Grassland birds have experienced population declines greater than any bird group in the United States due to the conversion of grassland habitat to alternative uses, natural succession, and intensification of land management practices. Surveys of grassland bird populations are necessary to understand and promote grassland management practices tailored to encourage bird population recovery. Current survey methods require researchers to walk through survey areas to identify nests. Traditional surveys are inefficient, can inadvertently destroy grassland habitat, and can interfere with reproduction. Small unmanned aircraft systems (sUAS), may provide an efficient, less destructive alternative to current methods. Research by the Smithsonian Conservation Biology Institute (SCBI) and James Madison University X-labs hypothesize that a sUAS could be used to locate grassland bird nests with comparable accuracy and fewer resources than traditional field surveys. Thermal and digital cameras mounted to a sUAS were used to determine count and location of grassland bird nests. Because grassland birds do not begin nesting until mid-May, preliminary testing was done with hand warmers, which are approximately the size and temperature of active nests. Warmers were distributed throughout a field with approximately 75% vegetation coverage. The sUAS scanned the field and images were captured at varying heights and speeds.
Results/Conclusions
In the preliminary testing, the sUAS survey successfully identified hand warmers. Height had negligible effect on hand warmer identification, while flight speed above 5 m/s negatively affected image resolution. Testing with SCBI will continue into the summer to survey live grassland bird populations using the sUAS method. Machine learning will be utilized to automate identifying nests within full field images. Our goal is to compare sUAS survey performance with walk-through survey performance for time and identification efficiency.