Interseismic coupling maps and, especially, estimates of the location of the fully coupled (locked) zone relative to the trench, coastline, and slow slip events are crucial for determining megathrust earthquake hazard at subduction zones. We present a physically motivated interseismic coupling inversion that explicitly incorporates locked zones with boundaries bordering an updip transition zone creeping at constant stress and a downdip transition zone with creep rate distributions consistent with updip propagation of the creep front into the locked zone. We show that the locked zone at Cascadia is west of the coastline and 10 km updip of the slow slip zone along much of the margin, widest (25-125 km, extending to ~22 km depth) in northern Cascadia, narrowest (0-70 km) in central Cascadia, with moment accumulation rate equivalent to a Mw 8.78 and Mw 8.89 earthquake for 300- and 500-year earthquake cycles. We find a steep gradient in creep immediately below the locked zone, indicative of propagating creep, along the entire margin. At Nankai, we find three distinct zones of locking (offshore Shikoku, offshore southeast Kii peninsula, and offshore Shima peninsula) with a total moment accumulation rate equivalent to a Mw 8.73 earthquake for a 150-year earthquake cycle. The bottom of the locked zone is nearly under the coastline for all three locked regions at Nankai and is positioned 0-5 km updip of the the slow slip zone. In contrast with Cascadia, creep rate gradients below the locked zone at Nankai are generally gradual, consistent with stationary locking.

Observations of fold growth in fold-thrust belt settings show that brittle deformation can be localized or distributed. Localized shear is associated with frictional slip on primary faults, while distributed brittle deformation is recognized in the folding of the bulk medium. The interplay of these processes is clearly seen in fault-bend folds, which are folds cored by a fault with an abrupt change in dip (e.g., a ramp-décollement system). While the kinematics of fault-bend folding were described decades ago, the dynamics of these structures remain poorly understood, especially the evolution of fault slip and off-fault deformation over different periods of the earthquake cycle. In order to investigate the dynamics of fault-bend folding, we develop a numerical modeling framework that combines a long-term elasto-plastic model of folding in a layered medium with a rate-state frictional model of fault strength evolution in order to simulate geologically and mechanically consistent earthquake sequences. In our simulations, slip on the ramp-décollement fault and inelastic fold deformation are mechanically coupled processes that build geologic structure. As a result, we observe that folding of the crust does not occur steadily in time but is modulated by earthquake cycle stresses. We suggest combining seismological and geodetic observations with geological fault models to uncover how elastic and inelastic crustal deformation generate fault-bend folds. We find that distinguishing between the elastic and inelastic response of the crust to fault slip is possible only in the postseismic period following large earthquakes, indicating that for most fault systems this information currently remains inaccessible.

It has been known for decades that the present-day shortening rates across the Western Transverse Ranges (WTR) in southern California are rapid, reaching 10-15 mm/yr near the heavily populated Los Angeles area. However, only recently have geodetic measurements of vertical motion in the WTR been sufficiently dense to resolve a tectonic vertical signal. In this study, we show that much of the geodetically-derived vertical velocity field in the WTR can be attributed to the interseismic signal of strain accumulation on reverse faults. We invert geodetic and geologic data for slip rate and interseismic coupling on faults using a kinematic model consisting of faults embedded in an elastic crust over an inviscid mantle. This method allows us to infer the permanent, long-term component of vertical motions from recoverable, short term motions. We infer that much of the geodetically observed 3-4 mm/yr of differential vertical motion across the WTR, involving subsidence along the Santa Barbara coastline and uplift of the Santa Ynez Range, can be attributed to recoverable elastic deformation associated with interseismic locking on faults dipping under the WTR. The sum of dip-slip rates across the WTR decreases from 10.5-14.6 mm/yr on the east side near Ventura, California to 5-6.2 mm/yr across the western side of the Santa Barbara Channel. The total moment accumulation rate in both the Santa Barbara Channel and the combined San Fernando Valley-LA Basin regions is equivalent to about two M=7 earthquakes every 100 years.