Research Letter #15

FIRE DANGER, GRASS CURING AND FUEL MANAGEMENT

It is often useful to be able compare the present with the past. Is this fire season worse than last, average or less of a problem than usual? It’s a common media role to ask how the fire season is shaping up and whether it is going to be a ‘bad’ one. This letter is about grassy fuels and grassland fire danger and how we can work out whether a season is different from others. It is not futuristic but it asks ‘Do we have to rely on memory for our knowledge of past seasons or is there a better way?’. It also asks whether or not we can alter grassland fire danger.

Grassland Fire Danger is worked out using the Grassland Fire Danger Index (McArthur 1966). This is the basis for the Grassland Fire Danger Ratings (indexes grouped into classes like ‘low’, ‘moderate’ ‘high’ etc.). The inputs required to work out this index (GFDI) are air temperature and relative humidity in a standard weather screen, windspeed at 10m height in the open and the degree of curing of the grassy fuel (‘curing’). Curing is the percentage of the grassy fuel that is dead. The Bureau of Meteorology works out values of GFDI routinely for their fire weather warnings.

The data for curing can come from various sources – observation (looking out the window), formal visual assessment at specified locations (e.g. by checking against a photographic guide), remote sensing (from aircraft or satellites) and computer modelling of pasture. Curing data are rarely archived. Data from satellite imagery can be retrieved retrospectively but are expensive and extend back a maximum of about 20 years. Models of pasture growth and senescence, giving an estimate of curing, can be run for periods as long as the weather data that drive them is available.

Models provide the opportunity to play "what if" games. What would happen to GFDI if we burnt the country every year, if we mowed it, if we grazed it or if we changed the grass species? What would happen if we changed from exotic pastures to native ones? The author and colleagues (Peter H.R. Moore and Richard H.D. McRae) have used models to try to answer these questions for the Canberra region. Here, only the results for fire danger, and its variation according to the type of grass present, are considered.

Two grass models, from a set of pasture models (see Donnelly et al. 1997 and Moore et al. 1997), were run using Canberra weather data. The results of decades of research have been incorporated into these models. How GFDI would be expected to vary for a landscape covered with the exotic grass Phalaris, compared with one covered with the native perennial Danthonia, was investigated. I hasten to add that the results are from models and should not be taken as reality. The challenge now is to check the results against the real world. Some checks have been made in a farming context but no checks have been carried out on the results of the models in the fire context, with one exception. The exception is that in our study we found a good statistical correlation between values from 5 years of formal visual assessment of Phalaris curing using a unique system (R. McRae) with those from the model for the same period.

The results of the models showed that:

  1. the grass species covering the landscape can have a marked effect on the fire danger index (Fig. 1 compared with Fig. 2);
  2. by using a ‘standard’ species in the model (Phalaris), we can see how the fire danger changes substantially from year to year, in this case the numbers of days per year with GFDI greater than 24 (Fig. 3);
  3. by using the average GFDI (or average curing) for each month for a long period (like the 50 years here), the time course of any one year can be traced on a graph so that any year can be compared with the average (e.g. Fig. 4);
  4. having predicted the biomass as well as the curing allows the effect of fuel load on fire intensity to be estimated.

Some of the take-home messages are that: to the extent that curing can be manipulated by altering pasture composition, so too can the GFDI; a series of low GFDI years could lead the unwary into complacency (e.g. the 5 years from 1958-59 in Fig. 3); extreme GFDI values can occur as late as April (Phalaris) (see the top lines in Fig. 1 and Fig. 2); and, the average of the maximum GFDI values per month (the middle lines in Fig. 1 and Fig. 2) are probably more useful as a guide to fire danger variation than the averages of all daily values per month (lowest lines in Fig. 1 and Fig. 2).

Modelled results should not be seen as the end of the process. They provide a means by which ideas can be explored, new research planned and management decisions evaluated.

Literature:

Donnelly, J.R., Moore, A.D. and Freer, M. (1997). GRAZPLAN: Decision support systems for Australian grazing enterprises. I. Overview of the GRAZPLAN project and description of the MetAccess and LambAlive DSS. Agricultural Systems 54, 57-76.

McArthur, A.G. (1966). Weather and grassland fire behaviour. Commonwealth of Australia Forestry and Timber Bureau Leaflet 100, 23p.

Moore, A. D., Donnelly, J.R. and Freer, M. (1997). GRAZPLAN: Decision support systems for Australian grazing enterprises. III Pasture growth and soil moisture submodels, and the GrassGro DSS. Agricultural Systems 55, 535-582.

Malcolm Gill
29 October 1999