The August 23, 2011 earthquake centered on Mineral, Virginia was felt as far north as Vermont, and even in parts of Canada – and researchers think there were warning signs in the air above. (Credit: USGS)

Some earthquakes occur underneath abnormally warm patches of atmosphere that form just hours or days before the Earth starts to rumble, according to preliminary reports from American and Russian scientists.

While geophysicist Dimitar Ouzounov of Chapman University in Orange, Calif., and colleagues have not yet succeeded in predicting earthquakes, they have linked past earthquakes to thermal and electromagnetic changes in the air.

In a poster session Monday at the American Geophysical Union Fall Meeting, the team presented supporting evidence from the August 23, 2011 magnitude-5.8 earthquake in Mineral, Virginia.

The scientists analyzed seven years of satellite images over the earthquake site, which were taken at the same time every morning. Focusing on the month of August, the researchers used computer algorithms to model the average heat radiating from the Earth each day.

Thermal radiation varied across days and seasons, but by comparing the same dates over several years, the researchers detected statistically significant spikes 8 to 12 kilometers (5 to 7.5 miles) above central Virginia on two days—August 11 and August 23. In data from August 23, the day of the quake, the team also observed electromagnetic changes higher up in the ionosphere at altitudes of about 400 kilometers (25 miles).

The pattern of changes follows a highly controversial theory. Proponents argue that before an earthquake, increased release of radon gas from the Earth’s surface heats the lower atmosphere, which in turn triggers electrical and chemical changes in the ionosphere.

In the Virginia dataset, the thermal changes preceded the earthquake by about seven hours — a warning window that Ouzounov believes could vary in different regions of the globe. The thermal “hotspot” appeared about 100 kilometers (62 miles) from the eventual epicenter, which is the limit of spatial resolution in their measurements. Still, Ouzounov says that more work is needed before atmospheric measurements could be used to predict quakes.

An even bigger obstacle, however, may be convincing other researchers that the theory makes scientific sense.

“There’s a lot of skepticism,” says Ouzounov. Many other hazards such as tornadoes, floods, and winds have strong visual indicators. But in the case of earthquakes, the research team faces the challenge of arguing for an invisible and unconfirmed communication between Earth and sky.

Malcolm Johnston, a research geophysicist for the U.S. Geological Survey (USGS), finds such a connection hard to believe. He and other USGS scientists measured radon levels throughout California for 20 years without finding any clear relationship with earthquakes. He suspects natural weather fluctuations, which can be quite large, could explain the anomalous thermal radiation readings over Virginia.

“That’s the biggest noise source, and a crucial correction that needs to be done,” says Johnston.

Ouzounov hopes to convince his critics with more data and, eventually, successful predictions. Ultimately, he envisions earthquake prediction involving many factors — some measured here on Earth, and others found high in the sky.

– Helen Shen is a science communication graduate student at UC Santa Cruz