Long period seismic waves from large, near-surface nuclear explosions

Seismic surface waves were well-recorded from the larger explosions of the U.S. test series detonated in the Marshall Islands during the spring and summer of 1958 and of the U. S. S. R. test series detonated at the Novaya Zemlaya test site during October 1958. In addition to waves of the fundamental Rayleigh mode, some unusual or unexpected waves were identified. These include: 1.) Love waves with lengths as great as 90 km at Hong Kong from the U. S. S. R. explosions, 2.) Love waves at Agra, India, from the U. S. S. R. explosions, 3.) waves of the 1st shear mode at Agra and Uppsala, Sweden, from the U. S. S. R. explosions, and 4.) Love, 1st shear and PL waves at Guam from the U. S. explosions. Group velocity data were derived for many paths and, for the U. S. series, these data are very accurate because they are based on exact knowledge of locations and origin times. For the U. S. S. R. explosions, revised locations and origin times based on a limited number of P-wave observations were used to determine group velocities. Using records from Hong Kong and Honolulu for two U. S. tests, one at Eniwetok and one at Bikini, oceanic phase velocities in the period range of 15 to 40 seconds were measured for the path between these islands. Using Brune's method for initial phase determination with a known phase velocity, an initial phase, φ0, lying between +3π2 and +π2 was determined for the U.S. explosion, Oak. According to Aki (1960), these values of initial phase are associated, respectively, with forcing functions of a downward impulse and an explosive impulse. Seismic magnitudes of 4.7 and 4.8 were assigned to the U. S. Oak and Poplar events on the basis of the surface wave data, while magnitudes of the larger U. S. S. R. tests range from about 4.0 to 4.5. The ratio of seismic energy as computed from the surface wave magnitude to the total explosive energy available (yield) is apparently greater for the U. S. explosions, assuming equal yield for the two shots compared. This suggests a higher altitude of detonation for the Russian events. The Fourier amplitude spectra of the surface wave trains are generally single peaked where a significant portion of the path traversed is oceanic, but for Uppsala, where the path is short and continental, a second peak occurs at periods of about 10 to 13 seconds. The ratios of the predicted amplitudes, based on the data of one station and on a reasonable dissipation factor, to the actual recorded amplitudes vary with azimuth by as much as a factor of 5. These variations may be explained by asymmetry at the source but might also be explained by instrumental and geologic factors. Long period components of P and S and multiples thereof were sometimes recorded from these events.