Wildfire cameras around Aspen appear effective but need time to prove worth | AspenTimes.com

Wildfire cameras around Aspen appear effective but need time to prove worth

Jake Andersen, deputy chief of operations at the Aspen Fire Department, shows off live feeds Wednesday from cameras installed around Pitkin County connected to artificial intelligence that scan for wildfires 24/7.
Jason Auslander / The Aspen Times

An experimental system that monitors the Upper Roaring Fork Valley for wildfires has already detected one fire start and so far proved to be a worthwhile tool, an Aspen fire official said Wednesday.

However, while the system did detect the lightning strike near Lazy Glen at the end of July, passers-by reported the smoke before the warning system because of a glitch, and questions remain about whether the Aspen Fire Department can afford the technology going forward, said Jake Andersen, AFD’s deputy chief of operations.

“When this came up, we were like, ‘Of course, why would we not?’” Andersen said. “Think about it — I have four fire lookouts (I can monitor) from my office or from my pocket. I can pull over and check my phone if I get an alert.”

In July, Pitkin County commissioners approved the pilot program that allowed a Silicon Valley-based company called Pano AI to install two cameras each on four communications towers owned by the county. The cameras atop towers on Aspen Mountain, Jackrabbit at Snowmass, the Williams tower near Gerbazdale and on Upper Red Mountain were installed about a month ago.

The cost of the $50,000 project was donated to Aspen Fire by a local Red Mountain homeowner who has been active in local wildfire mitigation efforts.

The cameras provide 360-degree views of the surrounding wilderness every minute and run 24 hours a day, seven days a week until the end of the fire season in October or November. They work mostly during daylight hours because it is a visual-based system, though the cameras would be able to see flames of a fire at night, Andersen said.

Technicians at Pano’s facility in Northern California receive the visual information first and any alerts when the artificial intelligence system detects smoke or fire. Those technicians then verify whether the video has actually spotted a newly started wildfire, and send an alert to Aspen Fire if they determine it is a blaze, he said.

The system did spot a lightning strike-started fire near Lazy Glen at about 3:30 p.m. July 30, but alerts about the fire were delayed in coming from Pano and drivers on Highway 82 spotted the smoke first and called emergency dispatchers, Andersen said. The problem has since been corrected.

That fire was extinguished the next day thanks to rain that followed the lightning.

However, the incident underscores the experimental nature of the program, he said. Pano’s artificial intelligence system needs to continue to learn what smoke looks like in this area so it can continue to make improvements and provide alerts sooner. Once the artificial intelligence learns an area, it can see wildfire starts better than human eyes, Andersen said.

Another benefit provided by the cameras is nearly instantaneous access to pictures of backcountry areas. For example, Andersen said he received a report Tuesday of possible smoke near Buttermilk Ski Area, and was able to pull up a live view and determine no smoke was visible.

Aspen Fire crews still went and checked the area — they found an overheated vehicle — but the cameras provided valuable initial intelligence, he said.

“If I can open (my computer) up and see something huge, I can alert Grand Junction (pilots) and get air support there,” Andersen said. “We can start to make some of these strategic decisions right away.”

The Pano system allows fire officials to send pictures of the fire and geographical coordinates to pilots and others who need to know exactly where it’s located, he said.

Parker Lathrop, director of operations for the Pitkin County Sheriff’s Office, used to do Andersen’s job until several months ago, and said the use of cameras to survey the backcountry is not new thanks to Aspen Skiing Co.-installed cameras atop each of the area’s four ski mountains.

“For years, that was my go-to,” Lathrop said.

The Skico cameras, however, take 10 minutes to provide a 360-degree view, while the newly installed cameras in Pitkin County provide the same view every minute, Andersen said.

The artificial intelligence component of the new technology is the real possible game-changer now, Lathrop said.

“I like the concept,” he said. “I want to see the product work, but I also don’t want to see fires. How it pans out (remains to be seen).”

One of the other concerns some area fire officials have had with the new system is privacy, Lathrop said. Some of the video from California played in Pano presentations showed people’s backyards, which spawned those concerns, he said.

Subsequent presentations have shown the same videos, but with people’s yards pixelated and unable to be seen, Lathrop said.

The cameras in Pitkin County are not in areas where people’s private property can be viewed, Andersen said.

At the end of the fire season. Aspen Fire officials will analyze the public benefits of the new system and decide whether they want to or can pay for it in the future. It’s not clear how much it might cost the fire district, and attempts to reach officials with Pano on Wednesday were not successful.

“It’s cool to be on the front end of this … and help it get better,” he said. “It makes fire services better.”

Support Local Journalism

Support Local Journalism

Readers around Aspen and Snowmass Village make the Aspen Times’ work possible. Your financial contribution supports our efforts to deliver quality, locally relevant journalism.

Now more than ever, your support is critical to help us keep our community informed about the evolving coronavirus pandemic and the impact it is having locally. Every contribution, however large or small, will make a difference.

Each donation will be used exclusively for the development and creation of increased news coverage.