We observed Rayleigh and Love wave dispersion characteristics between 5 s and 20 s using yearly-stacked cross-correlation functions (CCFs) of ambient seismic noise data, from the dense seismic broadband network consisting of 17 stations in the Ulaanbaatar region, Mongolia. The extracted empirical Green's functions of Rayleigh and Love waves between 136 pairs of stations showed substantially large signal-to-noise ratios (> 17) in the target periodic range, indicating clear trends of wave propagation up to ~270 km. We removed P-wave contamination, which exists as precursors, from the stacked CCFs using vertical-to-radial (ZR) and radial-to-vertical (RZ) components and succeeded in obtaining dominant Rayleigh-wave signals. The estimated Rayleigh and Love-wave phase velocities using the matched-filter frequency-time analysis showed little variation over the period between 6 s and 10 s, showing small subsurface heterogeneity in the Ulaanbaatar region. The estimated Rayleigh wave phase velocity map for 8 s indicated small (~0.05–0.1 km/s) velocity contrasts between the western and eastern parts, and this suggests the possibility of local-scale seismic velocity heterogeneity in the study area. The resulting velocity maps correspond well with the estimations derived with a sparse seismic array. We also found a feature which could extend our Rayleigh wave phase velocity dispersion to a much longer period, and shorter spacing stations, (which are typically excluded in seismic interferometry) could be included using the SPAC method. The combined use of seismic interferometry and SPAC method will make it possible to conduct a much more precise surface-wave tomographic study in the Ulaanbaatar region.
Keywords: Ambient noise cross-correlation, Ulaanbaatar region, Rayleigh wave, Seismic Interferometry.