Earthquake characteristics
The earthquake was initially reported as moment magnitude, Mw 9.0 (note that this is not the Richter scale or local magnitude scale, Ml, which is known to saturate at higher magnitudes.) In February 2005 scientists revised the estimate of the magnitude to Mw9.3.[6] Although the Pacific Tsunami Warning Center has accepted these new numbers, the United States Geological Survey has so far not changed its estimate of 9.1. The most recent studies in 2006 have obtained a magnitude of Mw 9.1 to 9.3. Dr. Hiroo Kanamori of the California Institute of Technology believes that Mw = 9.2 is a good representative value for the size of this great earthquake.[7]
The hypocentre of the main earthquake was at 3.316°N, 95.854°E ( ), approximately 160 km (100 mi) west of Sumatra, at a depth of 30 km (19 mi) below mean sea level (initially reported as 10 km). The earthquake itself (apart from the tsunami) was felt as far away as Bangladesh, India, Malaysia, Myanmar, Thailand, Singapore and the Maldives.
Indonesia lies between the Pacific Ring of Fire along the north-eastern islands adjacent to and including New Guinea and the Alpide belt along the south and west from Sumatra, Java, Bali, Flores, and Timor. The December 2004 earthquake actually occurred within the Alpide belt.[citation needed]
Great earthquakes such as the Sumatra-Andaman event, which are invariably associated with megathrust events in subduction zones, have seismic moments that can account for a significant fraction of the global earthquake moment across century-scale time periods. The Sumatra-Andaman earthquake was the largest earthquake since 1964, and the second largest since the Kamchatka earthquake of October 16, 1737.
Of all the seismic moment released by earthquakes in the 100 years from 1906 through 2005, roughly one-eighth was due to the Sumatra-Andaman event. This quake, together with the Good Friday Earthquake (Alaska, 1964) and the Great Chilean Earthquake (1960), account for almost half of the total moment. The much smaller but still catastrophic 1906 San Francisco earthquake is included in the diagram at right for perspective. Mw denotes the magnitude of an earthquake on the moment magnitude scale.
Since 1900 the only earthquakes recorded with a greater magnitude were the 1960 Great Chilean Earthquake (magnitude 9.5) and the 1964 Good Friday Earthquake in Prince William Sound (9.2). The only other recorded earthquake of magnitude 9.0 or greater was off Kamchatka, Russia, on November 4, 1952 (magnitude 9.0).[8] Each of these megathrust earthquakes also spawned tsunamis in the Pacific Ocean, but the death toll from these was significantly lower. The worst of these caused only a few thousand deaths, primarily because of the lower population density along the coasts near affected areas and the much greater distances to more populated coasts.
Other very large megathrust earthquakes occurred in 1868 (Peru, Nazca Plate and South American Plate); 1827 (Colombia, Nazca Plate and South American Plate); 1812 (Venezuela, Caribbean Plate and South American Plate) and 1700 (Cascadia Earthquake, western U.S. and Canada, Juan de Fuca Plate and North American Plate). These are all believed to have been of greater than magnitude 9, but no accurate measurements were available at the time.
Tectonic plates
The earthquake was unusually large in geographical extent. An estimated 1,600 km (994 mi) of faultline slipped about 15 m (50 ft) along the subduction zone where the India Plate slides under the Burma Plate. The slip did not happen instantaneously but took place in two phases over a period of several minutes. Seismographic and acoustic data indicate that the first phase involved a rupture about 400 km (250 mi) long and 100 km (60 mi) wide, located 30 km (19 mi) beneath the sea bed—the longest rupture ever known to have been caused by an earthquake. The rupture proceeded at a speed of about 2.8 km/s (1.7 mi/s) or 10,000 km/h (6,300 mph), beginning off the coast of Aceh and proceeding north-westerly over a period of about 100 seconds. A pause of about another 100 seconds took place before the rupture continued northwards towards the Andaman and Nicobar Islands. However, the northern rupture occurred more slowly than in the south, at about 2.1 km/s (1.3 mi/s) or 7,600 km/h (4,700 mph), continuing north for another five minutes to a plate boundary where the fault changes from subduction to strike-slip (the two plates push past one another in opposite directions). This reduced the speed of the water displacement and so reducing the size of the tsunami that hit the northern part of the Indian Ocean.[9]
The India Plate is part of the great Indo-Australian Plate, which underlies the Indian Ocean and Bay of Bengal, and is drifting north-east at an average of 6 cm/year (2 inches per year). The India Plate meets the Burma Plate (which is considered a portion of the great Eurasian Plate) at the Sunda Trench. At this point the India Plate subducts beneath the Burma Plate, which carries the Nicobar Islands, the Andaman Islands and northern Sumatra. The India Plate slips deeper and deeper beneath the Burma Plate until the increasing temperature and pressure drive volatiles out of the subducting plate. These volatiles rise into the crust above and trigger melt which exits the earth's crust through volcanoes in the form of a volcanic arc. The volcanic activity that results as the Indo-Australian plate subducts the Eurasian plate has created the Sunda Arc.
As well as the sideways movement between the plates, the sea bed is estimated to have risen by several metres, displacing an estimated 30 km³ (7 cu mi) of water and triggering devastating tsunami waves. The waves did not originate from a point source, as was inaccurately depicted in some illustrations of their paths of travel, but rather radiated outwards along the entire 1,600 km (994 mi) length of the rupture (acting as a line source). This greatly increased the geographical area over which the waves were observed, reaching as far as Mexico, Chile and the Arctic. The raising of the sea bed significantly reduced the capacity of the Indian Ocean, producing a permanent rise in the global sea level by an estimated 0.1 mm.[10]
Aftershocks and other earthquakes
Numerous aftershocks were reported off the Andaman Islands, the Nicobar Islands and the region of the original epicentre in the hours and days that followed. The largest aftershock, which originated off the coast of the Sumatran island of Nias, registered a magnitude of 8.7,[11] prompting debate among seismologists as to whether it should be classified as an aftershock of the December 2004 quake or as a "triggered earthquake" (which typically differs from an aftershock in that it is not located along the same fault line and may be as large or larger than the earthquake which triggered it).[12] This earthquake was so large that it produced its own aftershocks (some registering a magnitude of as great as 6.1) and presently ranks as the 7th largest earthquake on record since 1900. Other aftershocks of up to magnitude 6.6 continued to shake the region daily for up to three or four months.[13][14] As well as continuing aftershocks, the energy released by the original earthquake continued to make its presence felt well after the event. A week after the earthquake, its reverberations could still be measured, providing valuable scientific data about the Earth's interior.
The 2004 Indian Ocean earthquake came just three days after a magnitude 8.1 earthquake in an uninhabited region west of New Zealand's sub-Antarctic Auckland Islands, and north of Australia's Macquarie Island. This is unusual, since earthquakes of magnitude 8 or more occur only about once per year on average.[15] Some seismologists have speculated about a connection between these two earthquakes, saying that the former one might have been a catalyst to the Indian Ocean earthquake, as the two earthquakes happened on opposite sides of the Indo-Australian Plate. However, the U.S. Geological Survey sees no evidence of a causal relationship in this incident. Coincidentally, the earthquake struck almost exactly one year (to the hour) after a 6.6 magnitude earthquake killed an estimated 30,000 people in the city of Bam in Iran on December 26, 2003.[16]
Some scientists confirm that the December earthquake had activated Leuser Mountain, a volcano in Aceh province along the same range of peaks as Mount Talang, while the 2005 Sumatran earthquake had sparked activity in Lake Toba, an ancient crater in Sumatra.[17] Geologists say that the eruption of Mount Talang in April 2005 is connected to the December earthquake.[18]
Energy of the earthquake
The total energy released by the 2004 Indian Ocean earthquake was estimated early on to be as much as 3.35 exajoules (3.35×1018 joules). This is equivalent to over 930 terawatt hours, 0.8 gigatons of TNT, or about as much energy as is used in the United States in 11 days. A new seismic energy release estimate, however, generated September 30, 2005 using new data, placed the amount of energy released by the earthquake at the somewhat smaller figure of 1.1×1018 joules, which is equivalent to about 250 megatons of TNT. The earthquake generated seismic oscillation of the Earth's surface of up to 20–30 cm (8–12 in), equivalent to the effect of the tidal forces caused by the Sun and Moon. The shock waves of the earthquake were felt across the planet; as far away as the U.S. state of Oklahoma, where vertical movements of 3 mm (0.12 in) were recorded.[19]
Because of its enormous energy release and shallow rupture depth, the earthquake generated remarkable seismic ground motions around the globe, particularly due to huge Rayleigh (surface) elastic waves that exceeded 1 cm in vertical amplitude everywhere on Earth. The record section plot below displays vertical displacements of the Earth's surface recorded by seismometers from the IRIS/USGS Global Seismographic Network plotted with respect to time (since the earthquake initiation) on the horizontal axis, and vertical displacements of the Earth on the vertical axis (note the 1 cm scale bar at the bottom for scale). The seismograms are arranged vertically by distance from the epicenter in degrees. The earliest, lower amplitude, signal is that of the compressional (P) wave, which takes about 22 minutes to reach the other side of the planet (the antipode; in this case near Ecuador). The largest amplitude signals are seismic surface waves that reach the antipode after about 100 minutes. The surface waves can be clearly seen to reinforce near the antipode (with the closest seismic stations in Ecuador), and to subsequently encircle the planet to return to the epicentral region after about 200 minutes. A major aftershock (magnitude 7.1) can be seen at the closest stations starting just after the 200 minute mark. This aftershock would be considered a major earthquake under ordinary circumstances, but is dwarfed by the mainshock.
The shift of mass and the massive release of energy very slightly altered the Earth's rotation. The exact amount is not yet known, but theoretical models suggest the earthquake shortened the length of a day by 2.68 microseconds (2.68 µs, or about one billionth of the length of a day), due to a decrease in the oblateness of the Earth.[20] It also caused the Earth to minutely "wobble" on its axis by up to 2.5 cm (1 in) in the direction of 145° east longitude,[21] or perhaps by up to 5 or 6 cm (2.0 to 2.4 in).[22] However, because of tidal effects of the Moon, the length of a day increases at an average of 15 µs per year, so any rotational change due to the earthquake will be lost quickly. Similarly, the natural Chandler wobble of the Earth, which in some cases can be up to 15 m (50 ft), will eventually offset the minor wobble produced by the earthquake.
More spectacularly, there was 10 m (33 ft) movement laterally and 4–5 m (13–16 ft) vertically along the fault line. Early speculation was that some of the smaller islands south-west of Sumatra, which is on the Burma Plate (the southern regions are on the Sunda Plate), might have moved south-west by up to 36 m (118 ft), but more accurate data released more than a month after the earthquake found the movement to be about 20 cm (7.9 in).[23] Since movement was vertical as well as lateral, some coastal areas may have been moved to below sea level. The Andaman and Nicobar Islands appear to have shifted south-west by around 1.25 m (4.1 ft) and to have sunk by 1 m (3.28 ft).[24]
In February 2005 the Royal Navy vessel HMS Scott surveyed the seabed around the earthquake zone, which varies in depth between 1,000 m and 5,000 m (3,300 ft and 16,500 ft). The survey, conducted using a high-resolution, multi-beam sonar system, revealed that the earthquake had made a huge impact on the topography of the seabed. 1,500-meter (5,000 ft) high thrust ridges created by previous geologic activity along the fault had collapsed, generating landslides several kilometers wide. One such landslide consisted of a single block of rock some 100 m high and 2 km long (300 ft by 1.25 mi). The momentum of the water displaced by tectonic uplift had also dragged massive slabs of rock, each weighing millions of tons, as far as 10 km (7 mi) across the seabed. An oceanic trench several kilometres wide was exposed in the earthquake zone.[25]
The TOPEX/Poseidon and Jason 1 satellites happened to pass over the tsunami as it was crossing the ocean.[26] These satellites carry radars that measure precisely the height of the water surface; anomalies of the order of 50 cm (20 in) were measured. Measurements from these satellites may prove invaluable for the understanding of the earthquake and tsunami.[27] Unlike data from tide gauges installed on shores, measurements obtained in the middle of the ocean can be used for computing the parameters of the source earthquake without having to compensate for the complex ways in which close proximity to the coast changes the size and shape of a wave.
