Kevin Tolhurst is not prone to emotive outbursts. A scientist who has spent his career analysing bushfire behaviour, he has presented the most astounding evidence in the most dispassionate manner to the Bushfires Royal Commission.
But on the afternoon of February 7, when he saw the predictive map of the Kilmore East fire that one of his trainees had produced, he admits he was “horrified”.
The map showed Victoria’s most deadly fire front heading swiftly into the suburbs of Melbourne Eltham, Greensborough, St Helena, Warrandyte; heavily populated, leafy places, well within range of flame and flying ember.
If the prediction had come to pass, parks and gardens would have been alight and houses incinerated, Tolhurst says.
Power to the area would probably have failed and power poles collapsed; cars would have been unable to get out, fire trucks unable to get in. The fire would have raced up wooden fence lines, using them as wicks, setting gardens alight. Burning houses would ignite properties next door, causing a “mosaic” of fires for two days, causing confusion, panic and possibly death.
“Just the number of people that were likely to be impacted by the fire you get wall-to-wall traffic and everything jammed up it’s a paralysis effect,” Tolhurst told The Sunday Age.
That none of this happened was due to luck and meteorology. The prediction was based on a weather forecast of a cool wind change coming through at 9pm. As it happens, it came through earlier, at 6pm, and blew the fire towards Kinglake, where 38 people died.
The Black Saturday fires were what Tolhurst describes as a “blow-up” a combination of long and short-term weather, fuel loads, and the fire itself, which fed on and magnified its ferocity in a way only now being understood.
Long-term drought came first. Next was the late-January heatwave a week of 40-plus temperatures that “cured” the landscape so that even green leaves went crisp and ready to burn. Then came the area’s hottest day on record, 47 degrees, with average winds of 60 km/h.
One spark, and the fire was gone. Vegetation, dried and heated to 70 or 80 degrees in the sun, needed little encouragement from the flames to reach the 280 degrees needed for ignition.
Tolhurst estimates the Kilmore East fire started at 11.30am. That was 19 minutes before it was reported to the Country Fire Authority but, well before this, perhaps 10 minutes after it started, it was already out of control.
Fire-fighting equipment, even aeroplanes dropping water, can suppress fire only when its energy output is 3500 to 4000 kilowatts per metre. Within 10 minutes, this fire was already beyond that energy output and moving too fast to get around it.
At its peak, this fire would reach some 38 times that intensity 150,000 kilowatts per metre: unstoppable.
The next precondition was a trough of low-pressure air that preceded the cold front approaching from the Bight. Low air pressure means there is less resistance to hot air rising through the atmosphere. Along this trough, the hot air from this fire shot upwards, quickly forming a plume, like a chimney, promoting the efficient burning of fire.
At the height of the fire, that plume rose 8.5 kilometres into the sky. It sucked in air to replace the air shooting into the atmosphere and created a powerful wind at ground level, which fanned the fire. At its peak, according to Tolhurst, wind was rushing at 120 km/h towards the fire as strong as a cyclone and enough to snap off mature trees halfway up their trunks and pluck others from the ground.
It was enough also to fling burning, woody sticks and bark from local trees, called “firebrands”, into the sky. Some of these reached high enough to encounter the winds in the upper atmosphere. Some firebrands travelled 35 kilometres, further than ever recorded before, and the extreme heat of the day kept them alight for their entire journey.
But that is just one of three kinds of spotting that Tolhurst describes. The “high-density, short-distance” spotting the blizzard of embers that people caught in the fire have described so vividly is the most obvious.
But medium-distance spotting, of more than one to two kilometres, is a crucial, and so far unrecognised, aspect of fire behaviour. These embers cross highways (such as the Hume, early in the life of the Kilmore fire), lakes and rivers. They start new fires that create an “area of fire” often long before the main fire front arrives.
The junior fires they create are slaves to the huge wind created by the main plume, so these fires may actually burn in the wrong direction against the prevailing wind creating profound confusion and danger on the ground.
Arthur’s Creek fire captain David McGahy told the commission that though the fire was “coming from the north-west you would see pockets coming from the opposite direction”.
“(People) had a fire front coming from three sides (it) just became this monster that enveloped everything and you couldn’t predict where it was going to go.”
The fire front, Tolhurst now believes, does not act like a wave lasting 15 minutes a feature of the traditional model. Instead, this “area of fire” envelops a place for an hour or two in heat so intense that survival is tough, and firefighting impossible.
That is why, despite the average speed of the Kilmore East fire on Black Saturday, about 12 km/h, so many kangaroos were incinerated, even though they can move much faster than that. They were caught in the “area of fire”.
But the fire front itself can, for short periods, move devastatingly quickly. Tolhurst describes it as “pulsing” with flashes and fireballs, often in a circular motion like a hurricane, as the fuel-air mixture bursts into flame well out from the fire front.
Peter Olorenshaw, a policeman and farmer caught with his son watching their house burn down in the Bunyip fire, described the fire front on his property “swirling across the ground, 200 or 300 metres wide”. “The flames were quite random. They were what I would describe as angry and violent.”
The pulses are not just along the ground. Olorenshaw saw them 60 to 80 metres up into the air; Tolhurst says they reached up to 100 metres as wind swirled upwards, and downwards.
As the forest is devoured, the tree sap evaporates and is also pushed upwards, so that about 40 per cent of the fire cloud is water. As that water rises above the dew line, it condenses, becoming a stark white cloud with its own name: pyrocumulus.
The condensation taking place 8.5 kilometres above the ground releases stored heat, adding much more heat in a feedback loop that doubles the total temperature of the fire. This in turn strengthens the plume of air, sucking it like a chimney further into the atmosphere, intensifying the wind at ground level. So powerful was this cloud on Black Saturday that it caused its own thunderstorms.
“These thunderstorms produced lightning, which initiated a new set of fires,” Tolhurst said in his report. “Many of these new fires were in Melbourne’s water catchments.”
But there was one final ingredient the cool change. While people in Melbourne looked forward to the change as welcome relief from the day, it hit these fires from the side, turning these long, cigar-shaped strips pointing north to south into wide ovals of devastation over a much greater area.
“That’s the worst situation you can have,” Tolhurst told the commission.
Despite the cooler air and greater humidity after the change, “the fire then burns quite fiercely for a period of two, three, four hours after that”, causing 80 per cent of the total destruction.
These fires were devastating. Their behaviour was at the extreme end, but they were not freaks. They could, and will, happen again, with the continuation of the drought, the right conditions and another very hot day.
“From our records, we have to say this is a one-in-100-year event. But in terms of the current trends (with climate change) I’d say you could well see the same thing happening in the next 20 years,” Tolhurst told The Sunday Age.