Introduced grasses: triumph or Trojan horse?

David Bowman of the Northern Territory University argues that one of the most profound threats to the tropical savannas is developing under our noses yet little research is being carried out on the issue.

Gamba grass produces robust tussocks that can reach a height of 4 metres, producing fuel loads for fires four to 10 times that of native savanna grasses. By changing fire regimes, the grass may be irrevocably transforming the landscapes of northern Australia. Photo: Sam Setterfield

Over the past 50 years agronomists have undertaken a massive evolutionary experiment in northern Australia by introducing more than 450 grasses and legumes in the hope of discovering species that improve cattle production. Species from distant parts of the world have been brought into contact with north Australian savanna ecosystems in a flash of evolutionary time.

In 1994 Mark Londsdale reviewed the consequences of this program in a paper called Inviting Trouble ¹ . The reason for this provocative title was that the agronomy project has backfired badly.

He showed that of the 463 species that had been introduced between 1947 and 1985 only four species were found to be useful and had not become weeds while 60 species had become listed as weeds and 17 were listed as weeds despite being identified as being useful. Even useful exotic pasture plants come with a considerable risk that they will become environmental weeds.

Playing with fire

There are few ecosystems in northern Australia that are weed-free, so a reasonable person may accept the escape of a pasture plant as the price paid for the economic utilisation of north Australian savannas. However, this sanguine view overlooks the dramatic, irreversible ecological consequences that some exotic grass species can have on fire regimes. These consequences have been ably reviewed by the highly respected north American scientists Carla D'Antonio and Peter Vitousek ² .

They concluded that of all the ecological changes caused by invading grasses the most significant is due to the interaction between grass and fire. Grasslands are highly flammable given the abundance of quick-drying and well-aerated fine fuels. Further, perennial grasses can recover rapidly from fire because of underground buds.

The real danger comes from exotic fire-loving grasses that can trigger fires that are more intense, frequent and widespread. The cycle would unfold as follows: firstly the more intense fire feeds off the exotic grass and destroys many of the woody plants, which in turn increases the dominance of the following year's exotic grasses that suppress more native herbs and grasses. Eventually this grass-fire cycle can convert a diverse habitat with many different species and forms of plants to a grassland dominated by a few exotic grasses with little capacity to be recolonised by native species.

Such changes can also cause drier microclimates, further adding momentum to the grass-fire cycle. Furthermore, because some elements, such as carbon and nitrogen, are volatilised and lost in smoke while other nutrients, such as phosphorus, are made more chemically mobile and thus susceptible to leaching, nutrient cycles are disrupted with the consequent decline in overall stored nutrients for plants. These changes further reinforce the grass-fire cycle because the fire-loving grasses thrive on the temporary increased availability of soil nutrients.

Gamba grass threat

D'Antonio and Vitousek's model is relevant to understanding the future of northern Australian savannas. A number of exotic grass species are aggressively invading native vegetation here with consequent changes in fire regimes. Examples include parra grass ( Brachiaria mutica ) on freshwater floodplains, mission grass ( Pennisetum polystachion ) and gamba grass ( Andropogen gayanus ) in eucalypt savannas.

Since the early 1990s there has been a growing consensus among biologists and managers that gamba grass poses by far the biggest threat ³ . Why? The primary reason concerns the structure and size of this plant and how this influences wildfires. Gamba grass has robust tussocks that reach a height of 4 metres.

High densities of this grass produce standing crops of between 10 to 20 tonnes per hectare, four to 10 times more fuel than native tropical grasses4. Unlike the other invasive exotic grasses there are no equivalents to gamba grass in northern Australian savannas ⁵ .

The lifecycle of gamba grass is also different to native grass species. It flowers in the early dry season and continues to grow well into the mid-dry season, presumably reflecting extremely efficient use of water and root systems that can exploit moisture in the subsoil. This leads to a dangerous combination: great height and above-ground biomass in the late dry season. This creates fires for which there is no previous parallel. Indeed gamba can behave more like a short flammable tree than a grass; flames of these fires can reach the leafy canopy of eucalypts and defoliate or kill them. Gamba grass fires can be so intense that the efforts of fire fighters to contain blazes are thwarted.

The Bushfire Council of the NT considers the grass a major threat to public safety and in conjunction with other government departments has initiated a campaign urging landowners in the rural areas outside Darwin to control gamba grass by destroying infestations on their properties ⁶ .

Gamba grass has the potential to invade all tall tropical grass savannas in northern Australia given its broad ecological toleranceiv. It is able to disperse and establish itself tin uncleared native vegetation from nearby areas such as pastoral lands and unintentional infestations on roadsides and other areas of disturbance. Because gamba grass jumps property boundaries and hitches rides it is merely a matter of time before it is able to realise its geographic potential.

In sum, we have unwittingly pitted fire-tolerant northern Australian eucalypt trees against an extremely fire-adapted west African grass in a potentially fierce evolutionary struggle.

Garry Cook of CSIRO Wildlife & Ecology, is of the view that gamba grass and its associated intense, late dry-season fires have the potential to radically transform north Australian eucalyptus savannas, by causing the loss of biodiversity dependent upon a continuous tree and shrub layer6. Eucalypt savannas could become fire-promoting weed infestations.

Existing data on the effect of invasive grasses in northern Australia is remarkably sparse given their scope for fundamental and irreparable ecological change. Without more research it is difficult to know how best to tackle the problem of invading grasses generally, and gamba grass specifically.

Equally, it is difficult to know if a species actually has the capacity to destroy the tall-grass savanna ecosystem in northern Australia as has been suggested by Gary Cook ⁶ .

Given the current wait-and-see approach of land managers we will have an answer to this great evolutionary experiment in the next few decades.

Dr David Bowman is Principal Research Fellow Northern Territory University, Darwin.

Click here to view the reply Cut and dried: the issue is to manage disturbance

References

1. Londsdale W.M. (1994) Inviting Trouble: introduced pasture species in Northern Australia, Australian Journal of Ecology (19), pp 345-354.

2. D'Antonio, C.M. & Vitousek, P.M. (1992) Biological invasions by exotic grasses, the grass/fire cycle, and global change, Annual Review of Ecology and Systematics, (23), pp 63-87.

3. Cook, G. (1991) Gamba grass: impending doom for Top End Savannas? CSIRO TERC Newsletter 91 (12), p 2.

4. Barrow, P. (1995) The ecology and management of gamba grass (Adropogon gayanus Kunth.) Final report to the Australian Nature Conservation Agency, NT DPIF, Unpublished report.

5. Russell Anderson, CEO Bushfires Council of the NT, personal communication, May 1999.

6. Cook, G. (1997) A tale of two savannas: the effects of fire on vegetation patterns in West Africa and the Northern Territory.

Bushfire '97: Proceedings of the Australian Bushfire Conference 8-10 July 1997. CSIRO, Winellie and NTUPrint, Darwin, pp 45-50.