Marine mammal bycatch in New Zealand trawl fisheries, 1995–96 to 2010–11

Marine mammal species that are captured in New Zealand commercial trawl fisheries include common dolphin (Delphinus delphis), New Zealand fur seal (Arctocephalus forsteri), and New Zealand sea lion (Phocarctos hookeri). These incidental captures are recorded by fisheries observers when they are on- board vessels. For fisheries with sufficient observer coverage, these data, combined with fishing effort data, allow estimations of the total number of incidental captures via the development of statistical models. Here, we present estimates of the capture of common dolphin, New Zealand fur seal and New Zealand sea lion in New Zealand trawl fisheries, including the 2010—11 fishing year.

Common dolphin are frequently captured in the mackerel trawl fishery on the North Island west coast. In 2010—11, seven common dolphin were observed caught in six separate capture events in this fishery. Over the 16-year reporting period between 1995—96 and 2010—11, a total of 119 common dolphin were observed captured. A two-stage Bayesian hurdle model was built to estimate the total number of common dolphin captures and to identify covariates that were related to captures. The estimated total number of common dolphin captures in the most recent fishing year was 64 (95% c.i.: 26 to 116), more than twice the estimate of the previous year, 30 (95% c.i.: 7 to 68). The high capture estimate resulted from the influence of the observed event rate, which increased in the 2010—11 fishing year. Headline depth (distance of the headline below the surface) was confirmed as the covariate that best explained common dolphin captures, and this finding was supported by observer data, with the majority of observed captures occurring on tows with headline depths less than 40 m. The model results suggest that increasing headline depth by 21 m would halve the capture event probability.

Fur seal are captured in trawl fisheries encompassing a range of fishing areas and target species. In 2010—11, inshore fisheries were included in the fur seal bycatch assessment for the first time, following a recent increase in observer effort. In this fishing year, there were a total of 69 fur seal captures observed in trawl fisheries in New Zealand waters. In the 9-year period between 2002—03 and 2010—11, a total of 922 fur seal were recorded as bycatch by fisheries observers. The highest observed capture rate in 2010—11 was in southern blue whiting fisheries (8.33 fur seal captures per 100 tows), followed by ling fisheries (1.96 observed fur seal captures per 100 tows). The total number of estimated captures derived from Bayesian models was 376 (95% c.i.: 221 to 668) in the 2010—11 fishing year. This estimate is a decrease from the total 472 (95% c.i.: 269 to 914) estimated captures in 2009—10. The two covariates distance from shore and day of year were correlated with fur seal captures. The fur seal capture rate decreased with increasing distance from shore. At the same time, there was strong seasonal variation in fur seal captures, with a peak in August, and relatively high captures in July and September.

New Zealand sea lion have historically been frequently caught in the squid trawl fishery around Auckland Islands, and most observed sea lion captures between 1995—96 and 2010—11 have been in this fishery; however, in 2010—11, there were no observed captures in this fishery. In contrast, six male sea lion were observed captured in the Campbell Island southern blue whiting fishery, with one released alive and two caught in one net. Capture estimates for this fishing year revealed a total of 29 (95% c.i.: 17 to 43) sea lion captures in all trawl fisheries. This estimate is a decrease from 46 (95% c.i.: 32 to 66) captures in 2009—10.

Owing to the high number of incidental sea lion captures in the Auckland Islands squid fishery, this fishery uses sea lion exclusion devices (SLEDs) that are fitted to the trawl nets to allow sea lion to escape the net. On tows using SLEDs, the number and fate of sea lion that may have escaped the net are unknown. To account for this uncertainty, the Auckland Islands squid fishery model includes an estimate of the retention probability of sea lion on tows with SLEDs. The retention probability is then used to estimate the number of interactions, that is, the number of sea lion that would have been caught had no SLEDs been used. As the SLED design changed during the reporting period, estimates of the retention probability were derived by combining two models that included either a single or a split SLED retention probability.

The resulting number of estimated interactions in 2010—11 was 56 (95% c.i.: 4 to 233), with a corresponding strike rate of 3.5 (95% c.i.: 0.4 to 14.9) sea lion interactions per 100 tows. Both estimates were lower than those in 2009—10, but the large variation around the mean values highlights the uncertainty associated with these estimates. In addition to the dataset becoming more biased toward tows that used SLEDs, the decrease in observed captures in recent years makes it increasingly difficult to estimate the number of interactions and strike rate in the Auckland Islands squid fishery.