Techniques for monitoring Brachyramphus murrelets: A comparison of radar, autonomous acoustic recording and audio-visual surveys

Abstract

Conditions in Alaska, USA, pose a challenge for monitoring populations of Brachyramphus murrelets using standard survey methods, because of strong winds, 2 sympatric species, short nights, and variable nesting habitat. We tested 3 methods for monitoring Brachyramphus murrelets breeding in the Kodiak Archipelago, Alaska, in 2010-2012. In addition to standard audio-visual and radar methods, we tested - for the first time with murrelets in Alaska - the application of autonomous acoustic recorders for monitoring vocal activity. We completed 74 radar, 124 audio-visual, and 134 autonomous acoustic surveys, focused on presunrise activity peaks; this yielded 26,375 murrelet detections. Marbled (B. marmoratus) and Kittlitz's murrelets (B. brevirostris) could not be distinguished using combinations of radar and acoustic recordings; therefore, at-sea surveys will be required to determine localized species proportions. Of the 3 methods, radar sampled the largest area and detected silently flying murrelets, providing the most reliable data on local populations; however, radar identification of murrelets was unreliable in winds exceeding 18 km/hr. Audio-visual surveys were useful for species identification and to document behaviors associated with local nesting, whereas autonomous acoustic recorders allowed season-long monitoring of murrelet vocal activity. Within potential forest-nesting habitat of marbled murrelets, all 3 methods gave similar measures of presunrise murrelet activity, but only radar reliably sampled murrelets commuting between nest and ocean. Because of their low cost and flexible programming, automated sound recorders offer an affordable way to sample vocal activity prior to more intensive or expensive radar and audio-visual surveys. We recommend that population monitoring and habitat studies of Brachyramphus murrelets in Alaska include combinations of all 3 methods. Published 2016. This article is a U.S. Government work and is in the public domain in the USA.