Impact of distribution of seismic ambient noise sources on surface wave characteristics

The seismic ambient noise observation is widely used in site characterization studies to obtain the subsurface velocity structure because of the simplicity and low cost. Different single and array methods have been introduced in this field to study seismic ambient noise characteristics. In this paper, we investigated the effects of distributions of the seismic ambient noise sources on surface wave properties. In this regard, two datasets are simulated regarding omnidirectional and nonomnidirectional distributions of sources. In the former, it is assumed that the waves travel through the array from all even azimuths and for the latter, an arc shape distribution is considered. To process the data, the RTBF (Rayleigh three component beamforming) array signal processing technique is applied to retrieve Rayleigh and Love phase velocity dispersion curves and Rayleigh wave ellipticity. Furthermore, two wellknown single station methods HVTFA (horizontal to vertical time frequency analysis) and RayDec (random decrement technique) are implemented to extract the ellipticity curves. For all the applied methods, impacts of the two considered distributions of sources are investigated. Then, the results are compared to theoretical curves which are computed from a specific earth model used to produce synthetic noise data. Moreover, a data recorded in Ramsar (north of Iran), as a typical case of nonomnidirectional distribution of noise sources, selected to justify our assertion in processing synthetic noise data. Benefitting the mentioned array and single station methods, the characteristics of surface waves are studied and the shear wave velocity profile is estimated by joint inversion of retrieved dispersion and ellipticity curves.