The ABC reported that on January 6 and 7, 2016, a dark, dense and rotating plume of smoke grew over the West Australian towns of Waroona and Yarloop.What followed were intense ember attacks, which went on to destroy the town of Yarloop and kill two people.

The extreme behaviours of pyrocumulonimbus clouds and ember attacks have made some of Australia’s biggest bushfires significantly worse — not only due to the volatility of these phenomena, but because they are difficult to predict.

But after years of struggling to predict bushfire conditions, researchers believe they have built a helpful tool which may curb the impact of these events.

Spark Operational, developed by the CSIRO and the National Council for Fire and Emergency Services (AFAC), is a simulator that produces predictions, statistics and visualisations of bushfire spread.https://player.vimeo.com/video/151893991?app_id=122963VIMEOCloud develops over Waroona Fire

Similar to other simulators, it combines specific and localised weather information with topography, fuel loads, vegetation type and on-the-ground fire behaviour information to simulate and predict the path of a fire over six to 12 hours.

But Spark has the ability to easily incorporate new developments in science as they occur.

And with the increasing scale of bushfires in Australia, researchers and firefighters know they need every tool at their disposal to help protect lives and property fromsome of the more extreme fire behaviours of ember attacks and fire-generated thunderstorms.

Ember showers a major threat

Fire-generated thunderstorms form by intense updrafts from the heat rising from the fire — similar to the way a thunderstorm develops.

The danger is these thunderstorms can produce sudden and very powerful downbursts of wind that can quickly whip up fire behaviour, including spot fires and lightning.

Ember attacks are often linked to the powerful downbursts from the pyrocumulonimbus clouds, but have also been known to develop when there are gusty, erratic winds conditions in certain topography.

DFES WA rural fire division executive Mark Bowen said ember showers had caused fires to expand significantly, so predicting their likelihood and spread was important.

“In the past, we’ve had examples of fires in Sawyers Valley heading south-west and then jumping the Mundaring Weir — a whole dam — to start a fire on the other side,” he said.

The platform has been under development for six years, with Phase 1 of the rollout having just begun.

CSIRO Spark project lead Mahesh Prakash said while current fire simulators were helpful, there were limitations to their utility.

Dr Prakash said scientists often struggled to incorporate new prediction research into existing simulators, and had been stymied by the lack of a consistent national platform.

“The main thing is the consistency across all states,” he said.

“There is [also] a lot of scientists in CSIRO and universities and BoM who are working on new fuel models and ember transport models.

“Spark will be able to bring those into operational practice much sooner that the normal process it takes to get it onto operational systems.”

Accuracy of weather forecasts boosted

There are currently ember transport models used in simulators, but Dr Prakash said they relied on empirical data.

A “more robust”, physics-based ember transport model is being trialled on the Spark platform, through collaboration with the Bureau of Meteorology.

DFES WA director of bushfire technical services Jackson Parker said WA’s simulation tool, Aurora, had been a “game-changer” in fire prediction since its introduction a decade ago.

But he said Spark’s additional functionalities — particularly its ember transport model and potential to incorporate atmospheric effects, like fire-generated thunderstorms — would bring benefits.

Mr Parker said having a second modelling option would also strengthen the accuracy of forecasters’ predictions.

Fire storms a research challenge

Reliable models for predicting fire-generated thunderstorms are still in development, with several research organisations trying to better understand the physics behind their occurrence.

Bureau of Meteorology fire, heatwave & air quality team leader Bradley Santos said there were a few factors that made the storms so difficult to predict.

“The first one is that it requires an interaction of the weather and the fire and there’s also the complicating factor of the local variations of the weather due to topography and fuel,” he said.

While research has helped shed light on fire-generated thunderstorms in the past, there continues to be a significant gap around creating a prediction model that can tell if and when they will occur.

Spark will not be able to predict firestorms until there is sufficient research to support it, but Dr Prakash said the simulator would eventually be able to incorporate it — something other models could not do as quickly and easily.

“You’ve got research projects that take three to five years to reach a point where it can be operationalised, or in some cases, that research just stays there,” he said.

“Whereas now what we’ve got is a system that self translates that research into something that can be operationalised.”

The technology’s development has received funding from the Minderoo Foundation and is expected to become fully operational, with its additional functionalities, in the next two-to-three years.

Spark 2

The CSIRO have annoujnced the advent of the next generation of Spark:

https://research.csiro.au/spark/version-2-of-spark-is-coming/

This new version of Spark will be built on the Geostack platform for even more flexibility and modularity over the current system. New Spark will have web-based functionality and will be cloud-enabled.

If you wish to doenload the existing version you can do so on this link