Search

The snow leopard (Panthera uncia) inhabits one of the most challenging environments on Earth, referred to as the 'third pole'. Only a fraction of its vast range has been explored thus far, owing to myriad of barriers inflicted by the remote terrain and socio-ecological realities of the landscapes. Understanding distribution patterns of species is essential to devise practical management measures. This study aimed to understand the distribution pattern and factors influencing occupancy of snow leopard in the Pamir Mountain range through sign-based occupancy modelling. Our study confirmed that the Pamir range is a snow leopard stronghold, with occupancy estimated at 0.57 ± 0.02. The topographic features positively influenced the detection probability (p = 0.37 ± 0.005) of snow leopards. Occupancy was influenced by mean annual precipitation (β = -6.12 ± 1.8), density of roads (β = -1.61 ± 0.6) and water sources (β = 0.74 ± 0.4). Our findings underpin that sign-based distribution surveys provide vigorous scientific knowledge about elusive species and merit replication being used for other species. We propose to redefine the protected area boundaries based on ecological knowledge and encourage transboundary cooperation to safeguard snow leopards at a landscape scale. [ABSTRACT FROM AUTHOR]

Highlights: [•] This is first study to document habitat selection of brown bears in Himalaya. [•] Brown bears avoided higher elevations and steeper slopes and selected more productive areas. [•] Brown bears tolerated human infrastructures, like roads and camps, but strongly avoided grazing areas with high livestock density. [•] In Himalaya, brown bears prefer alpine meadows, which are experiencing increasing grazing pressure and agricultur al use. [Copyright &y& Elsevier]

Deosai Plateau in Northern Pakistan was designated a national park to protect the largest remnant population of brown bears in Pakistan. The natural resources of this high elevation (3500–5000m) park make a significant contribution to the livelihood of local and nomad communities. The present legislation excludes people from a park, which increases conflicts between management and local people. However, a pragmatic approach was adopted to involve people in conservation in Deosai. Community participation, achieved by recognizing rights and introducing incentives, reduced resistance against the conservation efforts, reduced grazing pressure in bear habitat and helped reduce poaching. The size of the brown bear population was set as an indicator of park success, and was monitored annually from 1993 through 2006. We observed a 5% annual growth of the brown bear population, suggesting that the conservation program has been successful due to a successful cooperation between an NGO, people, and the park management. The increase of the bear population is significant, because we observed an extremely low reproductive rate, due to late age of first reproduction (8.25years), a long reproductive interval (5.7years), and a small litter size (1.33). The reproductive rate of the Deosai population is the lowest yet documented for brown bear populations. Poor habitat quality, low quality food, high seasonality, and extreme weather conditions in the Himalaya probably explain the poor reproductive performance. Considering such low reproduction and known exchange of individuals with neighboring populations, we believe that the observed growth was a sum of reproduction and immigration. Brown bears are declining throughout South Asia and often have low-productive rates. Therefore, conservation efforts for brown bears in this region must target reducing human-caused mortalities, particularly of adult females. Involvement of people can increase efficiency in conservation, in addition to reducing cost and conflicts. [Copyright &y& Elsevier]

Abstract: Asian bears face major threats due to the impact of human activities as well as a critical lack of knowledge about their status, distribution and needs for survival. Once abundant in northern Pakistan, the Himalayan brown bear (Ursus arctos isabellinus) has been exterminated in most of its former distribution range. It presently occurs sparsely, in small populations, the Deosai National Park supporting the largest isolate. This decline might imply a reduction in genetic diversity, compromising the survival of the population. Using a combination of fecal DNA analysis and field data, our study aimed at assessing the size and genetic status of the Deosai population and give guidelines for its conservation and management. Using fecal genetic analysis, we estimated the population to be 40–50 bears, which compares well with the field census of 38 bears. The northern Pakistani brown bear population may have undergone an approximate 200–300-fold decrease during the last thousand years, probably due to glaciations and the influence of growing human population. However, in spite of the presence of a bottleneck genetic signature, the Deosai population has a moderate level of genetic diversity and is not at immediate risk of inbreeding depression. Gene flow might exist with adjacent populations. We recommend careful monitoring of this population in the future both with field observations and genetic analyses, including sampling of adjacent populations to assess incoming gene flow. The connectivity with adjacent populations in Pakistan and India will be of prime importance for the long-term survival of Deosai bears. [Copyright &y& Elsevier]

• We summarize existing deep learning-based image inpainting algorithms in three aspects, including inpainting strategies, network structures and loss functions. • We introduce the open source codes, popular used datasets, evaluation metrics, and application scenarios. • We compare the inpainting algorithms with released source codes. • We outline the challenges and possible future directions. Image inpainting has been widely exploited in the field of computer vision and image processing. The main purpose of image inpainting is to produce visually plausible structure and texture for the missing regions of damaged images. In the past decade, the success of deep learning has brought new opportunities to many vision tasks, which promoted the development of a large number of deep learning-based image inpainting methods. Although these methods have many similarities, they also have their own characteristics due to the differences in data types, application scenarios, computing platforms, etc. It is necessary to classify and summarize these methods to provide a reference for the research community. In this survey, we present a comprehensive overview of recent advances in deep learning-based image inpainting. First, we categorize the deep learning-based techniques from multiple perspectives: inpainting strategies, network structures, and loss functions. Second, we summarize the open source codes and representative public datasets, and introduce the evaluation metrics for quantitative comparisons. Third, we summarize the real-world applications of image inpainting in different scenarios, and give a detailed analysis on the performance of different inpainting algorithms. At last, we conclude the survey and discuss about the future directions. [ABSTRACT FROM AUTHOR]