T12A-01
Chemistry of Crustal Fluids in Continental Fault Systems: Results From Massive Fluid Production Test at KTB Drill Site
A new series of experiments at Germany's super deep drill site KTB is focusing on transport processes of energy and fluids in crystalline crust. The experiments continue to exploit the scientific potential of the 4.0 km deep pilot hole (VB) and the 9.1 km deep main hole (HB) (distance at surface is 200 m). Targets of particular interest are two major fault systems intersected by the boreholes at 3.9 km (both holes) and 7.2 km depths (HB). The first experimental phase was a long-term pump test in KTB-VB with production of 22,300 m$^{3}$ of $119\deg$ C hot (bottom hole temperature) saline fluids from the metamorphic basement. The fault zone at 4 km depth has a bulk permeability of 2$\times$10$^{-15}$ m$^{2}$ , at least. From fluid level observation in KTB-HB a weak hydraulic contact with KTB-VB is deduced. Seismic activity, recorded both at the surface and inside KTB-HB at depths down to 3.9 km, appears not to have increased by the pumping. Since the beginning of the pump test, the volume ratio of gas to water (surface conditions) varied between 0.95 and 1.05. The electrical conductivity of the Ca-Na-Cl fluid (63 gl$^{-1}$ TDS) was rather constant at 86 mScm$^{-1}$ as were the on-line values for pH (7.5) and Eh (-430 mV). Except forRn, concentrations of gases and dissolved constituents were constant throughout the production period of one year. The major portion of the formation gas phase is composed of nitrogen (66.2%$_{vol}$) and methane (33%$_{vol}$); He (0.59%$_{vol}$), Ar (0.12%$_{vol}$), and CO$_{2}$ (0.03%$_{vol}$) are present in trace quantities only. Furthermore, isotope ratios of helium ($^{3}$He/$^{4}$He = 6.2\pm0.3\times10$^{-7}$, R/R$_{A}$ = 0.47), neon ($^{21}$Ne/$^{20}$Ne = 0.0035), argon ($^{40}$Ar/$^{36}$Ar = 945), CH$_{4}$ (\delta$^{13}$C = -50\permil), nitrogen (\delta$^{15}$N = +1.5\permil), and strontium ($^{87}$Sr/$^{86}$Sr = 0.7094) were apparently constant during the pump test. However, $^{222}$Rn activity varied between 3800 and 6000 Bqm$^{-3}$ depending on the production rate. N$_{2}$ originates most probably from the release of NH$_{4}$-fixed N$_{2}$ from paragneisses which have \delta$^{15}$N values of +6\permil. Ne-corrected, He isotope ratios indicate small amounts of primordial He which reveal a maximum of 6% 'mantle He' contribution. I$^{-}$ and Br$^{-}$ concentrations of 17 $\mu$M and 6.9 mM suggest that the fluids have derived their halogens from organic-rich sedimentary rocks of marine origin. Neither thermophile nor hyperthermophile organisms were detected unequivocally. Conclusions on origin, evolution and model ages of the crustal fluids at KTB will be presented in comparison with fluid occurrences from other deep seated crystalline basements.
T12A-02
Injection-Induced Microseismicity at the German Deep Drilling Site (KTB) at 4km Depth
We analysed microseismicity induced during a recent hydraulic experiment at the German Deep Drilling Site (KTB), Germany. Water was injected into a continental fault system at $4\,km$ depth over a period of one year in 2004/2005. Injection rate was constant at $200\,l/min$, resulting in a cumulative volume of $85000\,m^3$ injected water. The injection was preceded by a one-year water production test in 2002/2003. Seismicity was monitored by a borehole sensor at $1.9/3.5\,km$ depth and a densely spaced near-surface seismic array. A total number of 3000 microearthquakes were detected by the borehole instrument, and 146 events could be located by using the near-surface seismic recordings.\\ We present the characteristics of the injection-induced seismicity with focus on high-resolution absolute and relative event location. We also discuss the correlation of seismicity and seismic reflectivity, the spatio-temporal evolution of the event distribution, and source mechanisms.
T12A-03
Modeling of Fluid Induced Deformation of the Upper Crust of the Earth: Tilt Investigations About the Large Scale Injection Experiment at the KTB/Germany
The injection experiment at the KTB started in June, 2004 with a medium injection rate of 180 liters/minute into the KTB pilot borehole (4000 meters deep). A tiltmeter array, consisting of five high resolution borehole tiltmeters of the ASKANIA type, was operating in the surrounding area of the KTB location from mid 2003 until September 2006. The tiltmeters have a resolution of better than 0.2 msec (about 1 nrad). The aim of the research project was to observe the induced deformation of the upper crust at kilometer scale and to interpret the observation by numerical modeling, together with the monitoring of induced seismicity in the area. We expect elastic as well as anelastic responses: Changes of the rheologic properties due to pore pressure increase will cause changes in the tidal parameters. Further we expect sudden changes of the drift curve as well as slow variations. For the separation of the induced drift signal it is necessary to eliminate locally induced interference, e.g. arising from groundwater variations. The ground water / pore pressure changes, observed at all stations show significant correlations with the recorded tilt signals. The reduction of these locally acting effects and also meteorological influences like barometric pressure changes or precipitation yield tilt signals, which are significantly correlated with the injection experiment: The hodograms, which describe the tip movement of the pendulum over ground, show a clear dominant drift away from the injection point for three stations. This corresponds with a bulge in the area where the injection takes place. The tilt amplitudes are in the order of some milliseconds. Parallel to the observations with the tiltmeter array we quantified the expected additional drift for different injection scenarios at each tiltmeter site, by numerical modeling using the program POEL. It can be demonstrated that the tilt signals caused by injection intervals of less than three days are not detectable by the tiltmeter array. But for long term injection phases of up to four months a maximum tilt effect of about 40 nrad is modelled. This first modeling result confirms the observed tilt signals. But which role plays the main geological features like the fault zones SE1 and SE2, the adjacent granite complex and the postulated Erbendorf body? In answer to this question and to clear up the whole injection process a comprehensive numerical poro-elastic modeling using the finite element method (software ABAQUS) is in preparation.
http://www.geo.uni-jena.de/geophysik/
T12A-04
Microseismic Evidence for the Interaction of Faulting and Fluid Flow During Hydraulic Fracture Injection
Microearthquake induced during hydraulic fracture treatments were precisely located using data from two borehole arrays of 3-component geophones. The treatments were conducted within interbedded sands and shales of the Upper Cotton Valley formation, located in east Texas. The microearthquakes occurred within narrow horizontal bands that correspond to sandstone layers that were specifically targeted for gas production. Double couple (DC), composite focal mechanism inversions indicate strike-slip faulting occurring uniformly along vertical fractures trending close to maximum horizontal stress direction. The banding of events and the slip-plane orientations are close to the reservoir's prevalent natural fractures, known to be isolated within the sands and trending subparallel to the expected hydraulic fracture orientation. Full moment tensor solutions were also attempted by amplitude inversion using higher signal-to-noise events. Significant non-DC components are possible including tensional crack components, but are often poorly resolved due to limited focal sphere coverage. Assuming Coulomb failure criteria, the observation of horizontal slip along fractures subparallel to maximum-horizontal stress implies a relatively high critical pore pressure. Thus, it is reasonable to expect fracture opening is accompanying slip and that the seismicity is directly associated with the activated fluid-flow paths. Faulting, in turn, appears to affect the fluid flow, as evident from the time-space patterns of seismicity. Anomalous event counts and moment release sometimes occur within dense clusters that delineate bends or jogs in the fracture zones. The dense clusters show location patterns diverging in time, suggesting the expulsion of fluid from compressive fault jogs. These jogs likely form choke points where the slip-induced loading tends to lock up and concentrate stress at the jogs, as evident by fewer but larger events populating the structures as injection proceeds.
http://www.lanl.gov/orgs/ees/ees11/geophysics/m icroeq/
T12A-05
Behavior of induced microseismic events with large magnitude
Hydraulic stimulation of geothermal and oil/gas reservoir is one of the conventional techniques used for enhancing the productivity from reservoirs. In most cases, the stimulation process induces microseismic events. Based on the activity, location, magnitude and source mechanism of such events, the 3D localization and characterization of the reservoir can be carried out with practical resolution. Typically, microseismic events from a reservoir have moment magnitudes of less than zero, and most of them are detectable only by downhole sensors with high sensitivity. However, it is known that some of these events have higher magnitudes and can be felt at the surface. These large events can be hazardous from an environmental point of view, while resulting in enhanced permeability in the reservoir at the same time. The authors have analyzed the spatio-temporal distribution, and source mechanism of such microseismic events having large magnitudes (big events) observed during the hydraulic stimulations at Australian hot fractured rock (HFR) site in the Cooper Basin (Asanuma et al., SEG Exp. Abst., 2004) and also at the European hot dry rock (HDR) site in Soultz, France (Asanuma et al., Trans. GRC, 2004). A comparison between the origin time of these big events and the hydraulic records showed that many of the big events occurred after the shut-ins at both the sites. Besides, during pumping, most of these events did not show a clear correlation to the wellhead pressure and the rate of pumping. In most cases, the source mechanism of the big events were consistent with the shear slip of a preexisting fracture. We have also found that some of the big events at the Australian site brought very clear extension of the seismic cloud into zones that were seismically silent before, suggesting that some kind of hydraulic barrier was overcome by these big events. The observational data also showed that the microseismic events at those sites originated mainly from a slip of asperities in existing fractures under a nearly critical stress state of the shear slip. The largest source radius estimated from the moment magnitude at the European site was around 300m. It is an accepted fact that the typical length of the existing joints in granite at the stimulated zone is around ten meters. Hence, we conclude that the source mechanism of these induced microseismic events at the European site cannot simply be interpreted as due to a slip of asperities alone, and an investigation into the correlation of the stress concentration at the joints with the magnitudes is also warranted.
T12A-06
An Episode of Fault-Valve Behavior During Compressional Inversion - the 2004 MJ 6.8 Niigata-ken Chuetsu Earthquake Sequence
Principal ruptures of the 2004 Niigata-ken Chuetsu sequence involved close-to-pure reverse slip on a 50- $60\deg$ WNW dipping fault system uplifting a Miocene rift basin on its hanging wall, consistent with geological evidence for ongoing compressional inversion in the region since the Late Pliocene. The M$_{J}$6.8 mainshock was followed by four other M$_{J}$>6 shocks (all with hypocenters in the 7-12 km depth range) and a rich aftershock sequence defined a criss-crossing network of reverse-slip rupture planes dipping 50-$60\deg$ WNW and 25-$35\deg$ ESE. The steep rupture planes dip close to the maximum for reverse-fault earthquakes. Assuming horizontal maximum compressive stress, they are poorly oriented for frictional reactivation (50- $60\deg$ to $\sigma$$_{1}$), lying close to the lock-up angle for standard coefficients of rock friction. In contrast, subsidiary ESE-dipping thrusts (25-$35\deg$ to $\sigma$$_{1}$) appear optimally oriented for reactivation. Reactivation of existing, unfavorably oriented reverse faults is thus competing with slip on optimally oriented thrusts, requiring high but probably variable fluid-overpressuring within the rock-mass. Supporting evidence for fluid-overpressures around the seismogenic structures includes direct measurements of overpressures in oil/gas boreholes within the >6 km thick sedimentary basin on the hanging-wall of the main fault, plus a range of geophysical measurements consistent with local overpressuring in and around the fault system. Minor postseismic effusion of anomalously warm, saline formation fluids was recorded from the aftershock area of this earthquake: similar discharges occur in the epicentral areas of previous earthquakes in the Niigata region. Compressional stress in the upper crust helps to contain overpressured fluids which may, in part, be derived from beneath the seismogenic zone where anomalous electrical and seismological characteristics also suggest distributed overpressuring in the mid-crust. The rupturing of unfavorably oriented faults in a fluid-overpressured compressional regime, together with the swarm-like character of the aftershock activity and the evidence of postseismic discharge â€â€œ are all suggestive of a fluid-driven, fault- valve episode with upwards migration of overpressured fluids.
T12A-07
Geomechanical Reservoir Characterization for Prediction of the Likelihood of Faulting Induced by CO2 Sequestration
Geologic sequestration of CO2 has been proposed as a key technology for reducing greenhouse gas in the atmosphere. However, leakage of CO2 from any potential reservoir could seriously compromise the goal of long- term storage. Therefore understanding the likelihood of leakage is a key step toward the feasibility of this technology. Because it has long been known that fluid injection can lead to fault reactivation, assessing the potential of induced fault slip and leakage of CO2 from any potential storage reservoir is critical prior to injection of large quantities of CO2. We have developed a geomechanical model in advance of a proposed injection experiment at Teapot Dome, WY in order to study in detail the processes of a CO2 leak through predicted faults. Teapot Dome is an elongated asymmetrical, basement-cored anticline with a north-northeast axis. It is part of the Salt Creek structural trend, located in the southwestern edge of the Powder River Basin. The anticline is compartmentalized in a series of oblique-slip faults, generally oriented NE-SW, although varying in geometry, displacement, and complexity. In particular, the fault network referred to as S2 shows a greater complexity and a range of geometries and azimuths that enhances its potential for leakage. Furthermore its surface expression show alkali springs and hydrocarbon samples within the fault veins and gouge. The S2 fault intersects the oil-bearing 2nd Wall Creek around 650m, which is close in terms of temperature and pressure to conditions of supercritical CO2 phase injection and oil miscibility. We developed a preliminary geomechanical model of the deeper Tensleep Fm. previously, utilizing observations of wellbore breakouts, drilling induced fractures, leak-off or mini-frac tests and other available geophysical data. We used this information to evaluate the potential for injection to induce slip on the S1 fault, approximately 2 km south of the present study area. These results are compared and improved for the shallower 2nd Wall Creek Fm. in the S2 fault zone, using additional observations of wellbore breakouts, drilling induced fractures, leak-off tests and other available geophysical data. Finally, we have used Quantitative Risk Assessment to evaluate the uncertainties in the geomechanical model parameters and the effect of these uncertainties on predicting the likelihood of induced fault slippage and potential leakage of CO2.
T12A-08
Fluid-Controlled Aftershock Patterns from Different Tectonic Regimes
I further explore the hypothesis that a significant contribution to aftershock generation results from the post- seismic degassing of trapped high pressure fluid sources at depth. In this scenario, high pressure fluids generated from earth degassing and trapped beneath the seismogenic layer are released by the co-seismic rupture of the hydraulic seal separating near-lithostatic pressure below from hydrostatic pressures above. The ensuing pressure pulse propagating through the newly-created damage zone triggers aftershocks by significantly reducing the effective confining stress. Different tectonic regimes are simulated by imposing the relevant regional stress and solving a non-linear diffusion equation, where the non-linearity results from the effective-stress dependence of permeabilty. Using the same model of crustal hydraulic properties, it is shown that the evolved fluid pressure field closely mimics aftershock patterns in normal, strike-slip and thrust environments. It is also demonstrated that Omori's Law, and controls on the decay rate, are a direct consequence of the flow properties of the medium and with the prevailing regional stress field. Since triggering amplitudes can be many MPa, this offers an alternative mechanism to static stress-triggering arguments. Model results show strong correlations in space and time (e.g. aftershock decay rates) with the Landers and Northridge aftershock sequences.