I recently started my second postdoctoral project under the guidance of Dr. Fred Pollitz and Dr. David Schmidt. Utilizing a geophysical modeling tool called VISCO2.5D (Pollitz, 2014), we have begun modeling early-stage postseismic deformation in areas surrounding the Noto Peninsula, Japan. We consider both viscoelastic (VE) relaxation and afterslip. As I make progress, I will upload some results here! Currently we only use GNSS data, but we plan to incorporate InSAR data into our analysis to enhance our models!

The focal mechanism solutions (Ekström et al., 2012) for the 2024 Mw7.5 earthquake and the 2023 Mw 6.2 earthquake are shown. The colored dots represent the aftershock hypocenters (from JMA) following the 2024 earthquake. The white triangles denote the locations of GEONET GNSS stations (from GSI). The dashed red box outlines the assumed fault plane with a strike of 051 and a dip of 35 (based on the USGS finite-fault model).

The GNSS time series data (black dots) and fitted functions (red) for stations (a) J252 and (b) I070. The locations of these stations are shown in the map above. To emphasize the postseismic signals, I removed the preseismic secular rate trends from both the data and the model fit.

(1) I determined postseismic velocities over a two-month period by fitting each component of GNSS daily timeseries (Blewitt et al., 2018) with a pre-seismic secular rate, amplitudes of annual and semiannual variations, steps, and a quadratic function (Pollitz, 2015) starting from January 2nd,2024.

(2) Varying a transient viscosity of a vertically layered viscoelastic model (30 km thick elastic layer, underlain by a viscous layer; A Burgers' body), we obtained a set of VE relaxation models. For forward models, we employed the VISCO2.5D (Pollitz, 2014) and the USGS finite-fault model as an input source.

(3) For each trial VE model, we inverted the residual velocities between modeled VE relaxation responses and observed postseismic GNSS velocities to estimate the afterslip distribution.

Data and Method

Horizontal postseismic signals from GNSS data (green) compared with model predictions (red). My preferred model integrates afterslip with VE relaxation. The color bar indicates the amount of afterslip on fault.

Preliminary Findings

(1) A joint VE relaxation and afterslip model is required to effectively describe the observed postseismic velocities.

(2) I derived a preferred model with a transient viscosity of 4.1×10E+17 [Pa s] and with weighting the horizontal GNSS data slightly more than the vertical data.

(3) The moment associated with afterslip amounts to ~ 9.2 % (~Mw 6.8) of the total coseismic moment release, and its spatial distribution fills the gap left by the coseismic rupture zone.

Acknowledgements

I used GEONET GNSS data (http://terras.gsi.go.jp/) processed by Nevada Geodetic Data (http://geodesy.unr.edu/). Focal mechanism solutions and hypocenter estimates are from Global CMT website (https://www.globalcmt.org/) and Japan Meteorological Agency (https://www.data.jma.go.jp/eqev/data/daily_map/index.html), respectively. The USGS finite-fault model is available at: https://earthquake.usgs.gov/earthquakes/eventpage/us6000m0xl/finite- fault. I plotted all geospatial images using the Generic Mapping Tools 6 (https://www.generic-mapping-tools.org/).