II. THE PARKFIELD DRILLING PROJECT

As explained above, this is a proposal to conduct a 4.0 km-deep fault zone drilling project into the San Andreas fault near Parkfield, CA. The reasons Parkfield was chosen for this project and an overview of geologic and geophysical knowledge of the area are presented below. The reasons why we are proposing a drilling project that penetrates the San Andreas to only 4.0 km and not the depth at which great earthquakes nucleate (~10 km) are fairly obvious: ultra-deep drilling into an active fault zone is extraordinarily expensive and difficult (see discussion in Hickman et al., 1995b) and before one could justify such a project, it is necessary to demonstrate that drilling and sampling in the fault zone is technically feasible and that the scientific objectives of such a project can be met.

By drilling to 4.0 km, we will reach temperatures of about 135°C (C. Williams, pers. comm., 1998). Although this is less than the critical temperatures of 200-300° C where some of the key questions (summarized above) about mineral stability, chemically assisted deformation mechanisms and fluid pressure compartmentalization might be addressed, and less than the depths at which major earthquakes typically nucleate, the scientific opportunities presented by drilling to 4.0 km are appreciable and allow us to address a number of first-order questions related to fault mechanics:

An important scientific benefit of conducting this project at Parkfield comes from the fact that by working at Parkfield we will be drilling in an area of active creep and microseismicity. If deep drilling into the San Andreas fault zone is eventually done, it is likely that it will be done in an area where the fault produces infrequent, but very large earthquakes (such as the sections that broke in the great 1857 and 1906 earthquakes). These sections of the fault are "locked" in the upper ~10 km during the interseismic period. Thus, by having conducted the pilot project in a creeping area we will have set the stage for trying to understand why some sections of the fault creep and other sections are locked. Further, by drilling in an actively slipping portion of the fault, during the monitoring phase of the experiment we will be able to study the nucleation and rupture processes of microearthquakes with near-field seismic recordings, investigate whether temporal variations of pore pressure occur during fault slip (creep and earthquakes) and study the processes responsible for shear localization.

 

Why Parkfield?

The objectives of this project require a geologic target with clear and attainable scientific goals that will also serve as a fair test of the technology required to drill a deep hole into the fault. Since we first decided to propose drilling a pilot hole, we focused on sites with shallow seismicity, a clear geologic contrast across the fault and good working knowledge of the geological and geophysical environment of the fault.

The requirement for shallow seismicity was key for two reasons. First, we would like to be able to conduct experiments within or very close to seismically active parts of the fault, for the reasons discussed above. Second, we can use the ongoing seismicity to tell us the precise location of the active trace of the fault. In a sense we will use the background seismic activity as "guide stars" to direct the fault zone crossing. To identify potential candidate sites we conducted a systematic search of the strike-slip faults in California, identifying all faults that met the shallow seismicity criteria. To our surprise, this criteria eliminated all candidate faults in southern California. In central and northern California only three fault segments met the criteria for reasonably complete geological and geophysical control. These were the Hayward fault near San Leandro, the San Andreas fault in the Cienega Road to Melendy Ranch region and the Middle Mountain region along the Parkfield segment.

We convened a workshop on the scientific goals, experimental design, and site selection for a ~3-km borehole from July 11-12, 1994, at the USGS in Menlo Park that was attended by about 45 people. Although all three potential sites had unique advantages, it became clear that the Middle Mountain site at Parkfield was the best place to conduct the proposed experiment because:

The latter point is a consequence of the Parkfield Earthquake Prediction Experiment (see Roeloffs and Langbein, 1994). A critical review of this experiment is the subject of the Hager Committee Report (1994) to the National Earthquake Prediction Evaluation Council, in which it was concluded that "Parkfield remains the best identified locale to trap an earthquake." While our goals are more modest, the considered judgment of those involved in the site selection process concurred with this conclusion.

An important new discovery about Parkfield that strongly supports its selection as the drilling target and adds a new scientific dimension to the experiment is the observation that the majority of the earthquakes there repeat in a characteristic manner. Although this was anticipated by us based on more limited studies (Ellsworth, 1996; Nadeau and McEvilly, 1997) have now completed a thorough documentation of this behavior beneath Middle Mountain. The upshot is that we will be targeting specific earthquake source zones with the drill hole, and have a very high expectation that the target earthquakes will repeat numerous times over the lifetime of the experiment.

One point that should not be overlooked is the difficulty of obtaining permission to drill holes of any kind in California. At Middle Mountain, however, the once off-limits southwestern approach to the fault had become accessible due to the inheritance of the land by a cooperative new landowner. Thus, logistical as well as scientific considerations also favored the Parkfield site. We have a signed agreement with the landowner at Parkfield that will allow us to carry out the proposed drilling and downhole measurements and then access the site for 20 years during the fault zone monitoring phase of the experiment.The San Andreas Fault at Parkfield

 

The San Andreas Fault at Parkfield

Geological and Geophysical Setting. The site we have identified for the 4.0-km-deep drilling project is located on the southwest flank of the central portion of Middle Mountain, about 10 km northwest of the town of Parkfield (Figure 6a). This site is located at the northwestern end of the rupture zone for the 1966 M=6 Parkfield earthquake, in the transition between the creeping and locked sections of the San Andreas fault. The San Andreas displays a wide range of behaviors at this site. At the surface, the fault is creeping at a rate of 1.8 cm/year, with most of the fault displacement localized to a zone no more than 10 m wide (Burford and Harsh, 1980; Schulz, 1989). Numerous earthquakes occur directly on the San Andreas fault in the depth interval from about 3 to 12 km (Figure 6b).
Figure 6
Figure 6. Earthquake locations at Parkfield from 1984-1990, togehter with the location of the proposed drillhole. (click for more information)
The shallow seismicity at Parkfield occurs in tight clusters of activity (Nadeau et al, 1994; 1995) that have remained spatially stationary for at least the past 20 years (Figure 7), although uncertainties exist in the exact location of the earthquakes, as discussed below.

An important feature of the microearthquakes beneath Middle Mountain is that they occur in families of repeating events. Individual earthquakes have been observed to recur numerous times using the U.C. Berkeley High Resolution Seismic Network (HRSN), at precisely the same location and with the same magnitude (Nadeau et al., 1994, 1995; Nadeau and McEvilly, 1997). Repeating sources of up to M=2 are located at drillable depths beneath the proposed drill site. Thus, a major goal of this experiment will be to drill as close as possible to one or more of these sources and to follow the build-up of strain and its release through multiple earthquake cycles during the monitoring phase of the experiment. A new-and controversial-prediction for these repeating events is the possibility that they may have localized stress drops in the kilobar range (Nadeau and Johnson, 1998). If confirmed, this will have wide-ranging implications for fault mechanics and the stress/heat-flow paradox.

Almost all events along this fault segment have right-lateral strike-slip focal mechanisms, corresponding to the geologic sense of movement on the fault (Eberhart-Phillips and Michael, 1993). Non-San Andreas type earthquakes close to the San Andreas include strike-slip, normal and reverse faulting mechanisms, the vast majority of which have P or T axis orientations in agreement with a north-south shortening and east-west extension within the fault zone. The few events that locate more than about 5 km from the San Andreas fault, however, commonly have P-axes oriented at a high angle to the fault, which is consistent with the regional framework of fault normal compression and a weak fault (Zoback et al, 1987). Also, recent heat-flow determinations by Sass et al. (1997) in 17 wells located within 10 km of the San Andreas fault near Parkfield-including a 1.6-km-deep well located 12 km from the proposed drill site-confirm previous conclusions (Lachenbruch et al., 1980) that there is no heat flow anomaly associated with the San Andreas fault in central California, indicating that the fault is sliding under low levels of resolved shear stress.

Beginning in the summer of 1994, members of the Site Selection Working Group for the San Andreas pilot project have conducted a number of relatively small-scale and detailed geophysical investigations to fill critical gaps in our knowledge about subsurface structure and microearthquake locations at the Parkfield site. A temporary seismic network was installed on and around Middle Mountain to calibrate crustal structure and study fault-zone guided waves near the proposed drill site (Li et al., 1997). This experiment included deployment of temporary seismic stations (to augment the permanent local networks), three chemical shot points and a 10-station REFTEK array crossing the surface trace of the San Andreas. Additionally, a small-scale seismic reflection survey was conducted by Peter Malin at Middle Mountain along two 2-km-long orthogonal lines west of the San Andreas fault (see Unsworth et al., 1997). The purpose of this survey, which employed a single vibroseis energy source and 128 receivers per line, was to determine the thickness of Tertiary sediments beneath the proposed drilling site. Two magnetotelluric (MT) profiles were also conducted at Middle Mountain to determine the electrical conductivity structure of the fault zone and its surroundings (Unsworth et al., 1997; M. Unsworth., pers. comm., 1998). Finally, Mike Rymer (USGS) has initiated detailed geologic mapping to ascertain the geometry and recency of faulting near the Parkfield drill site.

At Parkfield we have at our disposal some of the most accurately located catalog seismicity available anywhere. In spite of this, absolute locations for shallow seismicity beneath the proposed drill site derived by two independent analyses of Parkfield 3-D velocity structure (Michelini and McEvilly, 1991; Eberhart-Phillips and Michael, 1993; R. Nadeau, written comm., 1996) differ by up to one kilometer. Ellsworth (1996) performed a simultaneous inversion of arrivals from 6 earthquakes and three chemical explosions recorded by the temporary seismic array at Middle Mountain for earthquake locations and velocity structure. His analysis indicates that the shallowest events comprising a recurring cluster of small-magnitude earthquakes (M=1-2) occur at a depth of about 2.8 km below sea level, or 3.5 km beneath the proposed drill site. An independent analysis of these data by R. Nadeau and colleagues at Lawrence Berkeley Laboratory (written comm., 1996) place the depth of these events 300 m shallower. Uncertainties also remain in the epicentral locations of this cluster, which might be directly beneath the surface trace or about 1 km to the southwest. However, given the large (1.6 km) southwest set-back of the drill site from the surface trace of the San Andreas, it is likely that over time, small (M~1-2) earthquakes will be occurring within a few hundred meters, or less, of seismometers emplaced in the borehole (see Figure 9, below). The longitudinal cross-section of seismicity shown in Figure 7
Figure 7
Figure 7. Longitudinal cross-section of earthquake hypocenters. (click for more information)
was prepared to identify an optimal shallow earthquake cluster to use as a drilling target. The dense cluster of shallow seismicity extending from ~2.5 km to 4.0 km depth at the relative horizontal position of 3.5 km was chosen as the drilling target. As discussed below, calibration experiments will be performed while drilling is taking place at shallow depths to assure the accurate location of these earthquakes prior to drilling the inclined section of the well through the fault zone. If necessary, the well trajectory can be modified slightly to penetrate the fault zone at a position passing close to the microearthquake cluster.

The characteristics of the microearthquakes beneath Middle Mountain have been extensively studied by Nadeau and McEvilly (1997), particularly beneath our alternate drilling site located ~3 km south of the primary site (see Figure 7). Their work clearly demonstrates the repetitive nature of the earthquakes within our drilling targets. Recent analysis by R. Nadeau (written comm., 1998) of seismicity located directly beneath the primary site also reveals characteristic earthquake behavior, with recurrence patterns similar to those seen beneath the alternate site. At both locations, waveform correlations were used to show that the same very limited regions on the fault plane fail repeatedly in earthquakes of uniform size. The recurrence intervals between these events are regular, and are best described as a renewal process with memory. An important difference between the two sites is in the magnitudes of the largest repeating events: M~1 at the alternate site and M~2 at the primary site. Thus, we feel confident that either site can meet our science goals, but that the northernmost site offers the best chance for success in that it allows us to target larger-magnitude events.

Figure 8
Figure 8. Geologic map of the Parkfield site, with cross-section. (click for more information)
Figure 8 shows the general geologic setting at Parkfield. The map shown in Figure 8a was derived from Dibblee (1980), as modified through recent geologic mapping by M. Rymer (written. comm., 1996). The cross-section (Figure 8b) was drawn for the alternate drill site utilizing data from seismic reflection lines conducted immediately to the southeast by Unsworth et al. (1997). A similar, but more detailed, geologic and structural cross section will be prepared for the northern (primary) drill site using a 7-km-long high-resolution seismic reflection/refraction profile to be conducted in the Fall of 1998 (see below). Rocks of differing composition and origin, but not necessarily age, are found on opposite sides of the San Andreas fault zone in central California. The presence of such rock types juxtaposed across the fault in the Middle Mountain area results from prolonged post-early Miocene dextral slip of about 320 ± 20 km (Hill and Dibblee, 1953; Crowell, 1981). The two diverse basement terranes of central California are the Salinian block to the southwest of the San Andreas and the Franciscan Complex to the northeast. The Salinian block in the Middle Mountain area consists of Cretaceous granitic rocks (gr) and local metamorphic rocks unconformably overlain by a thin veneer of nearly flat-lying upper Cenozoic marine and non-marine strata and local volcanic rocks. The Salinian block is internally quite stable, with only minor folds locally developed in strata overlying the granitic rocks. The area immediately west of the San Andreas fault (the Gabilan Mesa) was elevated as a rigid block by motion on the San Andreas in late Quaternary time, in large part by slight southwestward tilt (Dibblee, 1980).

In contrast, the Franciscan Complex to the northeast of the fault (KJf; see Figure 8) is more heterogeneous and much more deformed than the Salinian block. Franciscan basement rocks are interpreted as a tectonic melange, consisting of Jurassic and Cretaceous graywacke, sandstone, shale, chert, greenstone and associated mafic rocks, all weakly metamorphosed. The Franciscan Complex is locally unconformably overlain by Cretaceous and Cenozoic marine sedimentary rocks, which also exhibit a high degree of faulting and folding.

The San Andreas fault is the most prominent structural feature in central California and especially in the Middle Mountain region. Comparison of sedimentary rock and their ages exposed near the fault indicate that the main trace has been in the same place for at least the past 2 Ma. Many other faults, active splays and older traces of the San Andreas are present both southwest and northeast of the main trace. In general, the older traces trend slightly more westerly than the main trace, possibly indicating minor tectonic rotations. More abundant faults and a higher degree of shearing are noticeable as one approaches the main trace of the San Andreas fault from either side, even within the relatively little-deformed Salinian block. In fact, at the proposed drill site on the west side of Middle Mountain, numerous faults are interpreted to cut upper Cenozoic and basement rocks. These faults trend subparallel to the San Andreas and are mapped to the southwest of the main trace by as much as 3 km (Figure 8a).

Figure 9 shows an electrical resistivity model of the upper crust through the primary drilling site derived from a recent MT profile conducted by M. Unsworth (pers. comm., 1998); similar results were obtained by Unsworth et al. (1997) through the alternate drill site (see Figure 8a). The resistivity structure at depth clearly defines the position of
Figure 9
Figure 9. Resitivity structure of the primary drill site. (click for more information)
the Salinian block to the southwest of the fault (the high resistivity block at about 1 km depth). This depth is in agreement with the depth of reflectors thought to represent the top of basement on the short reflection lines obtained close to the alternate drilling site (P. Malin., pers. comm, 1996; Unsworth et al., 1997). The MT model shows a ~800 m wide, near-vertical low resistivity zone that is correlative with, but somewhat offset from, the surface trace of the San Andreas fault. This low-resistivity zone presumably results from intense fracturing, hydrothermal alteration and/or high pore pressure within the San Andreas fault zone. The MT model also indicates a moderately resistive Franciscan formation to the northeast of the fault. Note that the P-wave velocity model shown in Figure 6b (after Michael and Eberhart-Phillips, 1991) also defines a near-vertical fault zone at depth, coincident with the background microearthquake locations. In the case of the seismic model shown in Figure 6b, resolution is too poor to define the properties of the San Andreas fault zone itself. The model basically indicates lower velocities east of the fault than west.

There are two sets of microearthquake locations shown in Figure 9 based on locations determined in 3-D velocity models by Michellini and McEvilly (1991) and Eberhardt-Phillips and Michael (1993). Note that the Eberhardt-Phillips and Michael locations are systematically further east than the Michellini and McEvilly locations. The hypocentral locations differ because the velocity models and location procedures used by the two groups differ. While the Eberhardt-Phillips and Michael locations are more consistent with the surface position of the essentially vertical San Andreas, the Michellini and McEvilly locations fall along the abrupt (likely fault-controlled) western edge of the broad fault zone. A detailed calibration experiment while the hole is being drilled at depths of ~1.5 and ~2 km (described below) will determine a more exact location for the hypocenters (see also Nadeau and McEvilly, 1997). Monitoring earthquakes with seismometers in the borehole after it is drilled (also described below) will yield extremely precise earthquake locations to guide subsequent coring operations. Given the width and offset of the electrical anomaly seen in the MT results (Figure 9) as well as uncertainties in absolute earthquake locations, we expect that the active trace (or traces) of the San Andreas fault may be encountered anywhere within the inclined section of the hole.

Anomalous Fault Zone Properties. It is not surprising that the physical properties of crustal rocks on opposite sides of the San Andreas fault should differ (Figures 6b, 8b, and 9), given the vastly different rock types juxtaposed across the San Andreas fault at Middle Mountain. Seismic wave velocities and densities in these rock assemblages differ significantly, leading to readily-observable changes in crustal properties at the fault (e.g. Wesson, 1971; Pavoni, 1973; Ellsworth, 1975; Aki and Lee, 1976; Walter and Mooney, 1982; Michelini and McEvilly, 1991; and Eberhart-Phillips and Michael, 1993; Thurber et al., 1996, 1997).

The lateral transition in crustal properties across the San Andreas fault, however, involves larger changes than can be readily explained on the basis of rock composition alone. The P-wave velocity contrast across the San Andreas fault near Parkfield, for example, is sufficiently strong that laterally refracted P-waves from the faster southwest side characteristically appear as first arrivals at seismographs located near the fault on the slower northeast side (McNally and McEvilly, 1977). Studies of the internal structure of the fault zone using explosion sources reveal a low-velocity fault zone with a width of a few km at most locations along the central San Andreas fault (Healy and Peake, 1975; Feng and McEvilly, 1983; Li et al., 1997) and other actively creeping faults (Mooney and Luetgert, 1982). Seismic P-wave tomography of the fault zone determined using local earthquake sources reveals that the region of anomalously low P-wave velocities extends well into the seismic zone in the Bear Valley/Cienega Valley region further to the north (Lin and Roecker, 1997; Thurber et al., 1997), and thus is believed to reflect unusual physical conditions within the core of the active fault zone. The anomalous nature of the crust is sometimes better expressed in the behavior of the Vp/Vs ratio than it is in either velocity field alone (see Julian et al., 1996, and Thurber et al., 1997). At Parkfield, the fault has an anomalously high Vp/Vs ratio of 1.9 (Michelini and McEvilly, 1991), which is the consequence of a much greater reduction in the S-velocity than in the P-velocity. Even higher Vp/Vs ratios (>2.1) were found by Thurber et al. (1997) within the San Andreas fault zone in the Cienega Valley.

The low seismic wave velocities within the San Andreas fault zone have also been observed more directly though the detection of trapped-mode and interface waves at Parkfield and elsewhere (e.g. Li and Leary, 1990; Li et al., 1990; Ben-Zion and Malin, 1991; Leary and Ben-Zion, 1992; Jongmans and Malin, 1995; Li et al., 1997). Indeed, these waves-which reflect very low velocities in the core of the fault throughout the seismic zone-have been observed even along faults that have no large-scale difference in crustal structure across them, and thus must reflect a characteristic property of active fault zones (Hough et al., 1994; Li et al., 1994). The width of the low-velocity fault zone is typically only a few hundred of meters or narrower.

Other crustal properties, notably electrical resistivity, correlate well with the observed velocity anomalies at the fault zone (e.g. Eberhart-Phillips et al., 1990, 1995). In addition to the major conductivity anomaly coincident with the San Andreas fault at Middle Mountain (Figure 9), fault zone guided wave studies conducted as part of the site characterization effort at Parkfield (discussed above) provide clear evidence for a strong low velocity anomaly within the fault zone at the proposed drill site (Li et al., 1997). This anomalous zone beneath Middle Mountain also has a profound effect on the propagation of S-waves, splitting them because of fault-parallel anisotropy in the shear wave velocity structure (Jongmans and Malin, 1995). By combining electrical and seismic techniques, it should be possible to place much tighter constraints on the nature of the anomalous fault zone properties than can be achieved with any single method (Eberhart-Phillips et al., 1995).

The coincidence of anomalously low seismic velocities within the fault zone with strong wave attenuation within it (Blakeslee et al., 1989) and the region of high electrical conductivity provide strong evidence for unusual conditions within the fault zone at Middle Mountain. Elevated fluid pressure within the core of the active fault zone (as proposed by Rice, 1992, Byerlee, 1990, Sleep and Blanpied, 1992, and others) represents one plausible explanations for these anomalies. Alternatively, these anomalies might merely represent a zone of intense fracturing and hydrothermal alteration at near-hydrostatic fluid pressures. One of the goals of the proposed drilling experiment at Parkfield is to calibrate these surface-based observations against in-situ measurements and core studies. This calibration, in turn, will allow for more reliable inferences of fault-zone properties and physical state from geophysical profiles elsewhere along the San Andreas fault and along faults in other tectonic environments.

 

The Expected M ~ 6 Earthquake

The site we have selected also lies virtually at the epicenter of the 1966 Parkfield earthquake (Figure 6). Thus, one final question about the seismicity at Parkfield deserves some discussion: What would happen should the anticipated Parkfield mainshock occur either before or after the drilling experiment is completed? Should the earthquake occur before the hole is either drilled or completed, we can anticipate an enhanced production rate of shallow earthquakes as part of the aftershock sequence, which would add to the return of the fault zone monitoring stage of the experiment.

Should the Parkfield mainshock occur after the hole is completed, it is likely that the coseismic displacement of the earthquake would extend through our fault crossings. This presents the possibility that we might observe the nucleation and initial rupture propagation of a M=6 earthquake at close range. While this is not an earthquake prediction experiment, the opportunity to make observations of seismicity, pore pressure, deformation and temperature directly within a fault zone preceding and during a moderate earthquake is a truly unique opportunity. If the M=6 Parkfield earthquake should occur during the lifetime of the experiment, we would make unique observations not only of preparatory fault zone processes but also of the dynamics of rupture propagation and the energetics of large-scale faulting.

As noted above, the dominant pattern is spatial stationarity of the seismicity (clusters), even through two Parkfield earthquake cycles (Nadeau et al, 1994; Cole and Ellsworth, 1995). The spatial stationarity of seismicity through other mainshock events elsewhere in the San Andreas fault system is the norm, and has been well-documented in many cases (e.g. Ellsworth, 1975). Furthermore, it is now known that repeating earthquake sources can be triggered into rapid repetition as aftershocks, when located near or within a mainshock rupture zone (Vidale et al., 1994; Ellsworth 1995). Thus, should the M~6 Parkfield earthquake occur during the lifetime of the experiment, we might also have an unparalleled opportunity to observe time-dependent loading and frictional behavior in the near field (Marone et al., 1995).

 

Fault Zone Heterogeneity

One of the most significant problems that could be encountered (and an often-cited criticism of the proposed experiment), involves the degree and scale of fault zone heterogeneities. For example, one could argue that essentially everything we propose to sample and measure varies so markedly across the fault that interpretation of the results will be impossible. This might be called the "small-scale" heterogeneity problem. As we have no data on the composition and properties of active faults at depth, we do not know if this problem exists, nor will we be able to assess its severity, without making the types of observations proposed. In fact, this is one motivation for conducting a pilot project before even considering deep drilling in the fault zone. With the proposed hole, we would be drilling into a very well-developed section of the fault which has likely accommodated ~100 km of relative plate movement. The observations of Chester et al. (1993) on outcrop exposures of what appears to be a previously active strand of the San Andreas suggest a very well-ordered picture of rock deformation, with increasingly deformed rocks as one progresses toward the "core" of the active fault zone. Whether or not this is also true of the San Andreas at Parkfield, the bottom line is that we simply will not know if fault zone composition, structures and properties are extremely complex, or chaotic, without conducting an experiment such as the one proposed. Should we discover that the structure, composition and properties of the active fault zone are incomprehensibly complicated and chaotic, while it would likely indicate that deep drilling is impractical, this finding would be a significant scientific discovery in its own right.

Another problem we face might be referred to as the "large-scale" heterogeneity problem. One could argue that even a profile of measurements across the fault zone is still a "point" observation, the significance of which may be difficult to assess in terms of overall fault behavior. While it is almost impossible to counter such arguments (one can always argue that regardless of what we find, conditions would be different elsewhere) the point of conducting the proposed experiment is that there are currently no unequivocal constraints on fault zone composition, properties or state at depth. Thus, whatever is learned will be a dramatic improvement in our ability to constrain existing hypotheses about fault zone processes as well as provide critically-needed constraints for future laboratory and theoretical studies. We propose to conduct the proposed experiment along a section of the fault whose behavior is well known, which will make interpretation of the results obtained as straightforward as possible. Moreover, by making geophysical measurements of fault zone processes on the core, in the hole and from the surface to the hole we will be able to extrapolate our observations away from the borehole.

Back to Top | Back to Table of Contents | Back to Section I | Proceed to Section III