Identification of 1D velocity structures in Kashiwazaki-Kariwa NPS based on dense earthquake observation and microtremor array measurements

Abstract

A dense earthquake array observation is expected to provide valuable information of a spatial variation of subsurface structures. A receiver function inversion is one of useful technique to estimate 1D velocity structure based on a three-component surface earthquake data. However, since receiver function inversions are highly non-unique, more competing models were identified from the inversion. Actually, also in this case by using our data, it was seen that the identified model from the receiver function inversion could not fully explain near PS logging model. Moreover calculated phase velocities were also not in agreement with each other. Therefore, in order to avoid the non-uniqueness of a receiver function inversion, we employed a joint inversion of receiver function data from a dense earthquake array observation and Rayleigh-wave phase velocity data from microtremor array measurements. In our joint inversion, relative weights of Rayleigh-wave phase velocity data against receiver function data were determined based on Akaike’s Bayesian Information Criterion. As a result, a revised model by joint inversion is in good agreement with near PS logging.