The Accuracy of Tsunami Predictions

Science’s best guesses about the location of seafloor features like seamounts or canyons or reefs and their dimensions come from soundings, which are physical measurements of the distance between the surface and the ocean bottom at a given location. Soundings are made by ships, but the process is expensive. In part because of the high price tag, only about 11 percent of the ocean bathymetry has been measured in this fashion.

Estimates of what the other 89 percent of seafloor look like are derived from altimetry measurements made by satellites of the height of the ocean surface. Satellites infer what the gravitational pull is at any given point; the greater the gravity, the higher submarine seamounts must be.

This method has been used over the years by researchers at Scripps Oceanography who supply ocean data to Google Maps, among other users, to fill in the blanks. Bathymetry data feed into what scientists call numerical models, or simulations that also rely on mathematics and hypotheticals” to estimate likely tsunami behavior. Errors in altimetry data can cause satellite-derived estimates of elevation to be off by several hundred meters.

“While satellite altimeters provide this global perspective on seafloor depth, they lack the accuracy and resolution that is obtained by multibeam echosounders aboard large research vessels such as [Scripps Research Vessel] Sally Ride,” said Scripps Oceanography geophysicist David Sandwell.

Sepulveda’s team created a new model by analyzing bathymetry data collected from several locations around the world and calculating how far off those data are from reality. The model they created then generates a margin of error estimate that can be used to inform a range of other oceanographic models, including tsunami propagation models.

They used the model to look at past tsunamis and found that the leading wave generally has a wavelength so large that any bathymetry errors do little to affect it. Trailing waves, which come minutes or hours later, have shorter wavelengths, placing them on a scale more comparable to the size of bathymetry errors. Those bathymetric features can magnify or attenuate the waves in a myriad ways, as can their interaction with normal breaking waves.

In Chile, many coastal towns are built around bays, which provide natural protection from storms most of the time. But when trailing tsunami waves strike, those same geographic features can focus the energy of the waves, creating waves that are larger than the first, and more localized. That was the case in 2010, where residents of the fishing village of Dichato, Chile recalled that it was the third tsunami wave that swept the town away, several hours after the 3:30 a.m. quake.

“The systematic study comparing detailed sea beam surveys of bathymetry and satellite derived bathymetry highlights the differences that can have a big impact for mitigating hazards from the secondary and trailing waves from tsunamis,” said study co-author Jennifer Haase, a geophysicist at Scripps Oceanography. “It can also be useful for many other ways that satellite derived bathymetry is used, for example understanding ocean currents.”

The study appears in the Journal of Geophysical Research Solid Earth.