New Mexico officials announce new AI wildfire monitoring network

Five new wildfire monitoring camera stations that use artificial intelligence to look for smoke have been installed in areas where urban life meets the wilderness near Albuquerque and Santa Fe as part of pilot programs the state is working on with two electric utility companies.

Within two years, more than 40 wildfire detection stations from San Francisco-based company Pano AI should be in operation, with three stations provided by the company, eight funded with state dollars and the remaining 32 paid for by Xcel Energy and Public Service Company of New Mexico, according to a news release and Pano AI staff.

As New Mexico’s climate is predicted to become more arid with less predictable monsoons in coming decades, wildfire may become a bigger concern in the Land of Enchantment. The five largest documented fires in state history occurred within the last 14 years, and the two largest fires — the Calf Canyon/Hermits Peak Fire in northern New Mexico and the Black Fire northeast of Silver City — took place in 2022.

For decades, New Mexico has relied on 911 calls and lookout towers to detect wildfire starts, said New Mexico State Forester Laura McCarthy.

“It worked for us because our weather was not so extreme and delays of a few hours in detection were not consequential. But it doesn’t work when it’s a red flag day, super windy, relative humidity less than 10% and it’s 98 degrees.”

The new camera stations will provide 24-hour monitoring of high-risk areas and notify emergency responders with precise GPS locations when there are confirmed wildfire incidents, she said.

The technology should help fire incident commanders make better decisions about how to deploy resources and protect public safety, McCarthy said. Every fire agency within the coverage area of the stations will be able to access the Pano AI platform and get training on the technology at no cost to their department.

Pano AI’s technology is in use in 17 U.S. states, including Washington, Utah and Wyoming.

Each wildfire detection station has two cameras that rotate to provide a 360-degree view of the surrounding area, said Kat Williams, Pano AI’s director of government development. The stations have visibility of at least a 10-mile radius. If two stations spot a potential fire, they can triangulate its location and provide a latitude and longitude. When a fire is detected, staff at Pano AI reviews the footage in an effort to avoid false positives and then send text or email alerts to relevant local emergency services.

Williams pulled up the live feed from one of the cameras near Albuquerque on Wednesday during a news conference at the Bernalillo County Public Safety Training Academy. She demonstrated how different layers could be viewed on maps of the area, giving emergency responders the ability to examine terrain, roads and other conditions before heading to a fire start. In an effort to protect privacy, any homes in view are pixelated from the video before it’s uploaded to Cloud storage.

Pano’s artificial intelligence program was trained on more than 3 billion images, said Peter Ambler, the company’s vice president of global government affairs.

New Mexico has tried using images to detect wildfires before, McCarthy said. In 2020 and 2021, the state participated in a free pilot program with a different company that used satellite imagery to detect fires. That technology was helpful, but it also got things wrong a lot, she said, sometimes misidentifying water vapor clouds as smoke.

Camera cost, locations

Pano AI representatives, state officials and staff from the two utility companies were unable to provide specific overall or annual cost estimates when the network of monitoring stations was announced Wednesday.

State legislators have allocated $300,000 to support the state’s 11-station pilot program, McCarthy said. New Mexico’s Senate Memorial 2 wildfire study group will recommend the Legislature allocate another $2 million during the 30-day session next year to expand the network of monitoring stations, she said.

Annual costs to maintain the system will likely be under $50,000 per station, Ambler said.

The state’s 11 detection stations will be in the East Mountains area, near Albuquerque and in Santa Fe County. The locations prioritize sites where lives and livelihoods would be lost in a wildfire, McCarthy said. Five of those cameras began operating within the last few months and the other six will be brought online through early 2026.

Xcel Energy plans to bring another 15 stations online this year and 15 more by the end of next year. The utility company operates in eastern New Mexico, and the cameras will be located in that region, focused on high-risk areas with homes and around infrastructure like poles, wires and gas pipes, according to Zoë Lees, regional vice president of regulatory planning and policy for Xcel Energy. Infrastructure that provides electricity is critical, Lees said, but it comes with some risk, especially with increasing extreme weather.

“So we feel like it’s prudent of us as a utility to make sure that we are implementing resiliency measures like installing Pano AI stations in the service territory to protect our communities,” Lees said.

PNM is starting a two-camera pilot program with stations located in the Santa Fe Basin. The utility is also exploring other wildfire prevention technology like engineering controls and new infrastructure fuses to reduce the risk of sparks, as well as new patrolling and inspection methods, said Wesley Gray, PNM vice president of operations and engineering.

The cameras aren’t New Mexico’s only tool for addressing wildfires. The state Energy, Minerals, and Natural Resources Department also has $33.5 million for tree thinning and is working to create tree thinning buffers around communities, McCarthy said.

The state’s 11-station pilot program “isn’t an expenditure of state government, state taxpayer dollars. This is an investment right back into the families, communities, land and what makes New Mexico so beautiful,” said Lt. Gov. Howie Morales.