The 2004 and 2005 Sumatra Earthquakes: Implications for the Lisbon earthquake

Abstract

The Sumatra mega-earthquake of 26 December 2004 (Mw=9.3) was the strongest earthquake in the world since the 1964 Alaska earthquake and the fifth strongest since 1900. The earthquake occurred at the interface of the India and Burma plates and triggered a massive tsunami that affected several countries throughout South and Southeast Asia. Three months later, on 28 March 2005, about 200 km south of this event but at a greater depth (28 km), occurred a magnitude 8.7 earthquake. This event was probably triggered by stress variations caused by the December mega-earthquake. In this work we describe the rupture process of both earthquakes, estimated from teleseismic broad-band waveform data provided by IRIS-DMC stations. The rupture direction and velocity were determined from common pulse durations observed in P waveforms using DIRDOP computational code (DIRectivity DOPpler effect). The modified Kikuchi and Kanamori method has been used to determine the slip distribution. For the mega- earthquake two segments of 150 km width (along dip) and 990 km total length with different azimuth were estimated, based on the subduction geometry, aftershock distribution and CMT. Results show that the rupture spread mainly to the North with an average velocity of 2.7 km/s. The focal mechanism shows thrust motion on a plane oriented in a NNW-SSE direction and a horizontal pressure axis in the NNE-SSW direction. The fault slip distribution shows the following pattern: 1) the rupture nucleated at the hypocenter as a circular crack breaking a shallow asperity of about 60 km radius during the first 60 sec; 2) after the initial break to the NNW, the rupture propagated during ~180 s and broke a middle large asperity centred at about 360 km from the epicentre; 3) finally, the rupture propagated further to the north and broke a third asperity centred at ~840 km from the epicentre during at least 110 sec. The maximum slip reaches 14 m in the central asperity and the total seismic moment is Mo=3.0x1022 Nm (Mw=8.9), which is less than the value given by the ESMC and USGS (the loss of seismic scalar moment was released in a third segment located to the north). The total source duration and rupture length are estimated to be above 350 sec and 990 km, respectively. For the earthquake of 28 March 2005, a rectangular rupture plane with 400 km length (along the strike direction) and 125 km width (along the dip direction) was obtained from the subduction geometry, aftershock distribution and CMT. Results show that the rupture spread during about 110s in the southwest direction with an average velocity of ~3.3 km/s. Most of the seismic moment was released at the break of two asperities: the largest one located at about 90km from the hypocenter, and the other one at 175 km from the hypocenter. These two asperities correspond on the surface to the areas most affected by the event (Nias Island). The maximum slip reaches 11.5 m in the largest asperity and the total seismic moment is Mo=0.82x1022 Nm (Mw=8.6). The focal mechanism shows thrust motion similar to this shown by the mega-earthquake. Probably, the 1755 Lisbon earthquake (Mw∼9.0) released as much or more energy as any seismic event of recorded history prior to 2004 December. Nevertheless, the location of the source, responsible for the Lisbon tsunami, is not well known; the epicentres suggested by various authors are separated by some hundreds of km. We compare the similarities and differences of these two mega- earthquakes (Sumatra and Lisbon) with the purpose of reducing the uncertainties relative to the location of the seismogenic zone responsible for the 1755 Lisbon earthquake. Lessons learned from the Sumatra earthquake, through scientific studies, should help to reduce the number of victims and damage during future earthquakes in Portugal.