Alan McArthur set up the framework for Australia's fire protection system and gave us confidence to predict bushfire behavior, to defend against it, to suppress it and to conduct controlled burns. He observed that temperature, relative humidity, rainfall effects and the process of dew formation influenced fuel moisture content of dead fuel. He said that a litter layer 75mm deep could absorb around 12- to 25mm of rainfall, and that the retained portion is subsequently evaporated away.

But when he discussed the effect of rainfall in detail, he applied it to the concept of fuel availability; the main effect of significant rainfall lies in reduction of fuel available for burning and incomplete combustion in the moist lower layer, due to smothering effect caused by water vapor. Strangely, he made no linkage of rainfall effects on fuel moisture content. Instead, he attributed the FMC changes to variations in atmospheric dryness. He then linked daily moisture content changes to fuel availability. Harry Luke and McArthur stated that when rain-wetted fuel has evaporated to reach around 15%, the diurnal adsorption/desorption process becomes critical in assessing fuel availability.

QUANTITY AND QUALITY

This statement leads us to another line of inquiry. Because the equilibrium moisture content varies from 3% to 19%, the implication is that a greater percentage of fuel burns at 3% than at 19%. However, if the drought factor is held constant, the McArthur Meter shows that fire behavior is more severe at 3%, but the range is explained by fuel moisture content differences, not fuel availability differences.

Did McArthur believe that the effect of recent rainfall is the same as a change in fuel moisture content? To answer this question we compare a revised drought factor derivation meter to Luke and McArthur's original (opposite). We can use the following conditions: drought index of 100, fine fuel load of 20 tons per hectare, temperature of 400°c, relative humidty of 10%, wind at 10 meters at 22kph.

Luke and McArthur's figure relates to a fuel load of 20t/ha. Using 22kph wind speed, rate of spread ranges from 1.5kph at 3% fuel moisture content to virtually zero at 19%.

If we hold atmospheric conditions dry — 400°c, 10% relative humidity — so that equilibrium moisture content is 3%, 20mm of recent rainfall causes drought factor to move from two to 10 in five days. After this rain, the litter bed is virtually saturated and self-extinguishing, let's say 15% to 20% FMC. On Day 1, when DF = 2, predicted rate of spread calculates at 0.2 × 1.5 = 0.3kph. This is equivalent to 6% FMC on the Luke/McArthur meter. Day 2 calculates at 0.5 × 1.5 = 0.75kph, but this relates to 4% fuel moisture content on that meter.

McArthur's logic is as follows: If the fuel load is 20t/ha, a drought factor of 2 means that 4t/ha (0.2 × 20) of the fuel load is available for burning at 3% FMC. Is this the top layer? If so, would it not travel at the same rate of spread as a standalone 4t/ha load at 3% FMC when the drought factor is 10? The meter agrees. In other words, one day after 20mm of rain, the top layer has dried out to 3% and a fire will run at around 0.3kph.

This means that McArthur believed that when rainfall reduced fuel availability, it had a different influence on fire behavior than fuel moisture content alone. The logic is difficult to explain.

Let's try to understand the derivation of this thinking. Unburned fuel was obviously observed after fires in higher moisture content fuel beds, but not in lower moisture content beds. The reason was ascribed to FMC differences. But the following explanation is also plausible. The flame phase clearly had different behavior in fuel beds of different moisture content. When they came to measure fuel load after the fire had extinguished, they found no fuel left in the drier beds (because the smolder phase consumed everything), but unburnt fuel in the moister beds. Their original conclusion was wrong because they had not separated the effect of the two flame phases. This omission is still apparent in recent research work and leads to incorrect conclusions.

There is also a leap of logic in McArthur's thinking. When wet litter dries out to around 16% to 20%, it comes under the influence of the daily atmospheric cycle. This may be so, but the fuel particles are continuing to dry to reach equilibrium with the atmosphere. But McArthur presents the daily drying cycle for very dry fuels in midsummer in his equilibrium moisture content theory. How does fuel move from 16% to 20% wetness to the very dry condition of the EMC table? Presumably via the DF scale. But as we have seen, they are not equivalent in effect.

What is the effect of the DF scale on fire behavior predictions? It probably tends to overestimate. Is that a bad thing? Probably not. Is it based on poor quality science? Yes. Does it help us in our search for the truth? No.

In hindsight, differentiation of rainfall and atmospheric dryness defies logic. When rain falls onto a fuel particle, the surface absorbs some moisture — the more rainfall, the more moisture that is absorbed by the surface layer, up to its saturation point. Rain causes the fuel moisture content of the particle and of the whole fuel bed to increase above its equilibrium moisture content for a given temperature and relative humidity. Rain in the previous 24 hours has been related to increase in fuel moisture content. With each rain-free day, the surface layers progressively evaporate, as fuel moisture content tends to return to its equilibrium.

FUEL LOAD

McArthur found that in very dry fuel beds, rate of spread is very sensitive to fuel moisture content. At constant wind speed and for FMC up to 7%, reducing moisture content by 2% will double the rate of spread. Firefighters need to be alert to this, because it can happen within a few hours. Rate of spread is inversely related to fuel moisture content to the power 2 to 2.5.

In comparison, for jarrah litter, rate of spread was inversely proportional to fuel moisture in the lab and to the power 1.5 in the field.

Logic says that fuel moisture content has a strong influence on flame height, which is related to the net heat in the flame mass. Higher-moisture fuel firstly absorbs heat during evaporation and secondly increases water vapor content in the flame mass, which tends to absorb radiation and cool the flame. However, McArthur suggested that the moister the fuel, the less the load that is consumed in the flame stage and therefore the smaller is the flame height. This explanation oversimplifies the true situation.

McArthur's early work confidently related rate of spread to fuel load in eucalypt-litter fuel beds, despite contemporary research that showed no such correlation in litter beds. Later it was determined that rate of spread is generally proportional to fuel load, but differences have been observed at the extremes. For example, for very fine fuels like grasses and reeds, rate of spread increases up to 1.5 times fuel load, and in very large or densely packed fuel beds, rate of spread is not affected by fuel load changes. On the other hand, Rothermel's rate of spread model is based on wind speed and rate of weight loss, rather than fuel loading per se. Rate of weight loss (dw/dt) is directly related to energy release rate (Er). In the field, the fire's rate of spread is determined from tables by adjusting fuel moisture, wind and slope, not fuel load.

Hard evidence supporting the link between fuel load and rate of spread is difficult to find, even in McArthur's work. McArthur presents only two examples in support of the statement that rate of spread in a homogeneous fuel bed is proportional to quantity of available fuel, and both are flimsy. It is amazing that the validity of this evidence has never been questioned.

1. Jarrah litter

   ROS measured at 24 minutes after start of nine simultaneous fires.

   Litter load (t/ha)	ROS (m/hr)
   2.5	12
   4.8	24
   9.6	48

2. Dry sclerophyll forest

   Two fires, same weather and slope.

   • Fuel load = 8t/ha, ROS = 14 m/hr, flame height = 0.3m
   • Fuel load = 16t/ha, ROS = 35m/hr, flame height = 1m

Other evidence shows a linear increase of rate of spread as fuel load increases in jarrah litter. But the maximum rate of spread is only 0.16mph, a low-intensity fire.

A critical question has been overlooked for each example. Does available litter load mean total fine fuel load measured on site beforehand, or quantity consumed by the flame phase or by the flame plus smolder phases? Until this is answered, the evidence is invalid, and McArthur's linear relationship between fuel load and rate of spread is unsupported.

Most early research work was done at low to moderate fire intensities. Intuitively, we can see that in a given fuel bed at zero speed, a higher fuel load will produce a larger flame and more heat and therefore tends to spread faster than in a lower fuel load. When wind is added to the system, it tends to take charge of rate of spread, and changes in fuel load become irrelevant. Slope can have a similar overriding effect on rate of spread.

Lately, the influence of fuel loading on rate of spread has been found to be minor. A recent Canadian study found that fuel load consumption had a minor effect on rate of spread, but only at low wind speeds.

HISTORICAL SIGNIFICANCE

Of the three factors that influence fire behavior, fuel load has long been believed to be the only one that forest managers have control over. McArthur developed a technique for safe control burning in the 1960s, and it was accepted because reducing fuel load caused a reduction in rate of spread and in fire intensity. It was a legitimate prevention measure that was quantified in Byram's intensity equation I = HWR, where halving the fuel load resulted in quartering of fire intensity. But as we saw above, fuel load is not proportional to rate of spread. This leads to the following question.

What changes in fire behavior are expected when fire travels from high fuel load to low fuel load? It has long been argued that prescribed burning is justifiable partly because fuel load was believed to be proportional to rate of spread. It has been shown in a number of cases that fuel reduction burning had a major role in stopping the spread of serious wildfires. They reported that FRB reduced the intensity of the fires or “stopped” the fires, and this allowed suppression to occur, but they did not describe the mechanism of fire behavior change.

The following notes summarize what happens to a fire that progresses from a high fuel load to a low fuel load.

1. Intensity reduces because of lower fuel load.

2. Rate of spread reduces due to factors other than fuel load per se.

   • If fuel load reduces but wind speed at fuel level does not change, rate of spread will not change, all other factors being constant.

   • If crown fire or shrub fire becomes a litter fire, rate of spread slows because of lower wind speed at ground level.

   • There is a remote possibility that rate of spread could reduce because of convection drawback by main fire.

3. Residence time reduces because of lower fuel load.

4. Flame depth reduces because of lower fuel load.

5. Flame height reduces because of lower fuel load.

6. Spotting from ground fuel reduces / nil.

7. Spotting from tree trunks reduces / nil.

8. Spot fire ignition in fuel-reduction burning area reduces potential.

Fire behavior moderates in fuel-reduced areas. The main benefits of fuel-reduction burning are lower flame height and less spotting activity. If rate of spread reduces, it is a bonus.

Denis O'Bryan graduated from Creswick School of Forestry and Melbourne University and began work as a forester. He has had more than 20 years of experience working in the Victoria government in fire protection at all levels, including hands-on firefighting, fire crew leadership, pre-season planning and preparation, training, statewide fire protection planning, and coordination of statewide firefighting operations. After leaving the government, he trained large numbers of volunteer and professional firefighters and tertiary students in basic and advanced fire courses. O'Bryan has consulted to the Victoria and federal governments in fire policy and planning issues, and has worked interstate. He currently serves as director of Red Eagle, a bushfire protection advisory service that provides objective fire risk assessment and management expertise and training.

Definitions are important

Available fuel load = fuel consumed by the flame. By definition, the quantity consumed can't be known accurately until the flame has passed over.

The definition of available fuel in Australia has been vague. It refers to the quantity of fuel that will burn or the proportion of fine fuel that will burn in a fire, but has never identified whether the fuel burns in the flame phase or the flame and smolder phase.

Fine fuel carries the flame. Fine fuel is defined as up to 6mm diameter in eastern Australian states and up to 10mm in western Australia. Fuel load estimates usually refer to total fine fuel load on site.