摘要

A method for predicting the signal strength of a tsunami at coastal tide gauges once the event is underway is presented. It is based on the calculation of a transfer function for a linear system whose input is obtained from Deep-ocean Assessment and Reporting of Tsunamis (DART) bottom pressure recorders (BPRs) and the output from coastal tide gauges. To perform the calculations, a data window for the input and output signals must initially be chosen that requires a compromise between signal purity and resolution. Limited experience suggests that windows in the range of 6-12 hours duration are a reasonable choice for 1-minute data. The method is illustrated with three examples, the first two taken from the tsunami generated by the Great Tohoku Earthquake of 2011 and the third from the tsunami generated by the Chilean earthquake of 2010. Magnitude spectra from the transfer functions in each case reflect the influence of oscillations present in the tide gauge data, a requirement that must be met if the transfer functions are to be of predictive value. The magnitude spectra for each of two pairs of DART BPRs employed in two cases were generally similar, suggesting that the results are robust in the sense that they are not strongly location dependent. This approach provides an alternate view of the forecasting process that should be useful in identifying spectral components in the coastal response that may be critical to the forecast. Once implemented, the method will be rapid, easily competing with other precomputed database methods. Finally, with further development the method could be used to augment other tsunami forecast methodologies that are presently in use.

  • 出版日期2014-5

全文