Figure 9. Slip rate and the corresponding frequency content for Scenario 5 (a and b) at 8.3 km and 22.0 km depth stations. Comparison between Scenarios 1 and 5 with frequency content at 8.3 km (c) and at 22.0 km (d), as well as comparison between Scenarios 2 and 5 with frequency content at 8.3 km (e) and at 22.0 km (f) are shown.
4 Discussion
Our numerical simulations on rupture scenarios reveal that the updip transition layer L from velocity-strengthening behavior near the trench to velocity-weakening behavior downdip suppresses rupture propagation toward the trench. With an employment of a conditionally stable layer d , total slip and rupture velocity significantly decreases, resulting in a longer rupture duration as d increases. As the low-velocity layer leads to a more compliant material near the trench, total slip is significantly higher in the scenarios with heterogeneous velocity structure (Scenarios 2 and 5). Imposing depth-varying friction promotes trenchward decrease in slip rate, as well as depletion in high-frequency radiation at shallow depth (Scenarios 3 and 4).
We identify that the slip distribution may highly depend on depth-varying rigidity, where a more compliant material leads to a larger total slip. In Figure 10, we further quantify the total slip distribution by normalizing all the Scenarios 2 through 5 over Scenario 1 (Figure 10). With L , d , as well as a uniform velocity structure employed, total slip ratio is less than 0.7 and more concentrated near the nucleation patch (Figure 10a-10c). In contrast, Scenario 5 has the largest slip ratio of 3 near the trench (Figure 10d). We summarize that while the amount of total slip is controlled by depth-varying rigidity, whereas the pattern of concentration is controlled by friction.