USING PHYSICAL MODELS AND WELL LOGS TO DETERMINE STRESS DIRECTIONS IN ANISOTROPIC ROCKS: IMPLICATIONS FOR THE SAN ANDREAS FAULT SYSTEM IN CENTRAL CALIFORNIA

Borehole breakouts are commonly used to infer stress directions in the upper crust. In California, breakouts demonstrated maximum horizontal stress directions at a high angle to the San Andreas fault, leading to the idea that the fault was unusually weak. However, the breakouts are commonly found in layered, sedimentary rocks. We investigate how this type of layer anisotropy may impact breakout formation using physical models and well logs from central California.

For the physical models, we use a cemented granular rock analogue composed of sand and gelatin. The material is cut into blocks and shortened uniaxially by 5%. To simulate a borehole, a hollow tube is drilled into the block. After the model rests for 30 minutes, the void is filled with molten wax to create a borehole cast. In addition to working with isotropic blocks, we create layered versions of the model material and vary the layer attitudes relative to the applied stress. Breakouts are observed in many of the configurations with a variety of cross-sectional shapes. Breakout directions, calculated by approximating cross sectional slices as ellipses, are fairly consistent in isotropic blocks. The breakout directions become more scattered when layers are misaligned with stresses; in these cases, some breakouts appear to nucleate near layer interfaces.

For the analysis of wells in central California, we reexamine those that include both dip and caliper logs to allow comparison of bedding directions to breakouts. We use slightly different criteria to determine breakouts than in previous studies, and we cannot always replicate the directions from the earlier work. In many cases, we do find breakouts that would suggest maximum horizontal stress directions at a high angle to the San Andreas fault. These directions, however, are also often parallel to the dip direction of layers in the borehole. Using the analogue models as a guide, it is possible that breakouts use layer interfaces to nucleate and may not necessarily reflect regional stresses.