2011 Kiwikiu (Maui Parrotbill) and Maui 'Alauahio abundance estimates and the effect of sampling effort on power to detect a trend
The Kiwikiu (Pseudonestor xanthophrys), also called the Maui Parrotbill, is an endangered, forest bird found only in high elevation, wet forest of the eastern portion of Maui Island. Recent surveys, conducted at five year intervals, have revealed wide variation in abundance estimates (Camp et al. 2009). Effective management and conservation requires accurate estimates of abundance, which is difficult for rare species such as the Kiwikiu because low density leads to few observations, resulting in low sample size and high uncertainty in abundance estimates. In addition to being rare, they occur in remote, difficult to access terrain, which makes them difficult to detect and further reduces the accuracy of counts.
The Maui `Alauahio (Paroreomyza montana), sometimes called the Maui Creeper, historically occupied the entire island of Maui (Gorresen et al. 2009). It has since been extirpated from much of its original habitat and now occurs in forested areas of East Maui where its habitat overlaps with that of the Kiwikiu. Though they share the same habitat, the `Alauahio is much more abundant—by more than two orders of magnitude—and occurs over a wider range than the Kiwikiu.
Both species appear to have no statistically significant population trend from 1980–2001, but abundance estimates vary widely from survey to survey and have wide uncertainties (Camp et al. 2009). Ideally survey design should result in estimates precise enough to be able to detect significant declines in abundance that may trigger management intervention.
We wished to improve the accuracy of Kiwikiu abundance estimates and the ability to detect significant trends in abundance. To that end, in 2011, repeated point count surveys were conducted across the Kiwikiu range, excluding Haleakalā National Park (Figure 1). The increased sampling effort increases sample size and improves the precision of estimates, and repeat samples also allowed us to partition within-year and between-year variation in surveys, increasing the statistical power to detect trends.