While much progress has been made over recent years in understanding the kinematics of San Andreas and other platebounding fault systems around the world, the physical and chemical processes controlling earthquake nucleation and rupture propagation remain a mystery. So in June 2004, the National Science Foundation (NSF) and US Geological Survey (USGS) began to drill a deep hole to install instruments directly within the fault zone near the initiation points of previous magnitude 6 earthquakes to reveal, for the first time, the nature of those processes.
The San Andreas Fault Observatory at Depth (SAFOD) project, launched as part of the US EarthScope, does not aim to predict earthquakes. Rather, one of its main objectives is to test the predictability of earthquakes, and it has already begun yielding valuable information.
SAFOD evolved from the Parkfi d Earthquake Experiment on the San Andreas Fault, in central California, and researchers' frustration that the huge amount of information obtained through field, laboratory and theoretical work, suffers from severe limitations - most notably inferences drawn from surface observations are indirect and subject to alternate interpretations.
For example, stress heterogeneities induced by fault slip can lead to considerable uncertainties in inferring past fluid pressures from observations of vein geometry in outcrop.
In research based on surface observations, a complex history of uplift and denudation may have severely altered, or even destroyed, evidence for deformation mechanisms, fault zone mineralogy and fluid composition during fault slip.
Drilling provides the only direct means of measuring pore pressure, stress, permeability, and other important parameters within and near an active fault zone at depth. It is also the only way to collect fluid and rock samples from the fault zone at seismogenic depths and to monitor time-dependent changes in fl uid pressure, fluid chemistry, deformation, temperature and electromagnetic properties at depth during the earthquake cycle.
The project team will be addressing a number of fundamental questions, including why plate boundaries such as the San Andreas are unique, why they are so narrow, why they persist for millions of years and what makes them so weak relative to that crust next to them.
Scientists will test rock and fl uid samples recovered from the fault zone to determine their compositions, deformation mechanisms, frictional behaviour and physical properties.
The borehole will serve as a long term geophysical observatory to directly monitor earthquakes, deformation, fluid pressure and ephemeral properties of the fault zone through multiple earthquake cycles.
Sampling, downhole measurements and long term monitoring will allow the team to find out more about the composition of fault zone materials and determine the laws that govern their behaviour; measure the stresses that initiate earthquakes and control their propagation; test hypotheses on the role of high pore fluid pressure and chemical reactions in controlling fault strength and earthquake recurrence; and observe the strain and radiated wave fi elds in the near fi eld of microearthquakes.
Many studies have already been carried out at the SAFOD site, including a high-resolution seismic reflection profile airborne and ground-based gravity and magnetic surveys, thermal and geochemical studies in shallow wells, magnetotelluric imaging, micro earthquake relocations and seismic tomography.
The team decided to drill down to 4km, rather than 10km - the depth where great earthquakes nucleate - because ultra-deep drilling into an active fault zone is extraordinarily expensive and difficult.
But even then SAFOD could never have happened without advances in drilling and instrumentation from the oil and gas industry. Scientists had originally planned continuous coring for the hole, but that proved unworkable. A petroleum industry specialist suggested drilling the hole and then using 'multilaterals' (satellite wells drilled from a single 'parent' well) for coring.
In phase one, in 2004, the main hole was rotary drilled vertically to a depth of about 2.2km. Then in phase two, the hole was directionally drilled at a 55º angle, deviating to pass straight through the fault zone to a final depth of about 4km. This phase finished in August, following the successful completion of a hydraulic fracture test, as the project team attempted to measure the least principal tectonic stress at a depth of over 3km on the north east side of the fault.
Phase 3 - continuous coring of the multilaterals - will get under way in mid 2007, acquiring cores and placing instruments up to the fault zone.
The team has used advanced logging while drilling techniques, collected spot cores and cuttings, and continuously sampled fluids and gases in the drilling mud.
The borehole's steel casing was cemented in so sensitive instruments including seismometers, strainmeters, and fluid and temperature gauges, could be installed.
A suite of fluid sampling, permeability and hydraulic fracturing stress measurements will be made through perforations in the casing.
After each test these will be sealed, with only one left open for fluid pressure monitoring.
Prime drilling contractor ThermaSource drilling engineer Louis Capuano says one of the main challenges was drilling a highly directional hole.
'We had to drill some 500m away from the fault to penetrate it at a depth of about 3km. We faced a stability problem due to uncertainty about what was in the fault. We prepared for the worst and took many safety precautions. It was a slow and expensive process.' The scientists are however very enthusiastic about the new research opportunities opened up by the project.
'We have drilled through a major plate boundary fault for the very first time. We are now sitting at the heart of the earthquake machine, ' says Bill Ellsworth, chief scientist for the US Geological Survey's Earthquake Hazards Team, in Menlo Park, California.
Ellsworth is one of the principal investigators for the SAFOD project, along with Stephen Hickman, a US Geological Survey geophysicist, and Mark Zoback, Professor of geophysics at Stanford University.
Zoback says the project team identifi ed the exact fault trace on 6 October through the deformation of the drill casing on the well. He says the fault moves continuously through creep and small earthquakes at a rate of 20mm to 40mm a year, and the casing is deformed by a few millimetres as a result.
'We have established access to the fault at depth. This allows entirely new science to be done that could not even be conceived of before.'