Skip to main content
U.S. flag

An official website of the United States government

Forecasting Earthquakes in California

Right-click and save to download

Detailed Description

A 99 percent chance of a magnitude-7 earthquake? That's the 30-year outlook for California, according to a new USGS State-wide earthquake forecast. Learn more in this interview with USGS geophysicist Tom Parsons.




Public Domain.


[Music fades in and out.]

Leslie Gordon

Welcome to CoreCast. My name is Leslie Gordon, and today we're going to talk about forecasting earthquakes in California. My guest is geophysicist Tom Parsons, with the U.S. Geological Survey in Menlo Park, California. Welcome Tom.

Tom Parsons

Thank you.


So Tom, what do we mean by forecasting earthquakes? Is that like forecasting the weather?


Well, that's good analogy. With earthquakes, we talk about a little bit longer time. We can't see them coming like we can see the weather coming, but we do rely on what's gone on before in our experience with past earthquakes: other ones we've seen historically or we can dig through the ground and find record of.

And using that information, we can give our best estimate of what we think the next period of time might represent, say 30 years is a typical period that we work with. And then we can say, we think there's a particular percent chance of a damaging earthquake striking a particular region.


So what is the current forecast for California?


Well, we've just completed a State-wide forecast . . . probabilities for every source we know of in California. There are different probabilities for different sizes of earthquakes. For example, for a magnitude-6.7 earthquake, which is the size of the recent Northridge earthquake, we think it's virtually assured in the next 30 years, about a 99.5 percent chance, of happening.

For larger earthquakes, they're less frequent, thus less likely. So for a magnitude-7 earthquake, it's still quite high for the whole State: We think that's about a 95 percent chance. When we get up to the higher levels, like magnitude-7.5, it's roughly 50/50 at a 46 percent chance and for the biggest-magnitude-8 earthquakes-the probability there is about 5 percent in 30 years.


So are some parts of the State more likely than others to generate moderate earthquake?


Yes. If we break the State into two equal pieces, and the dividing line roughly cuts through central California, it looks like the southern half of the State has slightly higher probabilities than the north half. It then again depends on size. The 6.7 threshold I mentioned is pretty much the same for Northern and Southern California-that's 93 percent in Northern California and 97 percent in Southern California. Those numbers are pretty similar.

When we get up to the bigger sizes, like 7.5 earthquake, then we find it's almost more than doubled in Southern California at 37 percent versus about 15 percent in Northern California.


And who cares about these forecasts? Why do we take the time to forecast earthquakes?


Well, I would think that anybody who lives in California would care. They want to be prepared for these when they happen. We know they happened a lot in the past. It's inevitable it's going to happen again in the future.

We want to make sure that our buildings are built properly to withstand what we expect to happen. We think the people should pay attention to this so they can make a decision whether or not to carry earthquake insurance on their house. If we do this in a State-wide fashion, we can get good opportunity to look at all the different regions of the State, and depending on where you live, you can look at this and assess that.


Who will use the results of this new report?


We've already merged this report with the National Seismic Hazard Map Program, which covers the entire country, but our California earthquake rate model is being used there. Those hazard maps inform the building code standards, so that's one application where they carry the probability out to hazard.

They can use the where and when kind of information that we get from probability and put that into a model that mimics the amount of shaking we expect at different parts all over the country. And then they use that information to derive the building codes. Additionally, they will be looked at by insurance providers to make sure the distribution of rates is fair across the State of California.


I know that I've seen in the past forecasts for earthquakes in San Francisco Bay area for example. What's new about this report?


This is the newest in the series of reports we call "The Working Group on California Earthquake Probabilities". These began in 1988 and have looked at various pieces of the State. Every 5 years or so, we assembled a different group under the same heading and re-assess the new data as it comes in. What's different about this report? You were right, there was most recently a report for the San Francisco Bay Area that was published in 2003. This report takes a lot of the methods from the most recent analysis but applies them across the entire State in a uniform way, which is something we've never had before.

As I mentioned before, with earthquake insurance rates, this enables us to look at that much more fairly because in the past, we've had sort of piecemeal reports that may looked at the Southern part of the State or the Bay Area or other pieces of particular faults. This way, we can look at it in a unified way and make sure that the rates are getting an even-handed view.


So you're with the USGS; are there other scientists involved? What goes in to putting together a forecast report like this?


It's a big job. There's a lot of data, especially when we're looking at the whole State. So we had representation from the U.S. Geological Survey, from the Southern California Earthquake Center, or SCEC, California Geological Surveys: CGS.

We've got a smallish working group of 6 or 8 people who are coordinating this effort, but then we've drafted a number of scientists, probably 40 more people, at various points in the analysis to look at various pieces of data, to review what we've done, and to provide their specialties as we've needed it. It becomes a collaborative effort, and we've involved people from universities, not just in California but nationally.


So Tom, in this new report of forecasting California's earthquakes, were there certain results or particular things that really stand out?


We do note that some faults are clearly more dangerous than others. One that stands out to us is the southern part of the San Andreas fault in Southern California. We calculate that that fault has a 59 percent chance of causing a magnitude-6.7 or greater earthquake over the next 30 years. We find that to be the most dangerous fault.

Another one that stands out is in Northern California: The Hayward-Rogers Creek fault system, which runs along the eastern side of the San Francisco Bay, has a 31 percent chance of generating a magnitude-6.7 or greater earthquake in 30 years.

These results, again, kind of highlight the need that we can't do much about this other than be prepared for them. And in both cases, these faults cross urban parts in Los Angeles or San Francisco Bay Area, so we are concerned about those.


So when you say there's a 59 percent chance of an earthquake on the San Andreas Fault, that's twice as likely as not, is that right?


Yeah, that's right. In 30 years . . . if you think about a coin flip, that's a 50/50 chance, so in the case of a 59 percent chance, that's higher. Now granted, this comes along with a great deal of uncertainty. Those people forecasting the weather have the opportunity to see hundreds if not thousands of weather cycles going by and can really have a much better chance at getting it right than we have.

In most cases, especially in California, there have only been people writing records here for a couple hundred years and we haven't seen a complete earthquake cycle for the largest events. This is why we use a probabilistic approach to begin with: It's a way for quantifying our uncertainty. So, in fact, for that southern San Andreas fault, we quote a range when we run through all the different possible models on all the different data that we have, we get a range of 22 to 94 percent probability for that fault, with the average being 59 percent.


Well, Tom, this was really interesting. In light of this forecast, I think I'll wear my galoshes tomorrow. [laughter; Tom: Good idea, I suppose.]

Thank you very much for taking the time to talk to us today, Tom. I really appreciate that.


You're very welcome. I'm glad to be here.

Scott Horvath

For complete transcripts or more information, visit CoreCast is a product of the U.S. Geological Survey, Department of the Interior.

[Music fades in and out.]



Resources related to this podcast:


Music Credit:

"anat.mp3", eatofu

Show Transcript