The northern end of the Bay of Bengal could be at risk of giant earthquakes and tsunamis in the coming decades, an Australian study concludes.
Thousands of people lost their lives in the December 2004 tsunami
Such events have been thought unlikely there, in contrast to the area further south where the 2004 tsunami began.
But the new work, published in the journal Nature, has found "compelling evidence" for tsunami-triggering earthquake activity to the north.
Geologists have said this warning should be taken "very seriously".
The area is densely populated, and more than a million of people could potentially be at risk.
The magnitude 9.2 earthquake that struck off the Sumatran coast on 26 December 2004 and the tsunami it generated killed thousands of people and left millions homeless.
It stemmed from a geological area known as a subduction zone.
Here, part of the Indian/Australian tectonic plate was slowly burrowing beneath a component of the Eurasian plate.
This created stresses in the upper plate, which were violently released in the form of a "locked-thrust fault" earthquake as it sprung back up, which in turn triggered the tsunami.
Since then, another stress point has been identified to the east of the 2004 epicentre, but the subduction zone further north along the Myanmar coast was thought to be of little concern.
Now, Phil Cummins, lead author on the Nature paper and a geologist at Geoscience Australia, believes this is not the case.
He said: "I reviewed the geological literature and found the evidence for a lack of tectonic activity along the Myanmar coast was not compelling."
Recent GPS data, he said, suggested that the plate boundary was at sea in this area, hidden below thick layers of sediment.
Dr Cummins added: "Although these GPS measurements are sparse, these show that there is active deformation near the Myanmar coast that is consistent with a locked thrust-fault offshore, which is the type needed to generate tsunami."
The geologist also looked at accounts of an earthquake that occurred in the area in 1762, which wrenched up parts of the coast by between 3-7m.
A computer simulation shows the havoc a tsunami could wreak
His computer simulation of the quake, which he believes would have measured magnitude 8.8, showed that a similar event today would have significant impacts.
"Such an earthquake would generate a large tsunami that could have a pronounced impact on the Chittagong coast and the Ganges delta," he said.
"The latter region is home to 60 million people living within just 10m of sea level."
Meanwhile, the quake itself could cause major damage to the region's largest cities, Calcutta and Dhaka. Overall, the simulation suggested that more than a million lives could be at risk.
Professor Richard Arculus, from Australian National University (ANU) in Canberra, said: "Phil Cummins' warnings should be taken very seriously.
"A few months before the devastating earthquake and accompanying tsunami triggered off northern Sumatra in late 2004, Phil Cummins published a perceptive analysis of historic events of this nature in the region.
"He warned that countries bordering the Indian Ocean, including the northern coast of Australia, were at significant risk, and the lack of a tsunami warning system analogous to that deployed in the Pacific was a serious issue.
"So his credibility with respect to tsunamigenic earthquakes is established."
Kevin McCue, a professor at Central Queensland University (CQU), added: "The message is alarming, perhaps justifiably, given the unexpected disaster that followed the great Sumatran earthquake and tsunami of 2004, a disaster of local, regional and global reach."
But he added: "Disaster planners might need more information than is given in the paper, particularly some quantitative measure of uncertainties in the science."
Phil Cummins agrees that more work is needed to confirm his analysis, and suggests this should take place before any drastic mitigation measures are considered.