Ewing, Tolstoy, and Press of Columbia University reported that “a striking correlation between the occurrence of a short-period earthquake phase (T phase) traveling through the ocean with the speed of sound in sea water and the occurrence of tsunamis has been observed.”
Their statements about the characteristics of T are incorrect in every essential detail.
For the Pacific Ocean, they list five tsunami between 1933 and 1946, of which the largest, on April 1, 1946, was generated by an earthquake for which no T was recorded. They neglect to mention the earthquake of January 23, 1938, near Hawaii, which produced the largest T recorded on the Pacific coast to date, but no tsunami. The importance of these outstanding exceptions, errors in reporting the data, and uncertainty concerning the actual number of T phases recorded on the Pacific coast combine to make the evidence for any value of T as a tsunami warning decidedly inconclusive.
In the Atlantic, the proposal that T be used as a tsunami warning reduces to an absurdity. Ewing, Tolstoy, and Press state that between 1939 and 1948 “20 Dominican Republic shocks produced T phases,” and that one of them was followed by a definite tsunami. Actually, more than 200 Dominican Republic shocks produced T within that span of years, and many in other Atlantic regions. With one minor tsunami among 200 to 250 T phases, the correlation is not impressive.