On November 8th, 2018, a downed power line kicked off the Camp Fire, a conflagration that spread over an area twice the size of Salt Lake City and produced a nearly 2.5 mile tall smoke plume that could be seen from space.
The Camp Fire destroyed thousands of homes and killed 85 people; its smoke did damage in less obvious ways. Particulates from the plume traveled across the continent, creating several days worth of smog and haze along its path. Inhaling those toxic particles is a general health hazard, and actively dangerous for those with asthma or other chronic conditions.
As climate change makes these destructive fires more common and more intense, new approaches are necessary to defend against them. Preventative measures can only go so far when a single errant spark can unleash chaos across the countryside; public health and safety responders need better tools to predict what a wildfire will do next so they can keep people out of harm’s way.
For Heather Holmes, professor in the Department of Chemical Engineering, a key component of that defense is hiding in the smoke.
She and her colleagues are providing fresh insight into the behavior of these smoke plumes, using new modeling techniques and data sources to help predict how they will spread. It’s a personal issue for Holmes; the Salt Lake Valley is on the front lines of wildfire country, and its bowl is the perfect shape for collecting atmospheric pollution.
“We have had great success in improving air quality in the U.S. since the 1970s but smoke from the large wildfires in recent decades has become a major issue,” says Holmes. “During these events, we have the worst air quality in the world.”
High AQI numbers and low visibility aren’t just annoyances, either.
“Breathing in wildfire smoke can be as harmful to your lungs as smoking cigarettes,” she says. “Warning vulnerable populations is the best defense we have, so getting a better understanding of smoke behavior is critical.”
Holmes and Jingting Huang, a former graduate research assistant in her lab, now a postdoctoral researcher at the University of Georgia, recently published a study describing their approach in the journal Atmospheric Chemistry and Physics.
“Knowing the height of the smoke plume generated from a fire at the initial stage is important for predicting the effect on air quality,” says Huang. “But it’s tough to get right because models and satellites are not always able to capture the small-scale details of the fire and smoke behavior.”
Predicting how smoke from a wildfire will spread starts by measuring the height of the initial plume; the higher smoke particles rise, the more they can laterally spread before falling back down to Earth.
But what may seem like a simple measurement can quickly become complicated. Each wildfire is unique in terms of size and strength, but also in the moment-to-moment weather patterns surrounding the blaze, and the topography of the land over which the plume travels.
The researchers’ latest study details how combining two types of datasets will enable them to develop a statistical data fusion model that can account for such factors. The key to the researchers’ model involves comparing two concurrent images of a given plume: one from above, and one from below.
Given the intensity of the 2018 wildfire season, several research flights were undertaken to get up-close images and other sensor data by traveling directly through the plumes. The researchers utilized laser-based LIDAR measurements to more accurately estimate plume height in a given location. Combined with spaceborne observations from recently released high-definition satellite datasets, the researchers’ findings illustrate how these two data streams can be used in new models that are more widely applicable to simulating the behavior of wildfire smoke.
In a hypothetical scenario, consider a small town near a forest in a mountainous region. During a particularly dry summer, a wildfire starts in the forest and quickly grows due to strong winds. A better plume model, with accurate smoke height and dispersion predictions, could forecast the path and concentration of smoke reaching the town.
Unlike current models, which might give a general idea of the smoke spread, this advanced model would provide detailed hourly updates on smoke levels, factoring in the complex interactions of topography, weather, and fire behavior. This would allow residents to see not only when the smoke will arrive but also how dense it will be and how long it is likely to linger.
By combining satellite data and climate models, local health officials could issue timely advisories for vulnerable populations, such as the elderly, children, and those with respiratory conditions, to stay indoors or evacuate if necessary before the worst smoke conditions arrive.
“The satellite data from this paper can be used in models that improve smoke prediction. Firefighting resources could be more effectively deployed by predicting the fire’s behavior and smoke spread, allowing authorities to prioritize areas at greatest risk,” says Huang. “Accurate data on smoke plume behaviors and their impact on air quality and climate can even inform more effective land management practices, urban planning, and environmental policies to mitigate the effects of future wildfires.”