A static arrest criterion classifies some fault-spanning ruptures as arrested

Wenjie Lyu 

A static arrest length and a fully dynamic rupture are compared, one rupture at a time, for 60 antiplane and 60 in-plane cases on one 120 km fault with linear slip weakening and one correlation length of the initial stress. The standard deviations were fixed before those ruptures were drawn. The comparison tests the rule, used to build earthquake populations, that a rupture stops where its static energy release rate falls below the fracture energy and that this rule decides which ruptures span the fault. In 22 of the 120 cases the two calculations disagreed on whether the rupture spanned the fault, and in all 22 the static criterion predicted arrest while the dynamic rupture spanned the fault. All 22 still spanned the fault when the spacing along strike was halved. Where both stopped, the dynamic length was longer in 41 of 42 antiplane cases and 38 of 40 in-plane cases, and the median ratio was 1.08 in each mode. The in-plane values depend on the grid across the fault: with half the nodes, the in-plane median ratio was 0.99 and seven of the twelve in-plane disagreements remained. Three-dimensional rupture and other slip-weakening distances were not tested.