Interplate Slip Following the 2003 Tokachi‐oki Earthquake From Ocean Bottom Pressure Gauge and Land GNSS Data

Abstract

Preseismic, coseismic, and 7.5 years of postseismic deformation of the 2003 M w 8.0 Tokachi-oki earthquake are modeled using land Global Navigation Satellite System (GNSS) data and two ocean bottom pressure gauges (OBP) using viscoelastic Green's functions. The postseismic slip distribution is shown to not overlap with the main shock or the source regions of past large earthquakes along the southern part of the Kurile trench. The preseismic locking is estimated in the coseismic and postseismic slip regions using the same plate interface geometry. The temporal evolution of the postseismic slip is described with a single logarithmic function. The slip is shown to decay faster updip of the coseismic slip region compared to the downdip extension. Sustained postseismic slip indicates that the interplate locking in the postseismic slip region had not fully recovered to the pre-2003 status in the 7.5 year analyzed period. Postseismic slip history also indicates rapid recovery of locking in the coseismic slip region. We examined the constraint of the OBP time series for the postseismic deformation modeling and found that the observed uplift at the OBP sites requires small postseismic slip near the trench, where a tsunamigenic M9 earthquake occurred in the seventeenth century. Uplift at the OBP sites does not constrain the absolute slip magnitude between the updip and downdip sides of the OBP sites but requires the slip in the updip side to be smaller than that in the downdip side. Our model describes transient postseismic deformation observed by both land GNSS and OBP.

Adjoint-based direct data assimilation of GNSS time series for optimizing frictional parameters and predicting postseismic deformation following the 2003 Tokachi-oki earthquake

“…The main slip location soon moved to the shallower side of the TA and continued to slip, exceeding 2 cm/year in the next 6 days, followed by a gradual decay until the end of the assimilation period. The time constant for decaying afterslip on the shallower side of the TA is longer than that on the northeastern side of the TA, which is consistent with the results of Itoh et al (2019) although they analyzed a much longer time series of 7.5 year of postseismic deformation. Figure 11 and Additional file 1: Figure S2 present the comparison of the theoretical displacements calculated by using the optimum parameters (red and blue lines) and those by using the first-guess parameters (pink and light blue lines) with the observed GNSS time series (circles).…”

“…Because we analyzed 30 days of data following the mainshock, these signals are mainly attributed to the afterslip on the plate interface rather than viscoelastic effect. This assumption is reasonable in view of Itoh et al (2019), which showed that the calculated surface displacement due to the viscoelastic effect of the 2003 Tokachi-oki earthquake at an inland station located at northern Hokkaido was, at most, a few mm in the first month following the earthquake, even in the case of a low viscosity (Additional file 1: Figure S13 in Itoh et al 2019).…”

“…Afterslip and the viscoelastic relaxation effect of coseismic slip reflect the recovery process following large earthquakes (Diao et al 2014;Sun and Wang 2015;Klein et al 2016;Itoh et al 2019). Itoh et al (2019) modeled the spatio-temporal evolution of afterslip for 7.5 years by considering the viscoelastic effect following the 2003 Tokachi-oki earthquake. Based on the spatial distribution of afterslip, they concluded that the interplate locking of the afterslip area had not fully recovered to the pre-seismic state after 7.5 years.…”

Compliant Volcanic Arc and Backarc Crust in Southern Kurile Suggested by Interseismic Geodetic Deformation

“…However, the two options give rise to very different coseismic deformation because the coseismic response of the underlying asthenosphere is elastic. To show this, we calculate coseismic displacements of the 2003 event using the parameters of the models in Figures 2c, 2g, and 1c and compare them with coseismic GNSS observations (Itoh et al, 2019; Figure 3). An arc rigidity as low as 10% (Figure 2c) leads to very unreasonable coseismic deformation (Figure 3b).…”

“…The subduction megathrust produces many large to great earthquakes, with the most recent one being the 2003 M w 8.0 Tokachi-oki earthquake (Figure 1a). The interpretation of the Global Navigation Satellite Systems (GNSS) data provided by Japan's GNSS Earth Observation Network System (GEONET) before 2003 suggests a high degree of interseismic locking of the megathrust, largely compatible with the rupture areas of previous earthquakes (Hashimoto et al, 2009(Hashimoto et al, , 2012Ito et al, 2000;Itoh et al, 2019;Loveless & Meade, 2010;Suwa et al, 2006). However, interseismic strain rates derived from GEONET velocities tend to show a higher rate of contraction along the volcanic arc in Hokkaido (e.g., Sagiya et al, 2000).…”