Forest bird mortality and baiting practices in New Zealand aerial 1080 operations from 1986 to 2009

We collated 48 surveys of individually banded birds or birds fitted with radio transmitters that were checked before and after 1080 poison (sodium fluoroacetate) baits were aerially distributed to control brushtail possums (Trichosurus vulpecula) in New Zealand forests. The surveys were associated with 34 pest control operations from 1986 to 2009 and covered 13 native bird species, of which four were kiwi (Apteryx spp.). Sample sizes ranged from 1 to 46 birds (median 15). In 12 cases a sample of 1 to 42 birds (median 13) was surveyed in an untreated area at the same time. In total, 748 birds were checked before and after operations and 48 birds disappeared or were found dead. In non-treatment areas, 193 birds were checked and four died. Surveys of kiwi, whio (Hymenolaimus malacorhynchos), kaka (Nestor meridionalis) and kokako (Callaeas cinerea) were grouped for meta-analyses. The 95% pooled upper confidence bounds for the point estimate of zero mortality were each less than 4% for kiwi, kaka and kokako indicating only a small risk of mortality during 1080 pest control operations. Prefeeding with non-toxic baits increased from 22% (1998-1999) to 79% (2007-2008) in 322 operations on public conservation lands but was used in only 9 (26%) of the operations during which individually marked birds were monitored. We caution that failure to observe bird deaths in small samples may lead to weak inference about zero mortality across a population, most surveys in the review did not involve prefeeding, and that 11 native bird species for which deaths were reported after 1080 operations have not been studied.