Customizing Deep Neural Networks for Call-Type Distinction in Eastern Meadowlark (Sturnella magna): Investigating links with Land Use, Behavioral Ecology, and Conservation
| dc.contributor.author | Cheeley, Lucy | |
| dc.date.accessioned | 2025-05-30T19:55:56Z | |
| dc.date.available | 2025-05-30T19:55:56Z | |
| dc.date.issued | 2025-04 | |
| dc.description.abstract | Grassland birds are experiencing steep population declines across North America due to habitat loss, agricultural intensification, and land-use change. The Eastern Meadowlark (Sturnella magna) has declined by over 75% since 1966 and serves as an important indicator of grassland ecosystem health. We utilized passive acoustic monitoring (PAM) and a custom BirdNET classifier to examine the spatial and temporal variation in Eastern Meadowlark vocalizations, songs, and calls across 45 grassland sites in New York State. We analyzed how vocalization frequency varied across management regimes, land-use types, regions, and the breeding season. A total of 143,305 songs and 11,489 calls were detected and analyzed using generalized additive models (GAMs). Results showed that song was significantly influenced by geographic region and date, and call was significantly influenced by date. Song rates increased in later parts of the season and were significantly higher in the Eastern region, while call frequency remained relatively stable with a slight increase at the end of the season. No significant effects were found for land cover, management regime, or protected versus working land status. These results show how PAM and call-type classification can provide insights into the differential roles of vocalization types in avian behavioral ecology. Future work should explore multi-year trends, additional sites, and juvenile versus adult vocalization patterns, and additional local and regional landscape variables. | |
| dc.description.sponsorship | USGS New York Cooperative Fish and Wildlife Research Unit and New York State Department of Environmental Conservation | |
| dc.identifier.uri | https://hdl.handle.net/1813/116994 | |
| dc.language.iso | en_US | |
| dc.subject | Sturnella magna | |
| dc.subject | Bioacoustics | |
| dc.subject | Conservation | |
| dc.title | Customizing Deep Neural Networks for Call-Type Distinction in Eastern Meadowlark (Sturnella magna): Investigating links with Land Use, Behavioral Ecology, and Conservation | |
| dc.type | dissertation or thesis |
Files
Original bundle
1 - 1 of 1
Loading...
- Name:
- L.Cheeley Honors Thesis.pdf
- Size:
- 825.35 KB
- Format:
- Adobe Portable Document Format