Abstract
Bird species richness can be a helpful indicator of a community’s biodiversity and health. The G. H. Gordon Biological Station has no record of species richness since 2014 and a baseline profile is necessary to assess changes in environmental health and to create monitoring systems to conserve biodiversity. This project updated species richness records at the G. H. Gordon Biological Station using automated recording devices (ARDs) to record bird vocalizations throughout this property and identified these vocalizations using BirdNET, a convolutional neural network created and trained by Cornell Bird Labs. These methods successfully identified 96 species present in lake, dense forest, prairie, stream, and wetland habitats. In comparison to previous surveys, 34 species were not shared. This discrepancy may have occurred because of the different methods used between past surveys and this project, or it may also indicate a change in bird species richness at the G. H. Gordon Biological Station. These methods proved to be an extremely useful tool for undergraduate research, as not a lot of experience or training was required to produce results. This study demonstrated how the use of technology allows undergraduates to participate in valid and useful experiments that produce reliable data and results.