Scatterer density estimation in the crust by seismic array processing

Abstract

SUMMARY Both the number density of scatterers and scattering strength are parameters characterizing the inhomogeneity of the Earth’s crust. An attempt is made to estimate the number density of scatterers (n-value) using small-aperture seismic array data. Many estimates of scattering coefficients have been made previously; however, in a medium with a scattering strength, the wave energy scattered by a medium containing strong scatterers with a low number density is identical to that produced by a high density of weak scatterers. Knowledge of the number density besides the scattering strength supplies important information on stress concentration in the crust due to the strong scatterer reflecting strong inhomogeneity. Therefore, the number density of scatterers is estimated with array signal processing. A seismic array can decompose the approach direction of scattered waves arriving at the array. With array processing, it is not only possible to estimate the energy arriving at the array, but also the semblance value. Temporal variation of the energy depends strongly on the scattering coefficient. On the other hand, the difference in number density of the medium appears in the temporal variation of the semblance value. As a first-order approximation, the semblance value is inversely proportional to the n-value. For practical purposes, an array response should be considered for n-value estimation. Comparing simulated and observed semblances, number density and scattering strength can be estimated from the time sequence of semblances and stacking power. This method was applied to seismic array data from the aftershock area of the 2000 Western Tottori earthquake (Mw 6.8). A rough estimate of number density, compared to a simulation, is 3 × 10 −2 km −3 .