used SLEDs and observed captures have decreased in recent years, it is becoming difficult to provide reliable estimate of the number of interactions and strike rate in the Auckland Islands squid fishery.