Some two million people in Bolivia face the risk of a magnitude 8.9 megaquake, 125 times stronger than the previously calculated potential maximum, according a study published Sunday.

The findings, reported in Nature Geoscience, came as a surprise, the researchers said.

"No one suspected that the previous estimates were too low," said Benjamin Brooks, a geophysicist at the University of Hawaii Manoa and lead author of the study.

Earlier calculations had set the most powerful expected earthquake for the region east of the central Andes mountains at magnitude 7.5, based in part on a relatively quiet seismic history.

But a careful analysis of Global Positioning System (GPS) data from the eastern flank of the mountain chain showed a buildup of stress consistent with a maximum magnitude of 8.7 to 8.9, Brooks said.

Tracking tiny changes in the location of GPS stations enables scientists to measure surface velocity to within a fraction of a millimeter per year.

The data showed that the area to the west of the Mandeyapecua thrust fault, which runs north-south, has moved far more than the area east of the fault, pointing to a dangerous accumulation of pressure.

A shallow section of the fault is locked in place over a length of about 100 kilometres (60 miles), the researchers found.

"Rupture of the entire locked section by one earthquake could result in shaking of magnitudes up to 8.9," Brooks said in a statement.

There is no way to know when such a quake might happen -- or if it will happen at all, he added.

A series of smaller temblors could release stress without unleashing a Big One.

To follow up, Brooks is probing the ancient seismic history of the region to determine the dates and sizes of past earthquakes, and to find out if one that size has ever occurred there.

"We hope that this information will be widely disseminated and considered in Bolivia by the people -- including the general population, engineers, planners, policy-makers -- who may be most affected," he said.

Several major earthquakes in recent years -- including the magnitude 9.0 monster off the coast of Japan in March, and an 8.8 quake in Chile in February 2010 -- have prompted some experts to reassess earlier forecasts.

"We probably should re-evaluate our estimates of the maximum sizes of earthquakes that could strike all fault areas," Ross Stein, a seismologist at the US Geological Survey (USGS), said shortly after the Japan quake.