Preliminary Geodetic Slip Model of the 2011 M9.0 Tohoku-chiho Taiheiyo-oki Earthquake

Data Set

Fig1_invertCO

Figure 1

Data from the Geospatial Information Authority (GSI) of Japan was processed by the Jet Propulsion Laboratory (JPL) and Caltech (ftp://sideshow.jpl.nasa.gov/pub/usrs/ARIA) to yield the three-dimensional coseismic displacement field over the Japanese Islands (Figures 1 and 2). The coast of northeast Japan moved eastward up to 4 meters, and the coastline generally subsided by about 0.5 meters (up to a maximum of 1.1 meters in eastern Miyagi Prefecture), due to the position of the coast approximately above the lower edge of the rupture area along the interplate boundary.

Fig2_invertCO-vert

Figure 2

Methods

The fault geometry is that of a single 700-km-long planar surface striking 195° and dipping 14° (Figure 3). The upper and lower edge depths are fixed at 3 km and 57 km, respectively. This geometry is based on the Global CMT solution and is similar to that adopted in recent seismic slip inversions. The location of the fault plane is guided by the slab contours of Huang et al. [2011]. The dip is thought to become progressively deeper from the trench towards land, where it may be 23° [Miura et al., 2005]

Fig3_backg

Figure 3

Distributed slip is represented with a distribution of continuous functions as employed by Pollitz et al. [1998]. These are Hermite-Gauss (HG) functions of position on the rectangular fault plane. Slip on the slab interface is related to static surface displacement using the source response functions calculated with the method of Pollitz [1996]. This yields theoretical displacements in a layered spherical geometry with a spherical harmonic expansion, and global Earth model PREM with isotropic elastic parameters, appended by crustal structure appropriate for the offshore region of northeastern Japan [Zhao et al., 2007], is used for this purpose.

Horizontal and vertical GPS data are inverted for distributed slip using weighted least squares. A small amount of damping of the squared gradient of the slip is used to regularize the inversion and determine the weighting coefficients of the HG functions [Pollitz et al., 1998].

Results

Fig4_slip-results

Figure 4

The resulting slip distribution is shown in Figure 4, corresponding to a ‘geodetic’ moment of 3.59×1022 N m. Slip is concentrated near the hypocenter and its updip extension. Corresponding fits to the horizontal and vertical GPS data are shown in Figure 1 and 2.

An example of the resolution of slip is shown in Figure 5. The vertical GPS data is important for obtaining resolution of shallow slip on the fault plane. Inversions utilizing only horizontal GPS tend to localize the slip deeper downdip and cannot adequately discriminate relatively deep slip from shallow slip. Moreover, the amount of slip required to fit the horizontal displacement field is substantially reduced when the locus of slip is deeper (equivalently, closer to the land observations). The moment is found to be reduced by 25% when only horizontal GPS data are considered.

Fig5_slip-synth

Figure 5

The broad slip maximum offshore Sendai is adjacent to the areas of peak ground acceleration onland (http://outreach.eri-u.tokyo.ac.jp/eqvolc/201103tohoku/eng/). The high-slip region coincides with the location of high seismic coupling inferred from interseismic crustal velocity measurements [e.g. Suwa et al., 2006; Hashimoto et al., 2009; Loveless and Meade, 2010].

Numerous small repeating earthquakes have occurred along the Japan Trench along the entire area between ∼ 35.5°N and 38°N, most intensely south of about 37°N [Uchida et al., 2009]. The rapid diminishment of coseismic slip south of this latitude (Figure 4) suggest that either this region had lower levels of absolute stress, thereby inhibiting rupture propagation, or that its frictional properties are different. The correlation of reduced slip with the areas of repeating earthquake would support the latter interpretation, suggesting that a large portion of the interplate boundary south of 37°N tends to exhibit velocity-strengthening behavior. In any case, postseismic GPS observations (available at ftp://sideshow.jpl.nasa.gov/pub/usrs/ARIA) suggest that the area that exhibits repeating earthquakes has also experienced afterslip of order 1 meter or greater in the days following the mainshock.

References

Separate image files (JPGs):

Preliminary GPS data provided by the ARIA team at JPL and Caltech. All original GEONET RINEX data provided to Caltech by the Geospatial Information Authority (GSI) of Japan.