Vasily Titov, Harold Mofjeld, Frank González, Eddie Bernard and Jean Newman, PMEL

Goal

Use numerical models and analyses of observed tsunamis to develop faster and more reliable forecasts
of tsunamis striking coastal communities in Alaska, California, Hawaii, Oregon and Washington.

Accomplishments

Tsunami Generation and Propagation

  •  The MOST (Method of Splitting Tsunami) generation and
      propagation codes were used to simulate the June 10, 1996 Andreanov
      tsunami. Model results agree very well with deep-ocean records.

  •  Studies using theoretical Airy waves show that even small shifts
      in the arrival times of later wave packets can dramatically change
      the times of maximum waves, suggesting a statistical approach
      to forecasting later wave heights in a tsunami.
Tsunami Inundation Modeling

  •  The MOST inundation code was tested against field data collected
      shortly after the July 12, 1993 Okushiri tsunami. Good agreement
      was obtained between the model simulation and field measurements.

  •  The very high maximum water velocity computed - more than 30
      knots - is consistent with field estimates. This tremendous current
      carried large floating objects (automobiles, small boats, logs, etc.) that
      became deadly battering rams; this was the principal mechanism
      responsible for the disastrous damage and loss of 230 lives.
Statistical Forecasting During Tsunamis

  •  Based on analyses of observed tsunami time series, amplitude
      envelopes for later tsunami waves can be predicted from the
      waves that have already arrived at a tide gauge site.

  •  Monte Carlo simulations, using wavelet-derived time series,
      verify the coefficients in the short-term forecasting algorithm.

  •  Tides and other background water levels can strongly affect
      the amount of tsunami inundation along the West Coast and
      Alaska; El Niños can elevate the background up to 0.3 m.
      Background water levels are less important in Hawaii.
Future Directions

Real-Time Forecasting

  •  The MOST model will be used to perform multiple-run sensitivity
      studies of two closely related, but separable, processes-offshore
      wave dependence on distant earthquake magnitude and position,
      and site-specific inundation dependence on offshore wave characteristics.

  •  The results will be organized into an electronic database, and
      associated software will be developed for analysis and visualization
      of this database, including the assimilation of real-time data
      streams; the goal is to provide guidance for forecasting
      and hazard assessment during actual tsunamis.

  •  Depending on the results of sensitivity studies, it may be feasible
      to use a supercomputer version of the MOST model to perform
      real-time computations for emergency guidance during tsunamis

Funding support provided by NOAA/OAR, DOD/DARPA and DOD/DUSD(S)