Paper Summary
We acquired data from a North Sea seabed seismic test measured by a 6-degree-of-freedom (6-DOF) sensor that provides three translational measurements (acceleration) along three axes and three rotational measurements of the angular motion of a particle about the axis. Portions of seismic records containing P, S and surface waves have different polarization motion patterns: linear, elliptical, or spherical and multiple energy arrivals can be present in a single seismic record. We applied polarization analysis to 6-C data to estimate elliptically polarized events and
filter them out from the vertical component data. The proposed filtering process allows for the detection and separation of seismic events, even if they are not of a pure state. Our results show that polarization analysis of 6-C data can, with optimized scaling and thresholding, successfully reduce surface waves and other events with elliptical particle motion polarization from the vertical component record. We demonstrate that direct measurements of the rotational energy help to overcome multiple arrivals problem in the polarization analysis.