Estimated capture of seabirds in New Zealand trawl and longline fisheries, 2002–03 to 2014–15

In New Zealand, captures of seabirds and other protected species are recorded by government observers when they are on-board commercial fishing vessels. This report addresses the question of how many seabirds would be reported caught if every commercial trawl and longline vessel fishing within New Zealand’s Exclusive Economic Zone carried an observer. Statistical models were used to scale up from observed captures, to estimate total captures across all commercial trawl and longline fisheries. Estimates were made for the 2002–03 to 2014–15 fishing years for trawl fisheries, and for the 1998–99 to 2014–15 fishing years for longline fisheries.

A consistent modelling framework was used to estimate the captures for ten species (and species groups): New Zealand white-capped albatross (Thalassarche steadi), Salvin’s albatross (Thalassarche salvini), Buller’s albatrosses (Thalassarche bulleri, combining both southern T. b. bulleri and northern T. b. platei subspecies), white-chinned petrel (Procellaria aequinoctialis), black petrel (Procellaria parkinsoni), grey petrel (Procellaria cinerea), sooty shearwater (Puffinus griseus), and flesh-footed shearwater (Puffinus carneipes). Estimates were also derived for seabirds grouped as either “other albatrosses” or “other birds”. A hierarchical mixed-effects generalised linear model (GLM) was used, fitted using Bayesian methods. For each model, the data were grouped by fishing method, target fishery, vessel class, spatial area, fishing year, and quarter of the year, and the capture rate was estimated within each of these strata from the observed captures. This capture rate was then applied to unobserved fishing effort to estimate total captures.

We estimated that a total of 4584 (95% c.i.: 3935–5477) seabirds were captured in trawl and longline fisheries during the 2014–15 fishing year, where the credible interval (c.i.) is the 95th quantile range of the posterior distribution. The total estimate included 2095 (95% c.i.: 1828–2414) seabirds captured in trawl fisheries, 1927 (95% c.i.: 1404–2746) seabirds captured in bottom-longline fisheries, and 561 (95% c.i.: 386–838) seabirds captured in surface-longline fisheries.

Among the seabird species, the highest number of total estimated captures duing the 2014–15 fishing year was of white-chinned petrel with 868 (95% c.i.: 578–1517) captures, followed by flesh-footed shearwater and Salvin’s albatross, with 575 (95% c.i.: 357–958) and 559 (95% c.i.: 386–818) captures, respectively. Other estimates included 416 (95% c.i.: 294–589) captures of New Zealand white-capped albatross, 307 (95% c.i.: 171–558) captures of black petrel, 439 (95% c.i.: 321–641) captures of sooty shearwater, 260 (95% c.i.: 183–381) captures of Buller’s albatrosses, and 239 (95% c.i.: 98–571) captures of grey petrel. In addition to the species estimates, there were 655 (95% c.i.: 475–912) and 561 (95% c.i.: 386–838) estimated captures for other birds and other albatrosses, respectively.

For seven of the ten modelled species groups, the total number of estimated captures decreased between 2002–03 and 2014–15. This decrease was primarily related to decreases in fishing effort over this period. Only Salvin’s albatross, white-chinned petrel and grey petrel did not show a clear decrease in total captures over this time period, and it was only for white-chinned petrel that the mean estimated captures in 2014–15 were higher than the mean estimated captures in 2002–03. In the large-vessel fisheries with sufficient captures to evaluate trends, there were no clear decreases in the capture rate of either albatrosses or petrels and other birds over this period. In large-vessel squid trawl fisheries, there was an increase in the capture rate of petrels and other birds over the reporting period, and in hoki trawl fisheries there has been an increase in the capture rate of albatrosses since 2006–07.

In many fisheries, it appears from the observer data that it would be possible to improve the mitigation of seabird captures. For example, streamer lines could be used by all small-vessel surface-longline fisheries, and the use of integrated weight line in large-vessel bottom-longline fisheries could be increased. However, improvements to data collection by observers are necessary to understand the use of mitigation measures across New Zealand’s fisheries.