Rayleigh wave ellipticity estimation from ambient seismic noise using single and multiple vector-sensor techniques

In this paper, we present three new methods to evaluate the ellipticity of polarized signals generated by seismic ambient vibrations. Classically, the estimation of Rayleigh wave ellipticity is performed using a spectrum ratio approach [15]. Here, we propose to use two single-sensor techniques (based on ellipse fitting [8] and random decrement [2]) and one multi-sensor technique (based on quaternion-MUSIC [12]) to substitute the existing algorithm. Rayleigh wave ellipticity is an important geophysical parameter because its estimation as a function of frequency leads to velocity vs. depth estimation, i.e. identification of the underground structure. We illustrate the three algorithms on simulation data and compare them to the existing approach.