IV. SCIENCE WORK PLANS: INTEGRATED SCIENCE TEAMS

In this Section we describe how the general scientific objectives discussed above will be met by the principal investigators who are part of the proposed experiment. Table I is a list of these scientists. As mentioned in the Foreword, should this project come to fruition, we fully expect many more scientists to become involved in project. Our overall approach to scientific management has been to create integrated teams of scientists who will address the key programmatic areas. The following teams have been assembled to assure that all key aspects of the science program are met:

We summarize the work plans of these different teams of investigators below and present summary budgets for scientists from U.S. universities in Section VI, along with budgets for drilling, coring and rig-related activities. Detailed work plans and budgets of the scientists seeking funding from NSF will be submitted as separate "stand alone" proposals to NSF, except for Mark Zoback, whose budget is included in this proposal. The USGS and DOE scientists (as well as those from other countries) will provide detailed work plans and budgets to their respective funding agencies.

 

Downhole Measurements

The downhole measurements team consists of:

Figure 13
Figure 13. Overview of the downhole measurement program as a function of measured hole depth. (click for more information)
As shown in Figure 13, the four principal types of downhole measurements to be conducted in the Parkfield hole are as follows.

Logging While Drilling. While drilling through the fault zone, measurements while drilling (MWD) and logging while drilling (LWD) will be carried by instrumentation located just above the bit. The MWD data are needed to assist the directional drilling operations and will start at a shallower depth to help the drillers "build angle" properly. The LWD measurements are to provide real-time measurements of geophysical properties. We plan to have three groups of measurements made, that are provided by Anadrill/Schlumberger or possibly other companies. The ISONIC tool will provide sonic velocity and natural gamma information, the ADN tool will provide density and neutron porosity information and the RAB tool will provide a resistivity image. A detailed list of the LWD measurements to be made (and corresponding costs) is given in Appendix B.

The logic for running LWD tools is twofold. First, if hole conditions are so bad that open-hole geophysical logging (described immediately below) is impossible to carry out, or severely restricted due to hole conditions, the LWD program will insure that a continuous profile of geophysical measurements are made through the fault zone prior to running casing. A possibly analogous case was encountered during ODP drilling through the active decollement in the Barbados subduction zone (Moore et al., 1995a). While open hole geophysical logging was impossible due to the extremely poor hole conditions, much was learned about the location and properties of the decollement from the LWD program. The other reason for conducting logging while drilling is to identify zones with anomalous composition or properties as they are being encountered. For example, if an overpressured zone is suddenly penetrated by the drill bit it should be indicated by anomalously high porosity, low sonic velocity and low resistivity. This is important to know while drilling is taking place, both for scientific and safety reasons (e.g., as an indication that the mud weight needs to be increased). The research proposed by David Goldberg and Gerardo Iturrino is to provide a real-time geophysical interpretation of these data with a particular emphasis on identifying what might turn out to ephemeral fault zone properties. Goldberg and Iturrino have appreciable experience with such measurements in the ODP program.

Open Hole Geophysical Logging. The open-hole geophysical logging program outlined in Figure 13 will be conducted in 3 phases that are coordinated with the drilling and casing program (Figure 12). An initial suite of logs will be run before casing the vertical hole being drilled to a depth of 2 km. A second suite of logs will be run before casing the hole at a true vertical depth (TVD) of ~ 3 km, just as the broad fault zone is being entered. A final suite of logs will be run when the well has been drilled to completion at a TVD of 4 km.

The logs to be run are intended to provide a comprehensive profile of the composition and properties of the fault zone. The resistivity (AIT), density (LDS), porosity (CNL), sonic (DSI) and geochemical (GLT) measurements to be made by Schlumberger will provide the information needed to characterize composition and physical properties. Renate Pechnig and Jürgen Wohlenberg will carry out a detailed analysis of the logs for reconstruction of a comprehensive lithologic profile. Pechnig and Wohlenberg carried out similar studies as part of the KTB project in Germany. Their proposed studies include a comprehensive check of log quality, comprehensive analysis of logs (closely correlating geophysical logs with cuttings and core) and a reconstruction of lithology. The geophysical logs are also of appreciable interest to the scientists carrying out the laboratory physical property measurements on exhumed samples and the fault zone characterization studies (described below). Ryuji Ikeda, Yoshihisa Iio and Kentaro Omura will analyze the logs to determine physical properties, especially the physical properties of fault zones. They have been recently conducting similar studies in shallow holes drilled through recently active faults in Japan. Mark Zoback and Steve Hickman will analyze the wellbore imaging data to characterize fractures and faults encountered in the borehole and to utilize these data as an integral part of the stress measurement program, as described in detail below. Zoback proposes to support two graduate students in this effort.

Colin Williams, Art Lachenbruch and John Sass will conduct a series of detailed temperature measurements (and corresponding thermal conductivity and heat production measurements on samples) to thoroughly characterize conductive and convective thermal regimes of the fault zone and surrounding crust. They will derive a profile of heat flow throughout the total depth of the hole and integrate the heat flow results with post-drilling temperature measurements in an effort to detect transient thermal signals associated with fluid flow and mass transport.

A detailed list of the open- and cased-hole geophysical measurements to be made (and the corresponding costs) is given in Appendix B.

Cased-Hole Logging. After the hole is cased and cemented three logs will be run (see Figure 13 and Appendix B). First will be an extremely precise well trajectory using a gyroscopic directional survey (GDS) log. Then, a cement bond log (CBT) and ultrasonic cement imager (USI) log will be run to assure that the cement has filled the annulus between the casing and borehole and that the cement and casing are well bonded. Effective cementing is required both for maintaining hole integrity over time and for a number of aspects of the science program, especially the measurements of stress, pore pressure, permeability and fluid sampling described in the next Section.

These measurements will be repeated mid-way through the monitoring program (about 1 year after the first phase of drilling is complete) and then again just before the coring program commences (about 2 years after the first phase of drilling is complete). In conjunction with analyses of seismic data collected by the monitoring array and the site characterization experiments (see below), these repeat logs will be used to help identify portions of the fault zone that are actively deforming and, hence, suitable for continuous coring in the second phase of the drilling program.

Stress, Pore Pressure and Permeability Measurements and Fluid Sampling. One of the most important aspects of the downhole measurement program involves the in situ stress, pore pressure and permeability measurements coupled with fluid sampling. The goal of these experiments is to conduct a series of measurements at different depths and positions with respect to the fault zone. Mark Zoback and Steve Hickman will be responsible for these tests.

The experimental procedures will start with a series of packer tests. After the borehole is cased and cemented at 0.3, 2.0 and 3.0 km, a 20 m section of hole will be drilled below the casing that will be used first for conducting a drill stem test (DST). A packer will be set in the casing and a valve will be opened to allow flow into the partially-evacuated drill pipe. The pressure build-up during the DST will enable us to determine pore pressure and permeability using standard well test procedures (e.g., Horne, 1996) and to obtain uncontaminated fluid samples. After the DST, each of the open hole sections will be hydraulically fractured to determine the magnitude of the least principal stress. This value can be determined with little ambiguity during a hydraulic fracture measurement (Zoback and Haimson, 1982) as minimum energy considerations cause the hydrofrac to propagate away from the well perpendicular to the least principal stress, regardless of the material properties of the formation being hydraulically fractured (Hubbert and Willis, 1954).

After conducting geophysical logging, open-hole hydraulic fracturing tests will also be attempted in the Salinian granite section of the well between 2.0 and 3.0 km depth using inflatable straddle packers. As was the case with the fractured granitic rock in the Cajon Pass borehole (Zoback and Healy, 1992), it is likely that we will be able to determine the least principal stress in these measurements. Finally, to make in situ stress and hydrologic measurements at different positions within the fault zone (as illustrated in Figure 10), drill stem tests and hydrofrac measurements will be made through perforations in the cemented casing using procedures common in the petroleum industry. The basic procedure for these tests is described in Section V. These tests will result in a complete profile of least principal stress, pore pressure and permeability measurements accross the fault zone to address the questions raised above. It will also result in a profile of relatively uncontaminated fluid samples (by "spiking" the drilling mud with a stable tracer such as fluorescene, the degree of pore fluid contamination by drilling mud can be easily determined).

While a profile of pore pressure and least principal stress measurements is sufficient to test the Rice (1992) hypothesis (Figures 2a, 10), we will also attempt to determine the full stress tensor along the entire well path. To accomplish this, we will utilize a series of techniques that Zoback and his colleagues have developed which combine knowledge of the least principal stress, pore pressure, vertical stress and observations of compressive, shear and tensile wellbore failures observed in borehole image logs (e.g., Moos and Zoback, 1990; Peska and Zoback, 1995; Zoback et al., 1993; Barton and Zoback, 1994; Brudy et al. 1997; Wiprut et al., 1997). The image logs we will be using for this purpose (see below) are the ultrasonic borehole imager (UBI) and electrical formation micro-imager (FMI). Such observations often make it possible to constrain the orientation and the magnitude of the maximum horizontal principal stress. As explained by Peska and Zoback (1995) and Zoback and Peska (1995) these techniques are especially effective in deviated wells. A detailed methodology and comprehensive suite of software routines known as SFIB (Stress and Failure of Inclined Boreholes) was developed by Peska and Zoback to accomplish this. SFIB is currently being widely used in the petroleum industry for this purpose. One technical note is that it is not necessary to assume that the principal stresses are in a horizontal and vertical plane. Still another method that will be used to assess the complete stress tensor is the measurement of rotations of breakouts (imaged with the UBI) and resulting from localized stress anomalies caused by slip on small faults penetrated by the hole (e.g., Barton and Zoback, 1994; Shamir and Zoback, 1992).

Hickman also has extensive experience making in situ stress and hydrologic measurements (Hickman et al., 1984; Plumb and Hickman, 1985; Hickman et al., 1985, 1988; Hickman and Rojstaczer, 1995; Hickman et al., 1997). Moreover, Hickman and Zoback have a long history of working successfully together on projects such as this (e.g., Zoback et al., 1980; Zoback and Hickman, 1982; Hickman and Zoback, 1983). In fact, the two have been recently conducting a comprehensive suite of stress and hydrologic measurements in an active fault zone in Dixie Valley, Nevada, using techniques that are similar in many ways to the measurements proposed in this study (Hickman et al., 1997, 1988; Barton et al., 1988; Morin et al., 1998).

A final set of permeability measurements will be attempted after the four continuous core holes are drilled. While these holes are being cored there will be packer in the main borehole that facilitates the coring process. Because this packer will seal off the main borehole, it should be possible to do a bulk permeability test of each of the cored intervals.

The direct costs associated with these measurements are principally those associated with rig expenses. The time required for these tests and the costs of packers, perforations, cementing, etc., are summarized in Appendix A.

 

Measurements on Core, Cuttings and Fluids

The team carrying the out broad suite of measurements on core, cuttings and fluids consists of:

This team is proposing 12 projects to study the mineralogy, deformation microstructures, physical properties, and constitutive behavior of rock samples recovered from the San Andreas zone and country rock at depth. They are also proposing 3 projects aimed at studying the chemical and isotopic composition of fault zone fluids and gases.

Mineralogy and Deformation Microstructures. Characterization of deformation mechanisms operating within the fault zone and the role of fluid-rock interactions in controlling fault zone rheology is being addressed by two closely coordinated proposals, working primarily at the mesoscopic (Chester et al.) and microscopic (Hacker et al.) scales. Chester, Evans, and Kirschner will be the primary research group responsible for ascertaining the overall structures and geochemistry of the cuttings and cores, and will be at the drill site during most of the coring operations. They will characterize the mesoscopic and microscopic structures (fractures, faults, cataclasites, microfractures, veins), mineralogy, petrology, major and trace element geochemistry, and stable isotope geochemistry by means of optical microscopy, X-ray fluorescence (XRF) and diffraction (XRD), and other techniques. Their observations and data will be used to determine the nature and extent of structural and mineralogical zonation across the fault; the mechanisms of deformation, recovery and shear localization; the extent of fluid-rock interactions in and across the fault; and questions related to open- versus closed-system behavior. Hacker, Boles and Chen will conduct a complimentary suite of microstructural analyses that include scanning electron microscopy (SEM), electron-microprobe, XRD and transmission electron microscopy (TEM) observations of solid phases at the grain and sub-grain scale; secondary ion mass spectrometry (SIMS) observations on fluid inclusions; and uranium-series dating of carbonate veins. These observations will be used to identify the nature and timing of diagenetic reactions and associated changes in fluid pressure, volume, and chemistry. In addition, their results will determine the grain-scale mechanisms controlling deformation during the interseismic and coseismic periods.

Two projects are focused specifically on crack sealing, solution-transport deformation and the geochemistry of vein-filling minerals. In the first of these projects, Kastner will conduct detailed analyses of the mineralogy, major and minor element chemistry, and isotopic composition of carbonate vein-filling minerals within and adjacent to the fault zone. This study will involve optical microscopy, cathodoluminescence (CT), SEM, XRD, and microprobe analyses of the geometry, chemical zonation, and timing of vein-filling episodes, augmented by mass spectrometry (MS) and inductively coupled plasma (ICP) atomic emission spectrometry for minor and trace-element geochemistry and isotopic ratios. Kastner will use U-Th-Radium techniques to provide high-resolution dating of carbonate veins ranging from 350,000 years to a few years in age and will examine P/T relationships, bulk chemistry and isotopes for both fluid and solid inclusions. In the second proposal, to be submitted to the French government, Boullier, Gratier, Pili and Sheppard will perform optical, SEM and CT observations of crack healing structures; microprobe, infrared spectroscopy (IR) and bulk chemical characterization of sealing minerals; fluid inclusion studies; stable isotope studies of sealing products and fluid inclusions; and laboratory experiments on the kinetics of dissolution/precipitation reactions under stress. This project is highly complimentary to the Kastner proposal, in that it deals with the kinetics of fluid-rock interactions, the interpretation of pressure-solution microstructures (stylolites, indented grains, etc.), and analyses of fluid inclusions. Taken together, these two projects will provide critical constraints on the nature and origin of fluids associated with the San Andreas fault zone, their physical and chemical roles in the faulting process, and the timing of fluid pressure transients and their possible relationship to the earthquake cycle.

The presence and relative abundances of hydrated clay phases (smectite, illite/smectite mixed-layer clay, and chlorite-smectite mixed layer clay) and their possible impact on fluid pressure generation and the seismic-to-aseismic transition will be the focus of a coordinated pair of proposals by Underwood and Brown. Using XRD and SEM/EDS techniques, Underwood will determine the mineralogy and chemistry of clays found in fault gouges and variously deformed and altered wall rocks collected from the Parkfield drill hole (cuttings and core) and from surface outcrops along different transects of the San Andreas fault system in central California. Brown will then conduct laboratory experiments on these samples at realistic in-situ temperatures and fluid chemistries to determine the stability of these clay phases, the hydration state of any smectites present, and the kinetics and physical manifestations of clay-mineral hydration/dehydration reactions. These experiments will be critical in addressing long-standing controversies regarding the presence or absence of smectites along the San Andreas fault zone at depths greater than a few kilometers and the effect that this might have on frictional properties (e.g., Figure 3).

A broad range of geochemical and microstructural investigations on cores and cuttings are being proposed by nine different investigators from Japan, headed by H. Tanaka (Ehime Univ.). Members of this project, who will be requesting funding from several different funding agencies in Japan, will conduct mineralogical and geochemical investigations (optical, SEM/EDS, XRD) and microstructrual observations to determine the mode and extent of hydrothermal alteration and deformation in the fault and wall rocks, the relative rates and pathways of fluid exchange within the fault zone, and the formation temperatures of fault rocks. One important aspect of this group's involvement in the Parkfield project is their experience in analyzing and interpreting cores from shallower fault-zone drilling experiments already conducted or underway in Japan.

Physical Properties and Constitutive Behavior. The physical properties and deformational behavior of fault-zone materials and country rock will be determined in several different rock physics laboratories in the United States and abroad. In most cases, measurements will be made on intact core samples at stresses, pore pressures, temperatures and pore fluid chemistries appropriate to the conditions from which the samples were recovered.

Christensen and Wang will be making acoustic measurements on cores at elevated confining pressure, fluid pressure and temperature to determine P- and S-wave velocities, attenuation, elastic constants and the magnitude and orientation of seismic anisotropy. Petrofabric observations will be used to differentiate between mineralogical and structural controls on seismic anisotropy, if it exists. This team will also conduct differential strain analysis to identify the geometry and impact of microcracking induced during core recovery and to relate these effects to the in-situ stresses. Finally, they will conduct extensive measurements of poroelastic properties at realistic in-situ conditions to help understand fluid pressure changes induced during seismically induced stress changes. These results will be "scaled up" through integration with physical properties determined from wireline logging, LWD, drill stem tests and surface-based geophysical imaging to provide critical input to structural and mechanical models for the San Andreas fault system.

Measurements of the frictional and fluid-flow properties of recovered core samples will be conducted by two groups, using complimentary state-of-the-art rock mechanics facilities at Brown University and the USGS. Both groups will obtain data at room temperatures and at realistic in situ conditions of temperature, normal stress, fluid pressure and pore fluid chemistry. Tullis and Goldsby will use a unique rotary shear apparatus to determine the frictional constitutive relations and permeability of fault rocks and gouge as a function of shear displacement, normal stress and direction of fluid flow, to a total displacement of up to several meters. These experiments will include velocity stepping and slide-hold-slide tests to determine important parameters in the rate- and state-dependent friction laws. Slip velocities will range from 0.001 mm/s up to 3 mm/s, the latter to test for possible weakening effects associated with rapid slip and shear heating. Lockner, Moore and Morrow (of the USGS) will use high-pressure and temperature triaxial deformation apparatuses to determine a broad range of rock strength and hydraulic parameters, including intact rock strength, frictional sliding strength, matrix permeability and complex electrical resistivity. They will also investigate pore pressure changes due to inelastic processes occurring during shearing and normal compaction, to better understand the physics of rupture propagation and fault interaction. Sliding experiments will be carried out to total displacements of about 1 cm, at rates ranging from 0.01 mm/s to 32 mm/yr. At these lowest rates, which are comparable to the long-term slip rate on the San Andreas fault, Lockner et al. will be able to study time-dependent processes associated with fault creep and long-term weakening. Both the Tullis and Goldsby and Lockner et al. groups will be conducting petrographic studies using SEM, TEM and optical microscopy to identify deformation microstructures, mineral assemblages and microfabrics before and after deformation.

In a related study, Rutter and Faulkner (University of Manchester, England) are proposing laboratory permeability measurements on recovered fault gouge as a function of core orientation, confining pressure, pore pressure, differential stress and temperatures up to 250° C. Their work is specifically focused on understanding the possible roles of low-permeability clay-rich fault gouge in controlling permeability anisotropy and the entrapment of fluids under high pressures in fault zones. Measurements will be conducted using both water and argon as the pore fluid, to determine the effects of different pore fluids on permeability. A key component of this study is comparison of the results obtained with similar measurements made by these investigators on surface samples from fault zones in southeast Spain.

Detailed knowledge of rock strength and failure mode is needed for the estimation of in-situ stress magnitudes using borehole breakouts. Toward this end, Haimson will be making measurements of rock strength and stress-strain behavior on cores recovered from the Parkfield hole. Haimson will use his unique polyaxial deformation apparatus, allowing him to investigate rock deformation and failure under true trial conditions (i.e., where the three principal stresses are not necessarily equal in magnitude), thereby better replicating the in-situ stress state. Experiments will be conducted under dry and wet conditions and at temperatures up to the in-situ temperature. These experiments will also be important in constraining the onset of dilatancy under triaxial stress conditions-an important parameter in some hydromechanical earthquake models.

Fluid Chemistry. Three teams of investigators will be conducting chemical and isotopic analyses of pore waters and gases. The first team, Kharaka and Kennedy (of the USGS and LBL, respectively) will determine the geochemistry of bulk water and gases encountered during drilling of the Parkfield hole, by measuring the concentrations of selected major, minor and trace elements as well as stable and radiogenic isotopes (e.g., 13C, 18O, 34S and 87Sr; 3/4He; 3H, 14C, 36Cl and 129I). Fluid samples will be obtained during drill stem tests (see below) using flow-through and closed down-hole samplers. Drilling fluids will be tagged to ascertain the extent of formation infiltration during drilling, so that these results can be extrapolated to "pristine" pore fluid conditions. The objectives of this study are to ascertain the origins, pathways and transport rates of pore water and solutes; the nature of water-mineral interactions within and adjacent to the fault zone; and the role of volatiles (CO2, CH4, N2 and noble gases) from mantle and deep crustal sources in the faulting processes.

Torgersen, Stute and Schlosser will use rare gas dating techniques on cores to determine the magnitude, direction and rates of fluid transport within the San Andreas fault zone. This involves measuring the concentrations of radiogenic elements (U, Th, K, Li) in the solid phases and the concentrations of rare gas daughter products (isotopes of He, Ne, Ar, Kr and Xe) in both the solid phases and the pore fluid. Results will be used to determine the apparent ages of the pore fluid and the enclosing rock and compared with model calculations to estimate the length/time scales of fluid-rock interaction. In addition to spiking the drilling fluids (see above), this team will employ secondary coring and vacuum packaging operations at the drill site to obtain relatively undisturbed samples of the matrix pore fluid. Step-heating experiments and fluid-flow modeling will also be performed to determine the rates of diffusive exchange of rare gasses between the solid and fluid phases.

Real-time analysis of gases dissolved in the drilling mud will be carried out by Jörg Erzinger of GeoForschungZentrum, Potsdam, who carried out similar studies in the KTB pilot hole and main borehole. The system he will use has the capability to do both automated measurements and automated sampling for subsequent analysis. The gases from a mud degasser will be run into an automatic gas mass spectrometer and gas chromatograph and quantitatively analyzed for N2, O2, Ar, He, CO2, H2S, SO2, CH4, C2H6, C3H8 and C4H10. A radon spectrometer will be used to detect 222Rn and 220Rn. Known quantities of pure and mixed gases are added to the mud system before being circulated into the hole for calibration purposes. Erzinger will also make a suite of geochemical measurements on the discrete samples collected through perforations. As mentioned below in the context of collecting, describing and archiving the core and cuttings, a second geologist is being contracted for through the mud logging company. After training by Erzinger, these geologists will be responsible for operation and maintenance of the automated gas analysis system. Because gases will be introduced into the mud column throughout the open-hole interval, this apparatus cannot substitute for the discrete spot sampling and analyses proposed by Kharaka and Kennedy, Torgersen et al. and Erzinger. However, the proposed real-time study will provide critical samples and analyses of ephemeral gas/fluid pockets penetrated during drilling that might otherwise escape unnoticed, and will provide essential guidance for decisions related to later fluid sampling and in-situ hydrologic testing.

Core and Cuttings Handling, On-Site Analysis and Sampling Protocol. As outlined previously, both drill cuttings and core will be acquired from the proposed hole. Geologists working for a commercial mud logging company will continuously monitor and record changes in cuttings mineralogy, mud chemistry, and drilling parameters (penetration rate, torque, pump pressure, etc.) during both rotary and core drilling and will bag and label cuttings for later analyses by interested investigators.

Three different types of core samples will be acquired during drilling. Eight spot cores will be collected during the main (rotary) drilling phase of the experiment: three in the granite country rock and five in the fault zone. These cores will be approximately 10-m-long and range in diameter from 12 to 17 cm. These spot cores will be supplemented by approximately 150 sidewall cores acquired below a depth of 2 km using a wireline-deployed coring tool (see Figure 13). These sidewall cores will be 1.9 cm in diameter and 5.1 cm long. Finally, at the conclusion of the short-term monitoring phase of this experiment, four 250-m-long continuous core holes will be drilled off of the main hole (see Figure 12), providing the bulk of the core to be used in the laboratory analyses described above. The diameter of these cores depends on the diameter of the casing off of which these sidetracks are drilled, and will be 6.4, 6.7 or 10.2 cm (the smallest core size would be necessitated by use of the 13 cm contingency casing string during completion of the main rotary hole through the fault zone). Both the spot cores and the continuous cores will be oriented by comparison of fractures and other features in the core with FMI and UBI logs (see Downhole Measurements Section).

Core handling and processing will utilize a newly refurbished mobile core lab and associated equipment to be supplied by the USGS Core Research Center in Denver (Figure 14).

Figure 14
Figure 14. Mobile core lab to be used at Parkfield site. (click for more information)

Much of the routine processing of spot, sidewall and continuous core will be performed by two people-a geologist and a technician-to be hired under the drilling contract (see attached Drilling Program and Cost Estimates), with on-site assistance from principal investigators associated with the Core and Cuttings Team. As illustrated in Figure 15, this processing will include cleaning, reorienting, and labeling the core; generating preliminary petrographic descriptions; photographing and scanning the core; and boxing the core for long-term storage at the Core Research Center. Core will be scanned using a digital color core scanner developed for the German KTB project. This scanner rotates the core over a scanning head, producing an "unwrapped" 360° image of the entire outer surface of the core. These images, along with drilling information and other data acquired during core processing, will be entered into a computer data base developed especially for this purpose by the International Continental Drilling Program (ICDP). This system-the Digital Information System (DIS)-and all other components of the mobile core lab shown in Figure 14, have been extensively repaired and updated for use in the Long Valley Exploratory Well during the summer of 1998, and should be thoroughly operational before they are needed for the Parkfield project. We have been in close contact with John Sass and Vicki McConnell, project leader and scientist in charge of core analyses, respectively, for the Long Valley project, to make sure that the sample handling and data archiving for the Parkfield drilling project will be as efficient and reliable as possible.

Figure 15
Figure 15. Protocol flow chart to be used on core and cuttings samples from the Parkfield hole. (click for more information)
In addition to this routine core processing, we anticipate that several members of the science team will be on site during drilling to prepare detailed petrographic and mineralogical descriptions of the core; prepare core and cuttings samples for later laboratory analyses; and conduct XRD studies on selected drill cuttings and rock flour. These investigators will be assisted by the contract geologist and technician, who will slab the core samples, scan the slabbed core, help prepare thin sections, and perform other tasks as needed. For example, co-PIs from the Chester et al. proposal will be on-site throughout the continuous coring operation to perform detailed geological and geochemical logging of the core as it comes out of the hole. They will be assisted in this effort by Jeff Hulen (Energy Research Institute, Univ. Utah), who has considerable experience in working with core recovered from scientific drill holes. Lodging (i.e., trailers) will be provided at the drill site free-of-charge for visiting scientists, which can accommodate up to four people at one time.

In order to meet the sample needs of the current science team, yet retain a sufficient quantity of core for later analyses by these and other investigators, it is clearly essential that we develop a careful yet responsive sampling protocol. Once the Parkfield project is underway, we will establish a sampling committee to review and evaluate requests for core, cuttings and fluid samples. The sampling procedures and protocol we employ will be based upon our experience with other large, international drilling projects (e.g., the KTB project) and will be developed in close consultation with the ICDP.

 

Geophysical and Geological Site Characterization

The site characterization team consists of:

We have assembled a science team divided into nine separate projects that will perform a suite of complementary investigations of the geophysical and geological environment of the San Andreas fault zone surrounding the borehole. These investigations have been selected to provide a high-resolution image of the earthquakes and physical environment of the fault zone and upper crust beneath the Middle Mountain drill site. Overall coordination between the projects, and scheduling of the field work will be provided by Bill Ellsworth. Figure 16

Figure 16
Figure 16. Timetable for conducting geological and geophysicals site characterization studies. (click for more information)

shows the relative timetable and phasing of the site characterization studies, which has been designed to maximize the coordination between experiments. Figure 17 shows the overall site coverage for the proposed experiments. Because our primary objective in these experiments is to image the crustal volume containing the drill hole, there is a high degree of overlap between the experiments and methods. Since each of the methods complements the others, the overlap will contribute to the development of a more complete and robust image of the subsurface structure.

These nine projects will address two distinct objectives. The first is to provide technical information that is critical to the siting and drilling of the borehole. The principal siting questions we need to resolve are 1) the location of the repeating earthquake cluster we will attempt to intersect with the borehole, and 2) the presence of shallow geologic complications at the drillsite that could be avoided or mitigated by a slight resiting of the drillsite. Our second objective is to create a comprehensive structural and physical
Figure 17
Figure 17. Schematic map of site characterization experiments. (click for more information)
model of the fault zone. Detailed knowledge of the physical environment of the fault will be essential to the interpretation of structures and properties observed in the borehole.

Although locations of earthquakes beneath Middle Mountain have been studied using fully 3-D velocity models obtained from local earthquake seismic tomography (e.g. Michelini and McEvilly, 1991; and Eberhart-Phillips and Michael, 1993), there are significant discrepancies between these results (Figure 9). The best estimates of the horizontal position of the targeted earthquakes at the drillsite differ by nearly 1 km between these two studies. We must resolve the true location of the target repeating earthquakes with an uncertainty of 100-200 m before deviating the borehole toward the fault if we are to have a high probability of hitting them.

We propose to resolve the absolute location problem using several methods while the vertical section of the hole is being drilled. McEvilly, Johnson and Rector propose to use the noise generated by the drill bit to measure the absolute traveltime to their network of permanent borehole seismic stations (High Resolution Seismic Network, or HRSN). By using the drill bit as a source, they will be able to determine absolute traveltimes from along the borehole to the HRSN stations, and thereby calibrate the 3-D structure to determine the absolute earthquake locations. Because of its long operational history, ultra-low noise recording environment and high sensitivity, the HRSN provides a key component in our strategy to resolve the absolute location of the target repeating earthquake cluster, and to monitor the evolution of the events with the cluster. Because the HRSN is currently not ideally configured for locating earthquakes in the target zone, McEvilly et al. propose to add two new borehole stations to the network. Replacement of the now obsolete and 13-year-old 16-bit digitizers with modern 24-bit digitizers is also requested, as we cannot afford to have the HRSN fail during the pre-drilling observation period or during the traveltime calibration experiments. McEvilly et al. also propose to deploy a linear array across the San Andreas fault and above the target cluster to image microearthquakes by migration.

Roecker and Thurber propose to solve the absolute location problem using a two-step approach. The first step will consist of a monitoring experiment, beginning 6-12 months before drilling, to collect traveltime data from shallow events in and near the target zone on a tight network of temporary (and permanent) stations surrounding the borehole on Middle Mountain. These data, together with existing explosion and earthquake data will be analyzed to construct candidate 3-D structures and earthquake locations. The second step will be to conduct traveltime calibration experiments when the vertical section of the borehole is being drilled. A borehole seismometer, provided by LLNL, will be briefly installed in the hole when it is approximately 1500 m and 2000 m deep. Small explosive charges will be detonated immediately beside the temporary and nearby permanent stations, and the traveltimes to the borehole will be measured. By reciprocity, these traveltimes are the same as if the source were in the borehole. Because the location of the borehole is known, we can use this "virtual earthquake" to test and refine candidate locations and perform joint hypocenter determinations for the target earthquakes as recorded by the temporary network. The results from the 1500 m calibration will be used to set the kick-off point for the hole, and the 2000 m calibration will be used to refine the trajectory and azimuth of the deviated hole.

In another project that addresses the first objective, Rymer, Catchings and Fuis will conduct a high-resolution seismic reflection/refraction profile in the fall of 1998 along a 7 km-long line that runs across Middle Mountain at the drill site. The very high fold and short group interval (10 m) of this explosion survey will provide detailed structural and physical information (Vp and Vp/Vs) in 2-D within the upper 1.5 km of the crust, and reflection information to several times this depth. We will use the results of this survey to select a drill site free of near-surface faults or other potential geologic complications that should be avoided in the drilling of the vertical upper 2 km of the borehole. This work will be performed by the USGS, with important assistance provided the Oyo Corporation.

Our second objective, to create a comprehensive 3-D model of the fault and its environment, will be addressed by all nine projects. Most of these projects will be conducted during the approximately 2-year-long hiatus in drilling between the completion of the main hole and the drilling of the core holes. At this time, we will deploy a multi-sensor seismic monitoring string in the borehole. The borehole array will contain 80 3-component sensor packages distributed along the inclined borehole, and is described in more detail in the Section on Fault Zone Monitoring below. By making simultaneous surficial and borehole recordings of both earthquakes and controlled sources, we will be able to study the fault and its environment in great detail. As shown in Figure 16, we will coordinate the timing of the active source experiments to field as many of them as possible at the same time to maximize the observation (and potential scientific return) of each experiment.

During the 2-year interval between the completion of the main hole and the drilling of the core holes, we will be studying the structure and properties of the fault zone using a variety of active source and passive seismological techniques. Roecker and Thurber are proposing to install a dense 2-D surface array and image the Vp and Vp/Vs structure using basically the same experimental design employed to study the San Andreas fault 100 km to the northwest (Thurber et al., 1996, 1997). Their work plan includes much of the explosion seismology work being proposed. Li and Aki will also field the same type of dense linear arrays they have used to study fault zone guided waves at many locations around the world. Li has previously successfully coordinated this type of study with Roecker and Thurber's active/passive tomography experiments (Li et al., 1997). The permanent stations of the U.C. Berkeley and USGS networks (Figure 17) will also be integral data sources for the tomographic studies, and Mc Evilly et al. Will be working in concert with Roecker and Thurber, and others in the development of 3-D models.

The complex structural relations and strong lateral velocity gradients we will encounter within the fault zone present many challenges for the creation of a unifying structural framework from the diverse data that will be collected at the surface and in the borehole. Martin Karrenbach and his colleagues at the University of Karlsruhe in Germany have already begun to address this problem by constructing 3-D structural models for wave-propagation studies. A preliminary seismic model of the drilling target volume beneath Middle Mountain that includes a low velocity fault zone has already been used to compute full P and S waveforms at frequencies up to 20 Hz. Karrenbach will be working with other members of the site characterization team to create integrated models from which specific predictions can be made and tested with data from other experiments including the monitoring array described in the next Section. An example of the wavefields for a seismic source located in the fault zone is shown in Figure 18. Fault zone guided waves can be clearly seen, as well as direct and refracted body waves.

Figure 18
Figure 18. 3-D representation of full waveform seismograms computed in a simple 3-D model of the San Andreas fault at middle mountain. (click for more information)
McEvilly et al. are also proposing to augment the 2-D array of Roecker and Thurber with a linear array of seismometers that will most probably re-occupy the refraction line of Rymer, Catchings and Fuis. Recording earthquakes along this line will permit us to tie their detailed shallow structure model to the 3-D seismic tomography. Prodehl and Ryberg will conduct a wide-angle seismic reflection survey using the roll-along procedure along a 50 km-long line centered on the drill site. This profile will give critically needed velocity constraints and detailed structural data to depths of 10-15 km, or to the base of the seismic zone. Klemper, Talwani and Hole are proposing to conduct a 3-D seismic reflection survey above the hole in the region of detailed geologic and geophysical mapping by Rymer and Arrowsmith, and by Malin. By scheduling this survey during the time when the monitoring string is in the hole, we will have detailed tomographic coverage from the 2-D shot pattern to the borehole array, which should provide critical constraints for the interpretation of the 3-D seismic reflection data.

Additional site characterization plans that do not need to be tightly coordinated with the controlled source work are those of Rymer and Arrowsmith, and Malin. Rymer and Arrowsmith will conduct detailed geologic and geomorphic mapping of the immediate environment of the hole, and relate surficial geology and processes to those encountered in the borehole. Malin proposes to conduct a detailed potential field geophysical survey of Middle Mountain in the same region. Together, these studies will provide a solid framework for interpreting the results of the seismic imaging experiments. Malin also proposes to field a seismic imaging array during the first drilling phase. He will conduct an "inverse" VSP with the array, using the drill bit noise as a source.

 

Fault Zone Monitoring

The fault zone monitoring team consists of:

The goal of the fault zone monitoring work is twofold: 1) to make in-situ measurements of deformation, pore pressure, seismic wave radiation and other relevant parameters in the nearfield of earthquakes, and 2) to select the optimal intervals for continuous coring through the fault zone, as described above. We expect to observe multiple earthquake cycles for repeating earthquakes (M ~ 2) in the target zone at distances of less than a few hundred meters to about 1.5 km over 20-year lifetime of the experiment. We may also observe the rupture of the fault in a large-magnitude event over this same time period.

Figure 19
Figure 19. Illustration of the manner in which fault offsets can be diagnosed through observations of well casing shear. (click for more information)

Following completion of the first stage of the drilling program, we will begin an initial 2-year-long period of intensive monitoring using a re-deployable monitoring array consisting of seismometers, accelerometers and a fluid pressure monitor (Figure 5). The removable monitoring array will be pulled out of the borehole after about 1 year to conduct a high precision borehole directional survey and ultrasonic cement imaging (USI) log-to identify any changes in casing shape or the cement bond integrity behind the casing-in order to determine if any of the faults crossed by the hole are actively creeping, or if broad-scale deformation is occurring. The USI tool (see Appendix B) uses a rotating acoustic transducer and receiver to measure the internal radius of the casing, the resonance properties of the casing itself and the acoustic impedance of the casing/cement interface. This log, which is automatically corrected for acoustic velocity and impedance of the borehole fluid, yields casing radius to an accuracy of ± 0.2 mm. The gyroscopic directional tool to be used has an absolute accuracy of about 0.1° in azimuth, with repeatability considerably better than this. Figure 19 illustrates the manner in which fault offset can be identified and interpreted in terms of rock strain by measuring changes in casing geometry. Although the casing shear in this example (Figure 19, right panel) is relatively large (~13 cm over a 4-m-wide shear zone), repeat measurements of casing ovality and trajectory over time using casing shape logs and gyroscopic directional surveys similar to those we are proposing to use have identified casing offsets as small as 1 cm over a 5-m-wide shear zone (B. Wagg, pers. comm., 1998; see also Wagg et al., 1997). Following the completion of coring, we will re-instrument the borehole with a modified borehole array that will be augmented with additional pressure transducers, a borehole strainmeter and possibly other sensors, as described below.

The construction and deployment of the borehole monitoring systems presents many technical challenges. We have assembled a team of scientists and engineers with extensive experience in the design, manufacture and installation of borehole instruments. The design team will be led by Phil Harben (LLNL), who has extensive experience in the design, fabrication and deployment of borehole instrumentation systems. Peter Malin (Duke) will play a leading role in the design and deployment of the re-deployable monitoring array. Malin has deployed a similar system at Parkfield in the 1.5 km-deep Varian well. Tom McEvilly (UCB) co-developed the Varian system with Malin, and in addition has extensive experience in the design and construction of borehole seismometers, particularly those that employ high-frequency piezoelectric sensors. Malcolm Johnston and Evelyn Roeloffs (USGS) will play a leading role in the design and deployment of strain and pressure monitoring instrumentation. They both have extensive experience with these systems at Parkfield and elsewhere.

The functional design of the removable monitoring array has been set by a combination of scientific and technical considerations, the latter as a consequence of extensive discussions with industry and the substantial experience of members of the design team with similar instrumentation. This monitoring array will consist of 80, 3-component seismic sensors of proven design for long-term deployment in deep boreholes. Thirty of the sensor packages will contain gimbaled 3-component seismometers with natural frequencies around 4.5 Hz. Our recent experience recording earthquakes at 2 km depth in Long Valley at hypocentral distances as short as 1.4 km (Figure 4) demonstrates that moving-coil geophones can detect kilohertz energy at close range. Fifty of the sensor packages will contain internal fluid-damped, 3-component accelerometers with (undamped) frequencies around 30 Hz. Overdamped accelerometers of this design have been in operation in the 1 km deep Varian well for over a decade, where they provide wideband acceleration response for recording in the nearfield.

We have elected against installing active electronic devices in the removable monitoring array, such as feedback accelerometers, because of the relatively high temperatures (> 100° C) we anticipate. We may, however, install a limited number of very high frequency sensors, either piezoelectric accelerometers or hydrophones to sample the multi-kiloHertz frequency band, if suitable high-temperature sensors can be identified. All data will be digitized at the surface using a dedicated data collection platform with high sample rate, high-resolution digitizers. We have considerable practical experience with recording data in this manner from borehole instruments at several km depth (e.g. in Long Valley, Varian). Formation pore pressure will be monitored in the open hole below the monitoring array with a packer and pressure transducer system.

The removable monitoring array will be deployed on coiled tubing and hydraulic pressure will be used to activate multiple hole locks that will hold the sensors in place. This type of locking mechanism has been used for some time in the petroleum industry, although the complexity of this system is admittedly at the edge of available technology. Our discussions with industry do not indicate that there are any insurmountable technical problems with this design. It is our intention to build and test a prototype of the instrumentation packages and packers at an early date in one of several available deep boreholes that will provide a real-world test of the design. Either the Long Valley Exploration Well or the Cajon Pass deep borehole would be suitable for this test, and both boreholes are under the control of the USGS.

The long-term monitoring system that will be deployed following the core drilling is being designed to operate for up to 20 years. Our experience with borehole strainmeters and seismometers at Parkfield has been uniformly excellent, where instruments of the designs being considered are now well into their second decade of operation. The main augmentation we intend to make to the removable monitoring array will be the addition of at least one strainmeter (of either the dilatometer or tensor design). Because the strainmeter must be installed in an open (uncased) hole, we plan to install it in either one of the core holes or in the bottom open hole section, if it is still accessible. We may also elect to install one or more clamping tiltmeters in the cased hole. Ultra-low noise borehole tiltmeters (1 nanoradian resolution) have been developed by a member of the design team, Roger Hunter (LLNL), who was awarded a DOE R&D 100 Award in 1997 for this instrument (Wright et al., 1998). Additional sensors that are being considered include thermister arrays for measuring transient heating from earthquakes, fiber optic Fabre-Perot strain interferometers, and electrodes for measuring differential resistance within the fault zone.

The scientific objectives of the fault zone monitoring program have been described in Section II of this proposal, and will only be summarized here. During the 2-year observation window between the two drilling phases we will make high-resolution, nearfield observations of earthquakes between M -2 and 5 in the frequency band from below 1 Hz to over 1 KHz. These data will provide critical constraints on the radiated seismic energy budget, dynamic stress acting during faulting, and the spatial organization of the locked zone itself. They will also be critical to the testing of the source model hypothesized by Nadeau and Johnson (1998) and to determine if there is a minimum earthquake size at Parkfield.

The long-term monitoring program that will begin after the core drilling is complete will extend the frequency band to periods of weeks with the addition of tiltmeters and strainmeters. This will permit us to critically examine the nucleation process of earthquakes in the target cluster, to document the interplay between interseismic deformation in the fault zone and the rupture of discrete patches in earthquakes, and to unravel the spatial connection between repeating earthquakes at a common centroid.

Should a M~6 Parkfield mainshock occur, we anticipate that the accelerometers will be on-scale during the passage of the rupture by the array. This will provide unique information on dynamic friction during faulting, and permit a direct measurement of the displacement weakening distance of the San Andreas fault. The time history of displacement of the fault will also have direct engineering application, because of the new importance being placed on source displacements in engineering design following the Landers and Northridge earthquakes.

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