Blog icon

By Chris McKay 14 December 2017 7 min read

Large-scale field experiments with fire can take years of planning.

FOR over 60 years CSIRO has been conducting research into bushfires in Australia. With fire being such a familiar and frequent presence in the Australian landscape, there have been no shortage of wild bushfires during that time to study and learn from. Not all fires are the same, so studying wildfires and the conditions under which they occurred can help us better understand fire behaviour.

The thing about scientists though is that they like to be able to test their hypotheses through experiments. The more variables you can control in an experiment and the more repeatable the experiment and the results are, the more those results can be trusted. When it comes to fire, being able to experimentally control the three major factors that contribute to bushfire behaviour—weather, vegetation and terrain—would give us much more reliable information about fire behaviour under a range of different conditions. This in turn would mean we could design more informed bushfire management solutions and build more accurate tools to predict their spread.

That might sound good in theory but how do you experiment with something as big and wild and seemingly uncontrollable as a bushfire? Well, it’s not easy, and it’s not cheap, and there’s a degree of risk to it, but CSIRO’s bushfire researchers have developed a method for doing just that.

Fire behaviour field experiments

Designing a good fire behaviour experiment is an exercise in control. When you’re dealing with variables like the weather, you need to get pretty clever with your strategy. CSIRO has been involved in large-scale bushfire field experiments since the 1950s and they involve a lot of careful planning. A single experiment like this can cost tens of thousands of dollars, so those involved are understandably keen to get the preparation and outcome absolutely right.

The latest field trial CSIRO was involved in was commissioned by the Queensland Fire and Emergency Services (QFES), was three years in the planning and was a collaborative effort with the Queensland and New South Wales Rural Fire Services and Victoria's Country Fire Authority. QFES engaged the research team to help answer questions about the role fuel load (amount of vegetation) plays in fire behaviour and fire danger in grasslands.

CSIRO’s bushfire behaviour research Team Leader, Andrew Sullivan, outlined the aim: “We hope to quantify exactly the role of fuel load, not just on rate of spread, but also on other aspects of fire behaviour such as flame height and flame residence time that contribute to the danger posed by a fire.

The story of that experiment will give you an appreciation of what’s involved.

The local Rural Fire Brigade were on hand to ensure the fire was contained to the test site and to put the fires out after testing. Experimental reference markers and thermocouple equipment for measuring flame temperature and structure can be seen in the foreground.


Probably the easiest factor to control is terrain. This is obviously defined by the site chosen for the experiment (which is dictated by the required vegetation type) but it’s not easy to find a suitable piece of land that you have permission to set fire to. In this case it needed to be grassland and because the key question the experiment was seeking to address was about fuel load, not terrain, the flatter the terrain the better, removing this as a variable.

Initially, a suitable site was identified in the Northern Territory but in an ironic twist the site was burnt out (twice!) when wildfires went through the area. QFES continued on with the search and eventually at the end of 2016 they found a suitably grassy area on Defence land, near the Amberley Airforce Base in Ipswich, and came to an agreement with the land managers to conduct the experiment there.


With the site chosen, the next phase of the planning focussed on the state of the vegetation that they would be working with when the time came. The experimental design called for grasslands with high or very high fuel loads. This involved planning a layout of plots at the site that maximised regions of heavier fuel loads and minimised regions that consisted of large amounts of weeds or low fuel loads.

A lot of the planning is also about safety and conservation of the site—they needed to keep the experimental site from burning accidentally until they wanted to use it. And when the time came to ignite the experiments it was important to ensure that no fire was able to escape to neighbouring areas. So, the first act was to cut a non-burnable buffer around the whole site area to both protect it and contain it.

Next, they needed to ensure that each experimental fire was containable under the conditions in which they were ignited. In May 2017 slashers were brought in to mark out distinct trial plots. Plots of 33 m x 33 m were laid out with a 3 m break in between them wide enough to allow a fire truck and firefighters to traverse and suppress the fire at the end of each experiment, before it spread to an adjacent plot. The orientation of the plots were such that the dominant prevailing winds could be exploited.

Finally, grass won’t tend to burn well unless it’s dead, so herbicide needed to be applied to ready it for the experiment, ensuring the grass was uniformly and completely cured before the experiments began.


While the team couldn’t control the weather, they could control the timing of the experiment. From the weather records they knew that August to September is historically a dry and windy period in Ipswich, more favourable for the burning required to meet the experimental design, so scheduled the trial for then.

When fire authorities are assessing grassland fire danger level ratings (as in the roadside signs you have probably seen), there are a few major factors that contribute to their assessments: wind speed, air temperature, relative humidity, grass curing state and rainfall. As fuel load was the focus of the experiment, the main consideration in this case was that these weather conditions were in the range where a wildfire could be expected to take place (rainfall is obviously not ideal if you’re trying to start a fire).

The team needed to monitor weather forecasts for the right stretch of days with an optimum set of conditions. And so it was that over a few sunny days in September, with strong winds from north-west or west, at greater than 15 km/h, dead fuel moisture contents (driven by air temperature and relative humidity) below 10%, and air temperature above 25°C, the trials took place.

One of the experimental plots soon after ignition. The fire is contained by fire breaks, buffers and fire crews standing by.

The experiment begins

Local Rural Fire Service crews were onsite to coordinate and apply fire suppression activities—that is, they were there to put the fires out when tests were complete and ensure the fire was contained to the test site. This was a valuable development opportunity for many in the crews, who don’t necessarily get the opportunity to train on a fire under challenging conditions, yet within a controlled test site.

Before any testing began, the crews conducted some additional burning out to increase the buffer around the plot area and ensure there would be no fire escaping from the test site.

Next, measurements were taken by the researchers on each of the plots that had been created to measure the fuel load. Before any fires were lit they placed sensors in and around each experimental plot to capture a rich range of data about the condition in which the fires would burn and the behaviour and spread of the fire. These included anemometers to measure wind speed at 2 m above ground level, thermologgers to measure the time of arrival of fire across each plot and thermocouples to measure flame temperature and structure. In addition, video cameras, both hand held and mounted on 5 and 10 m towers, were used to capture footage of each experiment.

With everything set to go, it was time to light the fires: over five days, 25 experimental fires were successfully completed.

A CSIRO researcher retrieves thermologgers following an experimental burn.

“The data that we captured through burning these plots will help us refine our knowledge of bushfire behaviour in grasslands and how it changes with fuel load,” said Sullivan.

“With these results we will have a better idea of the effect of fuel load on suppression difficulty and also fire danger.”

And it’s not just for science. Through the close relationship between the researchers and fire authorities, this data will be used to inform practical bushfire management strategies, which are constantly undergoing review and being adapted based on the latest research.

“Part of what Rural Fire Brigades do here in Queensland and around the country is not just respond to wildfires but also to mitigate risk,” said Barry Holbron from Queensland Rural Fire Service.

“Part of that system is the Fire Danger rating advice system, which rural brigades manage in their local districts through fire danger signs and other means. So this research we hope will feed into making that science more exact, so that the public can have more faith in the advice that they’re given by us.”

Contact us

Find out how we can help you and your business. Get in touch using the form below and our experts will get in contact soon!

CSIRO will handle your personal information in accordance with the Privacy Act 1988 (Cth) and our Privacy Policy.

First name must be filled in

Surname must be filled in

I am representing *

Please choose an option

Please provide a subject for the enquriy

0 / 100

We'll need to know what you want to contact us about so we can give you an answer

0 / 1900

You shouldn't be able to see this field. Please try again and leave the field blank.