Starting this month, 10 sensors placed in hilly areas east of Oakland, Calif., will be monitoring the air for signs of wildfires. The sensors, which can detect and measure gas, particulate matter and heat, are intended to help more quickly identify, locate and track blazes that could threaten the city.

With wildfires expected to increase in frequency and intensity, officials in fire-prone areas such as California and Oregon are testing out early detecting systems that promise to catch fires before they turn into raging infernos that endanger lives, homes and critical infrastructure.

While fire managers have been using satellite imaging and cameras to spot wildfires for years, that technology has gotten sharper and is part of a growing tool kit that now also includes artificial intelligence, drones and sensors.

“The sensors help level the playing field, if you will, between us and the devastating fire,” said Reginald Freeman, who until recently was Oakland’s fire chief. “We can be notified immediately. We could deploy our personnel and equipment immediately, which obviously increases the probability of lives and property being saved.”

Sensors and other types of early detection technology could be particularly important for catching fires that occur in remote areas, where they can often go unnoticed until they become much bigger.

“Fires that start in urban areas are easily spotted by someone, but who is monitoring the wilderness areas?” said Ankita Mohapatra, an assistant professor at California State University at Fullerton who published a review of wildfire early detection technologies in 2022.

But Michael Pavolonis, the Wildland Fire Program manager at NOAA’s National Environmental Satellite, Data and Information Service, said reliably detecting wildfires early will probably require more than sensors alone.

“Every technology and technique has strengths and limitations,” Pavolonis said. “There isn’t a single technique that can solve this problem.”

How do fire detection sensors work?

In California, the state’s fire agency, Cal Fire, is pilot-testing “ultra-early” wildfire detection sensors. Meanwhile, in neighboring Oregon, officials deployed two types of sensing technologies this year.

N5 Sensors, the Rockville, Md., company that provided sensors to the Oakland Fire Department, said it is working with four utilities around the country, as well as stakeholders in eight states and has two engagements in Canada.

Some sensors, like those deployed by Cal Fire, are designed to detect fires that just ignited or are in their early stages and to alert firefighters. They measure various gases, temperature, humidity and air pressure, and come “with built-in artificial intelligence to reliably detect a fire and avoid false positives,” according to Dryad, the German company behind the technology.

Certain types of sensors could also be used to more accurately predict where fires might happen based on the data they collect, said Zhaodan Kong, an associate professor in the mechanical and aerospace engineering department at the University of California at Davis. This information can be combined with knowledge about other factors that contribute to fire risk, such as the location of power lines, said Kong, who is researching technology for wildfire prediction and detection.

An inexpensive firefighting solution?

Catching wildfires early could not only save lives, but may also prevent millions of dollars in losses, experts say. The wildfire that devastated Oakland in 1991, for instance, killed 25 people and cost an estimated $3.9 billion in present-day dollars.

Networks of sensors can be simple to set up and generally inexpensive to run and maintain compared with the other types of wildfire detection technology, such as camera towers and drone fleets, Mohapatra said.

Depending on their function, individual sensors could range in cost from less than $10 to about $200 each, Kong said. He noted that a node, which includes multiple sensors and other technology, could cost about $1,000.

But he and other experts say one concern is the accuracy of readings. If sensors are too sensitive, that may lead to a number of false alerts, or they could not be sensitive enough and fail to detect some fires, Kong said, adding many of the sensing technologies for early detection of wildfires are still being developed and tested.

A layered approach

Reliably detecting wildfires early will probably require a combination of different technologies working together, experts say.

“The odds of a single technology coming along and solving the entire problem are not zero, but they’re not great — at least not in the near future,” said Cristina Davis, a professor at UC Davis.

Satellites, for example, can scan vast swaths of terrain and quickly produce images that can be monitored by AI to identify potential blazes. That has enabled fires to be detected “within the first 10, 20, 30 minutes of ignition,” Pavolonis said.

But satellites can have limits to their coverage and might not be able to identify very small fires, Mohapatra said. Cameras, which are often installed in towers, can more closely monitor conditions on the ground, but setting up these networks is typically an “immense initial cost,” she added.

And sensors, though often cheaper, can be hard to place in certain types of terrain, such as rocky areas at high altitudes, which typically aren’t easily accessible. Other locations might be inaccessible because they are on protected lands, Mohapatra said.

So, ideally, she said, there should be “multiple layers of monitoring that are fail-safe.” For example, after sensors detect a fire in its early stages, uncrewed aerial vehicles with cameras could be deployed to monitor the fire while satellites are repositioned to identify other hot spots and predict where the fire could spread.

The impact of having reliable early detection systems for wildfires would be tremendous, Kong said.

“It will not prevent wildfires, but it will prevent large fires,” he said. “If you can detect these wildfires at the early stage and you can alert the firefighters, they can hopefully quickly eliminate them as early as possible.”