Phase and group velocities and Q of mantle Love and Rayleigh waves of the first two modes and their azimuthal dependences for the 1963 Kurile Islands earthquake

Abstract

Phase and group velocities and Q of mantle Love and Rayleigh waves from the 1963 Kurile Islands earthquake ( M w = 8.5) were determined over 37 great circle paths by a time variable filtering technique, in a period range 100–500 s for the fundamental modes and 100–275 s for the first higher modes. The preliminary reference Earth model (PREM) explains reasonably well the average dispersion results for the fundamental Love and Rayleigh waves. There exists a small, but significant inconsistency between the observation and the model for the first higher Love and Rayleigh waves. The Q structure of PREM is inconsistent with the observation for the fundamental Love waves, but explains other observations reasonably well. The dispersion of each mode shows a clear azimuthal dependence from which the four azimuthal windows were established. The phase and group velocity measurements for each window were, in general, shown to be mutually consistent. The azimuthal variations are largest for the first higher Rayleigh waves, indicating strong lateral heterogeneity in the structure of the low velocity zone. The first of the four windows is characterized by the largest fraction of Precambrian shields and the second window by the largest fraction of normal oceans. A comparison of these two windows may give some insight into deep lateral heterogeneity between continents and oceans. The observed phase and group velocities of the first window are systematically higher than those of the second window for the fundamental Love and Rayleigh waves at periods up to 400 s, and for the first higher Love and Rayleigh waves up to 175 s. Their differences are greatest for the first higher Rayleigh waves and least for the fundamental Rayleigh waves. Although the fundamental Rayleigh waves show the least velocity differences, their persistence up to a period of longer than 300 s is in striking contrast with some of the pure path phase velocities derived earlier for continents and oceans. A set of models for continents and oceans. PEM-C and PEM-O are not consistent with our observation. The third azimuthal window is characterized by trench-marginal seas and the fourth window by mountainous areas, typically the Asian high plateaus from northern China to the Middle East through Tibet. A comparison of these two windows gives some information about deep structural differences between subduction zones and continental collision zones, both belonging to plate convergence zones. For the fundamental and the first higher Love waves, the phase and group velocities for the third window are markedly low, whereas those for the fourth window are somewhat comparable to those for the second window. Slow Rayleigh waves are evident for two windows, with the fourth window apparently being the slowest for the fundamental Rayleigh above 200 s and for the first higher Rayleigh. For the fundamental Rayleigh waves, the third window is very slow below 200 s, but becomes progressively fast as the period increases and tends to be the fastest window around 400 s, suggesting a deep seated high velocity anomaly beneath trench-marginal seas. The dispersion characteristics of the fourth window indicate a thick high velocity lid with an extensive low velocity zone beneath it. The shield-like lithosphere, coupled with an extensive low velocity zone, may be a characteristic feature of continental collision zones. The particle motion of the fundamental Love waves was found not to be purely transverse to a great-circle connecting the epicenter to a station. The departure from the purely transverse motion is systematic among different periods, different G arrivals (G 2, G 3,…) and different stations, which may be interpreted as being due to lateral refraction.