Some fires are so devastating they change not only entire communities forever, but the companies serving those communities.
Take the Tubbs Fire that savaged Sonoma County in 2017. That fire not only left permanent physical and emotional scars across countless neighborhoods, it sparked lasting innovation at the Pacific Gas and Electric Co.
“That day that changed us forever,” PG&E Chief Meteorologist Scott Strenfel said of Oct. 8, 2017, explaining that after “the massive wildfire events that happened not only in the North Bay, but in the Sierra, we really got on this mission to install weather stations and do the (Public Safety Power Shutoffs).”
So far, Strenfel said, addressing a virtual audience during a PG&E broadcast Wednesday, the company has installed more than 1,600 weather stations that help meteorologists like him track wind, temperatures and humidity across its service area in California, which he said is “70,000 square miles, mostly in Northern California, but also parts of Southern California.”
So many of these stations are needed, Strenfel said, because “a lot of weather can happen in those square miles, because of the state’s varied topography, which creates thousands of micro-climates that we need to be aware of and prepare for with accurate forecasts.”
To get forecasts that are as accurate as possible, Strenfel said all 1,617 weather stations are not only “sending out information every 10 minutes” on current conditions, but PG&E has also created machine learning models, such as the Fire Potential Index, which Strenfel described as “probably one of our most crucial models, as it’s predicting that if we were to have an ignition, what is the probability of that ignition becoming a large, catastrophic fire from any ignition source.”
To help the model determine that probability, Strenfel said a large amount of weather data is analyzed, including not only “really good data that takes us forward five days, but data that takes us back 30 years in time, hourly. We also have historical fires, and historical fuel moistures, and so some of the data scientists on the team were able to take those really robust historical data sets and compare them to the catastrophic fires in the past, and run them through an algorithm to come up with this model to project what the probability is based on all those factors. So we’re able to see, with very good detail, what the risk is across our territory, every single day, every single hour.”
So far this year, Strenfel said that Cal Fire has reported that there have been “generally more fires compared to 2024, and compared to fire year average,” but that less acreage has burned. According to a slide he presented, between Jan. 1 and Aug. 25, 2024, there were 5,132 wildland fires that burned 407,824 acres, while between Jan. 1 and Aug. 25, 2025, there have been 5,881 wildland fires that burned 387,002 acres.
“And we’re in the fire season now, for sure, and the only thing that gets us out of fire season is rain,” Strenfel said, describing PG&E as “very aware that things can change on a dime, and a heat event combined with a wind event can cause a catastrophic fire,” making the weather stations and weather prediction models crucial tools in the company’s efforts to reportedly “keep customers safe and prevent wildfires.”
“Ultimately, bringing in the best weather model information we can, as well as the best minds and scientists we can, is (so that we can) make the best decisions possible for our customers, including when and how long to do a PSPS (public safety power shutoff),” said Strenfel, noting that some of the company’s models help analyze what could have happened if the power had not been shut off. “Once the weather calms down (and the ‘all clear’ is given), crews start patrolling the lines with helicopters, looking for damage before energizing. And after each one of those patrols, they document every location of damage and hazard, and we’re able to simulate virtually, what could have happened if we didn’t do the PSPS.
“And while the models can’t prove 100% what would have happened, some of the tools and technology are definitely working,” said Strenfel, noting that he expected even more innovations to come. “None of this (technology) existed before 2017, so I am very optimistic and excited about the future. We have a rich foundation, and we’re able to innovate very quickly, as our job is to be able to provide the best forecast information so we can make the best decisions from them.”
And while Strenfel pointed out that some of the new tools are “making us safer and our hometowns safer, the weather is also changing rapidly. Absolutely these tools are making a difference, but climate change is also changing the risk, too. Therefore, we need to be very aware and on our game, so we are outpacing the risk of climate change that is also ramping up.”