AI is ready to help fight wildfires, and the 'WindNinja' is next
share
BOULDER, Colo. — Wildfire, and wildfire control and prevention, has become an increasingly hot-button issue in recent decades.
While the total number of fires occurring annually today is less than it was in the 1990s, fires since 2000 have burned more than double the average annual acreage burned in the 90s, according to a report from the Congressional Research Service last year.
This is due in part to climate change which has prolonged fire seasons and conditioned areas to be more fire prone through rising temperatures, drought and snowmelt, with needed snowpacks melting earlier in the season.
The same report noted that while more wildfires occur in the eastern and central states, wildfires in the West — including in Colorado — tend to burn significantly larger and inflict significantly more damage.
“[Wildfires] are a very complex problem, and we’re trying to develop new tools and technologies for before fires start,” said Zach Tolby, the director and lead scientist behind the Fire Weather Testbed, a fire weather technology accelerator that brings together scientists and sociologists to evaluate the latest in fire weather technology.
“And then after a fire, which is a whole 'nother set of problems,” Tolby said.
The National Oceanic and Atmospheric Administration (NOAA) launched the NOAA Fire Weather Testbed in June 2023 out of NOAA’s Global Systems Laboratory in Boulder to address mounting concerns about the ferocity of wildfires and the need for better research on wildfire technology.
A testbed is what NOAA describes as a “collaborative space,” sometimes physical and sometimes virtual, where researchers and end-users can test and analyze new technologies and systems in real-world scenarios.
NOAA already hosts about a dozen other testbeds working across the country, including the Hydrometeorology Testbed (flooding), the Hazardous Weather Testbed (thunderstorms) and the Hurricane and Ocean Testbed (hurricanes and cyclones).
About 20 evaluators with the Fire Weather Testbed convened in a wide rectangle room featuring four large television screens. They worked through wildfire simulations while implementing the first two technologies to come through the Fire Weather Testbed (brace for a storm of acronyms):
The National Environmental Satellite, Data and Information Service (NESDIS) Next Generation Fire System (NGFS) — otherwise known as the NESDIS Next Generation Fire System — and the National Weather Service’s (NWS) Integrated Warning Team.
NESDIS’s primary responsibility is to operate weather satellites, which play central roles in weather forecasting as well as monitoring and analyzing environmental conditions across the globe.
Circling the globe at around 22,000 feet, these geostationary satellites take new scans of the Earth every minute, sending information to meteorologists in the form of infrared bands that detect fires and track their growth and movement.
However, this produces an incredible amount of data, far too much for a meteorologist to be analyzing while attending to other responsibilities simultaneously.
The Next Generation Fire System introduces AI to this system. The algorithm helps meteorologists identify potential wildfires and instantly shares information about their location, size and movement.
“One of the biggest benefits is for areas with really sparse observability,” Tolby said, offering examples like remote woodland or mountain locations. “Wildfires are a lot easier to put out when they’re small, so I think that early detection of those wildfires and getting a foothold early is a really big advantage.”
Tolby said the Next Generation Fire System’s AI acts mostly as a “safety net” for identifying and monitoring fires, and that human meteorologists verify and distribute the collected information.
The NGFS works hand-in-hand with the Integrated Warning Team, the second wildfire technology to pass through the Fire Weather Testbed.
The Integrated Warning Team is more of a system than physical technology. It is a partnership between meteorologists and state forestry departments and land managers (such as fire chiefs and fire officials) to help share critical information about identified fires immediately so that officials can issue fire warnings early and attack fires safely.
When fires are detected in a given area, fire managers or emergency managers must call the National Weather Service and ask for a fire warning that they then put out to people living in areas that might be affected.
This multi-step process sometimes causes delays and miscommunication in scenarios when each minute makes a difference.
“We’ve seen over the last few years, unfortunately, some really tragic events of fast moving wildfires moving into populated areas,” said Tolby, “and in the past, there wasn’t a good mechanism to issue warnings for those fires.”
“What this Integrated Warning Team does is involves the National Weather Service from the beginning which allows us to speed up the process.”
Partnerships like these are already active in Norman, Oklahoma, and are planned to be put in place in more states across the western United States.
With the Integrated Warning Team in place, Tolby said that they have been able to reduce the time to issue a fire warning “from somewhere around an hour-and-a-half down to less than 10 minutes.”
Tolby hopes that these new technologies will help reduce the damage caused by extreme weather scenarios like those that have devastated Colorado, like the 2021 Marshall Fire which destroyed over 1000 homes not far from the NOAA’s office in Boulder.
The team is still writing up their final reports on the two most recent tests, which will then be passed on to NESDIS and the NWS to help them further improve and better implement the Next Generation Fire System and the Integrated Warning Team.
Tolby said that the Fire Weather Testbed are already underway on their next round of evaluations, and announced — in news breaking here on Rocky Mountain PBS — the arrival of the U.S. Forest Service’s WindNinja, an innovative wind measurement tool that more accurately answers which direction wind is blowing and wind strength from valleys to mountain tops.
Mixed in the nearly two dozen evaluators that participated in the Fire Weather Testbed were a handful of social scientists which Tolby described as essential to the process.
By matching NOAA meteorological scientists with sociologists, the team was able to evaluate each technology and system from the perspective of a scientist and from the perspective of the end user, according to Tolby.
“Especially in these very uncertain decision spaces, you can’t solely rely on technical or quantitative information, especially given the uncertainty and range of possible outcomes,” said Emily Wells, a social decision scientist from Colorado State University’s Cooperative Institute for Research in the Atmosphere.
Wells is one of the social scientists serving on the Fire Weather Testbed’s User Needs Assessment team. In short, she is responsible for talking with end users of the Fire Weather Testbed’s technologies in order to determine the questions that the Testbed needs to ask during evaluations.
She and her colleague, Dr. Ben Hatchett, have engaged directly with incident meteorologists, land managers and other fire partners to best understand how the Testbed’s technologies are operating in context and in real-time.
Wells said that fire partners have appreciated talking to researchers and being involved in the conversations that would end up impacting them and their field work.
“Having those conversations up front and at the forefront of product development can really help facilitate how these technologies are applied… and helping bridge the research gap to make the end users’ lives easier,” said Wells.
Wells is planning on completing firefighter training in the near future.
Before joining the Testbed, Tolby worked as an incident meteorologist (IMT), work which put him on the ground, side-by-side with firefighters battling blazes in the western United States.
“What I really realized is that the tools and technology we currently have really weren’t sufficient to answer the questions that incident commanders and operations folks out fighting a wildfire needed answered,” he said.
Fire strength, wind speed, wind shift and information on future weather conditions are all critical to fighting fires and necessary to ensure the safety of the firefighters themselves.
“There’s a lot of need right now for new tools and technology,” said Tolby, “and so I feel like we really have an opportunity to make some really positive change in society around wildfires and fire weather forecasting.”
“And to help transition things into the real world and make a real impact. So that’s really rewarding.”