Spatial and Temporal Variation in the Singing Rates of Two Forest Songbirds, the Ovenbird and the Black-throated Blue Warbler: Implications for Aural Counts of Songbirds.

Abstract

The most common method of surveying birds is the point count. There are two approaches to analyzing count data to produce species abundance estimates and monitor abundance changes through space and time. The first method uses count data as a direct index to bird abundance. Uncorrected index counts are currently the norm in ornithological research. The second method estimates probability of detection during each count so count statistics can be corrected for spatial and temporal variation. Probability of detection can be thought of as being composed of two components: an availability component and a detection given availability component. I designed an observational field study to examine the factors affecting variation in singing rates for two species of warblers over the breeding season, the breeding cycle, and in relation to territory density. I was also interested in determining the relative importance of the availability component of probability of detection during the detection process. I conducted field work in the French Broad Ranger District of Pisgah National Forest in the southern Appalachian Mountains of North Carolina. The Ovenbird and the Black-throated Blue Warbler were my study species. Work involved identifying individual males, tracking them, and recording samples of their singing behavior. To simplify rate recording, we developed a singing rate sampling program designed to be operated on a palm size computer (pda) that we could take into the field. We sampled at points during 2004, but moved sampling to plots in 2005. We searched for nests and monitored those found to fledging or failure, collecting singing samples for males associated with each nest. We performed variable circular plot, time-of-detection point counts on a weekly basis from the center of each plot to produce estimates of species density in 2005. We also spot mapped each plot to produce secondary territory density estimates. I was able to calculate true availabilities for both species across a range of count durations using program "R". Analysis of data indicated that average singing rates for both species declined significantly over the course of the breeding season. Ovenbird availability declined throughout the season while Black-throated Blue Warbler availability remained fairly constant. My work has confirmed the importance of considering both availability and detection given availability in any auditory sampling protocol.

Two methods of detection are available to us during a count: eyesight and hearing. We rely almost exclusively on bird vocalizations, especially songs, for detecting and identifying species during counts in forested habitats. I designed an experiment to quantify the relative importance of visual versus aural detections of birds in forested habitats and conducted it on 15 May 2006 in Umstead State Park, near Raleigh, North Carolina. The vegetation structure in the park is similar to that found at the Pisgah National Forest field sites where I collected singing rate data. Three teams of three observers each performed simultaneous 3-minute variable circular plot point counts at points along a park trail. In each team, one team member was blinded to all visual input, one was deafened to all auditory input, and the last team member had no handicap. We randomized detection methods among team members at each point. Non-handicapped observers detected 79% of the total number of birds detected at each point. Blinded observers detected 77%, and deafened observers detected only 3.5%. The auditory and visual detection processes are quite different from each other and can often compete for an observer's attention during a count. Attempting to use both simultaneously reduces the effectiveness of each, but by separating auditory counts from visual counts we reduce count complexity and improve observer effectiveness.