UK United Kingdom

Scientists and national park managers are failing northern Australia’s vanishing mammals

Conservationists should take heart that Australia is finally waking up to the biodiversity crisis in Australia’s north. It is an urgent problem: right now, a diverse assortment of our small mammals – bandicoots…

An early dry season fire in Kakadu National Park – are these fires burning up our mammals? Clay Trauernicht

Conservationists should take heart that Australia is finally waking up to the biodiversity crisis in Australia’s north. It is an urgent problem: right now, a diverse assortment of our small mammals – bandicoots, tree-rats, possums - are rapidly vanishing from northern Australia’s most iconic biodiversity strongholds.

Work in Kakadu National Park has shown that between 1996 and 2009 mammal populations crashed, with species richness and total abundance decreasing by 65% and 75% respectively.

Most alarming is that scientists can’t tell us why this is happening, let alone how to stop it.

Prominent conservationist Tim Flannery recently drew attention to the plight of northern Australian mammals in his Quarterly Essay and accompanying article in The Conversation (as have others writing in The Conversation and other media). Though valuable, these articles tend to oversimplify the complex and unresolved ecological questions that must underpin an appropriate management response.

Flannery asserts that “the main driver appears to be changes in fire regime, compounded by the presence of feral cats”. This overstates our current understanding.

Unlike most other conservation issues in Australia, the most immediate obstacle is lack of knowledge. We simply don’t know why mammals are declining or what management interventions could halt the decline.

Frequent fire hazards

There is widespread agreement that the mammal decline has something to do with fire. However, scientists have failed to answer the critical question ­– with a large budget, what type of fire regime should (and could) land managers implement to benefit small mammals?

The managers of Kakadu - the epicentre of the northern mammal decline - have been roundly criticised by conservation scientists for their excessive use of prescribed burning to prevent high-intensity wildfires. However, these same scientists have been remarkably silent on alternative management approaches.

There is evidence that high fire frequencies are detrimental to small mammals, demonstrated most significantly in the work by John Woinarski and colleagues. They found that in Kakadu, sites that had been most frequently burnt experienced the greatest small mammal declines over the period 1996-2009.

The challenge is understanding exactly what it is about fire regimes that has changed. The breakdown of traditional Aboriginal fire management - and possible increase in the size and intensity of fires - is often suggested as a trigger. However, in places like Kakadu this occurred many decades before mammal populations crashed.

A tinderbox landscape

There is a widely held view that there is too much fire in parks like Kakadu. However, calls for widespread reductions in fire frequency in the north tend to overlook the fundamental ecology of fire.

Fire is so prevalent in this region not because of people, but because of the intense monsoon climate. A reliable wet season promotes rapid grass growth, and a long dry season promotes drying of grassy fuels.

Abundant lightning at the end of the dry season ensures high fire activity. The effect of humans is to merely “jump start” the fire season, providing an ignition source earlier in the year.

It is inevitable that a large proportion of the landscape burns each year. For Kakadu, this figure is remarkably constant at around 45%.

The only “lever” available to land managers is whether burning occurs predominantly in the early (April­–July) or late (August–November) dry season.

This fact has driven the prevailing approach to fire management in northern Australia ­– that of extensive early dry season burning to pre-empt intense late dry season fires.

Fighting fire with fire

Because they occur under milder fire conditions, early dry season fires tend to be significantly less intense, smaller, and patchier than late dry season fires.

It is well known that many fire-sensitive plants (such as the northern cypress pine, and to a lesser extent sandstone heaths) benefit from an earlier fire regime. There is also evidence that this applies to small mammals too.

Based on satellite-derived fire mapping since 1980, we know that the proportion of Kakadu that burns each year has not changed over the last 32 years. However, it is clear that there has been a dramatic switch from late dry season to early dry season fires. Although such a shift in fire season is thought to be good for biodiversity, including mammals, there may be more insidious changes that have taken place.

For example, retired Kakadu ranger Greg Miles has suggested that a vicious grass­–fire cycle is now firmly established in Kakadu. The theory is that highly flammable speargrass (Sorghum species) increased in abundance following the breakdown of Aboriginal fire management, and now fuels fires of much greater intensity.

Even in the early dry season, speargrass fires tend to be of high intensity and low patchiness. Hence, abundant speargrass effectively negates the advantages of early dry season prescribed burning.

It is noteworthy that the introduced swamp buffalo was virtually eradicated from Kakadu in the 1980s, after more than a century in the park. Some suggest that this may have further promoted an increase in speargrass abundance and fire intensity. The only major change to have immediately preceded the mammal collapse in Kakadu was the eradication of the buffalo.

Whether a grass-fire cycle can be reversed is not clear. One option put forward is to burn heavily infested areas in the wet season, before the speargrass can produce seed. Another option that warrants objective consideration is grazing by exotic herbivores, especially as highly-flammable exotic grasses overtake the park.

We need a better understanding of how fire regimes have changed over the last century. Despite Kakadu having the best-described fire regimes in northern Australia, the available fire records are still relatively short (from 1980 on). These satellite-derived records are also limited to fire frequency, size and season. Recent changes in a broader spectrum of fire regime attributes, such as intensity and fine-scale patchiness, are highly uncertain.

Testing the critics' claims

Vague calls for “less fire” are unhelpful. Instead, we need to specifically identify what kinds of fire management work best for mammals at landscape scales. This means considering the intervals between fires, fire intensity, size and patchiness.

We also need to determine whether areas of long, unburnt vegetation are important for small mammals' survival, and how different species are affected by “mosaic” burning practices, in which fires of varying intensities, scales and times are burnt to create a patchwork or mosaic effect in the landscape.

Some researchers suggest that Aboriginal fire management would have resulted in such a mosaic, and the loss of this mosaic has triggered biodiversity declines. Although intuitively appealing, there is remarkably little direct evidence to support this hypothesis, especially in relation to small mammals.

The greatest challenge will be to translate these research findings into explicit management actions.

For example, Alan Andersen and colleagues described an approach to increasing the abundance of long unburnt areas, without actually decreasing the proportion of the landscape burnt each year. The trick is to decrease the randomness of burning, by concentrating prescribed burning on recently burnt areas, avoiding long unburnt areas.

If the current model of extensive early dry season burning is so obviously failing, then there is an onus on its critics to propose something new.

Unlocking answers from the land

Land managers need to do their bit too by facilitating collaborative research in conservation areas.

Sadly, national parks can be difficult places to conduct research in northern Australia, hindered by slow-moving bureaucracies and priorities other than biodiversity conservation.

The quantity and quality of research coming out of Australian Wildlife Conservancy properties in the Kimberley demonstrates how much easier it can be to conduct conservation research on private land.

Close collaboration between scientists and land managers is essential, ideally within an adaptive management framework. This is where new management interventions can be repeatedly tried, evaluated, improved or potentially abandoned.

Most importantly, unorthodox management interventions need to be objectively evaluated. These might range from grazing by large exotic herbivores (such as buffalo and cattle) to reduce fuel loads, to wet season burning of annual grasses.

As Tim Flannery points out, “things are now so dire that we cannot afford to persist with business as usual”.

Join the conversation

13 Comments sorted by

  1. Tim Beshara

    Executive Officer

    I am interested in the mesopredator release of reptiles ie goannas from lack of other higher-order predators. It makes sense that a rise the numbers of a diurnal and effective hunter would impact on small mammals.

    I am also interested in the fact that many winter arid zone burns were done for the purpose of goanna harvesting. But this practice has declined (both the burning and the hunting)

    Would the combination of changes to fire regimes, meso-predator release of reptiles and a decline in human harvesting of goannas have a significant impact on marsupial numbers?

  2. Andy Sheppard

    Theme Leader, Biodiversity at CSIRO

    As is so often the case in remote Australia, ecological understanding is so data limited. This article captures this issue well for fire management in northern Australia for biodiversity benefits. Here are a few comments based on my experience working with other teams operating more broadly across the north and my own experience in invasive species management.
    While there is general agreement around small mammal decline in the fire prone northern grass-tree savannas (there have been historical analyses…

    Read more
  3. Bernie Masters

    environmental consultant at FIA Technology Pty Ltd, B K Masters and Associates

    My limited understanding of northern Australian ecology asks two additional questions:
    1. if traditional Aboriginal burning practices ceased some decades ago, what evidence exists to show that the high small mammal diversity that is now crashing is not itself an artificial result derived from the changed burning regimes? In other words, is it possible that, under Aboriginal fire management (and hunting?) practices, mammal diversity was low and that the removal or alteration of these practices have…

    Read more
  4. Tim Beshara

    Executive Officer

    What this article (and the following comments) says to me is that we need to shift our management of natural areas to an 'active adaptive management' approach.

    We are not learning fast enough how to manage these areas and this results in a large cost of lost biodiversity and ecological function.

    The current framework for reserve management is barely adaptive management, let alone 'active adaptive management'. Plans of management are more tools for negotiating preservation vs recreation debates…

    Read more
    1. Greg Miles

      Conservation lobbyist

      In reply to Tim Beshara

      Well spotted Tim. And that Plan of Management only mentions the word "endemic" 3 times. And yet Kakadu and the adjacent Commonwealth IPA are home to about 90% of all the endemic plants and animals in the entire NT!

  5. Amber Hooke

    Environmental Scientist

    In my experience the biggest gap is listening and looking at the details. There is a fundamental difference between large scale incendiary burning programs, lighting roadside fires at lunch time in the late dry season and walking hand lit fire programs. Having experienced all of these at varying degrees, I can say that the walking method is an amazing thing to watch. Watching people following the fire as it trickles along the ground, and putting it out when it nears an animals nest or den or a flowering…

    Read more
  6. Greg Miles

    Conservation lobbyist

    Thanks for a very good item. It is great to see this issue get some exposure. It is vitally important in the north. I will take the plunge and offer a couple of ideas on how burning in Kakadu can be changed to benefit the environment. There are ways that fire can be managed differently in Kakadu. If there is a reason that nobody has been standing up and saying what it is - it is because only people with highly detailed knowledge of the landscape can do that. It is not possible for "conservation…

    Read more
  7. Black Knight


    Given the situation is dire, then let's end some of the self-privileging pretences.

    Instead of talking about introducing "large exotic herbivores'" we really need to look at the role played in Australian life management by the operation of the full range of First Peoples law.

    Their Ways aim at ensuring at ensuring the orthodox reproduction of a well-tempered cosmos.

    The removal from the ecosystem of the operation of indigenous forms of law - and the forced imposition of law and practices…

    Read more
    1. Greg Miles

      Conservation lobbyist

      In reply to Black Knight

      You are absolutely right Black Knight. Ahh but if only it could be done. In my experience, the damage done to our first peoples is so overwhelming and all encompassing that most of them no longer have the knowledge to reinstate 'how it was' even if they were given all the resources and tools and manpower they asked for. In terms of fire in Kakadu - since the park was declared in 1979 the fire regime has been modelled on traditional Aboriginal burning. But there are no living traditional owners who know how burning worked before the great population crash of the 1900's and before the spread of tens of thousands of landscape altering buffalo across the face of the land. And as you would rightly point out - this was all caused by whitefellas and their goods.

    2. Black Knight


      In reply to Black Knight

      Hi Greg,

      I had the privilege of working with senior Warumungu and Alyawarra lawmen in Central Australia in the 1980s.

      Based on my experiences I am more optimistic than you, I gather. Life has a way of ensuring important messages are kept alive, despite appearances.

      The "seed crop" of traditional land and life management knowledge may still be viable, even though the "crop" has been lost, and a long affirmation drought in place for two centuries.

      We need to direct serious resources to these healing tasks here in this country rather than looking for signs of life in distant galaxies etc.

      With the agreement of senior men, I would love to see some form of land and life management resource centre near (but not in) Tennant Creek, for example, which combines the best of both cultures on regard to caring for country, and mentoring young men.

  8. Caroline Copley


    I wish to point out some things although I have no experience in the tropics and am on a learning curve down here in the mallee about aridity.
    1. Not all small mammals behave the same and it is important to specify approaches of each. Rodents and ground marsupials need cover, so they will benefit from thick vegetation, so post-fire, successional vegetation will suit them, except until it grows back they will be vulnerable. However the reasons the possums are primarily coastal seems to be that…

    Read more