16 December 2010

Strengthening tsunami warning systems

Posted by mohi

Tsunami before and after

Sumatran coastline after 30 meter (100 feet) waves crashed on shore in the 26 December 2004 tsunami. Photo courtesy of NASA Goddard Space Flight Center.

Though it’s hard to imagine a huge wall of water sneaking onshore, tsunamis can catch coastal communities by surprise.

Two posters presented at “NH33A: Transmitting Hazard Science to End Users: What Works, What Doesn’t, and What’s Needed? I” discussed ways to strengthen tsunami prediction and warning that could help save lives.

Nick Holschuh of Carleton College told me about a new computer model that can predict a coastline’s tsunami risk based on data collected immediately after an earthquake, allowing tsunami warnings to be tailored to each unique tremor. Current warning systems issue alerts when a quake resembles one predetermined to cause tsunamis.

Predicting a tsunami is tricky. Scientists need to understand the physics of how seismic waves trigger ocean waves, how the tsunami spreads through the ocean, and how the shape of the coastline influences the way in which the giant wave breaks on shore. “To get accurate results from a model, you need to set a computer on it for days,” Holschuh says.

But the new model, nicknamed RIFT for “real-time inundation forecasting of tsunamis,” reduces the computational work. It keeps the complex calculations that describe the tremor and how the shaking moves water, but it simplifies how the budding tsunami travels through the ocean.

Holschuh modeled nine large tsunami-generating earthquakes from the past decade using RIFT. Considering that models always have inherent error, it accurately predicted places where the waves were large enough to warrant a tsunami warning.

Current warning systems only consider large earthquakes at subduction zones, where one tetonic plate slides underneath another. They pre-calculate the tsunami risk for certain areas given a range of quake locations and magnitudes. When a new quake hits, the system compares that tremor to those in its database and issues the appropriate warnings.

Because RIFT can calculate tsunami risk for each earthquake as it happens, it can be applied to earthquakes that originate from other types of plate interactions. Holschuh says RIFT is intended to supplement other tsunami prediction models prepared by NOAA.

The next step? Once you predict a tsunami, you need to be able to warn people in the affected areas as soon as possible.

Nathan Becker at the Pacific Tsunami Warning Center (PTWC) calculated how quickly the center could warn any country in the Pacific basin of a tsunami. He assumed the worst–an earthquake originating close to shore–and calculated the time for a tsunami to reach land. Then he subtracted the time it takes the center to collect enough data to issue a warning.

Assuming all seismometers are operational and the PTWC calculations are correct, almost 80 percent of the coastlines have more than an hour to prepare for the big wave. But for some places, the wave gets there before the warning.

Two issues contribute to the decreased warning time: either the fault that ruptured is very close to shore, or there are too few seismometers around the ruptured fault to gather enough information to issue the warning.

The seismometer issue can be fixed, and Becker’s results help target the placement of additional sensors to protect underserved coastlines.

But the areas where the fault lies nearby? “The laws of physics are not on our side sometimes,” Becker says.

The best protection is public awareness, says Becker. If you live on the coast, know the warning signs of a tsunami–a strong earthquake, changing ocean levels, and roaring from the ocean–and be ready to run to high ground.

–Melissae Fellet is a science communication graduate student at UC Santa Cruz.