A New Cluster Event Method for Accurate Determination of Attenuation and the Scaling Law between Corner Frequency and Seismic Moment

Seismic imaging of wave speed and attenuation is one of the most effective probes to reveal the mantle structure and geodynamic process in the earth’s interior. Seismic velocity and attenuation provide complementary information for the interpretation of physical properties of the mantle, whereas the latter is mostly sensitive to temperature and water content but insensitive to compositions and uncertainties in earthquake origin times. However, observed waveforms are dictated by both the source spectrum and attenuation of the medium. To isolate the influence of corner frequency fc from the determination of t* or quality factor Q is always a challenging task for attenuation studies. Here we present a new technique, the cluster event method (CEM), to simultaneously determine both fc and t* (or Q value) for all event-station pairs. With physically correct arguments that neighboring paths from clustered events share the same Q and each event only allows a single fc, we formulate a cost function to be minimized as the root-mean-squares of the differences between observed and predicted spectra. An optimization algorithm of simulated annealing is employed to search for global minimum of the cost function that yields optimal solutions of the path-dependent Q and corner frequencies of all the events. The CEM provides great improvements on resolving the tradeoff between fc and t* and investigating the relationship between fc and seismic moment Mo in the robust way. Furthermore, we expect to build more valid scaling law between fc and Mo by CEM for global earthquake datasets.