TS-CRC Student project - Vertebrate monitoring and re-sampling in Kakadu National Park

Charles Darwin University, Darwin

Michelle Watson

Summary | Change in vertebrate fauna | Response of mammal fauna to fire regimes  | Establish vertebrate fauna samples for baseline data  |


The rationale behind the project was the recent evidence that some populations of small mammals in the Kapalga portion of Kakadu National Park (KNP) had undergone a fairly rapid and in some cases severe decline during 1986–1993. As no comparable information exists for the rest of KNP, the existence (and extent) of any broad scale patterns of decline could not be known. In addition the relationship of this pattern to management practices could not be determined. The arrival of the cane toad ( Bufo marinus ) ,and the potential impacts of this arrival on some native species also acted as a catalyst for the commencement of this project.

Primary objectives for the first year of the project (2001) were to:

  1. Assess the change in the vertebrates (particularly mammal) fauna of KNP by re-sampling sites previously sampled during the 1980s and 1990s.
  2. Assess the response of the mammal fauna to varying fire regimes, through sampling a set of sites chosen to represent spatially and temporally variable fire regimes
  3. Establish a set of vertebrate fauna samples that will contribute to baseline data to be used in the assessment of cane toad impacts.

Change in vertebrate fauna during 1980s and 1990s

In all, mammal abundance data has been collected from 121 quadrats in Stage III of Kakadu National Park. Preliminary analysis of the data has shown no consistent pattern in the direction or magnitude of change in mammal populations in the southern half of the park.

Previous sampling in this area occurred between 1988 and 1990 and included 20 clusters each of 15 quadrats, two sites each of five quadrats, and another group of 70 quadrats placed in the Conservation Zone between Coronation Hill and El Sherana (Woinarski & Braithwaite 1990, 1991). Another set of 15 quadrats (all at KNP fire monitoring plots) was sampled in comparable manner in January-February 1996 (Woinarski & Griffiths 1996).

A number of species (including the common rock rat, pale field rat, Kakadu dunnart, northern brown bandicoot, western chestnut mouse, northern quoll and black footed tree rat) appear to have increased since the original sampling. However, of these species only the common rock rat, pale field rat, northern quoll and northern brown bandicoot have undergone statistically significant changes.

In contrast a number of species appear to have declined since the original sampling. These species include Kakadu pebble mound mouse, delicate mouse, sugar glider, sandstone antechinus, Leggadina and common planigale. Of these, only the decline in Leggadina is statistically significant. Other notable findings from the surveys to date include a very significant increase in the number of pigs since the original survey. In addition, surveys revealed the presence of the introduced black rat at Snake Plain and Black Jungle Springs.

We note that this analysis is very preliminary, and caution against over-interpreting the results. We have not yet attempted to consider the results in greater detail, but intend to see whether the degree and direction of change at any given quadrat can be related to fire history or other management factors.

Response of mammal fauna to varying fire regimes

Mammals were sampled in 12 large grids within Eucalyptus miniata —E. tetrodonta forests, with sites selected to represent contrasting fire regimes.

Fire histories were derived by collating and manipulating 21 years of fire data for KNP. A number of parameters were used to select sites that varied in the extent of spatial and temporal patchiness of burning over the 21 years of data. The selected sites included some that had been typically affected by small, localised fires and others which had been extensively burned or unburned.

At each site mammals were sampled in a grid covering approximately 9 hectares. Individual mammals were marked and released to provide for an assessment of population density. Sampling was carried out over a five-week period between June and August 2001. A broad range of environmental variables was assessed at each trapping location to allow an evaluation of the response of some key habitat factors (eg availability if tree hollows, understorey density, tree size classes and species composition) to fire history.

Although the data collected during this sampling is yet to be analysed it is evident from the raw data that there is an extraordinary level of variation among grids in both the numbers of animals and diversity of species caught. At this stage it appears that this variation can not be explained solely by the fire parameters investigated in this study. Further analysis of the data is required to explain these results.

Establish vertebrate fauna samples for baseline data

All sites sampled to date are amenable for use as monitoring plots in the assessment of cane toad impacts. It should be noted that cane toads were recorded at three sites surveyed during 2001 (Mt Evelyn, Birdie Creek and Sleisbeck Plateau).


Ms Michelle Watson
Ph D Student
Tel: 08 8944 8465

Fax: 08 8944 8455

26 Axehead Rd