Crustal and uppermantle velocity structure of the northern Korean Peninsula constrained by ambient noise cross-correlations

Abstract:

The seismic velocity structure of the northern Korean Peninsula has not been well known because available seismic data observed inside the region is very rare. Therefore, constructing a one-dimensional (1-D) reference velocity model will be a meaningful first step to understand the detailed velocity structure and also the tectonic evolution of the region. In this study, we use two-year long (2010-2011) ambient noise data recorded by three component broadband seismometers of KIGAM, KMA, F-net, and IRIS networks surrounding the study area. Ambient noise cross-correlations are calculated for station pairs sampling the study area. Multiple filter technique is applied to measure group and phase velocity dispersions of the fundamental mode Rayleigh and Love waves for period ranges between 5 and 70 s, which are sensitive to shear wave velocities in the crust and uppermost mantle. Recently developed inversion method using the Bayesian technique is applied to estimate 1-D models of shear wave velocities and corresponding uncertainties. Combined with the Markov chain Monte Carlo (MCMC) technique, the posterior probability density is evaluated for given data and the prior. Boundaries, velocities, and anisotropy parameters of layers are searched together on the assumption of the layered half-space model. The used method automatically adjusts the number of layers and degree of data fitting by the Bayesian Information Criterion (BIC) estimation and scaling of the data-covariance matrix, respectively. The estimated anisotropic S-wave velocity model in the crust and upper-mantle can help to investigate tectonic processes of the region. For practical use, our velocity model for the Northern Korean Peninsula can be useful to enhance monitoring power of the underground nuclear tests in given area.