Unmanned aerial vehicles as a useful tool for investigating animal movements

Determining animal abundance is crucial for assessing the effectiveness of management measures against pest animals (Sakata & Samejima, 2007). Meanwhile, investigating movements has become important conducting estimations unmarked since Rowcliffe et al. (2008) introduced a camera trapping method known as random encounter model (REM). This enabled researchers to conduct density targeting by using contact ratio between traps and targeted that randomly move at certain speed in given area. However, this requires an independent survey estimating (v, daily travel distance target species). For speed, (Rowcliffe al., 2016) GPS tagging (Kay 2017) are commonly used tools. Nevertheless, huge labour costs collecting cameras changing batteries. In contrast, invasive it physical capture animal. Unmanned aerial vehicles (UAVs) currently installed wildlife monitoring, often combined with thermal imaging (Bushaw 2019). Spaan (2019) investigated spider monkeys (Ateles spp.), Witczuk (2018) surveyed ungulates. Chrétien (2015) thermal-equipped UAVs detected multiple species such bison, fallow deer, wolves elks. They both visible infrared images obtain image data successfully these images. UAV evaluated non-invasive suitable surveys inaccessible landscapes (Old also mentioned about its challenges identification, high dependence on weather conditions (Witczuk 2018) vegetation height we assumed would alternative investigate movements, can record longer samples than traps; The recordable time be 1–2 min maximum (Nakashima 2018), where quadcopter video one flight. According producers, longest possible flight duration Inspire 1 (SZ DJI Technology Co., Ltd., Shenzhen, China; our research) 18 min. As August 2021, SZ Ltd released Matrice 300 series, which capable flying 55 maximum. Furthermore, cannot follow their once gets out detection zone, until battery run out. Fixed-wing periods compared (Burke Still, monitoring were more likely ability hover. Here, proposed UAVs, conducted case study then discussed advantage limitations method. We carried sampling wild boars (Figure 1). Wild medium-sized mammals found world-wide (Massei Genov, 2004). omnivorous damage variety crops (Ballari Barrios-García, 2014). To deal agricultural caused boars, Japanese Ministry Environment aims halve population 2023 (MOE, 2018). There need determine trends assess established/newly implemented research Ishibashi district Odawara City, Kanagawa Prefecture, Japan 2). hilly farming village, citrus fruit traditionally grown. major land use area orange farms, followed abandoned farmlands, coniferous forests deciduous forests. yearly harvest mandarin oranges City was 11,400 t, according city government 2006–2007 (Odawara 2020). Mandarin trees have 2.5 m, recommended planting 33 per 10a. November 2019. temperature month ranged from 0.5°C 23.7°C, average 12.8°C. cloudless throughout survey. suffers boar year, including crop terrain destruction. An increase farmlands cause situation allows enter human habitats (Honda, states that, 2016, total cost across whole approximately 27 million yen, 7,950 ha farmland damaged local farmers, appeared region 2010. measure counter persistent deteriorating problem, hunting practised sporadically 8 2019 10 2019, 21:00 5:00, collect movements. nocturnally intention allowing ambient drop because nocturnal district. equipped Zenmuse XT (FLIR Systems, Inc., Wilsonville, Oregon, United States) flown manually areas present. weight UAV, camera, 3,115 g. 1345T Ltd) propellers. diagonal dimension without propellers landing mode 581 mm. uncooled Vox microbolometer 640 × 512 resolution, 17 μm pixel pitch, 13 mm focal length 7.5 13.5 spectral band. GO (application, iPad Pro(Apple Cupertino, California, screen monitor flight, showed real time. reason, internet connection required during prepared eight batteries (TB48, Ltd), but had charge while flights. overcame issue carrying emergency 100 V AC output. flew 50 m ground. wanted avoid disturbing animals, able distinguish other body size shape. collected videos, 3), right above it, ensuring centre tablet's screen. Then let hover (until emitted low-battery signal). make easier calculate distances moved, recorded videos facing ground perpendicularly, if fixed-point sky. under permission Government, operations after sunset Japan. permit number flights 8118, 4352. carefully trained pilot complied Japan's aviation law. First, calculated actual width WA (km) each recording. values home point, elevation boar's location angle view 4). locations derived target, value constant first detected. records, ArcGIS Pro 2.4 (Esri, Inc.) catalogue camera. identified adult shape 11 boars' 26 (rate successful detection: 0.42). determined some cases could not detect soon take-off. period varied 820 s. continuously movement maximum, what methods achieve. estimated 0.006 0.16 km/hr (Table (v) parameter all 1.54 km/24 hr. Thus, revealed moving greatly among samples, activity patterns moment moved. observe three different patterns. defined ‘feeding’, ‘seeking’ ‘roaming’. When remained same position, digging feeding. ‘seeking’, seeking food, relatively slow-moving speed. ‘roaming’, harvesting or simply roaming around. sample depended videos. From (approximately targets), no marked reactions towards observed none seemed distracted UAV. heights unable identify piglets, resembled those small raccoons Nyctereutes procyonoides hares Lepus brachyurus. been reported previously (Burke, Rashman, remains controversial closer targets may led disturbance. masked palm civets Paguma larvata macaques Macaca fuscata long tails patterns, up grooming. Our applicability specific characteristics size, probability cover height, previous studies tall forest areas. solved colder seasons, when difference surrounding environment larger Longmore, even places boars. assume two climb trees, gap distinct. affected battery, (Mandujano 2017). take-off cases, much power thus consumed before begun recording samples. Although, managed time, trapping. One surveys' issues disturbance (Rebolo-Ifrán degree related (Brunton instance, Ditmer American black bears strong phycological responses 20 individuals. Considering this, carefully, avoiding unnecessarily low (50 minimum), so disturb Ultimately, Nakashima importance taking species' day into consideration means sufficient needed how spend pattern day. research, It note done activities influence animals' differed populations less (Ohashi 2013). moved unit gain index calculation targets. two-dimensional space they future studies, measuring path three-dimensional estimation accuracy. accuracy will improved. electrical wires, voltage lines utility poles area, restricted paths. disappear sight, due law Regarding facts, obstacles around ensure clear view. study, UAVs. cautious difficult-to-observe potential applying With improvements technology methodology earlier, tool performed government's growth project ‘Yosegi’; Yamagishi student support program; Research Development Fund (JPMEERF19S20507) Environmental Restoration Conservation Agency Japan; JSPS KAKENHI Grant Number (20H03014); SFC (Shonan Fujisawa Campus) Policy Support Organization. authors very grateful support. thank Ms Keiko Sasaki revising paper. Satoshi Sugita helping them appointments farmers Finally, residents' association letting public hall base camp declare conflict interest associated manuscript. All conceived ideas designed methodology; through survey; M.I. S.N. analysed data; primary author. contributed critically drafts gave final approval publication. peer review history article available https://publons.com/publon/10.1111/2041-210X.13829. findings openly Dryad Digital Repository https://doi.org/10.5061/dryad.79cnp5hxc. Video Please note: publisher responsible content functionality any supporting information supplied authors. Any queries (other missing content) should directed corresponding author article.