By Angus Dalton
NSW RFS Commissioner Rob Rogers and Assistant Commissioner Ben Millington demonstrate the Athena system at their headquarters in Sydney.Credit: Nick Moir
Save articles for later
Add articles to your saved list and come back to them any time.
Named for the Greek goddess of war, nourished by data from the CSIRO’s Black Mountain fire lab and propelled by AI, a new weapon lies in the headquarters of the authorities presiding over a state primed to burn.
The new fire modelling system Athena maps and predicts where bushfires are likely to spread, and automatically detects the lives and properties that may stand in a fire’s destructive path.
On Sunday, the start of the bushfire season, Athena’s map flashed with rippling purple pixels as the system simulated more than 85 blazes ripping through grass and bushland sapped by the hottest September on record.
After a trial run last year, the NSW Rural Fire Service is firing up a fully operational Athena for the first time to prepare for the worst bushfire season since Black Summer of 2019-20. At times during that season, 200 bushfires seared the state at the same time.
Athena’s ability to rapidly evaluate a fire’s severity, behaviour and threat to homes is designed to help the NSW Rural Fire Service decide where to deploy trucks, aircraft and personnel on those horror bushfire days.
The system highlights houses, nursing homes, schools and critical infrastructure such as powerlines at risk, and automatically ranks which bushfires require critical attention.
“There’s nothing like this, certainly in Australia,” NSW Rural Fire Service Assistant Commissioner Ben Millington said at their state operations centre.
“On a recent visit to the US we worked with the US Forest Service and Cal Fire and showed them a demonstration of this system. They were really impressed – they don’t have anything like this available to them.”
Athena also tracks firefighter trucks and aircraft while highlighting water sources, such as backyard pools, where air tankers can replenish their stores for firebombing.
Weather data, vegetation maps, fire plots from crews on the ground, infrared images from line scanner aircraft and fire behaviour models developed by the CSIRO all feed into Athena. The system will continue to learn and recalibrate its prediction power as fires spark across the state.
The Pyrotron: Wrangling data from the flames
Athena’s fire prediction is partly underpinned by the CSIRO’s fire behaviour model Spark, which uses data gleaned from Dr Andrew Sullivan’s decades of controlled incineration experiments.
Sullivan, the leader of the CSIRO Bushfire Behaviour and Risks team, has spent decades running fire trials to test how wind, moisture and vegetation type affect how a lick of flame can roar into a deadly blaze.
Many of those have taken place in the Pyrotron, a 29-metre steel tunnel with a two-tonne wind turbine that can hold batches of burning bushfire fuel under controlled circumstances to extract sterile data from one of nature’s most chaotic forces.
The Pyrotron wind tunnel is the centrepiece of CSIRO’s National Bushfire Behaviour Research Laboratory.Credit: Rhett Wyman.
Sullivan’s lab also runs field experiments. Those studies have yielded the fire danger rating system, used until its first overhaul in half a century last year, and made breakthroughs that revealed experts had drastically underestimated how rapidly large bushfires spread compared to smaller blazes.
The fire lab also developed a rule of thumb firefighters could use on the ground to roughly calculate the speed of fire spread.
“You can approximate the spread of the fire as being a fraction of the open wind speed,” Sullivan said. “In forests that’s about 10 per cent of the open wind speed, and in grass it’s 20 per cent.”
Decades of CSIRO fire field studies have helped shape prediction models and danger ratings, but variable weather conditions pose a challenge.Credit: CSIRO
Despite these findings, field studies on fire are hard to replicate because scientists cannot control the weather. Varying wind, rain and humidity dashes the chance to keep variables steady in the field, a problem the Pyrotron solves.
“It’s about providing a way to study the combustion of bushfire fuels in a safe and repeatable manner that builds statistical robustness in our results,” Sullivan said.
The Pyrtron is loaded with fuel with controlled variables, including moisture content, windspeed and leaflitter layout.Credit: Rhettt Wyman
With the Pyrotron, the fire lab has helped pinpoint the different ways fire behaves in eucalypt forests, wheat fields, paddocks and pine plantations, boosting researchers’ capacity to simulate bushfire spread with mathematical models.
“The fastest fire we’ve had in the Pyrotron over a five-metre length was 13 seconds, go to whoa. It was very, very fast and the wind speed wasn’t overly quick,” Sullivan said. “It was a bit of a blow because it was in standing grass that we’d been growing for months, and it was over and done with in a blink of an eye. Together, the field work and the laboratory work help us build the foundational knowledge necessary to improve our operational models that people like the RFS use.”
Fire behaviour predictor’s first time in the field
Dr Mahesh Prakash has overseen the CSIRO’s bushfire simulator Spark develop over six years, in collaboration with Sullivan’s team.
Once only useful for mocking up theoretical fires or planning hazard reduction burns, the model is now ready to map live bushfires and peer into the future of the flames.
“It’s transitioned from a research tool, to a risk assessment tool, to an operational tool. Now it can be used during a real bushfire. This is the first fire season where the agencies are going to use it in anger,” Prakash said.
To predict what a bushfire might do in 12 hours, Spark considers four main inputs: live weather data from the Bureau of Meteorology, the slope of terrain, the type of vegetation fuelling the fire, and the ignition point.
Nearby fires are drawn towards each other. Here, a Spark simulation (white line) is overlaid on actual fire behaviour in the Pyrotron.Credit: Data61, CSIRO
“It can predict how fast the bushfire is going, where it’s going, the flame height, intensity and radiative [heat] effect as well,” Prakash said. “Think of Google Earth and fire progression across that map. That’s the kind of visualisation we get.”
In January, the South Australian Country Fire Service used Spark to recreate a 45-hectare bushfire in steep terrain in the Adelaide Hills that injured two firefighters, to help analyse the consequences of the blaze.
“About 90 per cent of most of the behaviour that they saw in the actual fire was replicated by Spark,” Prakash said.
After Australia’s driest September and record heat, the senior research scientist is apprehensive about the summer ahead.
“I’m loving [the warm weather], but starting to get concerned,” he said.
“Spark has the ability to optimise where resources go, where they’re most needed. That has the potential to save lives.”
The Examine newsletter explains and analyses science with a rigorous focus on the evidence. Sign up to get it each week.
Most Viewed in National
Source: Read Full Article