We started our work to understand the earthquake potency in the Andaman region in August 2003, almost one and half years before the 2004 tsunami-earthquake struck the area. A moderate earthquake that hit the town of Diglipur at the northern tip of the Andaman group of islands in September 2002 had kindled our interest. The paper concluded, “…issue is the lack of a good database on the effects of tsunami waves, to which not only the coast of Andaman-Nicobar is exposed, but also the eastern coast of India – a threat that is generally underestimated.” There was a hint about the hidden threat from tsunamis in this region.
Following the intuition that the region would generate large earthquakes in future, our team at the Centre for Earth Science Studies in Trivandrum took up a project to understand more about the land-level changes along the Andaman and Nicobar Islands. We started GPS surveys in August 2003 in a campaign mode, which means the stations are not permanent, but the readings are taken at the same point occasionally. Casual observation of microatolls6 above the high-tide level caught our attention, but we could not recognise it as a sign of an impending earthquake. An experienced researcher exposed to the dynamics of subduction zones could have suspected that the stress was building up. We recognized them in hindsight and discussed the importance of such signals in a research paper published after the earthquake.
Coseismic ground level changes are generally associated with many great subduction zone earthquakes. George Plafker, an American geologist, found land-level changes associated with earthquakes in Alaska and Chile in the 1960s. Brian Atwater, a geologist at the United States Geological Survey and a pioneer in tsunami geology research, discovered red cider trees killed by land subsidence in the 1700 CE Cascadia earthquake in the northwest United States.
Kerry Sieh, a professor at the California Institute of Technology, and a team of researchers had been monitoring the microatolls to pick up evidence for past earthquakes, taking cues from Alaska and Cascadia. Their study took advantage of the sensitivity of corals to sea-level variations to identify past earthquakes and estimate recurrence patterns. However, the 2004 earthquake occurred while their studies were in progress; it was before they could identify any pre-earthquake process. Despite the missed opportunity, the earthquake gave researchers an impetus to study past earthquakes and tsunamis that might have impacted the region.
As the tsunami washed the coastal strips along the eastern Indian shore, it scoured the beaches and exposed ancient structures that had remained buried. The archaeological survey conducted excavations in Mamallapuram, where the scouring exposed the remains of a temple. During our surveys, we found the first evidence of a tsunami from the distant past within the temple’s remains. It was a layer of sand that looked out of place and not part of a usual beach sequence. In the laboratory, we discovered deep marine organisms not usually associated with regular beach deposits. We dated the sand layer with marine organisms as about 1000 years old, the first report of an earlier tsunami that visited the Indian shores prior to the 1881 tsunami.
Next, we surveyed Kaveripattinam (Poompuhar), the celebrated port city of the Chola kings (of the first millennium) on the Tamil Nadu coast. Known to the Roman author and naturalist Pliny the Elder (23-79 CE) as “Khaberis”, this site was a major centre of maritime trade until the tenth century CE. Here, the 2004 tsunami had travelled nearly 1 km inland. Trenches excavated in Kaveripattinam exposed evidence of a previous sea incursion in the form of tsunami deposits of the same vintage as at Mamallapuram. Interestingly, an early South Indian literary work Manimekhalai, by a Buddhist poet named Seethalai Saathanar, mentions the “angry” sea. He writes that the sea “swallowed” a part of the ancient port city of Kaveripattinam. Could this be an ancient tsunami? Is that the same event that deposited the layer of sand in the trench that we excavated? Some scholars believe that Manimekhalai (890 and 950) describes the sea as “angry”, as the sea was rough following a tsunami.
Tsunami deposits are found in faraway shores, but evidence for coseismic land-level changes is restricted to the rupture zone. During the post-earthquake surveys, we mapped the uplift and subsidence of land. We followed the ideas of George Plafker and Brian Atwater to explore subsidence locations in search of evidence of such events in the past. We spotted a line of roots sticking out to the stream at the low tide, about one metre below the ground. They belonged to a variety of tropical mangrove trees, Rhizophora, the most abundant species in these islands. The radiocarbon age of these roots turned out to be about 1000 years, in the same range as the date of the tsunami sand from Mamallapuram. Perhaps, we have been able to nail the predecessor of the 2004 earthquake and the tsunami. More proof was coming.
Around the same time, a team led by Brian Atwater was working in Phra Thong, a barrier island along the hard-hit west coast of Thailand. They provided evidence for at least three previous significant tsunamis in the preceding 2800 years. The youngest one occurred about 550-700 years ago. In a different part of the impacted zone, in the coastal swamps of Sumatra (Meulaboh), Katrin Monecke, a geologist formerly at Kent State University in Ohio, found tsunamigenic deposits dated 780-990 CE. Later investigations revealed more evidence for past tsunamis during the time intervals 770-1040 CE and 1250-1450 CE. These results led to the conclusion that the 2004 tsunami was not the first to have occurred.
The history of tsunamis around the globe will tell us that none of them can be the last. Perhaps a few hundred years would pass before the story repeats. But we never know. That is the nature of the earth. There are too many unknowns that are at play.
The 2004 tsunami triggered a fresh interest in similar events in the past. Most notable was the 1945 Makran tsunami of Pakistan, which had not created any significant damage in India. Another tsunami from 1881, triggered by a magnitude ~8 earthquake in the Andaman-Nicobar, was too small, 0.9 m in Port Blair and 0.8 m on the east coast of India, as recorded by tide gauges. An earthquake also occurred in the region in 1941, but apparently it did not generate any notable tsunami. The large earthquake of January 28, 1679, from the middle or north Andaman, falls in the same category of non-tsunamigenic earthquakes. Could there be other potential sources? A major earthquake of April 2, 1762, near the Arakan coast of Myanmar is believed to have generated a tsunami that reached the coast of Bengal. However, it is uncertain if the inundation was extensive along the Bengal coast (Bangladesh and West Bengal in India).
Excerpted with permission from The Rumbling Earth: The Story of Indian Earthquakes, CP Rajendran and Kusala Rajendran, Penguin India.