Risk of commercial fisheries to New Zealand seabird populations

This report presents a risk assessment of the effect of fishing-related mortalities on 70 of the seabird species that breed in the New Zealand region. The assessment covers all fishing by commercial trawl, bottom-longline, surface-longline, and set-net fisheries within New Zealand’s Exclusive Economic Zone. The risk was defined as the ratio of the estimated annual number of fatalities of birds due to bycatch in fisheries to the Potential Biological Removal (PBR), which is an estimate of the number of seabirds that may be killed without causing the population to decline below half the carrying capacity. The risk ratio for each seabird species is an estimate of the ratio between the number of fatalities due to fisheries bycatch, and the PBR.

When estimating the PBR for each species, a recovery factor, f, should be specified. This recovery factor is typically between 0.1 and 0.5. It allows some protection against errors in estimating the PBR, and a lower value allows for a shorter recovery time for depleted populations. In this report, the risk ratio is calculated using PBR₁: the value of the PBR calculated with f = 1. The values of the recovery factor for each species will be specified at a later stage. When interpreting the risk ratio, it should be borne in mind that species with risk ratios less than one may still be considered at risk, depending on the value of the recovery factor.

Estimation of annual fatalities was based on seabird captures reported by observers between the 2006–07 and 2010–11 fishing years. From these captures, the total observable captures were estimated (the number of seabird captures that would have been reported had an observer been on every fishing vessel). Not all seabirds that are killed are brought on-board vessels, and the observable captures were multiplied by a factor to account for these cryptic mortalities. The cryptic multiplier varied between 1.30 and 8.66, depending on fishing method and seabird species.

This study replaces a previous seabird risk assessment that used a similar methodology. Demographic parameters and distributions were updated for some species, the methods used for estimating observable captures were changed, and calculation of the PBR was modified to include an additional calibration factor. Because of the number of methodological changes that were made, changes in risk between the two assessments should not be taken as being due to changes in the fisheries.

In total, there were 15 100 (95% c.i.: 13 600 – 16 600) estimated annual potential seabird fatalities across the four fishing methods. The highest number of annual potential fatalities were in trawl fisheries with 9870 (95% c.i.: 8560 – 11 300) estimated annual potential fatalities. There were a total of 3560 (95% c.i.: 3040 – 4150) annual potential fatalities in bottom-longline fisheries. Captures in surface-longline fisheries were lower, with a total of 1340 (95% c.i.: 1170 – 1570) annual potential fatalities of all seabirds. The estimated fatalities in set-net fisheries were relatively low, with a total of 317 (95% c.i.: 228 – 460) annual potential fatalities of all species. The estimates of these annual potential fatalities depend strongly on the assumptions that were made about the extent of the cryptic mortalities.

Six species had a median risk ratio above 1 or an upper 95% confidence limit above 2 and may be considered as at “Very high risk”: black petrel, Salvin’s albatross, flesh-footed shearwater, southern Buller’s albatross, Chatham Island albatross, and New Zealand white-capped albatross. The risk ratio of black petrel was especially large, with a median of 19.9 (95% c.i.: 11.4 – 32.8), due to the combination of a high number of estimated annual potential fatalities (mean 1440; 95% c.i.: 1070 – 1900), and a low PBR₁ (mean 74; 95% c.i.: 47 – 117). Fatalities of the species with high risk ratios were mainly in poorly observed small-vessel fisheries.

Four species had a median risk ratio above 0.3 or an upper 95% confidence limit above 1, and may be considered as at “High risk”: northern Buller’s albatross, Gibson’s albatross, Cape petrel, and Antipodean albatross. Nine species had a median risk ratio above 0.1 or an upper 95% confidence limit above 0.3 (“Medium risk”), and another seven had an upper 95% confidence limit above 0.1 (“Low risk”). Among the 70 considered species, 45 had the upper 95% confidence limit of their risk ratio below 0.1, suggesting that commercial fisheries in New Zealand waters are unlikely to significantly impact the demography of these species.

The risk assessment methodology is not yet mature, and further improvements may be made in the future. Possible improvements include a better specification of the cryptic mortality. This improvement would require collection of data on cryptic mortality in New Zealand fisheries. While these kinds of data are difficult to collect, poor knowledge of cryptic mortality restricts understanding of the impacts of fisheries on seabird populations. For some species, the assessment could be extended to include fatalities in global fisheries, and to include broader impacts, such as bycatch in recreational fisheries. The risk assessment requires adequate observer data, and in small-vessel fisheries observer coverage remains low. Increasing observer coverage in these fisheries would greatly reduce the uncertainty associated with the current estimates of risk.