How many feral cats can be individually identified from camera trap images? Population monitoring, ecological utility and camera trap settings

The feral cat (Felis catus) is a key threat for many Australian native critical weight range animals (i.e. species of intermediate body mass between 35 and 5,500 g that are particularly susceptible to introduced predators) and estimates of cat abundance are required for assessing changes in population size. Camera trapping is a much used tool for monitoring and estimating population sizes, including with mark-resight techniques, for which the more robust estimators require individual identification. Many feral cats are individually marked, which potentially makes them suitable for such monitoring programmes. We sought to determine what proportion of cat images captured during a commonly used field deployment of camera traps could be individually identified, and whether aspects of camera trap deployment affected the rate of individual identification. Camera trap arrays were established in four conservation areas in south-west New South Wales, Australia, during 2017 (range 39-50 camera traps per site). The unlured camera traps were continuously deployed over 26 months, with five or 10 images captured per trigger. Where possible, cats were individually identified based on phenotypic characteristics. Over the deployment period (95,413 camera trap nights; CTN), we obtained 2.25 million images, of which 13,845 contained feral cats. Feral cat events (i.e. a series of images taken <5 minutes apart on the same camera trap) ranged from 0.004 to 0.047 events per CTN across the four conservation areas, with 85 individual cats identified. Depending on camera settings, few images could be assigned to a known individual (12.2-27.4% of feral cat events per site were of identifiable individuals). Minimum number known alive were 10-46 feral cats per site, with resultant quarterly densities ranging from 0.01 to 0.16 cats/km(2). With our current deployment, individual identification of feral cats was insufficient for estimating abundance or survival using individual mark-resight methods. Such deployment deficits limit the ecological conclusions that can be drawn from ours and similar studies.