Asperity erosion behavior in the Northern Japan Trench in the decade prior to the M9 Tohoku Earthquake
Abstract
The Northern Japan Trench has historically been thought to exhibit velocity-weakening areas of contact (asperities) and velocity-strengthening freely subducting areas. Asperities in this region have traditionally been thought to be constant in size. Then the M9 Tohoku rupture caused widespread coseismic slip throughout velocity-weakening and velocity strengthening regions, indicating that asperities may not be constant in area with time. GEONET GPS data further supports the idea that locking is changing; in the time period 1998 - 2009, there are significant accelerations in surface motions toward the trench. This acceleration is consistent with the concept of asperities shrinking during this timeframe (effectively reducing the coupling). We leverage the Cattania and Segall (2019) creep front propagation derivation in an asperity erosion model in which creep at constant stress accelerates around shrinking asperities inside which the plate interface is fully locked. Magnitude 6.5 and larger historical earthquakes and their inferred locations are mapped and sized based on time since rupture and magnitude. The inferred size of the M9 asperity was based on coseismic slip estimates, and we adjusted the M9 asperity area for the 1998-2009 time period by assuming shrinking began in 1600 AD when the previous M9 in the region occurred (i.e. the M9 asperity's hypothesized previous rupture). A MCMC inversion method is used to find the range of shrinking rates, constrainted by the Cattania and Segall physical model, that fit the GPS acceleration data. We find that the data can be explained only with an additional contribution from interseismic earthquake cycle mantle flow. We also explore a MCMC inversion without the physical constaint; we let the asperities choose their best shrinking path, and conduct a MCMC inversion with no earthquake cycle mantle flow contribution. We find that the Cattania and Segall (2019) model is able to create an estimated shrinking behavior capable of fitting GPS data nearly as well as the model that let asperities take whatever locking shape they need to fit the GPS data. This indicates that the Cattania and Segall (2019) physical model may correctly characterize locking behavior in the decade prior to the M9 Tohoku rupture.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2022
- Bibcode:
- 2022AGUFM.T32E0208B