For Detailed Maps of Its Service Area, CAL FIRE Turns to AI

The nation’s largest firefighting organization is using AI technology to accurately — and quickly — track building development, defensible space, evacuation routes and more.

The California Department of Forestry and Fire Protection (CAL FIRE) is using machine learning from Ecopia AI, together with GIS mapping from Esri, to provide detailed data related to buildings, roadways and other features, and the relationships objects in the built environment have with each other.

In the world of firefighting and protecting people and property, it matters how close buildings are to each other and what access there is to roadways or a water source, officials said.

“We’re looking at the proximity between buildings in communities, because some of the high standards of a fire prepared home are hard to meet, if you’ve got another structure in maybe the first 10 feet, or even 30 feet. So we’re really looking for ways to really improve community resilience,” Mark Rosenberg, CAL FIRE research data manager, said, noting the broader goals behind the detailed mapping project. CAL FIRE is in the middle of a three-year agreement with Ecopia.

The agency oversees some 31 million acres, or roughly one-third of California. Much of the area is rural or forestland, but it also includes about 1 million homes, businesses and institutions such as schools and hospitals.

Knowing the proximity of one building to another, or the nature of the defensible space around buildings or communities, helps inform firefighting strategies, officials said.

“This starts to allow us to delineate the difference between wildlands — truly wildland areas — and areas that are more developed and maybe have a different fire propagation issue,” Rosenberg said.

The CAL FIRE data mapping project took Ecopia about four weeks to develop detailed data for those 31 million acres of service area, Thomas Peck, a company senior associate, said.

“And that’s because we have that automation side of things which really does the heavy lifting for us,” he said.

This speed is a crucial component for an organization that needs up-to-date data every time a new fire season rolls around. Which is why Ecopia is providing what it calls “change detection,” which are changes to the landscape from one year to the next showing new development in an area, or how development has rebuilt after a fire.

“The main thing is being able to extract so many building footprints every single year in such a short timeframe. What we’re providing them each year now is the ‘new,’ ‘removed’ and ‘modified’ buildings,” Peck said. “It’s going to give them a lot of key information: after a fire, are people rebuilding? Are people completely tearing down their structures?”

Ecopia’s ability to make determinations around what objects are in the landscape is partially dependent on partnerships with other data providers.

“A large majority of the people that use our data are using it in the Esri product suite,” Peck said, indicating the company also works with imagery providers including Eagleview, Nearmap and Hexagon.

Initiatives like the AI mapping project are part of a larger strategy to place technology and data-based decision-making at the center of CAL FIRE’s operations, officials said, as the agency prepares for days that are hotter and drier, and leave wildlands and urban areas both more vulnerable to fire.

In recent years, the California Legislature has directed significant funding and effort toward CAL FIRE to grow its tools, research, and firefighting. In the last decade, the agency has “moved much more aggressively on a data-driven decision-making strategy. And location figures prominently in that,” Rosenberg said.

“The GIS data being current, and the use of dashboards, and the receptivity to using data is a big game-changer,” he said, reflecting on some of the technology leaps CAL FIRE has taken, helping it to become more effective. “These are capabilities we didn’t really have before.”



*This article was originally published by Government Technology, Industry Insider– California's sister publication.