Abstract: Photo-identification of individual snow leopards (Panthera uncia) is the primary data source for density estimation via capture-recapture statistical methods. To identify individual snow leopards in camera trap imagery, it is necessary to match individuals from a large number of images from multiple cameras and historical catalogues, which is both time-consuming and costly. The camouflaged snow leopards also make it difficult for machine learning to classify photos, as they blend in so well with the surrounding mountain environment, rendering applicable software solutions unavailable for the species. To potentially make snow leopard individual identification available via an artificial intelligence (AI) software interface, we first trained and evaluated image classification techniques for a convolutional neural network, pose invariant embeddings (PIE) (a triplet loss network), and compared the accuracy of PIE to that of the HotSpotter algorithm (a SIFT-based algorithm). Data were acquired from a curated library of free-ranging snow leopards taken in Afghanistan between 2012 and 2019 and from captive animals in zoos in Finland, Sweden, Germany, and the United States. We discovered several flaws in the initial PIE model, such as a small amount of background matching, that was addressed, albeit likely not fixed, using background subtraction (BGS) and left-right mirroring (LR) techniques which demonstrated reasonable accuracy (Rank 1: 74% Rank-5: 92%) comparable to the Hotspotter results (Rank 1: 74% Rank 2: 84%)The PIE BGS LR model, in conjunction with Hotspotter, yielded the following results: Rank-1: 85%, Rank-5: 95%, Rank-20: 99%. In general, our findings indicate that PIE BGS LR, in conjunction with HotSpotter, can classify snow leopards more accurately than using either algorithm alone.

Abstract: The cycle of livestock depredation and retaliatory killing constitutes a major threat to large carnivores worldwide and imposes considerable hardships on human communities. Mitigation efforts are often undertaken with little knowledge of ecological underpinnings and patterns of depredation, limiting conservationists' ability to develop, prioritize, and evaluate solutions. Carnivore detection and depredation data from interviews in affected communities may help address this gap, but such data are often prone to false-positive uncertainty. To address these challenges in the Pamir Mountains of Tajikistan we collected snow leopard, lynx, wolf, and bear detection and depredation reports from local communities via semi-structured interviews. We used a novel hierarchical multi-species multi-state occupancy model that accounted for potential false-positives to investigate carnivore site use and depredation concurrently with respondents' apparent vulnerability to that risk. Estimated false-positive probabilities were small, but failure to account for them overstated site use probabilities and depredation risk for all species. Although individual vulnerability was low, depredation was nonetheless commonplace. Carnivore site use was driven by clear habitat associations, but we did not identify any clearly important large-scale spatial correlates of depredation risk despite considerable spatial variation in that risk. Respondents who sheltered livestock in household corrals reinforced with wire mesh were less likely to report snow leopard depredations. Reducing depredation and retaliation at adequately large scales in the Pamirs will likely require a portfolio of species-specific strategies, including widespread proactive corral improvements. Our approach expanded inference on the often-cryptic processes surrounding human-carnivore conflict even though structured wildlife data were scarce.

Abstract: The main purpose of this work was to study the use of infrared trapping cameras to estimate Snow Leopard population size in a specific study area. This is the first time a study of this nature has taken place in China. During 71 days of field work, a total of 36 cameras were set up in Muzat Valley adjacent to the Tomur Nature Reserve in Xinjiang Province. We expended approximately 2094 trap days total. At least 32 pictures of Snow Leopards, 22 pictures of other wild species and 72 pictures of livestock were taken in the Muzat Valley. Meanwhile, 20 transects were run and 31 feces sample were collected. We also observed the behavior of ibex for 77.3 hours and found a total of approximately 264 ibexes in the research area.