Researchers are combining data on earthquake shaking intensity, crowdsourced traffic data from the Waze navigation app and police crash reports to find out if car crashes are more likely during seismic events.
So far, the data are too sparse to draw much of a broad conclusion, although the scientists did see some very slight increases in crashes in specific cases, said U.S. Department of Transportation researcher William Chupp at the Seismological Society of America’s Annual Meeting.
The recent availability of large traffic data sets and data from the U.S. Geological Survey’s ShakeMap product, which provides near real-time maps of shaking intensity, have made it possible to study this problem for the first time, said Chupp.
On the U.S. West Coast, where earthquake early warning is delivered through the ShakeAlert system, system operators also want to know more about the effects of shaking intensity to help them fine-tune the geographical extent of their alerts, he noted.
“They need to have answers and research to back up when not to alert, when to alert, how widespread impacts are, where there’s increased risk of crashes happening and everything that comes after earthquakes,” Chupp said.
He and his colleagues looked at the impacts of magnitude 4 and larger earthquakes within the contiguous United States between 2017 and 2023. They paired the location and the intensity of shaking from those earthquakes with Waze’s accident and road hazard data and police crash reports from state agencies. Altogether, their study covers 412 earthquakes and more than 50 million affected road segments.
“It's really hard to get enough data concentrated around certain periods of time to have solid evidence of increases in risk, especially at the time of an earthquake,” Chupp explained. Many earthquakes happen in places where there aren’t a lot of roadways and cars, and the shaking only lasts minutes.
The most interesting results have come from looking at the two hours after an earthquake, he said. “And what we've seen is that for the most part, there's not really enough data to show any sort of increase in frequency of crashes right after an earthquake. In some very specific circumstances, we can kind of see very slight increases here and there in areas that have high shaking.”
The researchers concluded that shaking below IV on the Modified Mercalli Intensity Scale— corresponding to “light” shaking usually felt indoors but not always outside—has a negligible effect on car crashes.
“In areas that see shaking below IV, there's no real difference between any other day of the week, basically, and the day that the earthquake happens,” Chupp said.
One of the biggest missing pieces in the researchers’ data sets is exposure, he noted. “To know with certainty if earthquakes or warning systems cause higher crash frequencies, we have to know how many people are driving on roads when an earthquake happens and at other comparable times when there was no earthquake. That level of detail is really hard to come by.”
The researchers did note some interesting impacts during specific earthquakes. During the 2023 Ojai, California, which also took place during the unusual Hurricane Hilary, “we had a reasonable amount of shaking close to the Los Angeles region, and we actually kind of see a big spike in crashes or crash reports during that day right around the time of that earthquake,” Chupp said.
Chupp and colleagues also analyzed the impacts of last December’s earthquake offshore of Cape Mendocino in California, which triggered tsunami warnings along parts of the U.S. West Coast.
“And what we saw there was actually some crashes that occurred not because of the earthquake or the earthquake alert, but because there were a lot of people that got in their cars near the coast and started driving, and there was just increased amount of traffic on the roadways,” Chupp said. “And anytime there's an increased number of cars on the roads, the likelihood of crashes happening just goes up.”