Abstract
We re-examined the slip distribution on faults of the 2004 Sumatra–Andaman (M 9.1 according to USGS) earthquake by the inversion of tsunami data with phase-corrected Green’s functions applied to linear long waves. The correction accounts for the effects of compressibility of seawater, elasticity of solid earth, and gravitational potential variation associated with the motion of mass to reproduce the delayed arrivals and the reversed phase of the first tsunami waves. We used sea surface height (SSH) data from satellite altimetry (SA) measurements along five tracks, and the tsunami waveforms recorded at tide gauges (TGs) and ocean bottom pressure gauges (OBPGs) in and around the Indian Ocean. The inversion results for both data sets for different rupture velocities (Vr) show that the reproducibility of the spatiotemporal SSHs and tsunami waveforms is improved by the phase corrections, although the effects are not so significant within the Indian Ocean. The best slip distribution model from joint inversion of SA, TG and OBPG data with Vr of 1.3 km/s shows the largest slips of 16–25 m off Sumatra Island, large slips of 2–11 m off the Nicobar Islands, and moderate slips of 2–6 m in the Andaman Islands. The inversion results reproduce the far-field tsunami waveforms well at distant stations even more than 13,000–25,000 km from the epicenter. The total source length is about 1400 km and the seismic moment is Mw 9.2, longer and larger than that of our previous estimates based on TG records.
Highlights
The 2004 Sumatra–Andaman earthquake (3.295° N, 95.982° E, depth = 30.0 km, M = 9.1 at 00:58:53 UTC on 26 December 2004 according to the United States Geological Survey [USGS]) is the largest event to have occurred so far in the twenty-first century, generating a devastating tsunami and causing severe damage and loss of human life (Satake, 2014)
Inversion results using only the satellite altimetry (SA) data show that the consistent slip distributions were estimated (Fig. 4) regardless of the assumed Vr: there are slight differences in the estimated slip amounts that depend on the Vrs, the slip distribution patterns are persistent within the assumed Vr ranges
In the case of Vr = 1.3 km/s (Fig. 4b), the slip distribution inverted from the SA data shows the large slips of 17–22 m off Sumatra Island, the moderate slips of 6–8 m off the Nicobar Islands, and small slips less than 4 m in the Andaman Islands, and indicates the total source length of about 1400 km
Summary
The 2004 earthquake was the first tsunami captured by satellite altimetry (SA) measurements (Gower, 2005; Smith et al, 2005). Using these instrumentally recorded tsunami data, previous studies estimated the slip distributions or rupture propagation on the faults of the earthquake (Arcas & Titov, 2006; Fujii & Satake, 2007; Gopinathan et al, 2017; Hirata et al, 2006; Lorito et al, 2010). While the inversion of only SA data indicated that the tsunami source extended to the Andaman Islands with a total length of 1400 km, such a long fault model with Vr of 1.0 km/s produced much larger tsunami waveforms than observed at Indian tide gauge stations
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
