www.sciencemag.org/cgi/content/full/337/6093/459/DC1

Supplementary Materials for

Isotropic Events Observed with a Borehole Array in the Chelungpu Fault Zone, Taiwan

Kuo-Fong Ma,* Yen-Yu Lin, Shiann-Jong Lee, Jim Mori, Emily E. Brodsky

*To whom correspondence should be addressed. E-mail: fong@ncu.edu.tw

Published 27 July 2012, Science 337, 459 (2012)

DOI: 10.1126/science.1222119

This PDF file includes:

Materials and Methods Figs. S1 to S7 Table S1 References

2

Materials and Methods

The data.

The paper analyzes the seismic data recorded by the Taiwan Chelungpu-fault Drilling

Project (TCDP) vertical borehole seismic array. The TCDP Borehole seismometers

(TCDP BHS) are velocity type short-period seismometers with Galperin deployment and

a natural frequency of about 4.5 Hz, damping of about 29% and sensitivity of 1.6 V/cm/s.

The continuous data with sampling rate of 1000 Hz were corrected for the instrument

response taking into account the Galperin angle and orientation before proceeding to

micro-event identification.

The continuous data for the first six months were initially inspected visually and used to

classify the events into different groups. The events with down-going waves recorded in

the vertical array appear to be related to human activity and were therefore discarded.

The events with identifiable P-wave and S-wave were further examined, and we focused

on those with travel time differences of S to P (ts-tp) of less than 2 sec. Another distinct

group of events had only impulsive P-waves without identifiable S-waves. These are the

I-type (I for Isotropic) events in this study.

After manual searches of the first six months of continuous records, we developed a

semi-auto-picking technique for the continuous data to increase the efficiency of event

identification. The semi-auto-picking technique uses the ratio of short-term average

amplitude (STA) to long-term average amplitude (LTA) for P-phase picking. We then

calculated the cross-correlation coefficients and delay times of P-phases between the

3

waveforms of the 7-level borehole seismometers to assure there was consistency of the

waveform shape and the upward propagation direction of the wave. After identification

of the event using the P-phases, the S-waves were picked manually to check if the ts-tp

was less than 2 sec or, classified as an I-type events if no identifiable S-waves were

observed. A comparison of the semi-auto-picking to the manual picking for the first six

months of data revealed that the semi-automatic method reached an accuracy of about

90%.

For the 12 months recording period, among 13232 seismic signals, including regional

earthquakes and teleseismic events, 2780 events with ts-tp < 2 sec were identified (9). To

distinguish I-type events, we cross-correlated the waveforms from the identified I-type

event of 200611201628 (Fig. 2c) with the 2780 seismic events. Among them, 88 I-type

events were identified for cross-correlation coefficients of > 0.6. The particle motions

for a regular earthquake (200702281350) and an I-type event (200611201628) are shown

in Fig. S1. The particle motion for the direct P-wave clearly shows the radial direction of

movement, for the I-type event, a small phase about 0.3 sec after the direct P-wave shows

transverse particle motion, indicating an S-wave, which is related to a P to S conversion

at the fault zone (Fig. S1b).The noise level of the borehole seismometers is about 10-7

cm/sec. The spectra of the I-type events have peak values at frequencies of about 10-

20Hz, and the signal is just slightly above the background noise level at high frequencies

(Fig. 2d).

4

Volume changes.

The volume change was calculated for ∆V=tr(Mij)/(3λ+2µ) (17) where the rigidity of λ

= µ =3.26 x 1010Pa and tr(Mij) is the magnitude of the diagonal terms of the moment

tensor. For the isotropic source of the I-type events, the tr(Mij) is the value of the seismic

moment.

5

Figure S1: (a) Particle motions of the P- and S- waves for a regular event recorded at BHS004. The particle motions of the P-wave and the P to S converted phase of the I-type reference event (200611201628) (b) at BHS004, and (c) at BHS001. Particle motions for a regular earthquake (200702281350) and an I-type event (200611201628). The direct P-wave clearly shows the expected radial direction of movement. 0.3 sec after the P wave for the I-type event, there is a small phase with transverse particle motion indicating an S-wave, which is related to a P to S conversion at the fault zone.

6

(a) (b)

Figure S2: (a) The regression of Mw-logA to logΔ, A: maximum P-wave amplitude in cm/sec, ∆: distance in km, using 80 identified local events [open circles, (9)]. Red line indicates the regression line, which yields the empirical relationship of Mw= log (A) +1.11 log(∆)+4.53 . (b) Comparison of Mw determined from the empirical relationship, and determined from the spectrum of the ω2 model (9). The correlation of these two values is about 0.94.

7

Figure S3: The dimensions and setting for the 3D finite-difference calculation. The fault is dipping 30 degree to the east and the velocity structure is from the logging data (right). The P-wave velocities range from 3.20 km/sec to 3.90 km/sec and, and S-wave velocities from 1.52km/sec to 1.94km/sec. The Kueichulin Formation has a low P-wave velocity of about 3.2-3.35 km/sec and S-wave velocity of about 1.58-1.62 km/sec for depths of 1300 m to 1800 m (bold red lines), which is about the depth of the Sanyi fault (11). The ray paths from a virtual source location to BHS1 corresponding to the direct P-wave (dark blue), PFM (light blue) and P to S converted phase at the depth of 1.8km (dark blue and pink), are shown for reference. In addition to the 7 level borehole seismometers (red triangles), pseudo stations (blue dots) every 150m in depth are also shown.

8

Figure S4: Calculated model seismograms of the vertical component for the 7-level borehole seismometers (top seven), and the pseudo stations (below red line) for the six components of the moment tensor, and an isotropic source, M11=M22=M33. The depths of the 7-level borehole seismometers are between STA007 to STA010. The direct P arrivals (solid dark blue line), S arrivals (dashed red line), and surface reflection waves (short dashed blue line) are marked. .

9

Figure S5: Snapshots at elapse times of 0.324, 0.396, 0.468, and 0.540 sec for the finite difference wave propagation model. The snapshots show the ray paths related to the direct wave, fault zone multiples, PFM, and reflected P to S wave from the depth of 1.8km (PS1.8). The phases from the fine structure of the fault zone were influenced strongly by the refracted and reflected waves at the depths of 1.3 km and 1.8 km of the Kuechulin formation, which has strong velocity contrasts above and below. The calculations are done in 3D, but the structural model is 2D. The 3D topography of the fault could have influence on the amplitudes of the PFM, and the converted phases.

10

Figure S6: Comparison of the three component waveforms of the calculated seismograms for a tensile failure mechanism (blue) with observations (black) at BHS1 and BHS4. The calculated seismograms for a tensile failure mechanism uses M11>M22=M33, with the ratios of 1: 0.5: 0.5. The synthetic seismograms can partially explain the P-waves, but, S-waves with significant amplitudes are seen in the calculated seismograms, especially for the horizontal components. Calculations for an isotropic source mechanism still give a better fit to the observations (Compare to Fig. 3c). These comparisons suggest that the I-type events have an isotropic expansion mechanism.

11

Figure S7 The cartoon illustrates a scenario for the occurrence of the I-type events. The combination of an impermeable fault zone with near zero stress overlaying a fluid rich permeable region presents an environment conducive to natural hydraulic fractures. The bold arrows indicate the vertical and horizontal stresses. The horizontal stress is almost equal to the vertical stress (σv~σh), yielding a near-zero shear stress on the fault. The circle with expansion arrows indicate the I-type events, which possibly occur in the existing open cracks of the damaged material. The solid to dashed arrows indicate the fluid/gas escaping to the overlying formation of the impermeable zone, which produces a transient over-pressurized fluid flow that causes natural hydraulic fracturing (I-type events).

12

Table S1. List of the time, location, incident angles at BHS1 to BHS5 for the 30 I-type events with clear first motions. The location is relative to the site of TCDP. The correlation coefficient of each I-type event to the reference event (200611201628) is shown. The distance of the event to BHS4 and the maximum amplitude are also listed, which were used to determine the magnitude of the events.

1The 2o variance in incident angles at borehole seismometers will yield a difference in location of about 8-10m in depth, 3-5m in the x direction, and 50-100m in the y direction. 2The incident angles of these events at BHS001 to BHS005 mostly have variations of about 10-20 degrees. The incident angle at BHS004 has a sudden change compared to the other sites, due to the thin lower velocity layer right beneath this seismometer. 3 The magnitudes of the I-type events were determined using an empirical relationship of the maximum amplitude of P-waves and distance, as compiled from the identified local events, using Mw= log (A) +1.11 log(∆)+4.53, A: cm/sec, ∆:km[Fig. S2, (9)].

13

References and Notes

1. R. H. Sibson, Implications of fault-valve behaviour for rupture nucleation and recurrence. Tectonophysics

211, 283 (1992). doi:10.1016/0040-1951(92)90065-E

2. C. T. Montgomery, M. B. Smith, J. Pet. Technol. 62, 26 (2010).

3. R. J. Davies et al., Hydraulic fractures: How far can they go? Mar. Pet. Geol.

10.1016/j.marpetgeo.2012.04.001 (2012).

4. K.-F. Ma et al., Slip zone and energetics of a large earthquake from the Taiwan Chelungpu-fault Drilling

Project. Nature 444, 473 (2006). doi:10.1038/nature05253 Medline

5. Y.-G. Chen et al., Surface rupture of 1999 Chi-Chi earthquake yields insights on active tectonics of

Central Taiwan. Bull. Seismol. Soc. Am. 91, 977 (2001). doi:10.1785/0120000721

6. H.-Y. Wu et al., Stress orientations of Taiwan Chelungpu-Fault Drilling Project (TCDP) hole-A as

observed from geophysical logs. Geophys. Res. Lett. 34, L01303 (2007).

doi:10.1029/2006GL028050

7. K.-F. Ma et al., Evidence for fault lubrication during the 1999 Chi-Chi, Taiwan, earthquake (Mw7.6).

Geophys. Res. Lett. 30, 1244 (2003). doi:10.1029/2002GL015380

8. Y.-J. Hsu, P. Segall, S.-B. Yu, L.-C. Kuo, C. A. Williams, Temporal and spatial variations of

post-seismic deformation following the 1999 Chi-Chi, Taiwan earthquake. Geophys. J. Int. 169, 367

(2007). doi:10.1111/j.1365-246X.2006.03310.x

9. Y.-Y. Lin, K.-F. Ma, V. Oye, Observation and scaling of microearthquakes from the Taiwan

Chelungpu-fault borehole seismometers. Geophys. J. Int. 190, 665 (2012).

10. Materials and methods are available as supplementary materials on Science Online.

11. A. T. Lin et al., Terr. Atmos. Oceani. Sci. 18, 22 (2007).

12. E.-C. Yeh et al., Core description and characteristics of fault zones from hole-A of the Taiwan

Chelungpu-Fault Drilling Project. Terr. Atmos. Ocean. Sci. 18, 327 (2007).

doi:10.3319/TAO.2007.18.2.327(TCDP)

13. J.-H. Wang, J.-H. Hung, J. J. Dong, Seismic velocities, density, porosity, and permeability measured at a

deep hole penetrating the Chelungpu fault in central Taiwan. J. Asian Earth Sci. 36, 135 (2009).

doi:10.1016/j.jseaes.2009.01.010

14. J.-H. Wang, Summary of physical properties measured at several boreholes penetrating through the

Chelungpu Fault in Central Taiwan. Terr. Atoms. Ocean. Sci. 21, 655 (2010).

doi:10.3319/TAO.2009.09.01.01(T)

15. V. Oye, M. Roth, Automated seismic event location for hydrocarbon reservoirs. Comput. Geosci. 29,

851 (2003). doi:10.1016/S0098-3004(03)00088-8

16. S. Carena, J. Suppe, H. Kao, Active detachment of Taiwan illuminated by small earthquakes and its

control of first-order topography. Geology 30, 935 (2002).

doi:10.1130/0091-7613(2002)030<0935:ADOTIB>2.0.CO;2

14

17. J.-J. Dong et al., Stress-dependence of the permeability and porosity of sandstone and shale from TCDP

Hole-A. Int. J. Rock Mech. Mining Sci. 47, 1141 (2010). doi:10.1016/j.ijrmms.2010.06.019

18. B. R. Julian, G. R. Foulger, Bull. Seismol. Soc. Am. 86, 972 (1996).

19. A. McGarr, An implosive component in the seismic moment tensor of a mining‐Induced tremor.

Geophys. Res. Lett. 19, 1579 (1992). doi:10.1029/92GL01581

20. S. J. Lee, H. W. Chen, K. F. Ma, Strong ground motion simulation of the 1999 Chi-Chi, Taiwan

earthquake from a realistic three-dimensional source and crustal structure. J. Geophys. Res. 112,

B06307 (2007). doi:10.1029/2006JB004615

21. W. Tanikawa et al., Transport properties and dynamic processes in a fault zone from samples recovered

from TCDP Hole B of the Taiwan Chelungpu Fault Drilling Project. Geochem. Geophys. Geosyst.

10, Q04013 (2009). doi:10.1029/2008GC002269

22. M. L. Doan, E. E. Brodsky, Y. Kano, K. F. Ma, In situ measurement of the hydraulic diffusivity of the

active Chelungpu Fault, Taiwan. Geophys. Res. Lett. 33, L16317 (2006).

doi:10.1029/2006GL026889

23. H. Tanaka et al., Terr. Atmos. Ocean. Sci. 13, 227 (2002).

24. C.-Y. Chen, National Taiwan University, thesis (2010).

25. R. H. Sibson, Frictional mechanics of seismogenic thrust systems in the upper continental crust:

Implications for fluid overpressures and redistribution. Am. Assoc. Petrol. Geol. Mem. 82, 1 (2004).

26. Y. C. Chan, K. Okamoto, T.-F. Yui, Y. Iizuka, H.-T. Chu, Fossil fluid reservoir beneath a duplex fault

structure within the Central Range of Taiwan: Implications for fluid leakage and lubrication during

earthquake rupturing process. Terra Nova 17, 493 (2005). doi:10.1111/j.1365-3121.2005.00636.x

27. F. Cappa, Y. Guglielmi, J. Virieux, Stress and fluid transfer in a fault zone due to overpressures in the

seismogenic crust. Geophys. Res. Lett. 34, L05301 (2007). doi:10.1029/2006GL028980

28. Y. Ben-Zion, J.-P. Ampuero, Seismic radiation from regions sustaining material damage. Geophys. J.

Int. 178, 1351 (2009). doi:10.1111/j.1365-246X.2009.04285.x

29. M. Bohnhoff, M. D. Zoback, Oscillation of fluid-filled cracks triggered by degassing of CO2 due to

leakage along wellbores. J. Geophys. Res. 115, B11305 (2010). doi:10.1029/2010JB000848

30.F. Cappa, Y. Guglielmi, P. Fénart, V. Merrien-Soukatchoff, A. Thoraval, Hydromechanical interactions

in a fractured carbonate reservoir inferred from hydraulic and mechanical measurements. Int. J. Rock

Mech. Min. Sci. 42, 287 (2005). doi:10.1016/j.ijrmms.2004.11.006

31. J.-H. Hung et al., Subsurface structure, physical properties, fault-zone characteristics and stress state in

scientific drill holes of Taiwan Chelungpu Fault Drilling Project. Tectonophysics 466, 307 (2009).