Why Australia’s climate shock needs a ‘Moon shot’ | Cosmos Weekly Taster

This article on Fire Shield and bushfire prediction first appeared in Cosmos Weekly on 24 September 2021. For more stories like this, subscribe to Cosmos Weekly.

The incredibly ambitious Apollo Moon shot program only succeeded because it had the backing of the American public.

It was a time of crisis. The Cold War was on the brink of turning hot. And there was a pressing need to prove something in what was a dramatic clash of civilisations.

Now the Minderoo Foundation philanthropic organisation has a ‘Moon shot’ for Australia, and it addresses another crisis.

It’s the need to find bushfires as quickly as possible, even as they ignite – and extinguishing them within the first hour.

Minderoo wants that ability within five years.

Space, Minderoo’s Adrian Turner says, is central to the race against the impacts of climate change. “If we pursue a civilian space program like Apollo, disasters have to be one of – if not the – primary focal points. As it cuts across the economy and the society and the environment, the whole country will come together behind it.

“For the sheer loss of life, economic impact alone – it’s got to be one of the top-tier focal points for any modern civilian space program. And we’re being very vocal about that.”

Portrait of adrian turner in a minderoo shirt
Adrian Turner: Credit: Minderoo Foundation

Turner is CEO of the Minderoo Foundation’s Fire and Flood Resilience Initiative. And for him, this stuff is personal.

“I walked out of a situation that we shouldn’t have walked out of,” he says.

In 2020, three days before he was due to talk to businessman and chair of the Minderoo Foundation Andrew Forrest on the subject of disasters, he experienced one himself.

Turner was with his two brothers and three others who had come to protect their property from a bushfire. It quickly turned into a firestorm. Soon three neighbours arrived seeking refuge.

“Being caught in those fires in New South Wales’ Kangaroo Valley, where three fronts converged, it was just a six-and-a-half-hour fight for our lives,” he says. “The next day, the RFS turned up and said, ‘we came expecting to find nine bodies’. It’s a life-changing experience”.

Now he’s determined to send Australia’s firefighting capability into the stratosphere and beyond.

Building resilience

Turner detailed the Minderoo Foundation’s Fire & Flood Resilience project last week at the 12th Australian Space Forum in Adelaide. In particular, he focused on the space-based Fire Shield component.

“The methodology we’ve decided to use is the Apollo missions’ methodology, where you start with an audacious outcome, quantify the timeframe for achieving it, and then break the problem down using first-order principles.”

It’s not a project, Turner says: it’s a program of work. But defining resilience isn’t simple.

You have to take into account the type of hazard, and exposure and vulnerability to hazards. The resulting Resilience Index identifies Australian regions at extreme risk.

“We haven’t made that public because we’re concerned it’s going to move property prices,” Turner says. Instead, the goal is to focus on how to lift their resilience to similar levels as the rest of the country.

And space is just part of the answer.

“It’s a combination of projects aimed at driving national systemic change combined with place-based interventions,” Turner says.

“We’ve set ourselves a goal of being able to put out any dangerous fire in an hour, and that’s audacious.”

“The first one is Fire Shield. And that’s focused on Earth observation. The second is healthy landscapes. The objective is to halve hazard exposure in Australia’s 50 most fire- and flood-prone regions.

“In the case of resilient communities, which is the vulnerability piece, that’s very place-based as well. And that’s about bringing the 50 most vulnerable communities on par with our 50 most resilient as informed by the Resilience Index.”

Eyes in the sky

The main challenge, Turner says, is evidence demonstrating risk and vulnerability, and also evidence proving that efforts to mitigate risk have tangible effects.

To kick off the resilience project, Minderoo sponsored a competition with DataQuest. First, it gathered satellite images from several fire seasons. Then, machine-learning researchers were let loose to glean as much information as possible from them.

The challenge isn’t just getting the data. It’s using it.

“They found ways to predict fire spread better, gained a better understanding of how a fire becomes an extreme event, and found that smoke plumes could signal the onset of such an event,” Turner says.

Similarly, a Data61 developed SPARK fire simulation tool received Minderoo backing to evolve into a fully operational national asset.

“But those models were less effective in the Black Summer fires because there was so much extra fuel load, and it was so dry that the fires burned with characteristics rarely seen in the past,” Turner adds. “So we need to ground-truth our models using Earth observation and satellite imagery to improve them.”

CSIRO’s SPARK program

When combined with a history of reference points provided by traditional photographs, up-to-the-minute data including multi-spectral imaging and LIDAR and RADAR pictures can paint an accurate threat assessment as fires evolve and move through terrain.

The challenge isn’t just getting the data. It’s using it.

Pulling it all together – and dissecting it through machine learning – results in an astounding depth of detail, encompassing such things as soil moisture levels, undergrowth density, dry fuel concentrations and tree canopy height and health.

Combined, it offers a high-fidelity vision of a landscape’s health.

“We can identify different vegetation types, for example, because they all have different burn characteristics,” Turner says. “It’s about finding the best way to apply finite resources to get ahead of a fire’s spread or make a decision about priority.”

Earth observation

The challenge now is finding the right tools for the job. And there are many different jobs.

The value of satellites was demonstrated during the Black Summer fires. But their availability was limited, and the ability to exploit their information was challenged by existing decision-support systems.

Now we have a better understanding of their strengths and weaknesses.

Satellite map with fire probability overlay in colour across the landscape
Probability of fire spread and intensity, based on SPARK’s calculations. Credit: Minderoo / CSIRO / Data61

Geostationary satellites don’t have high enough resolution. But they do offer the broad overview needed to establish necessary context. Sensors carried by near-Earth satellites and stratospheric balloons can provide the high-resolution needed for specific threat response. But getting them where they are required is problematic.

Turner says Minderoo is in talks with an international company that already operates its own constellation of low-earth-orbit (LEO) satellites. It’s also in discussions with the owner of a stratospheric payload platform.

“We’ll be testing different sensors and payloads against a set of criteria,” he says.

“What we really want to do is detect ignition points earlier. We even want to get ahead of the risk by predicting areas exposed to lightning strikes. When combined with a deep understanding of landscapes, we can get ahead of the curve by pre-positioning firefighting assets close to high-risk areas.”

The challenges are immense.

“What it’s doing is forcing non-conventional thinking. And you won’t get there in increments from where we are today.”

Not only must the entirety of Australia be scanned at both broad and high resolutions, but every stakeholder must also be engaged. They must speak a common language. And use standard measures. What’s the difference between a creek, a stream or a brook? Does that weed have the same name across the various states?

“We need common labels for things if we are to measure them in a structured way so that machines can interpret that data accurately,” Turner says.

Final countdown

“I feel it’s on track,” Turner says of the Moon-shot objective. “I mean, we’ve set ourselves a goal of being able to put out any dangerous fire in an hour, and that’s audacious. And I don’t know if we get all the way to that objective. But I do know that we’ll make a lot of progress from where we are today.”

Turner says this isn’t pie-in-the-sky thinking.

“We stand by that hour goal as the right aspiration,” he says. “What it’s doing is forcing non-conventional thinking. And you won’t get there in increments from where we are today.”

Computer-generated image of bushfire spreading through environment
A visualisation of SPARK’s bushfire prediction model. Credit: Minderoo / CSIRO / Data61.

He believes tangible demonstrations of the new technology and systems will begin within the next few months.

“I expect we will have demonstrated that we can detect early ignition from space with a high resolution. That we can understand fuel-load – also at high resolution. That we have predictive models that can take into account extreme and irregular weather conditions. That we can have high-resolution fire perimeter data faster. That we can see ground-based command systems are using that information to optimise their firefighting decisions.

“We will, I think, also have a stronger capability to be able to do analysis before, during, and after the fire, including remote assessment of damage.”

Meanwhile, the number of heavyweight players backing the project is growing.

“We’ve got Andrew and Nicola Forrest through Minderoo, which is a $2 billion philanthropy that’s pledged $70 million to jumpstart this. We’ve got 55 partners, including seven of the 11 top ASX companies.”

Big reinsurers have been donating the time of their best data scientists. A rear admiral is heading up the missions’ partnerships, including government involvement as well as ground operations. And non-profit and industry commitments continue to grow.

“The only way to solve this is through a public-private partnership,” Turner says. “And this missions’ model, we think, is really the only model that will drive the necessary system change.”


This article first appeared in Cosmos Weekly on 24 September 2021. To see more in-depth stories like this, subscribe today and get access to our weekly e-publication, plus access to all back issues of Cosmos Weekly.

Please login to favourite this article.