Estimating White-Tailed deer population sizes using unmanned aerial vehicles

To successfully manage wildlife populations, estimates of population sizes, adult sex ratios, and recruitment are essential. Aerial surveys are often used to conduct surveys for large mammals; however, they can be expensive, hazardous, and not always practical for small properties. Camera surveys and spotlight counts are alternative survey methods but are labor intensive and generally less accurate. Unmanned aerial vehicles (UAVs) are a quickly growing technology that have yet to be fully evaluated for wildlife population surveys. I evaluated preliminary UAV surveys conducted by contractors on 4 study sites during October 2018 – February 2019 using white-tailed deer (Odocoileus virginianus) as the focal species to assess detection and identification using thermal infrared and red green blue (RGB) optical sensors. Heat signatures were detected on thermal imagery and then species identification was confirmed via RGB optical imagery. I found deer and exotic ungulates were able to be detected in the thermal imagery, however, imagery quality was variable, and identification was difficult. Furthermore, UAV flight specifications lacked consistency, which contributed to the poor image quality. I conducted test flights at a 5th study site and developed a protocol for use in South Texas in order to enhance imagery quality and consistency and address issues identified in preliminary surveys. Finally, I applied this protocol to surveys I conducted across all 5 study sites in February – April 2020. Surveys were repeated ≥ 2 times at each site to evaluate variation between surveys. Distance sampling analyses were applied to the raw count data to develop population estimate sizes. UAV survey results were then compared to estimates from helicopter (n = 4 sites), spotlight (n = 1 site), and camera surveys (n = 3 sites). I found UAV surveys to be a viable method of surveying white-tailed deer in South Texas and the use of distance sampling methodology is recommended over the use of raw count data. Estimates from UAV surveys were comparable to helicopter, spotlight, and camera survey estimates. Additionally, UAV survey estimates were consistent across repeated surveys. Through the use of distance sampling, a number of factors that impact detection probability were revealed including hot spots, hilly terrain, and a high percentage of herbaceous and bare ground. Additionally, determining sex identification from the optical imagery proved challenging and further survey protocol development is needed in order to utilize UAV surveys for calculating sex ratios. A secondary optical camera or conducting surveys when bucks are in velvet may enhance sex identification efforts