Scientists are hoping that an observatory they are working to submerge in one of the world's most active faults -- California's San Andreas Fault -- will help yield answers to the causes of earthquakes and other questions that have perplexed them for decades.

The San Andreas Fault, which skirts the western edge of California, marks the boundary between the Pacific Plate on the west and the North American Plate on the east. As the two plates grind against each other, they create earthquakes.

In an effort to learn how to predict earthquakes, or if predicting them is even possible, U.S. Geological Survey and other scientists are working to plant an observatory 2.4 miles deep into a repeating earthquake area of the fault near the tiny town of Parkfield in central California.

Parkfield has experienced a number of moderate earthquakes at a magnitude of about 6 at fairly regular intervals -- 1857, 1881, 1901, 1922, 1934 and 1966. Scientists estimated that the next one would occur between 1988 to 1993, but it is overdue.

A magnitude 6.0 earthquake struck near Parkfield on Sept. 28, 2004, which scientists believe may have reduced the stress at the part of the fault where they are building the San Andreas Fault Observatory at Depth. Once finished, SAFOD will enable them to get real-time measurements of changes in temperature, rock deformation and fluid pressure that precede an earthquake.

Through the readings, scientists hope to learn more about the connection between fluid pressure and quakes, said Steve Hickman, senior research scientist with the U.S. Geological Survey and co-principal investigator of SAFOD. When the fault heats up, fluid -- mostly water -- swells and could be instrumental in triggering an earthquake or controlling the size of the earthquake, he said.

By observing the behavior of fluids, scientists may be able to recognize a pattern when an earthquake is about to occur. Such short-term earthquake prediction is the "Holy Grail" of geophysicists, said Hickman. At this point, "most scientists don't even know if [earthquake prediction] is possible," he added.

SAFOD also will enable scientists to retrieve samples of subsurface rocks and fluids for laboratory analysis.

A pilot hole was drilled in the summer of 2002 about a mile from the fault line to help guide construction of the main hole, which used the oil industry technique of directional or diagonal drilling. And in December 2005, project participants announced they had reached a milestone in the project -- drilling across the San Andreas Fault. They plan to start deploying instruments in January and taking core samples of the fault in the summer of 2007.

SAFOD is one component of the National Science Foundation's five-year $219 million EarthScope project, which is intended to help scientists observe and learn more about the changes in the North American continent and why earthquakes and volcanic eruptions occur.

Also under EarthScope is a project called the United States Seismic Array, which will create a grid of semi-portable seismometers to produce a high-resolution image of the Earth's crust and underlying mantle to help link structures from earlier continental formation with potential geologic hazards, according to NSF, which also funds the NewsHour's Science Unit.

The Plate Boundary Observatory will employ a network of Global Positioning System receivers and strainmeters covering the western United States and Alaska to record even the smallest movement across faults.

And the Interferometric Synthetic Aperture Radar, with the participation of the National Aeronautics and Space Administration, will use a satellite to detect changes in the Earth's surface before, during and after major earthquakes and volcanic eruptions.

EarthScope is intended to incorporate information across the Earth sciences and the Earth science community, including members of USGS, NSF, NASA, the Department of Energy, regional seismic networks, state geological surveys and 100 universities.

-- By Larisa Epatko, Online NewsHour