How NASA Is Using Machine Learning to Predict and Fight Wildfires

In dry places like California, wildfire season is known to occur during the warmer, drier months. But with climate change, things shifted. “What was once a four-month fire season now lasts six to eight months,” according to US Department of Agriculture. Which is worse: “Wildfire is year-round for much of the United States,” introducing the concept of “fire year.”

This sobering reality has inspired NASA to use its wealth of Earth observation data to help wildland fire management agencies predict and extinguish fires through the state. Much of this data comes from the agency Landsat satellites, which capture images of Earth’s terrain. They can also measure surface temperature and measuring natural materials that serve as wildfire fuels, such as dense trees or other vegetation that can ignite and fuel fires.

“We have an archive of satellite image data going back 40 years,” Christopher Potter, a researcher at NASA Ames, tells me as we sit in a room on the space agency’s Silicon Valley campus. “This gives us a very rich data set from which to build predictive models.”

As the forest fires become more and more frequent and severe, emerging technologies such as artificial intelligence and drones are being used to help fight fires and keep responders safe and informed. The California Air National Guard, for example, uses drones to track where the fires are spreading and to collect images in real time of disasters as well as pinpointing areas in need of resources. US respondents use software to predicted the movement of fire several days into the future, helping firefighters deal with these fires more effectively. And startups like Pano AI are using cameras paired with algorithms to place forest fires and warn customers, such as utility companies, before fires get too big. NASA, for its part, uses its wealth of data to build predictive models as well as airspace management tools to help drone pilots and agencies respond more effectively to disasters.

Along with the Landsat data, NASA also pulls data collected by agencies such as the California Department of Forestry and Fire Protection, which track the severity of fires from past years, including fatalities and property destruction. Using machine learning, NASA can quickly create predictive models of wildfires and how they might spread — along with what the resulting damage might be. These predictions can help firefighters delegate their resources.

Potter shows me the East Bay Hills, in the San Francisco Bay Area, as an example. Among the cluster of cities, including Concord, Fremont and Walnut Creek, different shades of red illustrate the density of buildings — and therefore the areas most likely to release huge amounts of hazardous gases if a fire were to break out.

Using this data, along with information on when the last wildfires were in the area, agencies can predict how large and dangerous the fire will be. The East Bay Hills is just one example, but NASA forecast models estimate that more than 247,000 acres and nearly 600,000 households could be in the fire line if this area burns.

“We can now tell you how much pollution will be released from it, what harmful gases will be released and where,” says Potter. The agency can estimate this up to several months in advance.

NASA’s wildfire data is available to anyone access online for free. The goal is for agencies like Cal Fire and the U.S. Forest Service to use these more accurate maps of high-risk fires to know immediately where to deploy their equipment as soon as something breaks out. This precision can make all the difference when putting out large fires.

“We’re not looking at entire parts of the state to get a forecast,” Potter says. “We’re looking for a few miles.”

Improving the communication of firefighting aircraft

NASA uses its research not only to predict wildfires, but also to fight them.

Using drones, responders can continue to fight fires even when it is unsafe for traditional aircraft and pilots to fly, such as at night or when there is heavy smoke. They can also perform prescribed burns, which eliminate dead brush that acts as fuel during wildfires without putting crews at risk. But there’s a challenge: It can be difficult for responders to see where these drones are flying.

To solve the problem, NASA is developing airspace management technologies that can help manned aircraft, drone operators and ground crews track and share locations and communicate information with each other. The initiative takes place within NASA’s enhanced capabilities for emergency response operations (ACERO) project. This allows response teams to receive more timely and accurate information and make faster decisions during an emergency response.

“Most communication is by radio, and most wildfires occur where there is no cell service,” said Kathryn Chapman, a researcher at NASA Ames. “It makes it much more difficult for different first responders to talk to each other and even know where the fire is.” … We have this opportunity to start trying to use technology to fill that gap.”

NASA created what it calls mobile air traffic control sets, which can help first responders visualize traffic, including aircraft altitude and location. This makes it safer to deploy drones and ensures that they don’t come into contact with anything else in the sky.

The kit I saw consisted of a yellow plastic container with a lid and handles, about the size of a large briefcase. It was equipped with an ADS-B (Automatic Dependent Surveillance-Broadcast) receiver to track aircraft, an iPad to display the location of local flights and a power plant.

The agency is testing these kits with US Forest Service pilots in Tennessee, Mississippi, Georgia, Florida and South Carolina, who use them during prescribed burns conducted by drones. The kits informed pilots of air traffic in the area and allowed them to more easily navigate the drones and conduct prescribed burns from a safe distance.

These airspace management technologies are still under development and will be tested in a series of low-visibility flight demonstrations with members of the wildland firefighting community. This will help ensure that the technology can work successfully in real-life situations – which, unfortunately, are becoming more common.

“As fires continue to get bigger and more intense,” Chapman says, “we’re giving people the tools to also scale their response.”