An automatic method for removing empty camera trap images using ensemble learning.

Camera traps often produce massive images, and empty images that do not contain animals are usually overwhelming. Deep learning is a machine-learning algorithm and widely used to identify empty camera trap images automatically. Existing methods with high accuracy are based on millions of training samples (images) and require a lot of time and personnel costs to label the training samples manually. Reducing the number of training samples can save the cost of manually labeling images. However, the deep learning models based on a small dataset produce a large omission error of animal images that many animal images tend to be identified as empty images, which may lead to loss of the opportunities of discovering and observing species. Therefore, it is still a challenge to build the DCNN model with small errors on a small dataset. Using deep convolutional neural networks and a small-size dataset, we proposed an ensemble learning approach based on conservative strategies to identify and remove empty images automatically. Furthermore, we proposed three automatic identifying schemes of empty images for users who accept different omission errors of animal images. Our experimental results showed that these three schemes automatically identified and removed 50.78%, 58.48%, and 77.51% of the empty images in the dataset when the omission errors were 0.70%, 1.13%, and 2.54%, respectively. The analysis showed that using our scheme to automatically identify empty images did not omit species information. It only slightly changed the frequency of species occurrence. When only a small dataset was available, our approach provided an alternative to users to automatically identify and remove empty images, which can significantly reduce the time and personnel costs required to manually remove empty images. The cost savings were comparable to the percentage of empty images removed by models.

The effective use of camera traps to document the northernmost distribution of the western black crested gibbon in China.

Gibbons represent a highly successful radiation of four genera and 20 species of Asian apes that, in response to recent habitat fragmentation and deforestation, are threatened with extinction. China has six species of gibbons, each of which is critically endangered. We present new biogeographical information on the distribution of the black crested gibbon (Nomascus concolor). Four subspecies of N. concolor have been described: three of them are present east of the Mekong River (Nomascus concolor jingdongensis, N. c. concolor and N. c. lu); and another is found west of the Mekong River (N. c. furvogaster). In addition, there has been speculation that gibbons exist in the Biluo Snow Mountains, between the Mekong and Salween basins. To clarify the biogeography of this species, from April 2011 to January 2012 and from January 2016 to September 2018, we conducted interviews with local villagers, completed line transect surveys, monitored gibbon calls, and placed 30 camera traps in the forest canopy. On October 30, 2016, we recorded gibbon's calls. On July 5, 2016, our camera traps obtained one image of a male gibbon, and on February 1 and 8, 2017, we captured two independent images of an adult female gibbon on Zhiben Mountains. Based on the black crest on the head, clearly visible in the photographs, the gibbons are N. c. furvogaster. Evidence from interviews and survey records indicate that N. c. furvogaster once was present in the Zhiben Mountains, at an altitude of between 2000 and 2700 m. Between 1990 and 2000, some 6-7 groups still existed in Caojian, Laowo and adjacent areas. Unfortunately, in the absence of an effective conservation strategy, the population was extirpated by hunters. The remaining forest in the Zhiben Mountains is highly fragmented, and most of the suitable habitat for gibbons has been lost. Therefore, we expect that this newly found gibbon population is under extreme anthropogenic pressure. It is imperative that further investigations of this gibbon population be conducted immediately, and that the local and national governments implement effective conservation plans, including educating the local communities to protect this critically endangered primate population.

Effects of reflective warning markers on wildlife.

Light pollution has become one of the top issues in environmental pollution, especially concerning how secondary light pollution, such as from traffic reflective materials, influences animal distribution and behavior. In this study, 15 camera traps were set up at sites with or without reflective warning markers (RWM) in coniferous forests on Cangshan Mountain located in Dali Prefecture, China. The results showed that the number of independent photographs and species at sites without RWMs were significantly higher than those at sites with RWMs. Significant differences were found between daytime and nighttime composition of bird species and non-flying mammals between two sites. This study found that RWMs had negative effects on wildlife, with the avoidance response of birds to RWMs being more obvious than that of animals at daytime. It is recommended that the use of reflective materials be carefully considered, especially in protected areas.

Black-and-white snub-nosed monkey (Rhinopithecus bieti) feeding behavior in a degraded forest fragment: clues to a stressed population.

Rapid global deforestation has forced many of the world's primates to live in fragmented habitats, making the understanding of their behavioral responses to degraded and fragmented habitats a key challenge for their future protection and management. The black-and-white snub-nosed monkey (Rhinopithecus bieti) is an endangered species endemic to southwest China. The forest habitat ranges from near-continuous to fragmented. In this study, we investigated the activity budget and diet of a R. bieti population that live in an isolated and degraded habitat patch at Mt. Lasha in Yunnan Province, near the current southern limit of the species. We used our data along with data from six other sites in more-continuous habitats across its range to model factors that predict stress, including feeding effort and time feeding on lichens against potential predictive parameters. Models showed feeding effort across all sites increased with increasing altitude and latitude, and with decreasing food species diversity. There was also a strong positive relationship between feeding effort and time feeding lichens. The Mt. Lasha R. bieti population exploited a total of 36 food species, spending 80.2% of feeding time feeding on lichens, Bryoria spp. and Usnea longissima. These figures are more comparable to those living in the north than those living in the mid- and southern part of the species' range. Given the models for feeding effort and time feeding on lichens, the unexpectedly high time spend feeding on lichens and feeding effort relative to latitude and elevation are suggestive of a stressed population at Mt. Lasha.