Crustal Stress Rotation Along the San Andreas and San Jacinto Faults: A Modeling Study With Constraints From Seismology, Geodesy, Topography, and Gravity

The active tectonics of the southern San Andreas transform plate boundary system respond to and contribute to the 4 D stress field throughout the region. We investigate the nature of this stress field in Southern California, with particular focus near the major strain-accumulating San Andreas and San Jacinto faults, by creating a forward model that incorporates observations from seismology, geodesy, gravity, topography, and earthquake rupture history. The forward model consists of three independent crustal stress field components: (1) a plate driving force of undetermined magnitude and orientation; (2) a heterogeneous fault loading stress accumulation along locked fault segments; and (3) spatial variations in crustal stress due to differences in topography. The forward model is then compared to the in situ stress field orientation inferred from earthquake focal mechanisms. We estimate the magnitude of the in situ stress field as that required to maintain its orientation in the presence of topography, which tends to resist the motion of strike-slip faults. Our results indicate that differential stress at seismogenic depth must exceed 40 MPa. To assess the orientation of the plate driving stress, we consider twelve independent segments of the San Andreas Fault System from Imperial Valley through Parkfield. We determine that along much of the central San Andreas fault, the maximum horizontal stress (SHmax) is oriented north-south (~0ºEofN), but that from Coachella to Imperial SHmax is rotated clockwise, oriented ~12ºEofN. Furthermore, SHmax along the San Jacinto and Superstition Hills segments gradually rotates clockwise from ~3ºWofN in the south to ~8ºEofN in the north. With these results, we are able to match the in situ stress orientation of most (≥85%) of the near-fault strike-slip areas to within 15º, comparable to the errors associated with focal mechanism determination. Creating a forward model consistent with so many different types of observations corroborates the credibility of the individual model components and reinforces the robustness of the inferred stress rotations. The underlying physical cause of these rotations may be related to differences in segment locking depth, earthquake cycle maturity, fault friction and strength, or the nature of the connecting segments.