Imaging the Subsurface with Ambient Noise Autocorrelations

Abstract

Abstract Autocorrelations created by stacks of near-offset traces from virtual source gathers are used to form an image of the deeper subsurface. We minimize the masking effects of the effective source time function by subtracting the survey-wide average autocorrelation from each trace. The result is a zero-offset reflection image of the subsurface generated by ambient noise correlation. The technique can be particularly useful for imaging the mid and lower crust, in which traditional seismic methods have penetration problems. We show examples from a one-component 3D survey and a three-component 2D profile. The 3D example shows the crust in the transition zone between the continent and the Inner Borderland in the Los Angeles, California, area, and for the first time, shows an image of the lower crust. The 2D profile provides both a P image and an S image of the basement interface in the San Bernardino basin in southern California.