ON 14 June, at 7.50 pm local time, an undersea quake rocked the coast of northern California, and the entire west coast of the US and Canada was instantly put on a tsunami alert.
The alert turned out to be a false alarm, but the patchy response to the scare is raising questions about whether the vulnerable coast will be ready if a big tsunami really does strike. And evidence of past tsunamis is showing that big ones do happen, and with ominous frequency.
Last Tuesday’s alert was triggered by a magnitude 7.2 quake that occurred about 140 kilometres off Eureka, a town about 400 kilometres up the coast from San Francisco. The alert was cancelled an hour after it was issued, once the seismic data revealed that the event was a “strike-slip” quake on a fault in the middle of the Gorda plate, in which two sides of the fault slipped horizontally past each other and didn’t displace enough water to cause a tsunami.
Advertisement
However, the west coast has reason to be on tenterhooks. If the quake had been along the Cascadia subduction zone that runs along the coast, 10-metre-high waves could have pummelled shores all the way from northern California to Vancouver in Canada. At this type of fault, one tectonic plate suddenly slips under another, displacing vast volumes of water. The resulting tsunami would have been comparable to last December’s Indian Ocean tsunami.
Had one happened last week, the tsunami warning system would have come up short in many places. “The response was variable depending on where you were on the coast,” says Lori Dengler, a geologist at Humboldt State University in northern California.
One problem was information overload. Two tsunami warning centres, one in Alaska and another in Hawaii, automatically issue a tsunami warning in the event of an undersea earthquake of magnitude 7 or greater. The warnings are sent out to the news media and to emergency services in major cities, and are also broadcast on a radio channel run by the US National Oceanic and Atmospheric Administration (NOAA).
While having multiple sources of information can be useful, it also creates confusion, says Dengler. In this case, the Alaskan centre first issued a warning for the US and Canadian west coast. This was followed three minutes later by a bulletin from Hawaii, saying that there was no risk of a tsunami. Although the Hawaiian bulletin clearly stated that its all-clear did not apply to Alaska, California, Washington, Oregon or British Columbia, the two seemingly contradictory messages created confusion among the emergency workers and the media. “It’s really time to re-examine the wording of these messages given the modern cyber-environment,” says Dengler.
There were other problems too, especially with disseminating the warning. For instance, Mike Dever of the Office of Emergency Services in Santa Cruz, California, was not notified directly. “I did not get a page from the tsunami warning centre, or a telephone call,” he says. “The on-duty fire chief heard about it when his mother called him, and my in-laws found out from their relatives in France.”
At least Dever eventually got the warning. Some regions along the US west coast cannot receive NOAA’s radio signals because the signals are blocked by mountains and they have to be contacted by phone from Seattle. But on Tuesday a downed phone line meant the message did not get through. Worse still, no one realised that these areas had not received the warning. “This showed us where we had a weak link in the chain,” says Brad Colman, of the National Weather Service (NWS) in Seattle.
But other things did work very well. In some areas, particularly Crescent City, California – which was hit in 1964 by a tsunami caused by a quake in Alaska – the dissemination of information “went like clockwork”, says Laura Furgione, NWS director for the Alaska region in Anchorage. She notes that 19 communities along the west coast are now designated “tsunami ready”, including Crescent City.
“The on-duty fire chief heard of the tsunami warning from his mother, and my in-laws found out from relatives in France”
However, the public’s response to the warning left a lot to be desired. “There were a lot of people standing around trying to confirm if it was real,” Furgione says. “The public did not react [quickly] enough.” Colman adds that some people actually went to the beach to watch for the wave, and drove rather than walked to higher ground, clogging the escape routes. Many also called the emergency 911 number for information, blocking up the lines. “I would give it a B grade on whether we got the right response,” he says.
For geologists like Dengler, the blanket warning along the entire coast creates unnecessary disruption, since knowing the geology of the quake can narrow down the threat. In this case, as the quake was not caused by a subduction slip, the only way it could have caused a tsunami was by triggering an underwater landslide. This would only occur in areas where tremors were felt, so Dengler says that the warning should have been trimmed to just parts of California and Oregon once the location and nature of the earthquake were known. She suggests that if geologists like her had been consulted, many regions could have avoided evacuation and that the alert could have been cancelled within half an hour.
Of course, if the Cascadia fault does rupture, the entire coast will go on high alert. The fault has not produced a major tsunami since 1700, which is before European settlers arrived in the Pacific north-west. By studying deposits in coastal marshes and Japanese records of that tsunami, geologists concluded in the 1990s that a magnitude 9 earthquake triggered the wave. But they were unclear how often such giant quakes occur.
Now Harvey Kelsey of Humboldt State University and his colleagues have reconstructed the longest continuous record of tsunamis in the region. They studied cores drilled from Bradley Lake in southern Oregon and found substantial amounts of sand and saltwater that have washed into the lake 12 times in the past 4600 years. The lake is 500 m from the coast and 5.5 m above sea level – just the right location to record tsunamis of 10 m or more from Cascadia quakes, but not smaller ones from more distant sources on the Pacific Rim, says team member Alan Nelson of the US Geological Survey in Denver, Colorado.
The record shows that tsunamis occurred every 390 years, on average, but tended to cluster in time. For instance, after a millennium-long hiatus, four tsunamis smashed the coast from AD 200 to 1000, and then there was another long gap before the 1700 tsunami (Geological Society of America Bulletin, vol 117, p 1009). The uneven timing of past Cascadia quakes leaves the outlook for the next big one unclear. “It could happen in the next several decades or in 600 years,” says Kelsey.
To try to improve quake forecasts, David Scott of Dalhousie University in Halifax, Nova Scotia, and his team looked at marshes and forests for subtle clues that could have predicted the magnitude 9.2 quake in Alaska in 1964. From the microfossils of foraminifera and algae, they deduced that some marshes in the region sank by up to a few centimetres as early as 2 to 10 years before the big quake. Cores drilled from the Oregon coast also showed similar changes before big quakes over the past 3000 years (GSA Bulletin, vol 117, p 996). Scott suggests sensitive GPS monitors or tilt meters could spot such minute changes in time to warn residents of an increased quake risk.
But knowing about a possible tsunami years in advance is no good if the warning systems don’t work properly on the day. Many are glad that last Tuesday’s false alarm provided a dress rehearsal for the real event. “It was a good drill,” says Vern Losh, director of emergency services in Sonoma County, California. Dengler agrees, despite her reservations about inadequacies in the system. “It was a very good event,” she says. “It didn’t hurt anyone and it pointed out problems.”
