IE11 Not Supported

For optimal browsing, we recommend Chrome, Firefox or Safari browsers.

Fire Tools on the Horizon: Data, Partnerships and, Later, AI

Networks of cameras and collaboration among public utilities, CalFire and research institutions have sprung up throughout the state to fight wildfires.

Leaders across the state are looking for new ways to stop wildfires, and while technology has been touted as a solution to the historic-sized events of the last few years, specific programs are harder to develop.

Networks of cameras and collaboration among public utilities, CalFire and research institutions have sprung up throughout the state. One example, AlertWildfire, grew out of the Nevada Seismological Laboratory's earthquake monitoring efforts and uses cameras to confirm fire locations. 

“We’re trying to have rapid response to carbon ignition,” geologist and UC San Diego professor Neal Driscoll told Techwire in an interview.

Nevada Lab Director Graham Kent and Driscoll see room for expansion of the system with sensors, weather stations and eventually, with more improvement, artificial intelligence. 

"We're doing huge data on a huge network," Kent told Techwire in an interview. "We put it up on the cloud so the public can see it. They can't control the cameras, but they can see their situational awareness."

Kent and Driscoll are working with counties to place cameras on existing towers. They also need more data centers to support and back up the network.

They want to place more than 1,000 cameras in the next two years, and know that while the cameras are used minimally on a daily basis, the site can be flooded with traffic during a fire. The cameras can be controlled by the Lab and CalFire officials from mobile devices.

“We’re really building a large data communications network," Driscoll said. "This communications network allows us to bring information about fires from back countries to command centers, where we can decide how we scale our response.”

The site is hosted on the Amazon Web Services S3 bucket, "so it can scale up," Kent said. "You can host tens to hundreds of thousands, potentially millions of people at once getting HD imagery."

While AI is still unable to keep up with all the climate differences seen in California's regions, the network can confirm 911 calls and create situational awareness for firefighters.

"We're still working with partners to make that AI better, but it's not even close," Kent said.

AlertWildfire is credited for helping to contain San Diego's Lilac Fire, which burned under the same conditions as the 160,000-acre Witch Fire, but was contained at just over 4,000 acres. San Diego fire officials were able to dedicate all resources to the fire because the rest of the county was clear, according to cameras.

Being able to correctly allocate resources can save at least $100,000 per fire, since each firetruck deployed costs around that much, according to Kent.

UC San Diego's supercomputer center and WIFIRE creates fire models that Nevada's lab and other decision-makers can use to allocate resources. The models can help predict how evacuations should take place and the patterns a fire will follow, within a minute of confirming that the fire exists, according to Kent.

A director at the center, Ilkay Altintas, told Techwire that data is one of the most important tools the state can use to fight fires.

Altintas also said the WIFIRE platform could integrate with any solution the state chose to use to harness that data.

 

The Nevada Seismological Laboratory site hosts AlertWildfire footage from the Zephyr camera, overlooking Lake Tahoe.


The Nevada Seismological Laboratory site hosts AlertWildfire footage from the Zephyr camera, overlooking Lake Tahoe.

Kayla Nick-Kearney was a staff writer for Techwire from March 2017 through January 2019.