Interpretation of non‐double‐couple earthquake mechanisms derived from moment tensor inversion

Abstract

Some of the proposed physical explanations for non‐double‐couple moment tensor inversion results for shallow‐focus earthquakes include source multiplicity, rupture on nonplanar fault surfaces, and tensile failure under high fluid pressure. Experiments using synthetic seismograms demonstrate that the first two effects can result in non‐double‐couple mechanisms due to the inadequacy of the source model used in the inversion; the latter explanation, however, implies the existence of non‐double‐couple mechanisms that are intrinsic to the source process. A waveform inversion algorithm for estimating seismic source parameters, in which the elements of the deviatoric moment tensor are allowed to have indepedent time histories, is applied to four approximately equal sized (M ∼ 6) shallow‐focus earthquakes, all of which had large non‐double‐couple components in their moment tensor solutions when the solutions were constrained to be point sources with step function time histories. The Coalinga, California, event of May 2, 1983, is shown to have had a complex source with a significant change in geometry, and the Yemen earthquake of December 13, 1982, is shown to have been a multiple event with little or no change in source orientation. Of the two largest events in the May 1980 Mammoth Lakes, California, sequence, the larger had a complex time history, but the smaller showed no evidence of source complexity. For these two events, tensile failure under high fluid pressure is the most likely explanation for the large non‐double‐couple components. For the larger event the complex time history could be explained by the existence of barriers to fluid flow which are broken during the rupture process. Thus all three of the previously mentioned explanations, either alone or in combination, may result in non‐double‐couple solutions being found.