Earthscope home
about activities data did you know? news site map

EarthScope > News > Researchers Drill Towards the San Andreas Fault

Researchers Drill Towards the San Andreas Fault
Agnieszka P. Baier

Aerial view of the drilling site. Credit: John Langbein, USGS.

On June 12, an international team of geoscientists started drilling a 2 km pilot hole near the San Andreas fault to unravel the mystery of how this fault works. The hole, located in Parkfield, California, is the first step in creating SAFOD--the San Andreas Observatory at Depth, which is part of EarthScope. Scientists hope that this experiment will help them understand in more detail how earthquakes occur and, one day maybe help them predict earthquakes.

In Parkfield, California you can walk up to a ranch fence and see how it has become crooked over time by the movement of the San Andreas fault (SAF). In aerial photography, the fault appears as a prominent scar in the ground. In fact, it is a thin feature in a much larger boundary between the North American and the Pacific plates. Nobody knows what the SAF is made of at depth, nor how it behaves or changes beneath the surface. Plate boundaries are important to understand because they are the places where most earthquakes originate.

Although the San Andreas is the best understood fault in the world and Parkfield is the best known place on this fault, the area still holds many mysteries. Southeast of Parkfield, the fault is virtually silent. No earthquakes have occurred on this section of the SAF for about 150 years. Near San Francisco and north of it, the fault has been locked since the 1906 rupture. However, between San Juan Bautista and Parkfield, the fault "creeps." It is weak and it never stores up enough stress to rupture in major earthquakes. It slides silently, releasing any built up stress in earthquakes not bigger than magnitude 5.5. Many small areas on the fault are sources of repeating, nearly identical earthquakes. Why does the fault behave this way? Scientists do not know.

The SAFOD pilot hole is located 1.1 southwest of the San Andreas fault near Parkfield, CA. The drill site is situated in a meadow in the rolling ranch land of Central California. Credit: Mark Zoback, Stanford University.

Earthquakes are one of the most powerful natural forces and researchers know very little about when earthquakes will happen and how large they will be. This is important to know if we want to reduce hazards associated with earthquakes and one day be able to predict them. However, earthquake prediction is much more difficult than weather forecasting, which we know is not always successful despite the assistance of advanced technology. "We are trying to predict a very chaotic phenomenon happening deep within the Earth where we cannot even look or measure that region directly," explains Clifford Thurber, Professor of Geophysics at the University of Wisconsin-Madison. Parkfield is a good example of how unpredictable earthquakes are. "Originally, researchers focused on Parkfield because they believed that earthquakes were happening there at regular intervals. The last one occurred in 1966 and, based on statistical analyses, the next one was forecast to happen by the early 1990s. It did not. Earth seems to be giving us some clues but it does not do so in any reliable way. Thus, researchers who attempted to predict earthquakes have had both remarkable successes and terrible failures," says Dr. Thurber.

The pilot hole is located 1.8 km from the San Andreas fault and will be drilled to 2.2 km depth using standard oil drilling equipment. Initially, the hole will be drilled with a rotary bit, a steel instrument with three cones or wheels with large teeth that rotate to grind up the rock. "A typical hole is larger at the surface and narrows with depth. The main part will be about eight inches in diameter," describes Dr. William Ellsworth, Chief Scientist for the Western Region Earthquake Hazards Team at the US Geological Survey in Menlo Park, CA. Dr. Ellsworth has been one of the driving forces behind the drilling program at Parkfield.

A typical drill bit has three cones (or wheels) with numerous teetch on each cone.

As the hole is drilled, fluids are pumped down into the hole to keep the bit cool and to flush the crushed rock up to the surface. Geologists will study these rocks and gas samples to determine conditions in the hole. After about a month, the hole will be nearly 2 km deep. At that point, drilling will stop. A variety of geophysical instruments will be lowered into the bore hole to take measurements.

Next, a steel pipe will be cemented into the open hole to keep it from collapsing and to prevent fluids in one rock layer from moving into another. When drilling resumes, a specialized drill bit called a coring bit will be used. It is a diamond-studded tool hollow in the center. During the conventional drilling, the rock is ground up and pumped out of the hole. However, the coring bit, while advancing, will leave an undisturbed piece of rock in the middle of the hole. Instead of chip-sized fragments, scientists will retrieve a continuous piece of rock as much as 60 meters long. According to Dr. Ellsworth, the coring should be completed by the end of July. Later, samples will be available to researchers all over the world.

A diamond studded coring bit as shown here is used to cut a solid core of rock during drilling. Scientists can learn much about how the rock formed and conditions it was under from the core.

One of the goals of SAFOD experiment is to understand the orientation and the magnitude of the forces and the stresses that drive the San Andreas Fault. "We have drilled holes and have used previously drilled holes to study earthquakes for over 10 years now," says Dr. Ellsworth. By placing underground instruments that record seismic waves and small shifts in the ground motion, scientists can minimize any noise or movements on the surface that interfere with instruments. The deeper the equipment is installed, the more accurate the recordings. "In the Parkfield area we have dozens of relatively shallow drill holes in which we have placed instruments. One of them, an abandoned oil well, is even a mile deep and in the mid 1980s we installed a variety of scientific instruments there. However, this hole is not near enough the SAF and thus cannot be used to answer a lot of unresolved questions about the fault. All these efforts have taken us close, but still not into the active fault zone and not deep enough," explains Dr. Ellsworth.

Critical to understanding the forces operating within the San Andreas will be the hydrofracture test, in which the researchers will pressurize the well with enough fluid pressure to split the rock locally. "Based on the direction that the crack opens up in the small part of the hole we will be able to determine the orientation of the forces that are driving the San Andreas Fault and also estimate their magnitude," explains Dr. Ellsworth.

Scientists have many models of earthquake processes, but are still uncertain whether they are correct or not. The pilot experiment will help the researchers build a comprehensive model of the San Andreas. "It will give us an idea about the magnitude of stresses and their orientations as we approach the San Andreas. It may eliminate some of the existing models of how the fault works. We know how strong rocks are. During laboratory experiments we can squeeze a rock under conditions at which earthquakes occur, so we know how much force we need to apply in the lab to slide a fault or break a rock. Based on these measurements, we can estimate how much heat is produced when the fault moves and what temperatures should be near a slipping fault. However, when we look at the San Andreas, we find that the fault's temperature profile does not fit this pattern. It implies that the SAF moves at lower stress levels than we encounter in laboratory rocks or at faults elsewhere. We want to understand what allows the SAF to move at relatively weak levels. This is critical to understand what parts of the fault might be important for generation of future earthquakes," argues Bill Ellsworth.

During the pilot hole experiment the scientists will also retrieve the rocks from the area adjacent to the fault and samples of fluids flowing through it. These fluids may be weakening the fault. Dr. Thurber explains: "Think about an air hockey table where you pump pressurized air out to keep objects floating around. The air reduces the friction between the objects and the table so the objects can fly around on the table. At depth, fluids act in the same. They may be pumping up the fault and spreading it by reducing the friction between the rocks."

Geologic cross-section through the San Andreas fault at the site of the SAFOD experiment. The SAFOD pilot hole is being drilled into fractured granitic basement rock (shown in pink) beneath near surface sandstone and shale (tan). The contour lines indicate the velocity Of seismic waves in the near surface and thus help identify the depth to the granitic rocks (Courtesy of Michael Rymer and Robert Jachens, USGS).

Dr. Thurber will continue studying the fault with a technique called seismic tomography. "We use seismic waves to look inside the Earth in the same way a CAT scan utilizes X-rays to create images of a human brain. Earthquakes usually generate most of the seismic waves, but sometimes we supplement them with small explosions in key points to better image the fault. A CAT scan shows density variations inside a human brain that enable the identifications of tumors. We look for small variations in the velocity of seismic waves traveling through the Earth. They tell us about the physical properties of the rocks. This information helps us locate the fault zones, determine how wide they are and how deep they stretch," says Dr. Thurber. He is particularly interested to learn about the relationship between the areas where earthquakes occur and the structure and types of rock in which they are happening. This information would help refine the models of how the fault works and learn about the conditions deep in the fault.

Using seismic tomography, Dr. Thurber and his collaborators have already created images of the San Andreas. These images picture the fault as a sharp boundary in the Earth between the North American and the Pacific plates. When imaged with seismic waves, the fault shows up as a region where seismic wave velocities are much slower than in the adjacent regions. Waves travel through different rocks with different velocities. Hard rocks allow these waves to travel faster and weaker rocks slow them down. "We still do not know the exact type of rock that the SAF is made of. Geologists can excavate the rocks on the fault's surface, but these rocks do not reveal the nature of the fault at depth and the areas inside the fault where earthquakes are occurring. We can obtain information on the seismic velocities which allows us to hypothesize what the fault's properties are, but we do not have any direct evidence of what the rocks are. This is why we need to drill," explains Dr. Thurber.

Cliff Thurber and his collaborators will also use the pilot hole to locate precisely the targets for the future SAFOD hole. "We need to know specific geographic coordinates with the accuracy of about 100 meters to make sure that when we drill the SAFOD hole, we will hit the SAF and will be within meters of an earthquake," adds Cliff Thurber. "We have already located small earthquakes using our surface instruments. The next step will be to set off explosions near our seismic stations on the surface and record them on the seismometer at the bottom of the pilot hole. Once we have done that for a number of our stations, we can use the seismic waves to tell the relative position of events in the Earth and to determine very accurately how far apart the real earthquakes are from the borehole."

Drill rig with casing pipes in the foreground

The next step in studying the San Andreas fault will be SAFOD, an experiment unprecedented anywhere in the world. The SAFOD's goal is to drill still a deeper borehole and install instruments directly within the San Andreas fault zone. "To go deeper means going to the depths where earthquakes occur. These phenomena do not start at shallow depths. Faults rupture at the surface, but the place where the forces are stored and released are deep within the fault. The SAFOD hole will go deep enough to get to the top of the interesting zone," explains William Ellsworth. The instruments will be set 3 to 4 km beneath the Earth's surface, a known area of persistent minor seismicity at depths of 3 to 4 km.

The SAFOD hole will drill straight down for about 2 km and then continue at an angle across the entire fault zone until relatively undisturbed rock is reached on the east side at approximately 3-4 km depth. It will be the first bore hole to penetrate where earthquakes happen, intersect the fault, and sample it at two different spots: where the fault is creeping and where it breaks in magnitude 2 earthquakes. But even the SAFOD hole is not the dream project for Dr. Ellsworth: "Ultimately, I envision drilling 9-10 km deep to the regions where the greatest earthquakes originate." Deep-drilling experiments in locations other than the SAF have been proposed. Some researchers want to sample the Chelungpu fault in Taiwan that produced the devastating Chi Chi earthquake in 1999. Also, the ocean drilling community proposed to use a drilling ship to study subduction zones of the Nankai Trough offshore Japan. Subduction zones, another type of plate boundary, are even more mysterious than the San Andreas fault.

The pilot hole is primarily funded by the International Continental Drilling Program, a consortium of national governments (including the U.S. government) headquartered in Germany. The experiment is also supported by the National Science Foundation (NSF) and the US Geological Survey (USGS). SAFOD is part of EarthScope, a partnership extending throughout the Earth science community, including more than 100 universities, NSF, USGS, NASA, Department of Energy, regional seismic networks, and state geological surveys.

To learn more about earthquakes and SAFOD, read our previous story on SAFOD or go to the press release by Stanford University and the USGS. Additional information about SAFOD can be found at


US Array | SAFOD | PBO | InSAR | EON
about | activities | data | did you know? | news | site map | home

Last update 21-May-2002
comments and questions