Understanding earthquake physics is a critical step on the road to forecasting the short-term behavior of faults with obvious consequences for hazard mitigation. This is a project whose goal is to examine the dynamics associated with earthquake rupture. The studies to be carried out will provide much more comprehensive constraints on the way that a major fault zone behaves. Specifically, the project will combine detailed imaging of the San Jacinto Fault (SJF) in Southern California using an array to characterize the fault zone in the subsurface. It will couple this with surface outcrop and mapping of the fault zone, paleoseismic analysis, GPS analysis of crustal deformation, and theoretical work on seismic propagation to understand how factors such as fault damage, juxtaposition of different rock types, and segmentation affect the behavior of the fault zone. The Principal Investigators bring together current ideas about the rupture process and outline an approach that may be able to provide a quantitative understanding of the evolution of fault zone structures and related deformation phenomena (seismicity, strain fields) in actively deforming regions. This approach requires a framework that accounts for faults with evolving geometrical and material properties, as well as time-dependent interactions between the seismogenic zone and underlying viscoelastic substrate.
This project contained two primary foci: earthquake history of the San Jacinto fault and development of fault zone architecture as a function of dynamic earthquake rupture processes. The earthquake history of the fault was developed at several sites, with three sites yielding long records: a 4,000 year record of surface ruptures at Hog Lake near Anza on the Clark strand; a 2,000 year record at Mystic Lake on the Claremont strand; and a 1500 year record at Clark Lake on the southernmost Clark strand. These results along with the paleo-earthquake data from several shorter records has been used to construct a moment-balanced earthquake history for the primary, longest elements of the San Jacinto fault which can be used to forecast the likelihood of the size and extent of future large earthquakes. The long San Jacinto records were also compared to long records from the San Andreas fault to test whether large earthquakes on one fault could jump to the other (cascade rupture model; UCERF3). It was found that some northern San Jacinto ruptures may match some San Andreas ruptures north of their intersection, but none of these could possibly involve the central San Jacinto fault, making very large SJF cascade earthquakes unlikely. The fault zone architecture of the Clark strand was studied in detail at Horse Canyon, as well as preliminary studies at Blackburn Canyon, Upper Horse Canyon, and Rockhouse Canyon. At Horse Canyon, an ~30 cm thick fault core is bounded by damage zones of variably width. Following development of several C# programs for evaulating elemental mobility patterns, volumetric strains, and changes in bulk mass, along with our detailed XRD studies, we discovered that mixed-layer illite/smectite had been converted to illite at the boundary between the NE damage zone and the fault core. In addition, following in parallel to this change, plagioclase was dissolved resulting in statistically significant losses of Ca and Na mass. The overall result of these processes produced a dark green to black tabular fault core rich in silica exhibiting the properties of ultracataclasite. Though the temperature at which illite/smectite is transformed into illite is not well constrained, most estimates would place temperatures at ~150 degrees C or higher. Thus, during dynamic rupture, the fault core likely was flushed with hot acidic fluids that ultimately were diffused outward into the adjacent damage zones. Significantly, geomorphic evidence suggests that such processes occur at a depth of ~500 m or less. Finally, under the microscrope, fragments of older fault core material is evident, an observation that supports the idea that the above processes occur cyclically. At Rockhouse Canyon, we recognize incipient pulverization (in situ shattering at the subgrain level) of weakly consolidated sandstone (alluvial fan deposits of the Pleistocene upper Bautista Fm) at very shallow, well-constrained depths of 120-130 m, whereas identical deposits at 30-50 m do not exhibit much evidence for pulverization. These observations indicate that fault zone processes that have generally been attributed to depths of several kilometers are occurring at significantly shallower depths of just hundred of meters.
