Predicting the evolution of the extensional step-over in the San Pablo bay area with work optimization

Abstract:

Field data and numerical modeling indicate that the releasing stepover in the San Pablo Bay area, between the Hayward and Rodgers Creek fault, presently seems to lack a strike-slip transfer fault. Analysis of gravity data suggests that only one high-angle normal fault may exist within the step, near the northern tip of the Hayward fault. To investigate a possible evolution of this fault system, we simulate this stepover with the numerical modeling tool Growth by Optimization of Work (GROW). GROW predicts the evolution of a fracture network by analyzing the gain in efficiency, or change in external work, produced by fracture propagation and interaction. We load the San Pablo Bay stepover models with dextral velocity and normal compression that reflects a range of seismogenic depths. The GROW analysis with overlapping starting fault segments separated by 5 km predicts that the Hayward and Rodgers Creek faults propagate toward one another following a gently curved path. The curved path of the fault segment representing the Hayward fault disagrees with the observed planar fault trace, which suggests that this fault may precede the southern propagation of the Rogers Creek fault. We explore various starting configurations that represent the potential geometry at the onset of interaction between the faults, such as different lengths of the two branches of the southern Rogers Creek fault. Throughout the development of this stepover, we analyze the evolution of external work, and change in external work (ΔWext) due to fault growth, interaction and linkage. Additionally, we use the distribution ΔWext at each increment of fault growth to produce probability density functions (PDFs). These PDFs describe fault propagation path forecasts that are defined by 90% confidence envelopes. The propagation forecasts facilitate analysis of the impact of anisotropy and heterogeneity on propagation path.