EARTHQUAKE GATES: LINKING RUPTURE LENGTH TO GEOLOGICALLY CONSTRAINED DYNAMICS OF FAULT COMPLEXITY, WITH EXAMPLES FROM THE ALTYN TAGH AND SAN ANDREAS FAULTS

Unusually large, rare, and unexpected earthquakes overwhelm mitigation measures and the societal capacity to respond, leading to a cascade of disastrous effects. In order to assess the potential for such rare events, an approach is required to calibrate how effectively geometrical complexities, such as stepovers and restraining bends along strike-slip faults, impede earthquake propagation. We define the term 'earthquake gate' to describe areas of fault complexity that halt earthquake ruptures conditionally as a result of proximal fault geometry, rupture direction, and prior earthquake history. We show how these systems may be modeled dynamically, and compare our results with geologic observations in order to assess how prior earthquake history and fault geometry control whether an earthquake gate is ‘open' or ‘closed' to a particular direction of rupture propagation. Along the Aksay restraining double bend of the Altyn Tagh fault, opposing slip rate gradients on two parallel fault strands result from distributed deformation that largely occurs as earthquake ruptures terminate within the bend. From this pattern of fault activity we calibrate a two-dimensional, multi-cycle numerical rupture simulation with visco-elastic interseismic deformation, from which we predict that less than 10% of events are able to breach this earthquake gate. The San Gorgonio Pass restraining double bend of the San Andreas fault is similar in size to the Aksay Bend, and exhibits a similar trade-off in activity between the Banning-San Bernardino strand versus the Mission Creek-Mill Creek strand. Multi-cycle numerical simulation of this system is more challenging at present due to the three-dimensional geometry of the bend. However, the overarching similarity of the San Gorgonio Pass bend to the Aksay bend suggests that its probability of thoroughgoing rupture may be low as well.