Document Type

Thesis - Open Access

Award Date


Degree Name

Master of Science (MS)

Department / School

Natural Resource Management

First Advisor

Kent C. Jensen

Second Advisor

Kinchel D. Doerner


yellow rail, sparrows, bird populations


Population status and habitat use of yellow rails (Coturnicops noveboracensis) (YERA), Nelson’s sparrows (Ammodramus nelsoni) (NESP), and Le Conte’s sparrows (Ammodramus leconteii) (LCSP) are poorly known, so systematic surveys of these elusive species are needed to inform conservation planning and guide management. A standardized protocol for monitoring secretive marsh birds exists (Conway 2009, 2011); however, these species call at night and may be missed during early-morning marsh bird surveys. I tested the effectiveness of autonomous recording units (ARUs) to survey these species by analyzing recorded vocalizations using bioacoustics software. I deployed 22 ARUs at 54 sites in northern Minnesota and eastern North Dakota, and conducted concurrent traditional broadcast surveys during May-June, 2010 and 2011. I compared ARU-based detections to the standard marsh bird monitoring protocol using a paired t-test, and used the robust design occupancy model in program MARK to estimate detection probabilities for each species by survey method. I found, on average, that ARUs detected 0.59 (LCSP), 0.76 (NESP), and 1.01 (YERA) fewer individuals per survey than were detected using the standard protocol. Detection probabilities using ARUs were on average 0.23 (YERA), 0.32 (LCSP), and 0.39 (NESP) lower than the standard protocol. Reduced detection by ARUs was likely due to the ability of human observers to detect birds at greater distances. ARUs may provide an effective means of surveying nocturnal secretive marsh birds if investigators correct for differential detectability from ground-based surveys. Arguably, reduced detectability may be outweighed by the increased spatial and temporal coverage feasible with ARUs, resulting in more cumulative opportunities for detection. I also identified factors that affect detection probability of YERA, LCSP, and NESP by manually scanning 3,035 three-minute ARU recordings for the presence/absence of these species. I related calling activity to hourly weather data from area weather stations. I used the generalized linear mixed models package in R, setting survey site as an a priori random effect to control for site to site variation, and using year, Julian day, precipitation, temperature, wind speed, atmospheric pressure, moonlight, and hours after sunset as fixed effects. The best supported model for YERA indicated highest detection probability occurred 4 hours after sunset, early in the field season, with lower wind speeds and no precipitation. The best supported model of LCSP detection indicated that this species has low detection through the night, but with highest detection at 8 to 10 hours after sunset, early in the field season, with lower wind speeds, and with increased moonlight. The best supported model for NESP indicated highest detection during periods with low wind speeds, no precipitation, and brighter moonlight; with peak dates for detection being from late May through the first two weeks of June. I recommend that existing standard protocols, or protocols developed for using ARUs to monitor these species, should incorporate these results to provide more reliable information on estimates of population trends, habitat uses, and distributions of these species of concern.

Library of Congress Subject Headings

Yellow rail -- Monitoring -- Methodology -- Evaluation
Sparrows -- Monitoring -- Methodology -- Evaluation
Bird Populations


Includes bibliographical references



Number of Pages



South Dakota State University


In Copyright - Non-Commercial Use Permitted