Evaluation of autonomous recording units for detecting 3 species of secretive marsh birds

Abstract

Population status and habitat use of yellow rails (Coturnicops noveboracensis), Nelson's sparrows (Ammodramus nelsoni), and Le Conte's sparrows (A. leconteii) are poorly known, so standardized surveys of these species are needed to inform conservation planning and management. A protocol for monitoring secretive marsh birds exists; however, these species regularly call at night and may be missed during early morning surveys. We tested the effectiveness of autonomous recording units (hereafter, recording units) to survey these species by analyzing recorded vocalizations using bioacoustics software. We deployed 22 recording units at 54 sites in northern Minnesota and eastern North Dakota, USA, and conducted traditional broadcast surveys during May-June, 2010 and 2011. We compared detection probabilities between recording units and standard monitoring protocols using robust-design occupancy models. On average, recording units detected 0.59 (SE ¼ 0.11) fewer Le Conte's sparrows, 0.76 (SE ¼ 0.15) fewer Nelson's sparrows, and 1.01 (SE ¼ 0.14) fewer yellow rails per survey than were detected using the standard protocol. Detection probabilities using the standard protocol averaged 0.95 (yellow rail; 95% CI ¼ 0.86-0.98), 0.93 (Le Conte's sparrow; 95% CI ¼ 0.78-0.98), and 0.89 (Nelson's sparrow; 95% CI ¼ 0.56-0.98), but averaged 0.71 (yellow rail; 95% CI ¼ 0.56-0.83), 0.61 (Le Conte's sparrow; 95% CI ¼ 0.42-0.78), and 0.51 (Nelson's sparrow; 95% CI ¼ 0.19-0.82) using recording units. Reduced detection by recording units was likely due to the ability of human listeners to identify birds calling at greater distances. Recording units may be effective for surveying nocturnal secretive marsh birds if investigators correct for differential detectability. Reduced detectability may be outweighed by the increased spatial and temporal coverage feasible with recording units. Ó 2015 The Wildlife Society.

Use of autonomous recording units increased detection of a secretive marsh bird

“…) and others reporting them as either inferior to (Hutto and Stutzman , Sidie‐Slettedahl et al. ) or equal to human observers (Holmes et al. , Alquezar and Machado , Van Wilgenburg et al.…”

“…, Sidie‐Slettedahl et al. ). Compared to point counts, ARUs may also be better able to detect elusive or threatened species that seldom vocalize because they can be left in the field for long periods (Celis‐Murillo et al.…”

“…Despite these potential benefits, use of ARUs can still present challenges (Hutto and Stutzman , Sidie‐Slettedahl et al. ). For example, determining relative abundance of a species is difficult with an ARU, whereas the relative locations of vocalizations determined during point counts can be used to more accurately determine the number of individuals present at a site (Sidie‐Slettedahl et al.…”

“…For example, determining relative abundance of a species is difficult with an ARU, whereas the relative locations of vocalizations determined during point counts can be used to more accurately determine the number of individuals present at a site (Sidie‐Slettedahl et al. ). Direct comparisons of recording units to human observers have revealed mixed results, with some authors suggesting that detection rates were superior with ARUs (Acevedo and Villanueva‐Rivera , Celis‐Murillo et al.…”

“…), these programs can yield high rates of false positives and false negatives, making manual analysis a preferred alternative for some studies (Sidie‐Slettedahl et al. , Stiffler et al. ).…”