AI-Driven Predictions Key to Combatting Rise in Wildfires

Artificial intelligence-based technologies, including simulators, weather prediction models and autonomous systems feature heavily in the report, and they provide the key to unlocking safer and more efficient defenses against the exponential rise in wildfires.

PCAST contends that critical technological and organizational aspects of wildfire response are stuck in the past century, but the report also outlines how AI-based predicting and tracking can handle the wildfire activity overwhelming fire departments across the country. In fact, the report refers to this kind of real time wildfire progression tracking as the “Holy Grail of firefighter safety.”

DISCOVER: The current landscape for AI in state and local government.

Drones Provide Eyes in the Sky for Firefighters

In its list of recommendations, PCAST highlights the need to develop autonomous detection, assessment and containment systems for wildland fires. It also urges federal agencies to engage with the private sector, which could help accelerate the impact of emerging technologies.

Drones and other autonomous systems will increase the efficacy of the nation’s overall wildfire response and provide a new way to protect firefighters on the ground while they tackle active fires.

The U.S. Forest Service is already working toward using this technology, including two new types of drones. One can fly over vast forest land to spot fires, and the other can then come in and set controlled fires to deprive the larger fire of the vegetation that fuels it.

Virtual Worlds Make the Real One Safer

While drones and other autonomous systems operate in the real world, there are other AI-driven technologies that can help through the virtual world. Digital twins replicate real-world environments and systems in a virtual platform, and they will play a big role in all of this.

Modeling predictions of wildfire behavior in a simulated platform will dramatically change the way federal, state and local emergency responders face these increasingly frequent blazes. Simulation labs with the platforms to support digital twin environments allow fire officials study real-world wildfire behavior in real time. A digital twin can suggest the best actions to combat the blaze in a photorealistic environment.

To create these environments and ensure the accuracy of these digital twin suggestions, large, varied data sets are necessary. These data sets are made up of  real-time fire data collected by sensors, as well as information on topography, vegetation and other variables. This level of analysis is only possible through planning and orchestration models that rely on AI supported by the processing power of a graphics processing unit.

LEARN MORE: 5 questions state and local agencies can ask about AI.

Predicting Fires Helps Stage Resources Where Required

Federal agencies already are doing work on this cutting edge. The USDA Forest Service Missoula Fire Sciences Laboratory and the Colorado Division of Fire Prevention and Control both use an advanced visualization and virtual-world simulation platform to process a fire’s magnitude and forecast its progress. This helps federal, state and local agencies predict where to direct resources on the ground to keep everyone as safe as possible.

Beyond that, Forest Service researchers in Montana are recreating historical fires using real data to help develop and train new AI models. This allows engineers to compare how historical fires behaved relative to both the Rothermel surface fire spread model — the standard created at the Missoula lab — and newly designed models.

At the GPU Technology Conference, March 20-23, Lockheed Martin discusses its Cognitive Mission Manager program, which creates digital twins of wildfire incidents by fusing real-world data from satellites, aircraft and ground assets via NVIDIA’s Omniverse.

The dramatic increase in wildfires presents danger to U.S. land, residents and first responders, but we have the technology available to make fighting these fires safer and more effective. AI-driven systems, especially those already being prototyped in the private sector, will provide the edge needed to better fight these fires as well as to predict their paths of destruction and even snuff them out before they rise up.