Charting Seismic Effects On Water Levels Can Refine Earthquake Understanding

Just last November, the magnitude 7.9 Denali earthquake in Alaska was credited with sloshing water in Seattle's Lake Union and Lake Pontchartrain in New Orleans, and was blamed the next day when muddy tap water turned up in Pennsylvania, where some water tables dropped as much as 6 inches.

But the relationship between seismic activity and hydrology is not well understood and is ripe for serious examination by scientists from the two disciplines, said David Montgomery, a University of Washington professor of Earth and space sciences.

He and Michael Manga, associate professor of Earth and planetary science at the University of California, Berkeley, reviewed evidence of changes in stream flow and water levels in wells following earthquakes dating as far back as 1906, when a quake estimated at magnitude 7.7 to 7.9 struck San Francisco. Montgomery is an expert in surface hydrology and Manga is an expert in subsurface and aquifer hydrology.

The scientists found that, generally, an earthquake's effects on water depend on the distance from the epicenter, the magnitude and the geologic conditions at the location where changes to a well or stream are noted. They also found that effects on wells and aquifers were likely to be recorded at substantially greater distances from an earthquake's epicenter than are changes to stream flow.

"Put the two together and what it says is that the stream-flow response is a completely different beast than the water-well response," said Montgomery, lead author of a paper documenting the findings that is being published in the June 27 edition of the journal Science.

Montgomery said the new analysis provides a framework for understanding the broad range of earthquakes' effects on hydrology, and should help guide the study of links between seismology and hydrology.

Montgomery and Manga found that a mild earthquake, around magnitude 3, could generate effects on subsurface water, such as in wells, as far as about 10 miles from the epicenter. But effects on well water from a magnitude 9 quake could be observed more than 6,000 miles away. In fact, the latter scenario played out in the 1964 Alaska earthquake that registered 9.2.

"Wells in South Africa, clear on the other side of the world, responded," Montgomery said. "They didn't respond much, mind you, but the observations corresponded with the Alaska earthquake."

In examining changes in surface water related to seismic activity, the scientists found that the maximum distance from the epicenter at which effects were noted corresponded closely with theories about the maximum distance from the epicenter that liquefaction could be expected in an earthquake of the same magnitude. In addition, those maximum distances were far less than for subsurface water. For example, a magnitude 9 quake produced surface water changes only as far as about 750 miles from the epicenter. Montgomery noted that stream flow changes could be detected at much greater distances if they were, in fact, occurring that far away from the epicenter.

When an earthquake occurs, well-water levels can change as energy from the quake compresses the rock containing the water, thus forcing water out of its pores. Similarly, the flow of streams on the surface can increase as the aquifer is compressed, or either liquefies or settles during strong shaking, and water rises to the surface, Montgomery said.

"It's like squeezing a sponge because you're reducing the pore space and the water comes out. It has to go somewhere," he said.

Changes to surface and subsurface water could be related to each other at very close distances from the epicenter, Montgomery said, but even then different processes control them. That becomes more evident by the way they react at greater distances.

"One gives us a window on connections between hydrology, seismology and deformation of the Earth's crust," he said, "and the other gives us a better picture of connections between hydrology, seismology and geology at the surface."