Kevin W. Brinck
Richard J. Camp
Chris Farmer
Sheldon M. Plentovich
Paul C. Banko
P. Marcos Gorresen
2016
<p><span>Both of the 2 passerines endemic to Nihoa Island, Hawai‘i, USA—the Nihoa Millerbird (</span><i>Acrocephalus familiaris kingi</i><span>) and Nihoa Finch (</span><i>Telespiza ultima</i><span>)—are listed as endangered by federal and state agencies. Their abundances have been estimated by irregularly implemented fixed-width strip-transect sampling from 1967 to 2012, from which area-based extrapolation of the raw counts produced highly variable abundance estimates for both species. To evaluate an alternative survey method and improve abundance estimates, we conducted variable-distance point-transect sampling between 2010 and 2014. We compared our results to those obtained from strip-transect samples. In addition, we applied state-space models to derive improved estimates of population size and trends from the legacy time series of strip-transect counts. Both species were fairly evenly distributed across Nihoa and occurred in all or nearly all available habitat. Population trends for Nihoa Millerbird were inconclusive because of high within-year variance. Trends for Nihoa Finch were positive, particularly since the early 1990s. Distance-based analysis of point-transect counts produced mean estimates of abundance similar to those from strip-transects but was generally more precise. However, both survey methods produced biologically unrealistic variability between years. State-space modeling of the long-term time series of abundances obtained from strip-transect counts effectively reduced uncertainty in both within- and between-year estimates of population size, and allowed short-term changes in abundance trajectories to be smoothed into a long-term trend.</span></p>
application/pdf
10.1650/CONDOR-15-214.1
en
Cooper Ornithological Club
State-space modeling of population sizes and trends in Nihoa Finch and Millerbird
article