Calif. county's mountain cameras link to statewide system to spot wildfires
The cameras are linked to a growing network that spans five western states and a supercomputer in San Diego that may someday use AI to spot fires
By Richard Halstead
The Marin Independent Journal
MARIN, Calif. — The Marin County Fire Department has installed six new, high-definition cameras at several mountaintop locations in Marin.
The cameras are linked to a growing network that spans five western states and a supercomputer in San Diego that may some day use artificial intelligence to spot fires quicker and predict their course.
“Really, what we’re building is a big data communications network,” said Neal Driscoll, professor of geology and geophysics at the University of California at San Diego.
UCSD, along with the University of Oregon and the University of Nevada, Reno, has formed a consortium, AlertWildfire, to help provide firefighters with access to state-of-the-art pan-tilt-zoom fire cameras and associated tools.
The new Marin cameras are located at Mount Tamalpais East, Mount Tamalpais West, Big Rock Ridge between San Rafael and Novato, Barnabe Peak East and Barnabe Peak West in the western San Geronimo Valley, and Mount Vision near Inverness.
Another camera will be installed on Mount Burdell within the next month and at least two more are planned over the next year.
“Our goal is to have any fire in Marin to be visible to at least two of the cameras,” said Marin County fire Deputy Chief Mark Brown, “so we’re going to continue to build this network and enlarge it over the years.”
The public can monitor conditions at these locations along with firefighters at alertwildfire.org/northbay.
Marin County Fire installed the county’s first fire cameras in 2014. The new cameras will replace that equipment, and the new system will provide the public with greater access to the pictures taken. Formerly, once firefighters spotted a fire and began manipulating the camera to take a closer look, public access was blocked. This will no longer be the case.
Brown said the county has used human fire spotters stationed atop Mount Tamalpais and Mount Barnabe since the 1940s or 1950s. Brown said human fire spotters are still used at those locations during the fire season during daylight hours.
“The cameras aren’t intended to replace human fire lookouts,” Brown said, “just to augment our ability to gather intelligence on an incident.”
Marin’s former camera system, developed by EnviroVision Solutions, was supposed to use mathematical algorithms to detect smoke, but Brown said that feature never panned out.
“The background work to make that detection system work was super intensive and required constant work,” Brown said. “We never had an instance in the four years we were using the cameras where they detected a fire prior to a human being.”
He added, “We really don’t have that much of a detection system problem. Fires are being reported to us quickly by the public.”
Brown said the cameras can help pinpoint the location of fires more effectively, since often 911 callers provide only a rough approximation. He said the cameras’ greatest value, however, is in providing intelligence on fires once they’re located.
Brown said using the cameras, Marin’s Emergency Command Center can anticipate the incident commander’s needs, moving resources to the fire before they are even requested and beginning evacuations, if necessary, before the incident commander asks for them.
Pacific Gas and Electric Co. picked up the $207,000 cost of installing Marin’s first camera system in 2014 and is paying for the new cameras, each of which cost $5,000.
“The staff at PG&E’s Wildfire Safety Operations Center in San Francisco plans to operate and use the cameras as part of its mission to help the company’s preparedness and response to wildfires,” said PG&E spokeswoman, Deanna Contreras.
Contreras added, “PG&E’s Wildfire Safety Operations Center team will monitor conditions across our system and evaluate whether to temporarily turn off electric power lines, as a last resort, in the interest of public safety, if extreme fire danger conditions are forecasted.”
UCSD’s Driscoll said there are currently 13 North Bay cameras plugged into the ALERTWildfire website and 21 more cameras are funded for the North Bay.
“This is important because it gives us a density of cameras that optimizes their functionality,” Driscoll said. “We can use the cameras to triangulate and determine the location of a fire.”
Driscoll said the ALERTWildfire system’s connection to the supercomputer in San Diego provides some tantalizing possibilities for future high-tech firefighting and other disaster response efforts.
“As we’re going forward, we’re putting two cameras on each tower, including one that spins every minute,” Driscoll said. “This is to help with artificial intelligence.”
He said the hope is that machine learning will eventually allow the system to detect fires before 911 calls come in.
A computer model, dubbed WIFIRE, has been developed at UCSD to predict the growth of wildfires once they start. The model incorporates information feeds on current weather conditions along with data on topography and fuel types.
Driscoll said once supplied with a ignition location, the computer “can give you predictions for two hours, four hours, or six hours.”
Driscoll said this big data approach to fighting fires may also be applied to other calamities.
“This can be used for many hazards,” he said, “ such as early earthquake warning and extreme weather and water events such as atmospheric rivers.”
Copyright 2018 The Marin Independent Journal