COMPUTATION AND APPLICATION OF TSUNAMI TRAVEL TIMES

The Tsunami Warning System uses tsunami travel time information to define the extent of warning and watch areas and to provide estimated arrival times to the public and emergency management officials in potentially affected areas. In order to decrease warning response time, the West Coast and Alaska Tsunami Warning Center has pre-computed travel time values for approximately 700 locations around the Pacific, Atlantic, Indian and Arctic ocean basins.

The method used is an application of Huygen's Principle, which states that all points on a wavefront are point sources for the production of secondary spherical waves. The travel times are computed using the shallow-water wave speed approximation on a finite-difference grid with 4 arc minute spacing, except near the warning point where the increment is 30 arc seconds. The results are then subsampled and stored at 12 arc minute resolution. Because the calculations are reversible, travel times are computed to each grid cell within the ocean basin starting from the point of interest, for example a tide gauge or a DART buoy. In the Pacific, the travel times to distant, seismically active coastal regions are adjusted to account for the fact that basin-wide tsunamis are usually generated seaward of the continental shelf.

The data are stored as three-dimensional arrays in six netCDF files; totaling 1.5 gigabytes. The travel times for both the Pacific and the Atlantic oceans are divided such that the values for locations within the WC/ATWC area of responsibility are in one file and those for sites in other regions of the basin are in a second file. The average computational time for a wavefront is more than 4 minutes in the Pacific and Atlantic oceans and nearly 2 minutes in the Indian and Arctic oceans, but the travel time values from an epicentral location to each warning point can be extracted from one of the files in less than 5 seconds.