Graduation Year

2023

Graduation Month

December

Document Type

Thesis

Degree Name

Bachelor of Science

School or Department

Wildlife Biology

Major

Wildlife Biology – Terrestrial

Faculty Mentor Department

Wildlife Biology

Faculty Mentor

Victoria Dreitz

Faculty Reader(s)

Victoria Dreitz, Zachary Cheviron, Michael Musick

Keywords

false-positives, survey simulation, error rates, songbirds, multispecies abundance surveys, identification cues

Subject Categories

Biology | Life Sciences | Ornithology | Research Methods in Life Sciences

Abstract

Errors in wildlife field data threaten to bias resulting abundance and occupancy estimates if not properly accounted for or minimized. Methods to account for false-positive errors in wildlife data have not been as thoroughly developed as those for false-negative errors despite false-positives being present across diverse wildlife taxa and study systems. The calibration method to account for false-positives involves assessing the field detection method to determine how often false-positive errors occur in the field data. Rates can then be incorporated into estimations based on the field data to improve estimation accuracy. This study presents an application of the calibration approach for multispecies avian abundance surveys of seven songbird species (Brewer’s Sparrow, Chestnut-collared Longspur, Horned Lark, Long-billed Curlew, Thick-billed Longspur, Vesper Sparrow, Western Meadowlark) in the sagebrush steppe and grassland ecosystem of eastern Montana. Completion of simulated avian surveys resulted in estimation of species-specific false-positive rates as well as examination of how rates may change with the availability of different identification cues. Visual identification cues (video of birds) were always available but auditory identification cues (bird vocalizations) were not always present. Approximately 15% of focal species identifications were false-positives (SD= 0.36). False-positive rates varied significantly between 15 out of 21 focal species pairs, ranging from rates of 0.003 to 0.402 (SD=0.054, SD=0.49). The availability of bird vocalizations in tandem with bird visuals did not differ significantly from false-positive rates based on visuals only (p=2e-16). These results suggest that among these species and study system false-positive rates are primarily a product of similarities in species morphology rather than vocalizations.

Honors College Research Project

Yes

GLI Capstone Project

no

Download

Share

COinS
 

© Copyright 2023 Amelia K. Evavold