Quantifying the movement, behaviour and environmental context of group‐living animals using drones and computer vision

Journal of Animal Ecology, Journal Year: 2023, Volume and Issue: 92(7), P. 1357 - 1371

Published: March 21, 2023

Abstract Methods for collecting animal behaviour data in natural environments, such as direct observation and biologging, are typically limited spatiotemporal resolution, the number of animals that can be observed information about animals' social physical environments. Video imagery capture rich their but image‐based approaches often impractical due to challenges processing large complex multi‐image datasets transforming resulting data, locations, into geographical coordinates. We demonstrate a new system studying wild uses drone‐recorded videos computer vision automatically track location body posture free‐roaming georeferenced coordinates with high resolution embedded contemporaneous 3D landscape models surrounding area. provide two worked examples which we apply this approach gelada monkeys multiple species group‐living African ungulates. how simultaneously, classify individuals by age–sex class, estimate individuals' postures (poses) extract environmental features, including topography trails. By quantifying movement while reconstructing detailed model landscape, our opens door sensory ecology decision‐making within

Language: Английский

---

Large‐scale and long‐term wildlife research and monitoring using camera traps: a continental synthesis

Biological reviews/Biological reviews of the Cambridge Philosophical Society, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 17, 2025

ABSTRACT Camera traps are widely used in wildlife research and monitoring, so it is imperative to understand their strengths, limitations, potential for increasing impact. We investigated a decade of use cameras (2012–2022) with case study on Australian terrestrial vertebrates using multifaceted approach. ( i ) synthesised information from literature review; ii conducted an online questionnaire 132 professionals; iii hosted in‐person workshop 28 leading experts representing academia, non‐governmental organisations (NGOs), government; iv mapped camera trap usage based all sources. predicted that the last would have shown: exponentially sampling effort, continuation trends up 2012; analytics shifted naive presence/absence capture rates towards hierarchical modelling accounts imperfect detection, thereby improving quality outputs inferences occupancy, abundance, density; broader scales terms multi‐species, multi‐site multi‐year studies. However, results showed effort has reached plateau, publication only modestly. Users reported reaching saturation point images could be processed by humans time complex analyses academic writing. There were strong taxonomic geographic biases medium–large mammals (>500 g) forests along Australia's southeastern coastlines, reflecting proximity major cities. Regarding analytical choices, bias‐prone indices still accounted ~50% this was consistent across user groups. Multi‐species, multiple‐year studies rare, largely driven hesitancy around collaboration data sharing. no repository Atlas Living Australia (ALA) dominant sharing tabular occurrence records. ALA presence‐only thus unsuitable creating detection histories absences, inhibiting modelling. Workshop discussions identified pressing need enhance efficiency, scale management outcomes, proposal Wildlife Observatory (WildObs). To encourage standards sharing, WildObs should promote metadata collection app; create tagged image facilitate artificial intelligence/machine learning (AI/ML) computer vision space; address identification bottleneck via AI/ML‐powered image‐processing platforms; commons suitable modelling; v provide capacity building tools Our review highlights while investments monitoring biodiversity position global leader context, realising requires paradigm shift best practices collecting, curating, analysing ‘Big Data’. findings framework broad applicability outside meet conservation objectives ranging local scales. This articulates country/continental observatory approach also international collaborative networks.

Language: Английский

---

A novel vibration energy harvester based on eccentric semicircular rotor for self-powered applications in wildlife monitoring

Energy Conversion and Management, Journal Year: 2021, Volume and Issue: 247, P. 114674 - 114674

Published: Sept. 10, 2021

Language: Английский

---

Camera trapping in ecology: A new section for wildlife research

Ecology and Evolution, Journal Year: 2023, Volume and Issue: 13(3)

Published: March 1, 2023

Ecological research is undergoing a substantial transformation. Camera trapping—"capturing" photograph remotely, allowing observation of wildlife separately from the observer—has been around for over century. However, it emerged as substantive mode sampling occurrence only about three decades ago (Kucera & Barrett, 2011; O'Connell et al., 2011) and now rapidly improving innovating, changing face ecology (Burton 2015). With repeated made possible across space time, limited by logistics resources, observations can be gathered analyzed at unprecedented spatial temporal scales. engineering relatively inexpensive camera models that do not require costly support systems (such those needed satellite telemetry), traps also serve to democratize research. trapping has consequently spread global south developing countries (Agha 2018; Cremonesi 2021; Galindo-Aguilar 2022). Many private citizens run their own traps; networking these citizen scientists have yielded great insights will continue so (McShea 2016). are being employed Indigenous peoples ask questions on traditional territories (Artelle Fisher 2021), an important step towards meeting principles United Nations Declaration Rights Peoples (Gilbert, 2007). Camera-trap spans ecological hierarchy, with applications animal behavior (Caravaggi 2017, 2020) such diel activity (Frey 2017; Rowcliffe 2014), populations (Bischof 2020; Gardner 2010), species' distributions (Rich Tobler 2015), communities (Ahumada Wittische 2021). adequate inferential logic analysis, more complex processes species interactions discerned (Beirne Clare 2016; Niedballa 2019). The field rich planting seeds new ideas. In fact, though largely used mammals, expanding taxonomically include vegetation (Seyednasrollah 2019; Sun herptiles (Moore Welbourne 2020), avifauna (Jachowski 2015; Murphy 2018). Software advanced in-step hardware. Converting images numerical data easier custom software, much open-source (Greenberg Young Processes automatic identification developed greatly speed up image classification process "big data" (Duggan Shepley Conceptual advances, frameworks understanding how detections sample underlying processes, paving way sophisticated (Glover-Kapfer Hofmeester Tremendous discoveries lay in future. Networking arrays different landscapes—even globally, similar weather networks (Steenweg 2017)—will allow macroecological scale never before (Chen 2022; Magle Rich 2017). Notwithstanding, await small focal studies too—these foundations inference. We endeavors Ecology Evolution's section Trapping Ecology. journal's mandate author-friendly, without gatekeeping assessments importance barrier, makes us place welcomes both small-scale autecological large-scale syntheses. This philosophy help authors work read scientific community—we believe this Section goal. first volume featured its camera-trapping study (Fisher 2011), paper desk-rejected several other journals "interesting but improbable" among fare. Editorial team gave chance, 100 citations later, continues stimulate debate (Stuber Fontaine, Since then, we published s camera-trap studies. eagerly anticipating many papers dedicated Section, Evolution plans forefront proliferation research, platform thought debate. Jason Thomas Fisher: Conceptualization (equal); writing – original draft (equal). None. No available.

Language: Английский

---

Comparison of Camera Traps, Edna, and Visual Encounter Surveys for Threatened Species Detection

Published: Jan. 1, 2025

Language: Английский

---

Comparison of camera traps, eDNA, and visual encounter surveys for threatened species detection

Journal for Nature Conservation, Journal Year: 2025, Volume and Issue: unknown, P. 126948 - 126948

Published: April 1, 2025

Language: Английский

---

Ungulate occurrence in forest harvest blocks is influenced by forage availability, surrounding habitat and silviculture practices

Ecological Solutions and Evidence, Journal Year: 2023, Volume and Issue: 4(2)

Published: April 1, 2023

Abstract Forest harvesting causes habitat loss and alteration can change predator–prey dynamics. In Canada, forest has shifted the distribution abundance of ungulates (deer, elk moose) that prefer early seral forest, resulting in unsustainable caribou predation by shared predators (bears, cougars wolves). Long‐term solutions for recovery require management to reduce ungulate prey species within ranges. Silviculture practices applied after directly affect amount forage available harvested areas, therefore influence distribution, but few studies have completed detailed assessments on how specific treatments site preparation, planting stand tending use harvest blocks. We used camera traps, silviculture data, GIS‐derived disturbance vegetation data collected at field sites investigate occurrence blocks west‐central Alberta, Canada. compared seasonal investigated site‐specific characteristics, surrounding density, fine‐scale influenced Deer, moose was higher summer winter. Elk, white‐tailed deer with greater availability species. Moose a lower road density area, further from seismic lines proportion area. younger Mule densities lodgepole pine, mule also had been tended. Our study provides information response methods linking wildlife forestry practices, providing practical scientific inform sustainable forestry. Translating this research into landscape decisions could benefit boreal biodiversity, including threatened like caribou, culturally economically important deer, moose.

Language: Английский

---

Ecological Insights From Camera Trapping Span Biological Taxa, and the Globe

Ecology and Evolution, Journal Year: 2025, Volume and Issue: 15(2)

Published: Feb. 1, 2025

Camera trap research has grown to encompass the globe, with applications in terrestrial, marine, and aquatic environments. Insights on plant, invertebrate, vertebrate communities are rapidly expanding our knowledge of ecological systems. A technological revolution was launched by advancements camera trapping (Kucera Barrett 2011; O'Connell, Nichols, Karanth 2011). "Trapping" an observation a species its environment at fixed place time allowed scientists sample widening range taxa ecosystems. With application appropriate sampling designs statistical models (Burton et al. 2015), have been able answer questions never before possible. Mammal ecologists were early adopters, taking advantage passive heat-in-motion detectors developed 2000s homeotherms (Finn 2005; Vercauteren, Smith, Stevenson 2005). The development larger memory cards long-lasting power enabled use repeated time-lapse photography for ectotherms vegetation, now insights derived from traps span biological hierarchies. In this special issue Ecology Evolution, we invited camera-trap papers around globe showcase depth breadth applications. We particularly aimed representation global south, which rich heretofore relatively untapped potential compared north (Steenweg 2017; Agha 2018; Fisher 2023; Mugerwa 2024). some great versatility scientific they offer (Figure 1). provide unique into species' behaviour, as allow observations without intrusive effects in-person observer (Caravaggi 2017, 2020). Research aspects such parental investment offspring care, can shed new light. Antarctic, Adélie penguins (Pygoscelis adeliae) reproduction varies environmental factors, but also degree maternal building nests, limited resources (stones) (McLatchie revealed that occupying nests earlier, resulting earlier clutch initiation, associated high-quality increased reproductive success. Larger advantageous successful breeding, these more likely be created built Environmental factors mediated relationship, innovative study illustrates what emerge future combines behaviour heterogeneity. Behaviour is important part risk avoidance, many studies showing humans impose perceived mammals, even via non-consumptive activities recreation (Taylor Knight 2003; Larson 2016). British Columbia, Canada, Fennell (2023) examined spatiotemporal avoidance eight large mammals alpine protected area. They observed spatial co-occurrence between ungulates consistent human shield hypothesis (Berger 2007), not expected consequent segregation carnivores humans; instead, (and herbivores) temporally displaced recreationists (Fennell 2023). Scaling up (sensu Steenweg (2017)) landscape 10 landscapes across Canadian west, Granados conducted similar analysis. used hierarchical quantify influence (roads logging) ungulate carnivore site use. Across vast heterogeneous space, found support hypothesis, positive negative responses disturbance (Granados signal temporal ran contrary predictions, overlap people deer road density. Such macroecological studies—made possible networking arrays (Gallo 2019; Fidino 2021; Barnas 2024)–are revealing much about large-scale departures phenomena smaller scales. Territorial defence another behavioural mechanism minimised, sometimes means scent-marking. Previously, most scent-marking focused transmitting individual, receivers harder observe; filled gap. northern Botswana, latrines scent-marked African wild dogs (Lycaon pictus) classify behaviours sniffing (less aggressive response) or overmarking (aggressive (Claase Wild exhibited "dear enemy" response competitors, less neighbours than strangers. This changed increasing size responding pack, switching strongly neighbour scent marks ("nasty hypothesis") Animals' activity pattern component illuminated traps, continuously through diel cycle (Frey 2017). Rhode Island, USA, Mayer multi-state occupancy-modelling framework informed data investigate how 14 mammal responded anthropogenic development. All respect magnitude season, illustrating plasticity trait, adaptation (Mayer body complexity species-human interactions their changing contexts, marked challenge coming century. One greatest strengths ability observe multiple syntopic simultaneously (Rovero Zimmermann 2016)–as opposed GPS collars, typically restricted one two species. Thus, community composition like before; being done regions previously rarely researched. Nepal, Regmi native occurrence forest cover livestock detections while declining proximity settlements. provided 15 studied species, including blue sheep (Pseudois nayaur) giant flying squirrels (Petaurista magnificus) relationships diverse understudied region (Regmi South Africa, live uneasy coexistence, competing subject intraguild mortality. placed elephant (Loxodonta Africana) carcasses association shared peak periods black-backed jackals (Lupulella mesomelas) spotted hyaenas (Crocuta crocuta), indicating resource sharing (Honiball Conversely, analysis showed lions (Panthera leo), suggesting latter dominates scavenging opportunities. Parsing apart dominance dynamics key area investigation facilitated Serengeti National Park, Tanzania, van den Bosch aardwolves (Proteles cristata) aardvarks (Orycteropus afer)—both nocturnal insectivores—to examine competitors. fact, multispecies occupancy modelling high evidence commensals: increase food accessibility (van Mozambique's Gorongosa Grabowski, Phillips, Gaynor (2024) explore patterns niche partitioning among mesocarnivores: large-spotted genet (Genetta maculata), civet (Civettictis civetta), honey badger (Mellivora capensis), marsh mongoose (Atilax paludinosus). Statistical analyses no and, indicated civets mongooses, competition syntopy different diets (Grabowski, Differing outcomes ecosystems reveal ways strategies affect coexistence. Many modern contain invasive major threat biodiversity (Rosenzweig 2001). Tianjin, China, Li generalised additive mixed see (Canis lupus familiaris Felis silvestris) three indigenous (Siberian weasels, Mustela sibirica; Amur hedgehogs Erinaceus amurensis, Tolai hares, Lepus tolai) densities along gradient urbanisation. Densities cats, urbanisation, green spaces urban areas predictors density (Li 2023)–a critical finding when growing (Seto Simkin 2022). urbanisation effect Australia Alting dingoes dingo). sought dispersion suggests concentrated should shrink home-range sizes. Both corroborated dingoes, subsidies (Alting 2024), repercussions Australian non-native diversity (Johnson, Isaac, 2007). Prior predator–prey ecology largely radiocollar predation rates predator prey (or both) observed. discern until capability bridged design how–for example–anthropogenic extraction affects predators space time. Boczulak deployed western Alberta, Canada wolves lupus) avoid features extraction, presumably due risk—unless occur those patches, point switch strong selection features. interaction novel availability understanding respond change (Boczulak cougars (Puma concolor) eastern Cougars positively only prey—particularly snowshoe hares (Lepus americanus) (Gaston Invasive white-tailed (Odocoileus virginianus) played discernible role, defying predictions difficult it extrapolate conclusions within same guild. Scavenging, case interactions, prime research, yield competition. northeastern Norway, Lacombe asymmetric Arctic fox (Vulpes lagopus) red vulpes) sites supplied carrion. tended occupied heterospecific, either clearly benefitting. Geographic variation rodent (prey) abundance affected dynamics, yielding delicate interplay potentially cold systems (Lacombe There inherent trapping, so threatened endangered Mortality always concern, moreover, catchability often declines rarity. non-invasive sampler rare United Kingdom (UK), Shannon, Valle, Shuttleworth monitor squirrel (Sciurus vulgaris) populations, image-capture rate correlated well live-capture rate. identify characteristics greater inform conservation planning, crucial UK's deforested (Shannon, Climate generate wholesale changes distributions (Pereira, Navarro, Martins 2012), devoted adapt (Boutin Lane 2014). Among pelage primary variably snowy environments, arise whether phenological mismatches snow may occur. Stokes network mountain hare timidus) coat colour climatic gradients. Mountain higher latitudes altitudes retained winter white coats longer did lower altitudes. Moreover, coastal climates inland variable conditions (Stokes spans marine well, continuous video feed commonly "trap" (Willis, Millar, Babcock 2000; Whitmarsh, Fairweather, Huveneers Bulger, Volpe, 2019). Red Sea coral reef, Lilkendey feeding mechanics foraging energy expenditure herbivorous brown surgeonfish (Acanthurus nigrofuscus) yellowtail tang (Zebrasoma xanthurum). Here, "camera trapping" included remote underwater stereo processed AI-driven recognition, classification, 3D tracking. exerted grazing pressure reef spite low biomass. Brown specialist strategy, tangs generalist differentiation maintaining efficiency. combination assisting technologies showcases revolutionary biotic sensing environments (Lilkendey Exciting opportunities lay land-water interface. Sullivan, Rittenhouse, Vokoun cold-water patches riverine These refuges aggregators fish seeking thermal refuges, there become susceptible Avian mammalian pervasive concentrate (Sullivan, southern Mexico, Delgado-Martínez ephemeral water bodies birds seasonally dry tropical forests. surface pools tree holes recorded behaviour. Terrestrial preferentially bodies, whereas arboreal scansorial small medium common bodies; complementary sources thus facilitate gamma (Delgado-Martínez Finland, Holopainen artificial (with eggs) wetlands ground-nesting boreal ducks. Predation surrounded agricultural land forested land. Shoreline had further away contribute duck population Finland (Holopainen summary, work illuminates importance spanning interface, ripe discoveries. Homeothermic focus taxonomic camera-based inquiry expanded greatly frequent timelapse photos. plant phenology (Hofmeester 2020; Sun 2021) plant-insect (Naqvi 2022) very holds exciting potential. California, Simokat pollinator endemic Encinitas baccharis (Baccharis vanessae). focal compare efficacy methods. discovered attended insect groups. Focal underreported approximately half images too low-resolution visually Cameras benefit recording activity; however, dominated Lepidopterans (Simokat study, Gao ground-facing imagery soil-dwelling invertebrate China. quantifying Formicidae, Diplopoda, Gastropoda, Araneae, Coleoptera, Orthoptera, Chilopoda, Oligochaeta. Like scale problems (Levin 1992), abundance, richness, all quite sensitive frame (shooting area), authors effective protocol (Gao As image resolution increases gets cheaper, sure accelerate near future. field nascent endeavour, having newly evolved natural philosophy biogeography parentage past It still developing coherent theoretical framework, suffers scale, middle-number problems, lack unified theories, general laws (Peters 1991; Levin 1992; Lawton 1999; Allen Starr Nonetheless, march on, daily inexorably accumulate. plays role discoveries fast world. networks researchers citizen pooling data, scaling local continental scope 2017) efforts Snapshot USA (Kays 2022), Canada's WildCAM 2023), Safari (Pardo 2021). will scales help us concepts system function. Evolution promote endeavours continuing publish highlight camera-trapping papers, adding virtual online. Jason T. Fisher: conceptualization (lead), project administration software visualization (supporting), writing – original draft review editing (lead). Thanks & editors-in-chief, Drs. Moore, Andrew Beckerman, Gareth Jenkins, especially Marcus Lashley Arley Muth, contributions revisions. thanks, associate editors reviewers who gave generously make peer-review process successful. Jeff Dixon illustrated Figure 1 under direction licence JTF. funded JTF work. author declares conflicts interest. nothing report.

Language: Английский

---

A cautionary tale comparing spatial count and partial identity models for estimating densities of threatened and unmarked populations

Global Ecology and Conservation, Journal Year: 2022, Volume and Issue: 38, P. e02268 - e02268

Published: Aug. 18, 2022

Population monitoring is critical to wildlife conservation, but density estimation difficult for wide-ranging, unmarked species inhabiting remote habitats. Furthermore, recent investigations into with camera trap data has revealed models be potentially unreliable, prompting cautious application and continued model development. Two related approaches increasing appeal include spatial count (SC), which infer latent identities from the pattern of detections, partial identity (SPIM), additionally leverage covariates (e.g., sex, antler point count, presence GPS/radio collar). To assess performance models, we applied SC SPIM threatened boreal caribou in Canada, are declining have few rigorous estimates across their broad distribution inform conservation efforts. In particular, focused on two spatially proximate ranges northern Alberta, Canada that differ estimated demographic trends, disturbance histories, abundances predators apparent competitors. Estimates varied over a 4 year period (2016 – 2019), were higher region more stable reported growth rates less anthropogenic (mode estimates: 155 225/1000 km2 vs. 19 96/1000 km2). However, differed by modeling approach had low variable precision, hindering inferences about population status trajectories. Simulations suggest may been biased precise. likely underestimated mistaking detections neighboring individuals as recaptures single individual, although also overestimated inflating assignment probabilities non-existent individuals. Findings highlight need explore how grouping dynamics non-independent movement violate assumptions reduce ability distinguish We advocate investigation accuracy approaches, ecological sampling conditions appropriate different coordination efforts analyses improve inferences.

Language: Английский

---

Multispecies modelling reveals potential for habitat restoration to re‐establish boreal vertebrate community dynamics

Journal of Applied Ecology, Journal Year: 2021, Volume and Issue: 58(12), P. 2821 - 2832

Published: Sept. 7, 2021

Abstract The restoration of habitats degraded by industrial disturbance is essential for achieving conservation objectives in disturbed landscapes. In boreal ecosystems, disturbances from seismic exploration lines and other linear features have adversely affected biodiversity, most notably leading to declines threatened woodland caribou. Large‐scale needed, yet empirical assessments effectiveness on wildlife communities remain rare. We used 73 camera trap deployments 2015 2019 joint species distribution models investigate how habitat use the larger vertebrate community (>0.2 kg) responded variation key line characteristics (line‐of‐sight, width, density mounding) following treatments a landscape oil gas development northeastern Alberta. proportion explained was low comparison type season, suggesting short‐term responses were relatively weak. However, we found that with consistent restored conditions predicted support an altered composition, reduced wolf coyote, thereby indicating will result contact rates between caribou these predators. Synthesis applications . Our analysis provides framework assess predict emerging efforts. With growing importance species, recommend longer‐term monitoring combined landscape‐scale comparisons different approaches more fully understand direct critical investments. Only combining rigorous multispecies large‐scale restoration, effectively conserve biodiversity within rapidly changing environments.

Language: Английский

---