19th Annual Student Research Conference
Interdisciplinary
Quantitative Identification of Northeastern Missouri Bats via Acoustic and Standard Surveys
Rachel O. Van Amburg* and Joshua B. Kelly
Dr. Scott Burt and Dr. Jason E. Miller, Faculty Mentors
To understand and improve the effectiveness of quantitative methods of bat identification based on recorded echolocation calls, a library of representative echolocation calls was first assembled. A total of 189 known call sequences for seven species and several hundred unknown call sequences were collected during Summer 2005 through acoustic samplings of bat populations from seven locations in Northeastern Missouri. All recorded calls were cleaned using a set of standard protocols and call parameters were extracted using Analook software. This data was used to create a classification method based on discriminant function analysis. Preliminary analysis shows that the summary statistics for the known calls were found to differ significantly from those collected by other investigators in the region as well as those in the literature. In addition, classification rates achieved using discriminant functions were compared to those achieved using classification trees.
Keywords: Bats, Echolocation, Classification, Discrimin. Function, Analook
Topic(s):Mathematical Biology
Presentation Type: Oral Paper
Session: 13-1
Location: OP 2210
Time: 8:15