Population indices for wild carnivores: a critical study in sand-dune habitat, south-western Queensland

Populations of feral cats, foxes and dingoes were assessed at four sites in the northern Simpson Desert from May 1995 to October 1996 using track counts. Counts were based on presence/absence of footprints on sandplots placed randomly throughout each of the four sites, with ail habitats being sampled in accordance with their availability. Sandplots were repositioned between sampling periods so that data were temporally independent. This unbiased method was used to assess the reliability of more conventional indices of carnivore abundance based on spotlighting and counts of footprints on roads, which were run concurrently with the random sandplot counts. The sandplot data were also used to assess habitat use of the three study species. Counts on regularly spaced sampling plots along roads were correlated with random sandplot counts for foxes, but not for cats. Conversely, an index of activity based on counting individual tracks along roads was correlated for cats, but not for foxes. Spotlighting counts appeared to have little relationship with random sandplot counts for either species, with random sandplot counts and spotlight counts of foxes being correlated for only one of the four sites and no correlations being apparent for cats. The sandplot data showed that cats used dune crests preferentially over other habitats, while foxes preferred both dune crests and roads. Dune crests are probably used as natural runways by both species. Dingoes used roads preferentially over other habitats. It was concluded that differences in the use of roads and other runways by the carnivores can potentially lead to biases in indices collected along roadways. In relation to the random sandplot counts, foxes were overestimated by road counts and spotlighting relative to cats, the latter bias being related possibly to temporal differences in the activity patterns of the two species.