As the number and size of wildfires in the West and across the country grow, scientists warn that the increased mingling of housing and nature put more property and people at risk. One blaze 25 years ago demonstrated how serious those risks can be.
On Oct. 20, 1991, the hills across the bay from San Francisco were cloaked in smoke. Flames coursed up the steep slopes of Oakland and Berkeley, California, feeding on overgrown vegetation parched by years of drought. Along those slopes were houses. Most had yards clogged with plants. Some were built on overhanging hillsides. Many were built with wooden shingles and other flammable materials. They stood no chance against the blaze.
Officials had known about these risks for decades. The Oakland Hills area had burned three times previously – all under similar conditions. In 1923, a wildfire destroyed 640 buildings. Another fire claimed 39 buildings in 1970 and flames destroyed five more in 1980. Efforts to reduce the risk such as firebreaks and clearing dead trees had been planned, but not implemented. When the Oakland Hills Firestorm hit in 1991, about 3,000 homes burned and 25 people were killed. Damages from the firestorm were among the highest ever recorded – $2.7 billion in today’s dollars.
Jan Null was the lead forecaster at the National Weather Service’s offices in Redwood City, California, at the time of the fire. He was watching Steve Young lead the San Francisco 49ers to victory over the Detroit Lions before he went to work the night shift. Broadcasters kept cutting away from the game to show the growing plume of smoke over the Oakland Hills. The fire was rekindled from a blaze that firefighters thought they had extinguished the previous day. Null got to work and began supplying crucial weather information to officials charged with battling the fire.
Now retired, Null works as a consultant and teaches at San Francisco State University.
Reveal from The Center for Investigative Reporting asked him about the role weather played in the 1991 firestorm. We edited some of Null’s responses for length.
Can you explain the importance of weather conditions before and during a wildfire?
The behavior of a fire is influenced by three factors: fuels, topography and weather. The weather segment is a function of wind, humidity and atmospheric stability. The combination of these factors causes how hot and fast a fire burns, with high winds, low humidities and dry fuels being the most volatile.
The importance of the weather prior to a fire is [related to] the time of year and then that translates into how dry the fuels are. In California, our Mediterranean climate is defined by a summertime drought. We basically have no precipitation from May through well into October in most years. So by the time we get into September and October, all the vegetation is dried out.
The situation was exacerbated by the fact there was a very hard freeze the prior year, in December 1990, and that led to a lot of the fuels dying.
At the time a fire starts, the really relevant conditions are the wind speeds, the temperature and the humidity. Again, the humidity goes to the dryness of the fuel. The temperatures also go to the dryness of the fuels and the wind speeds go to what the spread of the fire is. If we’d had that same Oakland Hills fire without any wind, we wouldn’t be talking about it now.
What weather patterns played a strong role in the 1991 Oakland Hills Firestorm?
The weather pattern over the western United States was dominated by high pressure over Nevada, eastern Oregon and Idaho, which created warm, dry winds blowing from the Northeast to the Southwest. This pattern most often occurs in the fall months and has come to be known as the Diablo winds.
What are the Diablo winds?
The Diablo winds are basically the Northern California equivalent of the Santa Ana winds. They are most often in the fall months when we have high pressure instead of being off the coast and causing a seabreeze, high pressure moves on into the Great Basin area.
Air flows from high pressure to low pressure. So now we have high pressure over the over the high deserts of the Great Basin blowing toward lower pressure along the California coast.
The average elevation of the Great Basin is about 4 or 5,000 feet. As air flows from a higher elevation to a lower elevation, it warms by compression. You add more atmospheric pressure around it as it gets down near the surface.
As it warms, it dries out and then as it flows over mountain ranges, it gets compressed by the what’s called the Venturi effect. So you’re just, like, if you have a nice, slow-moving river going along then it goes into a canyon, it gets squeezed and it starts flowing faster.
It’s the same thing if you put your thumb over the nozzle of a hose. You know it squirts out a lot faster and farther. Well, the same thing happens when you get this airflow over these mountains. It can’t go through the mountains so it gets squeezed over the ridges. You get acceleration and you get faster winds. The worst place for one of these fires to start is along the ridge lines.

What contributed to the rapid spread of the Oakland Hills Firestorm?
A lot of it is a combination of all the things that we’ve been talking about. One was the winds. The others were the very low humidity and the very dry fuels.
If you have wind speeds in the 30- to 40-mile-an-hour range, it is almost impossible for firefighters to keep up with that. I know from my discussions talking to the emergency services at the time, the main strategy was to control the flanks of the fire as much as possible because they knew they weren’t going to be able to get out in front of it. You can’t get in front of a fire that’s moving at 30 miles an hour toward you.
What happened that helped crews in their efforts to contain the fire?
As Sunday went into Monday, the high pressure began weakening and moving to the east, so the Diablo winds died off. On Monday the 21st, cooler marine air was moving back into the area. So you had the combination of not having the strong winds to drive the fire and also the better firefighting conditions.
Since 1991, California has seen thousands of fires. What changes in weather patterns since then have influenced how fires start and behave?
I think the differences are relatively small when we’re looking at something on the scale of a fire. As we’re looking at these things historically going forward we will probably be seeing the fire season possibly being a little bit longer and also maybe the amount of acreage being greater.
Then in the last five years here in California you have to factor in things like drought. You look at the drought as a separate entity. But during drier years, you’re going to have more acreage getting burned on average. Then you step back and look and say, “OK are we having more droughts? Are they longer, are they more impactful?”
So there’s multiple moving pieces here. I don’t think you could say that any given fire was the result of climate change. But I think if we look at the pattern of fires then that’s when we’ll be able to see what the impacts are.
What advances in meteorology are improving our ability to anticipate and respond to these types of events?
I remember going to both the Lexington Fire, which was in the Santa Cruz Mountains back in ’85, and a fire out in Point Reyes in the mid ’90s. At that time, the internet was becoming a little more prevalent and they were able to get forecasts and weather data at the scene. And you could write computer models on a laptop to see what the wind flow patterns will be in and around the fire.
[Today] there are better micro- and mesoscale forecast models. So we are able to look at what wind patterns over just an area that is burning during the firefighting process. The incident meteorologists are able to give better briefings about what the conditions are and what they’re going to be in the short term.
It’s important for the firefighting team to have that data. [So they can say] “here are the winds you should expect on the western flank of the fire. Here are the ones on the eastern flank of the fire. This is what the evolution is going to be with the winds over the next 24 hours.” That data is then used by the fire behaviorist and the incident commander to put together how they’re going tackle a fire.
What lessons came out of the Oakland Hills Firestorm?
The largest issues identified afterwards were mostly related to firefighting. These included better communication between agencies, standardization of equipment and ensuring defensible space around homes.
Learn more about wildfires in our recent investigation.
Eric Sagara can be reached at esagara@cironoline.org. Follow him on Twitter: @esagara.