engleski

Review of camera trap features for successful lynx identification

Eurasian lynx (Lynx lynx) individuals have unique coat patterns that enables researchers to identify individuals from photographs. This feature is useful in numerous ecological studies and camera traps became one of the most important tools for lynx research and monitoring. The aim of this study was to analyze how the choice of the camera trap model and its spatial-temporal arrangement affect the quality of lynx photographs for identification of individual animals. Research was conducted on 789 photographs that were taken in lynx natural habitat in Croatia and Slovenia in the period from 2007 to 2020. Twelve different camera trap models were used. Photographs were classified into three categories in terms of quality for identification. The following features were recorded for each photograph: photo trap manufacturer and model, date and time, location of camera trap (marking site, forest roads), flash type and whether it is used, focal length, ISO (sensitivity of digital camera sensor to light), resolution, speed trigger, aperture, length and width of the animal's body and body position (lateral or other position) on the photo, how much and what proportion of the body was photographed (length of the animal divided by the width of the photo in pixels), photo quality and color of the photo (color or black-white). Our analysis showed that photos taken at lynx marking sites had significantly better quality than those at forest roads, as it was possible to identify lynxes on 97.6% of the photos taken at marking sites (p<0, 001). Also, we confirmed that for animals in motion on forest roads there is a statistically significant difference in the quality of the photographs with respect to flash activity. On daytime photographs from forest roads identification was not possible in 8.1% of cases, while on night photographs it rises to 35.2%. The success of lynx identification can be improved by using camera models that have resolution higher than 1270 pixels. Also, if the lynx is present on more than 50% of the photo surface, identification will be even more successful. Considering all the parameters of camera traps that were included in this research, the best models for lynx identification are Cuddeback's C1, C2, C3, E1 and E2 models and Sunplus SPCA553 model.

camera traps, camera features, identification, Eurasian lynx, Lynx lynx

nije evidentirano