Fire, cats and the Kimberley’s declining mammals

HSF 13-7 | Amount: $ 63,000 | Project Leader: B Murphy | Project Period: Jul '13 - Jul '16

A project undertaken at the Research Institute for the Environment and Livelihoods, Charles Darwin University, and supervised by Brett Murphy


Until just 20 years ago, the mammal fauna of northern Australia’s vast savanna landscapes was considered secure; the catastrophic declines of central and southern Australia had not occurred in the north. However, work in Kakadu National Park since the 1990s has shown a massive, ongoing decline in mammal diversity and abundance. Recent evidence based on Indigenous knowledge has shown that these declines extend well beyond Kakadu, across the Top End of the Northern Territory. We are unable to implement an effective management response because the underlying drivers remain highly uncertain. Predation by cats and changes to fire regimes, and their interaction, have been strongly implicated, but it is unclear why northern mammals didn’t decline earlier – cats were present for well over a century, and the cessation, or at least attenuation, of Aboriginal fire management occurred many decades ago.

Figure 1. The black-footed tree-rat (Mesembriomys gouldii) is one of northern Australia’s threatened mammal species understood to be disadvantaged by frequent fire. (Photo: Hugh Davies)
Figure 2. An early dry season fire in Kakadu National Park, where small mammals are in rapid decline. (Photo: Clay Trauernicht)

The strongest evidence that fire also plays a role in the northern mammal decline is recent work from Kakadu, showing that between 2001 and 2007 mammal declines were greatest at the most frequently burnt sites. Similarly, results from CSIRO’s Kapalga fire experiment suggested that a regime of annual fires reduced mammal abundance relative to unburnt controls. Critically, both of these studies have indicated that the predominant season of fire (i.e. early vs. late dry season) has little effect on small mammals – both early and late fires seem to be equally detrimental. This finding is starkly at odds with the prevailing fire management paradigm in northern Australia, that of extensive prescribed burning early in the dry season – under mild fire conditions – to prevent extensive, intense late dry season fires.

Prominent ecologists have advocated increasing the ‘patchiness’ and heterogeneity of fires to favour small mammals, despite little direct evidence linking them. Hence, one of the critical knowledge gaps is whether mammals are disadvantaged by some ‘types’ of fire. We expect that fire attributes associated with reduced shelter (e.g. high intensity, low patchiness) would facilitate predation by cats and be detrimental to mammals. This question is of enormous applied significance, because varying the frequency and intensity of fire through prescribed burning is one of the few tools available to land managers in these vast savanna landscapes.

Our study focusses on the remaining intact mammal assemblages in isolated parts of Australia’s far northern savanna landscapes to examine the nexus between cats, fire and small mammals.

We will address four key questions:

  1. Do the characteristics of a single fire or fire mosaic (e.g. intensity, patchiness, size) affect the abundance of cats and the abundance and diversity of small mammals?
  2. Does cat abundance affect small mammal survival and recolonisation in burnt areas?
  3. How do interactive effects of fire and cats on small mammals vary across the north–south rainfall gradient?
  4. What are the fire management approaches that minimise the impact of cats, and maximise the persistence of small mammal in savanna landscapes?
Progress and Results

In 2015/2016, the project’s fieldwork was completed. At 88 sites across Melville Island, which were surveyed for mammals in the early 2000s, we conducted live trapping of small mammals (5 nights trapping at each site) and camera trapping (5 cameras × 5 weeks observation at each site). We undertook detailed assessments of vegetation and habitat at these sites. We now have information on changes in small mammal distribution and abundance over a 15-year period, as well as the current distribution and abundance of feral animals (cats, cattle, horses) and dingoes.

We completed data collection associated with a fire frequency experiment, where three fire frequency treatments (no fire, annual fire, triennial fire) have been applied since 2009. For two years we deployed 72 motion-sensor cameras in the 18 experimental fire treatment areas (3 treatments x 6 replicates), and these were recently retrieved.

So far, we have used our data to address two questions: (1) whether the threatened brush-tailed rabbit-rat (Conilurus penicillatus), a species which has declined dramatically across the mainland Top End, has also declined on Melville Island; and (2) whether feral cats and recent fire activity have influenced the pattern of decline. We used single-season occupancy models to investigate correlates of the current distribution of C. penicillatus, and dynamic occupancy models to investigate correlates of C. penicillatus local extinction. Our results show that C. penicillatus, which once occurred more widely across the island, is now restricted to parts of the island where feral cats are rarely detected and shrub density is high.

Our results suggest that feral cats are driving C. penicillatus towards extinction on Melville Island, and hence have likely been a significant driver in the decline of this species in Australia generally. The impact of feral cats appears to be mediated by vegetation structure. The ongoing development and implementation of methods to effectively reduce feral cat densities, coupled with the management of landscape processes to maintain shrub density, through fire management and the removal of large exotic herbivores, will contribute substantially to conserving this threatened species. We demonstrate that the distribution of species can be strongly influenced by top-down factors such as predation, thereby highlighting the importance of including biotic interactions when investigating the distribution of predation-susceptible species.

Our results for the fire experiment are preliminary, but suggest that the species richness of mammals is highest in frequently burnt areas (but burnt by low-intensity, patchy fires) – compared to long-unburnt areas (8 years without fire). However, there is evidence that for key species known to be sensitive to cat predation (such as the threatened black-footed tree-rat, Mesembriomys gouldii), lower fire frequencies are preferred in areas where cats are present. Again this points to an interaction between fire activity and cat predation.