Human expertise combined with artificial intelligence improves performance of snow leopard camera trap studies

Bohnett, Eve; Faryabi, Sorosh Poya; Lewison, Rebecca L.; An, Li; Bian, Xiaoxing; Rajabi, Ali Madad; Jahed, Nasratullah; Rooyesh, Hashim; Mills, Erica; Ramos, Saber; Mesnildrey, Nathan; Perez, Carolina M. Santoro; Taylor, Janet A.; Terentyev, V. P.; Ostrowski, Stéphane

doi:10.1016/j.gecco.2022.e02350

Cited by 5 publications

(1 citation statement)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Due to the secretive nature and low density of the forest musk deer, coupled with their partial habitats (such as extremely steep slopes) which are difficult for humans to access [13,14], it is challenging to observe or survey all of the individuals using traditional methods like transect or strip methods [10,15]. Recently, an increasing number of studies have proposed using analysis methods based on camera traps to estimate population Diversity 2023, 15, 1071 2 of 12 sizes [16][17][18]. Remote cameras are non-invasive, function continuously over protracted periods, save manpower, and can monitor animals day and night [19,20].…”

Section: Introductionmentioning

confidence: 99%

Integrating Species Distribution Models to Estimate the Population Size of Forest Musk Deer (Moschus berezovskii) in the Central Qinling Mountains of Shaanxi

Gao,

Bu,

Zheng

2023

Diversity

View full text Add to dashboard Cite

Understanding the population size of animals is crucial for formulating scientific management policies, especially for endangered species. The central area of the Qinling Mountains in Shaanxi is a vital area for forest musk deer, but research is insufficient and estimates of its population size are lacking. In this study, we constructed a species distribution model for the forest musk deer in the central Qinling Mountains in Shaanxi using topography, land-use, and bioclimatic variables alongside forest musk deer occurrence data. The Time-to-Event (TTE) model was employed to estimate the population density of forest musk deer in the selected survey area. By utilizing the suitable habitat area provided by the species distribution model, the population density in the central Qinling Mountains of Shaanxi was estimated by extrapolating from the survey area. Our estimate of the population size of forest musk deer in the central Qinling Mountains of Shaanxi is approximately 2722 ± 788. Similar population estimation methods could be more widely applied, especially in areas with limited survey resources.

show abstract

Section: Introductionmentioning

confidence: 99%

Integrating Species Distribution Models to Estimate the Population Size of Forest Musk Deer (Moschus berezovskii) in the Central Qinling Mountains of Shaanxi

Gao,

Bu,

Zheng

2023

Diversity

View full text Add to dashboard Cite

show abstract

The influence of bait and camera type on detection of a spectrum of medium-sized Australian mammals

Goldingay

2024

Aust. Mammal.

View full text Add to dashboard Cite

Camera-trapping programs are most effective when informed by knowledge of the influence of components of the camera setup such as bait and camera type. I investigated the influence of two baits and three camera types on detection of eight species of medium-sized mammals across a 10-month survey. The spotted-tailed quoll (Dasyurus maculatus) was the only species that favoured a meat bait (sardines) over a peanut-butter bait. The long-nosed potoroo (Potorous tridactylus), Parma wallaby (Notamacropus parma), long-nosed bandicoot (Perameles nasuta) and swamp wallaby (Wallabia bicolor) favoured the peanut-butter bait over the meat bait. The northern brown bandicoot (Isoodon macrourus), mountain brushtail possum (Trichosurus caninus) and feral cat (Felis catus) showed no preference. Camera type influenced weekly detection in three species. The long-nosed bandicoot had lower detection at a Reconyx white-flash (WF) camera compared with a Reconyx infra-red flash (IR) and a Swift wide-angle (Sw) camera. The mountain brushtail possum had higher detection at a Sw compared with WF and IR. The feral cat had higher detection at a WF compared with IR and Sw. The findings have implications in relation to selection of bait and camera type that can lead to more effective monitoring, whether a program is directed at one of these species or the spectrum of species.

show abstract

Survey techniques and impact mitigation for the Endangered northern quoll (Dasyurus hallucatus) in the semi-arid landscapes of the Pilbara

Dunlop,

Moore,

Cowan

et al. 2024

Aust. Mammal.

View full text Add to dashboard Cite

Improvements in survey techniques for threatened species gives quantifiable confidence about their presence or absence at a given location, enhancing our understanding of patterns of distribution and abundance. This is particularly important for legislatively protected threatened species that may be at risk of disturbance. Survey techniques vary in detection confidence, resource investment, and invasive impacts to individuals. We review the appropriate applications of techniques in surveying for the endangered northern quoll (Dasyurus hallucatus), including the effort required to be 95% confident of detecting presence and monitoring change in population trends in the Pilbara bioregion. The outlined protocols indicate best practice for effective and efficient northern quoll monitoring while protecting the welfare of the animals being monitored, and are relevant to Environmental Protection and Biodiversity Conservation Act requirements. We also provide suggestions to mitigate impacts on animals and habitat, and describe future directions and emerging techniques for the monitoring of northern quolls and other endangered species. This information is targeted at those interested in monitoring northern quolls in a field setting, including researchers, environmental consultants, Traditional Owners, and land managers.

show abstract

Human expertise combined with artificial intelligence improves performance of snow leopard camera trap studies

Cited by 5 publications

References 30 publications

Integrating Species Distribution Models to Estimate the Population Size of Forest Musk Deer (Moschus berezovskii) in the Central Qinling Mountains of Shaanxi

Integrating Species Distribution Models to Estimate the Population Size of Forest Musk Deer (Moschus berezovskii) in the Central Qinling Mountains of Shaanxi

The influence of bait and camera type on detection of a spectrum of medium-sized Australian mammals

Survey techniques and impact mitigation for the Endangered northern quoll (Dasyurus hallucatus) in the semi-arid landscapes of the Pilbara

Contact Info

Product

Resources

About