Can AI help fight wildfires?

Large forest fires fueled by climate change have burned over 1 million acres in California and Oregon, marking a particularly destructive start to wildfire season and hastening the need for new solutions.

It’s no surprise that much of that focus is on artificial intelligence, a burgeoning and hopeful technology that has also stirred fears and anxiety about its potential consequences, such as eliminating human jobs and industries.

AI is already used to mitigate the threat of wildfires nationwide, helping communities in the drought-hit West and places like Maui find, avoid, fight, and recover from them.

AI can help detect the first signs of smoke, spot fires from satellites, and predict where fires may start and how they will burn in certain terrains. On July 23, the Biden administration announced a $20 million investment that will improve detection, tracking, and provide public safety data.

The federal government maintains an online tracking system that offers a trove of live wildfire data neatly displayed on a map. It can show the location and size of fires, including associated smoke plumes and air sensor readings, acting as early warning systems for residents in neighboring communities and states.

While AI has been useful for residents, there are reservations about its effectiveness on the front lines, where humans remain the best form of defense against wildfires.

Max Alonzo, a 12-year wildland firefighter veteran and current Secretary-Treasurer of the Washington, D.C.-based National Federation of Federal Employees, told Reckon that he doesn’t believe AI has made much of a difference and has detracted from major pay and welfare disputes within the federal government’s wildland firefighting service.

“Artificial intelligence cannot pay our men and women enough,” said Alonzo, whose organization is leading a multi-year effort to secure a permanent pay increase and benefits for wildland firefighters. “It can’t raise the budget to create housing in rural areas for our land management employees. It can’t address the mental health issues we see with our men and women deployed to these fires for months at a time.”

Alonzo said he’s unaware of any formal AI training or planning within the service. However, he did note that as the threat from climate change has grown, the resources needed to fight wildfires have not and are becoming increasingly difficult to maintain.

“I’m sure there is a place for AI,” he added. But I have not seen where it can really make any positive impact at this point.”

The average wildland firefighter’s pay is typically around $34,000 a year but was significantly increased in 2021 by a temporary $20,000 bump that expired in Sept. 2023. However, with predictions that 50% of firefighters would quit without a new pay deal, the increase was extended by another year.

The U.S. House passed a $330 million pay increase, and there are encouraging signs that the plan will pass the U.S. Senate. But money is just one of the issues. High rates of suicide, homelessness and cancer remain major points of contention as pay negotiations continue.

Big wildfire season

Nearly 30,000 wildfires have burned around 4.7 million acres in 2024, higher than the 10-year average, according to the National Interagency Fire Center (NIFC).

The Durkee Fire in Oregon’s Blue Mountain region is 86% contained as of Monday after burning around 300,000 acres since July 17. Started by lightning strikes, it’s one of the largest wildfires in state history. Dozens of other large blazes in the state are close to collectively surpassing the massive and destructive 2020 season, which burned around 1.2 million acres.

California’s Park Fire has burned close to 400,000 acres but was allegedly started on purpose. It’s the fourth largest in state history and is around 30% contained.

It was hoped that heavy summer rains would deter drought conditions, but a triple-digit heatwave in July left large chunks of the West vulnerable. Smoke from the fires has spread to northern Canada, the Gulf of Mexico, and the Atlantic Ocean, according to the NIFC.

AI has shown some promise in mitigating the worst of wildfires and helping plan ahead.

In the aftermath of the Maui wildfires, which were started by downed powerlines but exacerbated by dry conditions, Michigan State University researchers produced highly detailed maps that allowed them to track how the fires started and spread. The maps can help with remediation efforts and assess future risks, like the location of forests and vegetation in relation to at-risk communities. Other AI technologies under consideraation in Hawai’i enable officials to make hyper-local wildfire predictions using rainfall, soil moisture and wind speeds. According to scientists at the University of Hawai’i, dry brush and high winds are strong predictors of wildfires.

Although some AI wildfire technology may be years away from reaching its full potential, combining it with advances in robotics, software, and climate change research could help with more than just detecting and fighting fires, such as monitoring flooding.

Here are some ways technology is helping:

1. Early detection and monitoring

• Satellite imagery: Satellites equipped with thermal sensors and infrared cameras can detect hotspots and monitor wildfire spread in real-time.
  
• AI-supported wildfire sensors: Ground-based sensors enhanced with artificial intelligence (AI) analyze data in real-time to detect early signs of wildfires, such as smoke, temperature spikes, and unusual atmospheric conditions.
  
• Drones: Unmanned aerial vehicles (UAVs) provide high-resolution images and real-time data, even in remote or inaccessible areas. They can also detect heat signatures and map fire perimeters.

2. Prediction and risk assessment

• Machine learning: Algorithms analyze vast amounts of data, including weather patterns, vegetation moisture levels, and historical fire data, to predict wildfire risk and behavior.
  
• Weather forecasting models: Advanced meteorological models predict conditions conducive to wildfires, such as high winds and low humidity, enabling better preparedness.
  
• GIS mapping: Geographic Information Systems (GIS) map high-risk areas by analyzing topography, vegetation, and human activity, helping to allocate resources more effectively.
  
• Flooding sensors: Post-fire sensors can monitor areas at risk of flooding due to vegetation loss and altered landscapes, providing early warnings and risk assessments.

3. Communication and coordination

• Incident management systems: Integrated systems like the Incident Command System (ICS) help coordinate responses by providing a common platform for communication and resource allocation among multiple agencies.
  
• Mobile apps: Apps like “Wildfire Info” and “Fires Near Me” provide real-time updates and alerts to the public, helping communities stay informed and safe.

4. Suppression and containment

• Aerial firefighting: Advanced aircraft equipped with infrared cameras and water or retardant dropping systems, such as I4F foam, are used to combat fires. These include helicopters, drones, and fixed-wing planes.
  
• Robotics: Ground-based firefighting robots can navigate hazardous terrains to create firebreaks, clear vegetation, and apply fire retardants, reducing risk to human firefighters.
  
• Firefighting equipment: Innovations like fire-resistant drones and autonomous vehicles enhance firefighting capabilities and safety.

5. Post-fire analysis and recovery

• Remote Sensing: Post-fire, remote sensing technology assesses damage, maps burned areas, and monitors vegetation recovery.
  
• Data Analytics: Analyzing data from past wildfires helps improve future responses and strategies. This includes understanding fire patterns and the effectiveness of suppression techniques.