[SSS05-P02] Heterogeneity of direct aftershock productivity of the main shock rupture
キーワード：direct aftershock, finite-source ETAS model, background probability
The epidemic type aftershock sequence (ETAS) model is widely used to describe and analyze
the clustering behavior of seismicity. Instead of regarding large earthquakes as point sources, the
finite-source ETAS model treats them as ruptures that extend in space. Each earthquake rupture consists
of many patches, and each patch triggers its own aftershocks isotropically. We design an iterative algorithm
to invert the unobserved fault geometry based on the stochastic reconstruction method. This model is
applied to analyze the Japan Meteorological Agency (JMA) catalog during 1964–2014. We take six great
earthquakes with magnitudes >7.5 after 1980 as finite sources and reconstruct the aftershock productivity
patterns on each rupture surface. Comparing results from the point-source ETAS model, we find the
following: (1) the finite-source model improves the data fitting; (2) direct aftershock productivity is
heterogeneous on the rupture plane; (3) the triggering abilities of M5.4+ events are enhanced; (4) the
background rate is higher in the off-fault region and lower in the on-fault region for the Tohoku earthquake,
while high probabilities of direct aftershocks distribute all over the source region in the modified model;
(5) the triggering abilities of five main shocks become 2–6 times higher after taking the rupture geometries
into consideration; and (6) the trends of the cumulative background rate are similar in both models,
indicating the same levels of detection ability for seismicity anomalies. Moreover, correlations between
aftershock productivity and slip distributions imply that aftershocks within rupture faults are adjustments
to coseismic stress changes due to slip heterogeneity.
the clustering behavior of seismicity. Instead of regarding large earthquakes as point sources, the
finite-source ETAS model treats them as ruptures that extend in space. Each earthquake rupture consists
of many patches, and each patch triggers its own aftershocks isotropically. We design an iterative algorithm
to invert the unobserved fault geometry based on the stochastic reconstruction method. This model is
applied to analyze the Japan Meteorological Agency (JMA) catalog during 1964–2014. We take six great
earthquakes with magnitudes >7.5 after 1980 as finite sources and reconstruct the aftershock productivity
patterns on each rupture surface. Comparing results from the point-source ETAS model, we find the
following: (1) the finite-source model improves the data fitting; (2) direct aftershock productivity is
heterogeneous on the rupture plane; (3) the triggering abilities of M5.4+ events are enhanced; (4) the
background rate is higher in the off-fault region and lower in the on-fault region for the Tohoku earthquake,
while high probabilities of direct aftershocks distribute all over the source region in the modified model;
(5) the triggering abilities of five main shocks become 2–6 times higher after taking the rupture geometries
into consideration; and (6) the trends of the cumulative background rate are similar in both models,
indicating the same levels of detection ability for seismicity anomalies. Moreover, correlations between
aftershock productivity and slip distributions imply that aftershocks within rupture faults are adjustments
to coseismic stress changes due to slip heterogeneity.