Why We Do PSPS: PG&E Meteorology Runs a ‘What If’ Simulation After Each Event to See the Impact of Avoided Wildfires

By Matt Nauman

The late August Public Safety Power Shutoff (PSPS), PG&E’s first in two years, was a relatively small-scale event. It affected 3,900 in seven counties, and almost 90% of those customers had their lights back the same day they were turned off as a last-resort response to severe weather.

Post-event, PG&E does a thorough evaluation of the PSPS and looks to see what lessons can be learned.

As part of that, PG&E's Meteorology team takes verified information about locations when damage occurred or hazards were found during patrols, factors in the weather at the time of the de-energization and runs a fire-spread model. 

For this PSPS, one incident of damage was confirmed, on a circuit near Round Mountain and the Pit River not far from Burney. (The photo above shows that damage: A floating conductor, or a conductor that is no longer resting in its insulator. It was damaged as a result of a broken tie wire when the line was de-energized.) Had that powerline not been de-energized, a spark could have led to an ignition that ultimately could have started a 4,700-acre fire that would have burned in an area with 86 buildings and where 65 people live. 

“PG&E will not take any chances with customer safety. PSPS continues to be a tool that we use as a last resort to keep customers safe,” says Mark Quinlan, PG&E’s senior vice president of Wildfire & Emergency Operations. “We know that losing power disrupts lives, especially for those who rely on power for medical devices or medications. That’s why we’re finding ways to reduce the impact of PSPS outages, without compromising safety.”

One example of that is PG&E is refining existing artificial intelligence and machine-learning models to precisely target PSPS in areas experiencing severe weather.

“Running a fire-spread model based on where we saw damage or hazards is a pretty standard play that we execute after every PSPS,” says Scott Strenfel, PG&E’s senior director of Meteorology & Fire Science. “We look at what could have been if we did not de-energize.”

PG&E works with Technosylva, whose proprietary fire-spread modeling is also used by Cal Fire, the CPUC and other utilities.

And although this first PSPS of 2023 didn’t affect many customers, the memory of one of the 2019 PSPS events where nearly a million customers were de-energized remains fresh on Scott’s mind.

In that instance, more than 400 incidents of damage or hazards were found. The fire-spread modeling showed the PSPS had prevented fires that would have burned more than 3 million acres, a footprint with nearly 258,000 buildings and 421,000 people.

Here’s how the process works. Once PSPS de-energization happens, crews wait for the weather all-clear signal to be given to start their aerial and ground patrols of the affected circuits. If crews find damaged equipment or perhaps a big tree limb in a line that would have caused an outage, they note the precise location and take photos. That information is then reviewed and verified by the Incidents & Intelligence unit in the Emergency Operations Center.

From there, Meteorology plugs in the location and the exact weather at that exact location at that exact time and “drops a match on it,” Scott says. In this case, he ran a 24-hour fire-spread model. He does note that the modeling doesn’t account for any suppression activities and, importantly, “we don’t know for a fact that this would have caused an ignition.”

Still, it does provide a very visual sense of the benefit of PSPS. And whether it’s one incident of damage or a found hazard, or 400, it’s a great tool to understand what could have happened.

“The value is trying to prove one of the hardest things that’s possible – trying to prove a negative,” Scott says. “This is an established way that shows what could have been if we had not had a Public Safety Power Shutoff.”