Supplement Material Readme file: Manuscript number: 2013GL055606 Title: Numerical simulations on megathrust rupture stabilized under strong dilatancy strengthening in slow slip region Author: Yajing Liu Affiliation: Department of Earth and Planetary Sciences, McGill University Figure Captions: Figure S1. Dependence of maximum coseismic slip and downdip rupture limit (top and bottom panel respectively in each subfigure) on the drainage parameter in the updip seismogenic zone ($U_{SZ}$) and SSE region ($U_{SSE}$). (a) $U_{SZ}=0.01$, (b) 0.1, (c) 1.0, (d) 10.0. $U_{SSE}$ varies from 0.01 to 10 for each fixed $U_{SZ}$. For each simulation case, thin black bar shows the variation between 10 to 90 percentile, thick black bar shows 25 to 75 percentile, red circle shows the median value. Horizontal dashed lines show the levels of the two source properties without dilatancy. Figure S2. (a) Slip history and (b) slip budget, with low dilatancy $\epsilon/\beta=0.1$ MPa at all depths. The five earthquakes, EQs 1-5, have maximum coseismic slips of 14, 10.8, 10.3, 4.8 and 23.9 meters and intervals from the respective preceding earthquake are 323, 281, 279, 262 and 454 years. The slip budget is calculated for a cycle that consists of interseismic slip prior to EQ1, coseismic and postseismic slip of EQ1 in (a). Figure S3. Extremely small dilatancy effect case: $(\epsilon/\beta)_{SZ}=0.025$ MPa in the updip seismogenic zone, $(\epsilon/\beta)_{SSE}=0.05$ MPa in the slow slip event region between downdip 140 and 215 km. Drainage parameter is $U=1.0$ uniformly along the entire fault. (a) Slip history on the fault. Black lines are interseismic slip every 50 years. Red lines are coseismic slip every 20 seconds. A median coseismic slip of 22 meters is calculated for earthquakes with slip larger than 4 m (here, 24.8, 15.5, 23.9, 22.0, 18.7 meters respectively), and the median of their rupture downdip limits is 238.65 km. This is further downdip than the no-dilatancy case, due to larger coseismic slip in general. (b) Maximum slip rate on the fault in $\log_{10}(V_{max}/V_{pl})$. Vertical dashed line represents the cutoff threshold for ``coseismic" slip. (c) Maximum slip rate for a 50-yr SSE periods.} Figure S4. Extremely high dilatancy effect case: $(\epsilon/\beta)_{SZ}=0.175$ MPa in the updip seismogenic zone, $(\epsilon/\beta)_{SSE}=0.5$ MPa in the SSE region. $U=1.0$. (a) Slip history on the fault. Black lines are interseismic slip every 50 years. Red lines are coseismic slip every 20 seconds. A median coseismic slip of 7.7 meters is calculated for modeled earthquakes with slip larger than 4 meters (here, 7.5, 7.5, 9.1, 6.2, 8.6, 8.0, 8.5, 6.2 meters respectively). The median of their rupture downdip limits is 142.4 km. (b) Maximum slip rate on the fault in $\log_{10}(V_{max}/V_{pl})$. Vertical dashed line represents the cutoff threshold for ``coseismic" slip. (c) Maximum slip rate during a 3-yr period containing the final rupture sequence EQ 1 to EQ 3. Figure S5. An example of spatiotemporal relative distribution of short-term and long-term SSEs under heterogeneous dilatancy conditions.