About three hours after a magnitude 8.8 earthquake, rocked Chile on Feb. 27, a buoy off Peru’s coast recorded a wave just over a foot tall heading toward Hawaii and the West Coast of the United States. The buoy was one of nearly 50 Deep-ocean Assessment and Reporting of Tsunamis (DART) sensors deployed around the world in tsunami hot zones.
The catalyst behind developing the system was an evacuation of the Hawaiian Islands following warning of an imminent tsunami in the late ’80s that didn’t cause as much damage as expected. However, the evacuation cost the state $60 million. “Really the main driver was to eliminate these false alarms because nine out of 10 earthquakes did not generate a tsunami that could be damaging,” said Chris Meinig, leader of the Engineering Development Division at the National Oceanic and Atmospheric Administration’s Pacific Marine Environmental Laboratory (PMEL).
Photo: A map of the DART stations. Courtesy of the NOAA.
The current generation of DART buoys includes a tsunometer on the ocean floor that measures wave height and water pressure at 15-second intervals and then predicts what the wave height will be in the next interval. The sensor on the ocean floor then sends the measurement to the buoy on the surface via an acoustic modem. If the difference between the actual measurement and the predicted measurement exceeds a certain threshold, then the buoy sends a report of the deviation to a tsunami warning center via satellite link. From the warning center the data is sent to partner countries including Australia, Chile and Indonesia.
Under normal operating conditions, the buoy reports its measurements every six hours so technicians at shore stations can check the data and ensure that the buoys are working properly.
Improvements in the Technology
In the late ’90s, just being able to measure and report the wave height to a station in Hawaii or Alaska was considered state of the art. Now the system uses a global satellite network to report the data, which can be ingested into real-time models for decision-making.
The next generation of data models will allow for site-specific forecasts, Meinig said. “Now we can get site-specific forecasts,” he said. “Instead of saying, ‘Well, the whole West Coast has a problem.’ We can now narrow it down much like you would for a hurricane forecast.”
Keeping the sensors operational can be a challenge. “We’re working in some of the worst ocean conditions, whether they’re off the southern tip of New Zealand or in the Aleutian Islands where it is extremely difficult to even service these,” he said. “You’ve got a very small servicing window. And just keeping the buoys alive is an extreme challenge, not just from the weather but also from fishing, either intended fishing or unintended fishing, [and] vandalism.”
To improve the maintenance of the sensors, PMEL is working on a second generation of smaller, more rugged, less expensive sensors incorporating an all-in-one design that’s deployed in minutes as opposed to the six to eight hours it currently takes.
Following the magnitude 8.8 quake, Hawaiian officials decided to evacuate residents from coastal communities, issuing a tsunami warning and activating sirens at 6 a.m. on Feb. 27. “The models said there probably wouldn’t be very big tsunami, but the Hawaii Civil Defense decided to evacuate anyways,” said Lauren Koellermeier, PMEL outreach coordinator. The tsunami prediction models are 80 percent accurate, she said.
“While we weren’t expecting a huge tsunami in Hawaii, it was on the order of something that could have caused damage,” said Chip McCreery, director of the Pacific Tsunami Warning Center. “If you look at the historical data, you find that there are lots of earthquakes that are on that order of size that have produced more problems actually than this tsunami produced.”
McCreery said the readings from the DART buoys confirmed that evacuating was the standard action to take based on data about the earthquake. “In this particular case, for this event, actually the earthquake parameters were pretty good,” he said. “From the initial guess from the earthquake the forecast model didn’t change a whole lot based on the deep-ocean readings. So in this case it was more or less used as a kind of a confirmation of the forecast.”
He said the data generated from the sensors following the recent earthquake will be helpful for making improvements to the models.
[Photo courtesy of the NOAA.]