Subsections

Using Geodetic Data to Understand Postseismic Processes following the Sumatra-Andaman Earthquake

Kelly Grijalva, Roland Bürgmann, and Paramesh Banerjee (Earth Observatory of Singapore)

Introduction

Investigations of postseismic deformation are often plagued by ambiguities between fundamental deformation mechanisms that can be expected to contribute to the deformation, including viscous flow, localized afterslip and poroelastic rebound. Previous studies have explained the near-field postseismic deformation following the Sumatra-Andaman earthquake with afterslip or poroelastic rebound, and the far-field postseismic deformation with viscoelastic relaxation from both the 2004 $M_w$ 9.2 and 2005 $M_w$ 8.7 events. We aim to fit both the near and far-field data, spanning years 2005-2007, with a combination of postseismic mechanisms. We utilize campaign GPS data from Gahalaut et al. (2008) and continuous GPS data from the regional networks: NICT, SuGAr, and THAI.

Poroelastic Rebound

The strain field resulting from a coseismic dislocation produces changes in pore fluid pressure in the brittle upper crust. The subsequent decay of the excess pore-fluid-pressure gradients will lead to fluid flow and poroelastic deformation. We approximate the fully relaxed poroelastic response by subtracting the undrained solution for coseismic deformation from the drained solution for coseismic deformation. Based on earlier studies (e.g. Masterlark, 2003), the crust is assumed to be fluid-saturated down to  15 km depth. However, Ogawa and Heki (2007) propose that the downgoing slab releases fluids into the mantle wedge in sufficiently high quantities, with sufficiently large pore pressure diffusivities, to contribute to the poroelastic rebound during the early postseismic period. We therefore test a range of earth models, with undrained Poisson's ratio values 0.05 above the drained value for the top 15 km to 60 km of the lithosphere (Figure 2.5a).

Figure 2.5: a) Modeled surface deformation due to a) poroelastic rebound and b) afterslip on the Sunda megathrust. The modeled displacements are compared with geodetic measurements over the year 2005.
\begin{figure}\begin{center}
\epsfig{file=grijalva10_1_1.eps, width=8cm}\end{center}\end{figure}

Figure 2.6: Comparison of a) 2005, b) 2006, and c) 2007 geodetic data with the preferred viscoelastic relaxation model and a combination model. The coseismic rupture planes for the Sumatra-Andaman earthquake are outlined in black (Banerjee et al., 2007).
\begin{figure*}\begin{center}
\epsfig{file=grijalva10_1_2.eps, width=14cm}\end{center}\end{figure*}

Afterslip

We test afterslip models that range in depth from 10 km to 75 km, spanning the coseismic rupture planes and the downdip portion of the megathrust (Figure 2.5b). In general, the horizontal deformation is oriented towards the trench and magnitude of the deformation in the far-field increases with increased afterslip plane depth. When the afterslip plane is placed trenchward of the GPS site, there is vertical subsidence. Conversely, when the GPS site is located trenchward of the afterslip plane, there is vertical uplift. Afterslip can be optimized to fit the near-field vertical and horizontal data with 1-2 m of slip. However, in order to fit the far-field sites in Thailand, it is necessary to have approximately 10 m of afterslip at 60-75 km depth.

Viscoelastic Relaxation

Deep-seated transient postseismic relaxation can produce time dependent deformation exceeding that from the earthquake itself in the intermediate-to-far field range. We test a range of earth models with the lithosphere-asthenosphere boundary (LAB) increasing from 50 km to 80 km in depth. Increasing the LAB decreases the magnitude of the surface deformation. The viscoelastic models can be optimized to fit the far-field data, but they poorly fit the near-field data. Our preferred LAB is 60 km because it is similar to previously published results (Pollitz et al., 2006) and the subsidence in the Nicobar Islands helps to offset the modeled uplift from poroelastic rebound and afterslip. We find that our viscoelastic model fits the rate of decay at the Thai sites best with a transient asthenospheric viscosity of $2.5 x 10^{17}$ Pa s and a steady-state viscosity of $5 x 10^{18}$ Pa s.

Discussion

We discovered that the vertical component is more important than the horizontal component for diagnosing the near-field postseismic processes. For the Sumatra-Andaman earthquake, poroelastic rebound always has uplift in the near-field, afterslip can have either uplift or subsidence depending on the slip location, and viscoelastic relaxation always has both uplift and subsidence. Our preferred postseismic model for 2005 includes a combination of the 15 km-poroelastic rebound model, the 60 km-viscoelastic relaxation model, and between 0.5-1.5 m of afterslip (Figure 2.6a). For 2006 and 2007, we did not include poroelastic rebound in our preferred combination model. The 2006 combination model includes between 0.7-1.0 m of afterslip (Figure 2.6b) and the sparser 2007 data suggest that the afterslip location is continuing further downdip along the megathrust with cumulative slip magnitude of between 0.5-2.0 m (Figure 2.6c). The afterslip magnitude does not appear to have decayed significantly during our observation period from 2005-2007. Overall, we find that the near-field and far-field data can not be fit by just one postseismic process and the vertical component is especially necessary for properly diagnosing the near-field postseismic deformation.

Acknowledgments

This work is supported by the National Science Foundation grant EAR 0738299. The GPS data is provided by the Tectonics Observatory, LIPI and SOPAC. The modeling programs are provided by Fred Pollitz.

References

Banerjee et al., Coseismic slip distributions of the 26 December 2004 Sumatra-Andaman and 28 March 2005 Nias earthquakes from GPS static offsets, Bull. Seism. Soc. Am., 97, S86-S102, 2007.

Gahalaut et al., GPS measurements of postseismic deformation in the Andaman-Nicobar region following the giant 2004 Sumatra-Andaman earthquake, J. Geophys. Res., 113, 10.1029/2007JB005511, 2008.

Masterlark, T., Finite element model predictions of static deformation from dislocation sources in a subduction zone: Sensitivities to homogenous isotropic, Poisson-solid, and half-space assumptions, J. Geophys. Res., 108, doi:10.1029/2002JB002296, 2003.

Ogawa, R. and K. Heli, Slow postseismic recovery of geoid depression formed by the 2004 Sumatra-Andaman Earthquake by mantle water diffusion, Geophys. Res. Lett., 34, doi:10.1029/2007/GL029340, 2007.

Pollitz et al., Post-seismic relaxation following the great 2004 Sumatra-Andaman earthquake on a compressible self-gravitating Earth, Geophys. J. Int., 167, 397-420, 2006.

Berkeley Seismological Laboratory
215 McCone Hall, UC Berkeley, Berkeley, CA 94720-4760
Questions or comments? Send e-mail: www@seismo.berkeley.edu
© 2007, The Regents of the University of California