EMPIRICAL GREENS-FUNCTION ANALYSIS - TAKING THE NEXT STEP

An extension of the empirical Green's function (EGF) method is present ed that involves determination of source parameters using standard EGF deconvolution, followed by inversion for a common attenuation paramet er for a set of colocated events. Recordings of three or more colocate d events can thus be used to constrain a single path attenuation estim ate. I apply this method to recordings from the 1995-1996 Ridgecrest, California, earthquake sequence; I analyze four clusters consisting of 13 total events with magnitudes between 2.6 and 4.9. I first obtain c orner frequencies, which are used to infer Brune stress drop estimates . I obtain stress drop values of 0.3-53 MPa (with all but one between 0.3 and 11 MPa), with no resolved increase of stress drop with moment. With the corner frequencies constrained, the inferred attenuation par ameters are very consistent; they imply an average shear wave quality factor of approximately 20-25 for alluvial sediments within the Indian Wells Valley. Although the resultant spectral fitting (using corner f requency and kappa) is good, the residuals are consistent among the cl usters analyzed. Their spectral shape is similar to the the theoretica l one-dimensional response of a layered low-velocity structure in the valley (an absolute site response cannot be determined by this method, because of an ambiguity between absolute response and source spectral amplitudes). I show that even this subtle site response can significa ntly bias estimates of corner frequency and kappa, if it is ignored in an inversion for only source and path effects. The multiple-EGF metho d presented in this paper is analogous to a joint inversion for source , path, and site effects the use of colocated sets of earthquakes appe ars to offer significant advantages in improving resolution of all thr ee estimates, especially if data are from a single site or sites-with similar site response.