Quantifying fine-scale resource selection by introduced feral cats to complement management decision-making in ecologically sensitive areas

Abstract

The feral domestic cat (Felis catus) is considered to be one of the most damaging introduced predators, responsible for the decline and extinction of numerous native species. Advanced satellite technologies enable the study of resource selection by small mammals at fine-scales through remote data. These tools can improve understanding of the spatial ecology of introduced predators in ecologically sensitive areas, such as where cats pose a threat to native species and where improvement of predator control methods is required. We studied fine-scale resource selection by feral cats in the ecologically important New Zealand braided-river ecosystem, where they pose a risk to endangered native ground-nesting birds. We collected 34 location datasets from 21 cats fitted with lightweight global positioning system-collars, and extracted landscape variables from a resource map created using very high spatial resolution satellite imagery (Quickbird) and object-base imagery analysis for image classification. We modelled second-order seasonal and annual resource selection functions and characterized landscape composition of highly-used areas using compositional analysis. At a population level, cats generally selected fine-scale landscapes that are important for their primary prey rabbits (Oryctolagus cuniculus), and for refugia. An external validation of the annual model using data from cats tracked in an independent study showed a positive correlation with model predictions. Individual cats also visited habitats used by native ground-nesting birds, and thus pose a threat to them through secondary predation or individual specialization. Cat control operations should therefore focus around areas of concentrated ground-nesting bird activity and in areas identified as high-use by cats.