Summary of the capture of seabirds in New Zealand commercial fisheries, 2002–03 to 2013–14

Fisheries worldwide interact with non-target and protected species, including interactions that result in mortality. In commercial fisheries in New Zealand’s Exclusive Economic Zone, government fisheries observers document the incidental captures of protected species on-board fishing vessels. These independently recorded data, together with records of fishing effort, provide the basis for the development of statistical models that are used to estimate the total number of incidental captures in commercial fisheries. The present study provides an update of previous assessments by including data from the 2013—14 fishing year to estimate the total number of seabird captures in commercial trawl, surface-longline and bottom-longline fisheries in New Zealand waters. In addition to the data update, the present study updated the statistical models by applying a unified framework that allowed direct comparisons across species groups and fisheries.

Based on the unified modelling framework, the present study estimated that a total of 5075 (95% c.i.: 4547—5726) seabirds were incidentally captured in commercial fisheries during the 2013—14 fishing year. This total estimate included 2277 (95% c.i.: 2041—2542) seabirds in trawl fisheries, 2137 (95% c.i.: 1722—2745) seabirds in bottom-longline fisheries, and 659 (95% c.i.: 523—835) seabirds in surface-longline fisheries.

There were ten species or species groups distinguished in the modelling framework. For individual species, the highest number of total estimated captures was of white-chinned petrel (Procellaria aequinoctialis) with 653 (95% c.i.: 414—1131) captures, followed closely by flesh-footed shearwater (Puffinus carneipes) and Salvin’s albatross (Thalassarche salvini), with 637 (95% c.i.: 495—822) and 623 (95% c.i.: 462—884) captures, respectively. Other estimates included 490 (95% c.i.: 380—627) estimated captures of New Zealand white-capped albatross (Thalassarche steadi), 392 (95% c.i.: 276—552) captures of black petrel (Procellaria parkinsoni), 385 (95% c.i.: 306—488) captures of sooty shearwater (Puffinus griseus), 298 (95% c.i.: 224—398) captures of Buller’s albatrosses (Thalassarche bulleri, combining both southern T. b. bulleri and northern T. b. platei subspecies), and 207 (95% c.i.: 114—354) captures of grey petrel (Procellaria cinerea). For the two seabird groupings, there were 1022 (95% c.i.: 820—1269) and 659 (95% c.i.: 523—835) estimated captures for other birds and other albatrosses, respectively.

This study used a unified modelling approach, with the same model structure for each of the ten modelled species groups. We  recommend that this same structure is used for future modelling, as it can be applied relatively straightforwardly to new data. There were some convergence issues with some of the models, however, and we recommend increasing the length of the chains in the statistical modelling to improve the accuracy of the estimates.

A comparison between the results from this study and the estimates of observable captures from the seabird risk estimation highlights some differences. Most notably, estimates of captures of flesh-footed shearwater and grey petrel were higher when estimated using this method than when estimated in the risk assessment. We recommend that, for the seven species that are the same between both models, the capture estimates made using this approach are used as the observable captures in the risk assessment. We also recommend that the uncertainty in the observable captures in the risk assessment is inflated to account for the differences in the estimates between the two approaches. These differences indicate that there is a structural uncertainty in the risk-assessment modelling that is not currently being accounted for.

Total estimated captures of seabirds in all trawl and longline fisheries fell from from 9185 (95% c.i: 8390—10096) in 2002—03 to 5075 (95% c.i.: 4547—5726) in 2013—14. Over this period the direct impact of New Zealand fisheries on seabirds, as measured by the mean number of estimated annual seabird captures, has decreased by 45%. This decrease is associated with both decreases in fishing effort and in changes in seabird capture rates. While there was a decrease in total seabird captures over the twelve years, however, there has been no change in the total estimated seabird captures since 2008—09.

Estimated captures were highest in small-vessel fisheries, which were typically poorly observed. The fishing-method-vessel-size group with the highest total number of seabird captures was small-vessel bottom longline, with estimated captures of 1612 (95% c.i.: 1331—1953) seabirds. Because of low observer coverage in small-vessel trawl and bottom-longline fisheries (in many years, less than 2%), the estimates were uncertain, and the model was unable to estimate trends in estimated captures in these fisheries (other than from changes in fishing effort). In order to better understand and manage seabird bycatch in New Zealand fisheries, it is important to increase observer coverage in small-vessel trawl and bottom-longline fisheries.