Enumerating White‐Tailed Deer Using Unmanned Aerial Vehicles

Abstract

The white‐tailed deer (Odocoileus virginianus) is an ecologically important species in forests of North America. Effective management of forests requires accurate, precise estimates of deer population abundance to plan and justify management actions. Spotlight surveys in combination with distance sampling are a common method of estimating deer population abundance; however, spotlight surveys are known to have serious drawbacks such as high costs and sampling biases. Therefore, we tested the effectiveness of enumerating deer from unmanned aerial vehicle (UAV) flights, conducted 1 and 6 March 2018, to develop population and density estimates in 2 United States National Parks: Harpers Ferry National Historic Park (HAFE) and Monocacy National Battlefield (MONO). Concurrent spotlight surveys at MONO enabled us to compare estimates obtained by the 2 methods. Deer density estimates by 4 observers of UAV‐obtained thermal imagery from HAFE were 94.5 ± 3.9 deer/km2. Concurrent UAV and spotlight surveys at MONO found 19.7 ± 0.5 deer/km2 and 6.4 ± 4.9 deer/km2, respectively; suggesting that spotlight surveys may significantly underestimate deer densities. Despite the logistical challenges to UAV operation, our findings demonstrate that UAVs will become an invaluable tool for wildlife management as technology improves. © 2021 The Wildlife Society. This article has been contributed to by US Government employees and their work is in the public domain in the USA.

Cumulative effects of human footprint, natural features and predation risk best predict seasonal resource selection by white-tailed deer

“…The recovery of woodland caribou is dependent on the reduction of habitat loss within their ranges and predation pressure by grey wolves 69 www.nature.com/scientificreports/ ecosystems by over-browsing, which reduces plant growth, forest understorey diversity, and consequently animal diversity 20,70 . Estimates for deer abundance are difficult to ascertain across terrain and cover types 71,72 and whether carrying capacity for deer has been met in this system is not known. Anthropogenic landscape change facilitating deer invasion may entrain future landscape and biotic change where deer populations are on the rise.…”

Interaction between transect design and animal distribution in distance sampling of deer

“…Such information can come from complementary, contemporaneous sampling by other methods (McShea et al 2011). Such methods include aerial surveys, pellet‐count surveys, and trail camera surveys (deCalesta 2013, Howe et al 2017, Preston et al 2021). Each of these alternatives possesses the key advantage that they can be conducted independently of roads, and thus can be located randomly or systematically to produce unbiased sampling of the study area.…”

“…If road based spotlight surveys are to be used to enumerate deer, then researchers and land managers might consider the inherent biases in this method, the potential interplay between road locations and deer distribution, and the potential benefits of supplementing road surveys with additional methods. A recent pilot study used unmanned aerial vehicles (UAV) to estimate deer density at 2 NCR parks and found that UAV surveys detected significantly more deer than concurrent road surveys (Preston et al 2021). The flight paths used in that study are similar to the systematic and random survey types we evaluated in our simulations.…”

Eyes on the herd: Quantifying ungulate density from satellite, unmanned aerial systems, andGPScollar data

“…However, satellite imagery has not been commonly used to identify ungulates (but see Yang et al [2014] and Duporge et al [2020] on elephants) or in temperate regions (Xue et al, 2017). Similarly, UAS data sets allow for a remote assessment of locations and numbers of animals, offer higher resolution, and provide more control over survey timing and conditions (Barasona et al, 2014; Preston et al, 2021). GPS collar technology has also improved, with longer battery life, higher fix rates, and lower spatial error providing finer temporal resolution (Kays et al, 2015).…”