BUSHFIRE smoke is a toxic brew that includes carbon monoxide, nitrogen oxides, organic chemicals, black carbon, carbon dioxide, and miniscule particles that can penetrate the lungs and enter the bloodstream. It’s known to trigger asthma and other respiratory problems in exposed populations and, occasionally, can prove lethal.
Until recently, scientists and fire agencies across Australia were limited in their ability to predict exactly what substances were present in plumes from bushfires in different areas, under different conditions.
Predicting which population centres would be hit by plumes presented another challenge. There just wasn’t enough known about the chemistry and physics of smoke plumes from the burning of Australian vegetation.
Then, in February 2006, something unexpected happened.
Fires broke out on Robbins Island, off the northwest coast of Tasmania. An easterly wind happened to blow the smoke 20 km westwards, smack-bang into an array of highly specialised air-quality monitoring equipment set up at Cape Grim Baseline Air Pollution station. The array had been set up for another, ongoing international experiment running at the same time.
The fire may have been bad news for Robbins Island – home to a unique herd of ‘saltwater’ Wagyu beef cattle – but it did have a positive effect: the accidental experiment provided CSIRO scientists with detailed data to validate a smoke-plume forecasting model that was under development.
More accurate modelling
Sarah Lawson wrote up the Robbins Island smoke-monitoring findings for the peer-reviewed journal Atmospheric Chemistry and Physics.
She says that getting this data was a crucial step in validating CSIRO’s chemical transport model, now being used by Bureau of Meteorology and the Victorian Department of Environment, Land, Water and Planning for smoke forecasting.
“To measure all the chemicals in smoke, you need a lot of specialised instruments running,” notes Lawson. “It’s rare for all of them to be running at the same location during or after a bushfire
“From our measurements, we were able to determine emission factors of several pollutants in the smoke – hydrogen, phenol and toluene – that had never before been recorded in an Australian bushfire.
“The most important outcome from that monitoring work has been getting data to improve the accuracy of the smoke-forecasting model. It gave us the opportunity to test the modelling framework for smoke emission rates, smoke transport and smoke chemistry to make the model much more accurate.”
Changing chemistry
Another key feature of the smoke model is that it simulates chemical reactions within a plume.
Previous research has shown that plumes from fires can change rapidly, within a matter of hours. Highly reactive molecules are destroyed, particles coagulate and become more oxygenated, and new compounds form.
“That’s why having accurate information on the different chemicals and pollutants is so important in the model, because the composition of the plume – and the background emissions from cars, factories, and so on – makes a big difference to the chemical reactions occurring within it,” adds Lawson.
The smoke forecasting framework also includes the Bureau’s weather forecasting model to help meteorologists, fire agencies and other prospective users predict how and where smoke and emissions will be transported.
Plumes from the January 2016 Tasmanian fires, for example, travelled as far north as Brisbane, and researchers were able to use satellite data to validate the plume model’s transport component.
It was a reminder that, while the most damaging air pollution occurs close to a fire’s source, smoke plumes can also affect other cities and areas within a region, and even across the globe.
Smoke warning systems for Victoria
Martin Cope is part of a team at the CSIRO Climate Science Centre responsible for developing the CSIRO plume modelling framework.
These researchers are working with a broader team of meteorologists, land managers, ecologists and epidemiologists who, collaboratively, have been shaping the research model into a predictive tool that will enable authorities to issue health warnings around planned burns and bushfires.
Currently, Cope is working with the Bureau of Meteorology and the Victorian Department of the Environment, Land, Water and Planning (DELWP) to develop a system called AQFx.
AQFx is being used by the Bureau for smoke forecasting during bushfire seasons, and by staff at the Victorian State Control Centre responsible for issuing community advice about the DELWP’s prescribed burns.
“The planned burn conundrum is that you want to reduce the risk of catastrophic fires, such as Victoria’s 2009 Black Saturday fires, by using fire to reduce the build-up of flammable fuel loads,” explains Cope.
“But in the process of doing that you generate smoke, which is also a health risk to the population.
“To be able to generate robust smoke forecasts, it’s important to accurately characterise behaviour of smoke emissions and how emissions vary over time, because they change with fuel type and with burning conditions – how dry the fuel is, how fine or coarse the fuel is, the humidity, wind speed, whether there’s been rainfall or not.
“We know from epidemiological studies that smoke can lead to health effects. So what you’re trying to do is to help fire managers balance one risk, the risk of catastrophic fire occurring, against the smoke risk to human health.”
The particles in smoke that presents the biggest risk to human health are very small particles less than 2.5 microns in diameter – much smaller than the diameter of human hair, and so fine, that they can be breathed deep into the lungs and can also pass into the bloodstream, as Cope points out.
“These particles have been shown to increase risk for respiratory and cardiovascular problems. Research done overseas and here shows that fine particles generated in smoke have a health risk factor comparable to fine particles generated by motor vehicles.”
“The challenge for DELWP is that fire managers often have to work within narrow windows of opportunity. They need stable atmospheric conditions with higher levels of fuel moisture, running for two to three days at a time.
“That set of circumstances may only occur a few times a year (typically in autumn), so then they have to try and run as many prescribed burns as they can.”
Detection sensors in schools
DELWP and CSIRO are collaborating on a school-based ‘citizen science’ program in Victoria to help fire managers broaden the smoke monitoring network.
Fabienne Reisen from the Climate Science Centre has developed a curriculum and low-cost smoke sensor package that fits within the STEM curriculum for Grade 6-8 students.
The module shows students in some regional Victorian schools how to set up and program particle sensors, as well as teaching them some background science. These smoke sensor packages are set to be deployed in the next burn season, during autumn 2018.
Cope and his colleagues are also working on a system that will refine 24- to 48-hour forecasting accuracy in response to real-time inputs from satellite and ground-based sensor data. The system will include social media feeds and citizen science measurements.