Today’s real-time wildfire data helps prepare for a changing climate

As the risk and damage of wildfires continues to climb, states are turning to future-facing data to help plan for and combat unruly burns.

Historical wildfire data on fire patterns from decades ago doesn’t capture how climate change has intensified the frequency and severity of burns in recent years. And when outdated data is less relevant to real-time risks, wildfire management agencies could find themselves unable to confidently plan and budget for the necessary resources to address incoming fire events. 

With state governments shelling out more money every year to subdue wildfires, policymakers should “consider using forward-looking data to better anticipate wildfire costs and focus on long-term wildfire planning,” according to a recent report from The Pew Charitable Trusts. “Doing so would minimize the disruption of supplemental appropriations during times of crisis and address overall risk.” 

That’s easier said than done.

The number of fires, acres burned and costs are trending higher, and fire managers are concerned, said George Geissler, chair of the National Association of State Foresters and a Washington state forester. Fire management agencies have fixed budgets, so “a large fire season can significantly impact not only the agency’s budget, but the overall state’s budget,” he said.

Another complication is that firefighting “is one very small piece of what we do in the fire world,” said Gregory Dillon, director of the Fire Modeling Institute. “Fire management is a much bigger picture, where we have a lot more decision space to think proactively about … [wildfires] and get out ahead of it.” 

Recent strides in tech and data capabilities can help state agencies looking to implement future-facing wildfire management efforts, experts say. Rapid response or early detection tools such as cameras or sensors that monitor and track wildfires in real-time can provide more accurate insights into the fire’s behaviors, Geissler said. That data, in turn, can help agencies improve data models used to project future fire behavior and better determine the budget they need for upcoming fire seasons. 

Over the last year, wildfire-prone states like California, Colorado and Washington deployed artificial intelligence-enabled cameras to spot early signs of a burn, such as smoke, to help firefighters stamp out fires before they escalate. 

Local governments are leveraging innovative tools too. California’s Santa Clara County officials approved the use of sensors in August to analyze air temperature and particulate data to detect budding wildfires in the area. And in October, city council officials in Austin, Texas, decided to deploy AI cameras starting this summer to provide first responders with data visuals such as the triangulation of wildfire ignition sites and the identification of at-risk structures.  

Data tools that inform first responders of a fire’s path, its speed or current weather patterns, for instance, can help agencies more effectively plan for where and how fires progress. Take a fire originally moving southward. Real-time data from remote sensors that pick up a change in wind speed and direction can let firefighters know that the fire will likely veer to the east. 

Agencies should also consider other information like the terrain surrounding a community and the type of vegetation in the area, which impact a fire’s progression, said Frank Frievalt, director of the Wildland-Urban Interface Fire Institute at California Polytechnic State University. Since heat rises, for example, a fire is likely to burn faster uphill, especially if the ground is covered in dry, flammable grass, so it’s critical for agencies to understand more than just data on the wildfire itself.

“There’s certainly a lot more interest in fuels data [with] the way we can do remote sensing or use satellite resources,” he said. “There’s no doubt we have access to a lot more data than we’ve ever had.” 

Those data insights can help agencies more effectively dispatch first responders in low-risk areas or proactively station air tankers carrying fire retardant where the fire is projected to be, Geissler said.

As climate change’s effects on the frequency and severity of wildfires are still unclear, it’s critical that agencies collect and analyze as much data as they can to improve the accuracy of data models and projections for future wildfire management efforts. 

“It’s very easy to have two or three mild fire years [and] for legislators to think, ‘Oh, I don’t have to invest in that,’” Geissler said. 

But with more comprehensive data on wildfire behavior, he said, wildfire agencies can advocate for more funding if data in a community shows a heightened risk of fire compared with previous seasons or if historical data fails to capture the scope of current wildfire risk. 

More states, Geissler said, are experimenting with new technology to address common questions plaguing agencies like: “How do we use the information we have to make daily, weekly or monthly projections on what we need as far as resources and funding? As fires occur, what is the most efficient way to fight a fire? What’s the safest way to put our firefighters in there to notify the public?” 

But when it comes to predictive wildfire management tools, agencies should err on the side of caution, Dillon said. Managers should ensure that they use quality data to train models, then test and validate the tools for accuracy. Data projections, for example, can be compared with actual field information to determine if projections adequately reflect observed fire behaviors. 

Resources like artificial intelligence and machine learning used for wildfire activity projections, are “only as good as the training data and information that you feed it,” he said.