Deformation rates in northern Cascadia consistent with slow updip propagation of deep interseismic creep

Abstract

Interpretations of interseismic slip deficit on the northern Cascadiamegathrust are complicated by an enigmatic 'gap' between the downdip limit of the locked region, inferred from kinematic inversions of deformation rates, and the top of the episodic tremor and slip (ETS) zone. Recent inversions of global positioning system (GPS) and tide gauge/leveling data for shear stress rates acting on the megathrust found a ~21 km locking depth with a steep slip-rate gradient at its base is required to fit the data. Previous studies have assumed the depth distribution of interseismic slip rate to be time invariant; however, steep slip-rate gradients could also result from the updip propagation of slip into the locked region. This study explores models where interseismic slip penetrates up into the locked zone. We consider the creeping region, corresponding to the gap and the ETS zone, as a quasi-static crack driven by the plate velocity at its downdip end. We derive a simple model that allows for crack propagation over time, and provides analytical expressions for stress drop within the crack, slip and slip rate on the fault. It is convenient to expand the non-singular slip-rate distribution in a sum of Chebyshev polynomials. Estimation of the polynomial coefficients is underdetermined, yet provides a useful way of testing particular solutions and provides bounds on the updip propagation rate. When applied to the deformation rates in northern Cascadia, best-fitting models reveal that a very slow updip propagation, between 30 and 120 myr-1 along the fault, could explain the steep slip-rate profile, needed to fit the data. This work provides a new tool for estimating interseismic slip rates, between purely kinematic inversions and full physics-based modeling, allowing for the possibility for updip expansion of the creeping zone.