A major earthquake 2,500 years ago caused one of the largest rivers on Earth to abruptly change its course, according to scientists. The study, published this week in Nature Communications, showed that a previously unknown earthquake in the Bengal delta of present-day Bangladesh rerouted the main channel of the Ganges River in a process known as river avulsion.
"Earthquakes pose widely recognized and well-studied geohazards," said Liz Chamberlain, lead author of the study and an assistant professor at Wageningen University & Research. "Similarly, it is clear that river disturbances like dam collapse can be catastrophic. But it was not previously confirmed that earthquakes could drive river avulsion (rerouting) in deltas - especially for an immense river like the Ganges."
Bangladesh hosts one of the most densely populated landscapes on the planet. A modern recurrence of such an earthquake could bring flooding and widespread softening of sandy deposits, known as sediment liquefaction, as well as ground shaking. "This could impact millions of people living in the Ganges floodplain today," said Steve Goodbred, coauthor on the study and a professor at Vanderbilt University, USA.
Sand dikes in sediment reveal ancient earthquake
The earthquake and river avulsion were reconstructed by studying buried sediments. Sustained shaking during an earthquake can inject pressurized sand into overlying mud layers. This phenomenon is recorded in the soft sediments of deltas as vertical sand dikes, or "seismites". Modern examples are found around the world.
The scale of the earthquake was determined using relationships to sand dike width and distance to earthquake origin, and its timing was established with optically stimulated luminescence dating performed at the Netherlands Centre for Luminescence dating.
"In this case, we found extensive 30-40 cm-wide sand dikes breaching a 3-4 m thick mud cap. The scale of these sand dikes, and their position more than 180 km from the nearest likely seismogenic zones, indicate a prehistoric earthquake likely in the range of M 7.0-8.0" said Mike Steckler, a coauthor on the study and research professor at Lamont-Doherty Earth Observatory of Columbia University. "This was a big event for the Bengal delta."
Landscape reconstructions to peer into the past
Adjacent to the sand dikes, the scientists discovered a 1.5-km wide abandoned river channel, also dated to 2,500 years. "The scale, grain size, and chemistry of the channel deposits show this was the main pathway of the Ganges River and that its abandonment was immediate," said Goodbred. Chamberlain noted that landscape reconstructions can be a puzzle with many unknowns. "However, reconstructions also provide the great benefit of peering into the past," she said. "This allows us to identify Earth surface processes and dynamics happening over very long timescales and low-recurrence intervals that are omitted from modern and historic records. A prime example is this earthquake-triggered river avulsion".
Better prediction of natural hazards and their impacts
Though infrequent, low-recurrence geologic hazards can cause significant damage, for example, the deadly 2004 Indian Ocean tsunami impacts to Indonesia. "In Indonesia, archives of prehistoric tsunamis existed but had not been sufficiently interrogated by scientists to spur the needed awareness or preparation prior to the 2004 recurrence," said Chamberlain. "In the case of our study, the new knowledge comes in advance so it can be helpful for awareness and hazard mitigation planning."
Syed Humayun Akhter, a coauthor on the study and Vice-Chancellor of Bangladesh Open University, compared the findings to the 1762 Arkan earthquake in Bangladesh, estimated to be M 8.5 or greater. "Large earthquakes impact large areas and can have long-lasting economic, social, and political effects," Akhter said.
"This study clearly shows how basic research can improve our ability to predict natural hazards and their impacts," said Justin Lawrence, a program officer at the U.S. National Science Foundation, which funded this work.