Developing accurate survey methods for estimating population sizes and trends of the critically endangered Nihoa Millerbird and Nihoa Finch.

Abstract

This report describes the results of a comparative study of bird survey methods undertaken for the purpose of improving assessments of the conservation status for the two endemic passerines on the Island of Nihoa—Nihoa Millerbird (Sylviidae: Acrocephalus familiaris kingi ) and Nihoa Finch (Fringilidae: Telespiza ultima ; also referred herein as millerbird and finch)—both listed as endangered under the Federal Endangered Species Act (ESA) and Hawai`i Revised Statutes 195D. The current survey protocol, implemented since 1967, has produced a highly variable range of counts for both the millerbird and finch, making difficult assessments of population size and trend. This report details the analyses of bird survey data collected in 2010 and 2011 in which three survey methods were compared―strip-transect, line-transect, and point-transect sampling―and provides recommendations for improved survey methods and protocols. Funding for this research was provided through a Science Support Partnership grant sponsored jointly by the U.S. Geological Survey (USGS) and the U.S. Fish and Wildlife Service (USFWS). Point-transect surveys indicated that millerbirds were more abundant than shown by the striptransect method, and were estimated at 802 birds in 2010 (95%CI = 652 – 964) and 704 birds in 2011 (95%CI = 579 – 837). Point-transect surveys yielded population estimates with improved precision which will permit trends to be detected in shorter time periods and with greater statistical power than is available from strip-transect survey methods. Mean finch population estimates and associated uncertainty were not markedly different among the three survey methods, but the performance of models used to estimate density and population size are expected to improve as the data from additional surveys are incorporated. Using the pointtransect survey, the mean finch population size was estimated at 2,917 birds in 2010 (95%CI = 2,037 – 3,965) and 2,461 birds in 2011 (95%CI = 1,682 – 3,348). Preliminary testing of the line-transect method in 2011 showed that it would not generate sufficient detections to effectively model bird density, and consequently, relatively precise population size estimates. Both species were fairly evenly distributed across Nihoa and appear to occur in all or nearly all available habitat. The time expended and area traversed by observers was similar among survey methods; however, point-transect surveys do not require that observers walk a straight transect line, thereby allowing them to avoid culturally or biologically sensitive areas and minimize the adverse effects of recurrent travel to any particular area. In general, pointtransect surveys detect more birds than strip-survey methods, thereby improving precision and resulting population size and trend estimation. The method is also better suited for the steep and uneven terrain of Nihoa