Research ArticleGEOPHYSICS

A laboratory nanoseismological study on deep-focus earthquake micromechanics

See allHide authors and affiliations

Science Advances  21 Jul 2017:
Vol. 3, no. 7, e1601896
DOI: 10.1126/sciadv.1601896
  • Fig. 1 XMT images of faults and correlation with relocated AE events.

    (A) View of 3D tomographic reconstruction of the recovered sample, with the bulk tuned to 85% transparent, revealing sets of conjugated fault traces (dark lines). (B) Cut-away view of the same image showing conjugated faults. The faults are digitized and plotted in (C) and (D). For 3D representations of (A) and (B), see movies S1 and S2, respectively. (C) Group 1 events superimposed with digitized faults to show their spatial correlation with fault 1. (D) Group 2 events superimposed with digitized faults showing correlation with fault 2. 3D representations of spatial correlation between group 1 and 2 events and the faults are shown in movies S3 and S4, respectively.

  • Fig. 2 Examples of synthetic waveforms (red curves) based on moment tensors from the grid search.

    Events 277 and 288 are from groups 1 and 2, respectively. Only the first P-wave arrivals are modeled, which show good agreement with observed waveforms after low-pass–filtered at 3 MHz (black curves). Complex waveforms beyond the first P-wave arrival are most likely due to the narrow bandwidth of the sensors and wave reverberations in the instrument. The relative amplitude difference between the horizontal channels 3 and 4 (CH3 and CH4, respectively) is controlled by the strike of fault planes (also see fig. S5).

  • Fig. 3 Relation of AE event counts (in log scale) versus moment magnitude.

    Dark gray histograms are from the 593 events above the detection threshold (MW between −9.5 and −8.0). Light gray histograms are based on the refined event detection algorithm based on waveform cross-correlation for events below the threshold and scaling of saturated events based on duration (see text for details). Detectable magnitude range has doubled. This yields a more tightly constrained b of 1.5. The relation for b = 1.0 is also plotted for reference (dashed line).

  • Fig. 4 Event evolution throughout the deformation/transformation process.

    (A) History of group 1 events projected along the strike of fault 1, whose average trace is shown as the red dashed line. Most of the large events after 1700 s are located in fault 1. (B) Group 1 events projected onto fault 1. Size of the events is proportional to their MW to minimize overlap. (C and D) Similar plots for group 2 events and fault 2. (E) Moment magnitudes of 5818 events detected from the continuous record in the time sequence as they occurred. Superimposed are pressure (blue), strain (red), and differential stress (green). Stress peaked near 1700 s and then dipped when large AE events stared bursting. In addition, plotted are the b values over 600-s time windows during the deformation/transformation process. A continuous decrease trend in b value is evident. After the 1700-s mark, b decreases rapidly from 1.5 to 1.0, suggesting that later events are preferentially distributed in a planar fashion.

  • Fig. 5 Variation of ISO strength (ζ) and event rate as a function of MW and time.

    (A) Events below (red circles) and above MW ≈ −8.5 (blue circles) display different spread in ζ. Dashed lines are a guide to the eye. (B) The deformation/transformation process is divided into four regions based on event rates (given on top of each colored region, in number of events per minute). Event rate reaches maximum (24.5/min) between 1500 and 2100 s, presumably due to the nucleation and propagation of NSBs. The critical point is indicated by the vertical arrow. Note the rapid tightening of ζ values after ~2000 s (dashed lines are a guide to the eye), followed by a “quiescent period” between ~2150 and 2550 s (event rate drops to 4.0/min), before the sample failed shortly after ~2700 s.

  • Fig. 6 Low-magnification SEM BSE images showing the distribution of faultlets and fault segments.

    Maximum compressive stress is approximately vertical. Pyroxene grains are the bright and round crystals against the dark gray olivine matrix. Spinel has ~8% higher density than olivine and appears slightly brighter. (A) Area corresponding to the white box in fig. S2A. The two macroscopic faults (faults 1 and 2) are those indicated in Figs. 1 and 4. Numerous faultlets (thin and dark traces; a few are indicated by white arrows) are present, with lengths of <0.2 to 0.3 mm primarily subparallel to either fault 1 or fault 2, in conjugated relations. Their orientations are consistent with AE event distribution with time, as shown in Fig. 4. (B) Area corresponding to the white box in fig. S2B. Tips (white arrows) of two fault segments (faults a and b) are connected by a long NSB. Near horizontal cracks are due to decompression.

  • Fig. 7 BSE micrographs showing the presence of long NSBs connecting visible fault segments.

    (A) Image corresponding to the white box in Fig. 6B, showing the long and continuous NSB in the NW-SE direction (red arrows). Note a bifurcated of NSB (white arrows), about 20° away from the main branch. Near-horizontal cracks were due to decompression. (B) Image corresponding to the red rectangle in fig. S2B. The “gap” between the upper left and lower right fault segments (white arrows) is shown in (C) and (D), each consisting of two SEM images. (C) An NSB runs from upper left to lower right (indicated by the red arrows). One pyroxene grain has been sheared into two halves, with an offset of about 30 μm (white double-headed arrow). (D) The same NSB in (C) continues into this image. Overall, this NSB is more than 150 μm long, connecting the two fault segments in (B). (E) Composite TEM image from an FIB foil cut at the white line in (D), showing that the NSB is curvy at depths, and passing through olivine grain G2 and merging into the boundary between G3 and G4.

  • Fig. 8 BSE micrographs showing NSBs initiating from stress concentrations.

    (A) Two LVRs with NSBs radiating from the corners (white arrows). The NSB originated from the LVR on the right has sheared the pyroxene (pyx) grain with an offset of ~3 μm. (B) Two NSBs radiating from two pyroxene grains. The originating points are indicated by the white arrows.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/3/7/e1601896/DC1

    fig. S1. Top view of the experiment setup.

    fig. S2. Low-magnification SEM images showing cross sections of the entire samples of the experiments.

    fig. S3. Example waveforms of AE events in groups 1 and 2.

    fig. S4. An example of grid search result.

    fig. S5. Histograms of ISO parameter ζ, CLVD parameter χ, and strike φ of 593 moment tensor solutions (light gray).

    fig. S6. An example of cross-correlation template searching results for smaller AE events.

    fig. S7. Waveforms of events with different moment magnitudes.

    fig. S8. Orientation relations between olivine grains and NSBs.

    fig. S9. BSE micrographs of recovered samples.

    movie S1. XMT image of the recovered sample D1247.

    movie S2. XMT image of the recovered sample D1247.

    movie S3. Relocated AE events in group 1 superimposed with the digitized faults, as shown in movie S2.

    movie S4. Relocated AE events in group 2 superimposed with the digitized faults, as shown in movie S2.

    movie S5. Group 1 events viewed along the strike of fault plane 1, throughout the deformation/transformation history.

    movie S6. Group 2 events viewed along the strike of fault plane 2, throughout the deformation/transformation history.

  • Supplementary Materials

    This PDF file includes:

    • fig. S1. Top view of the experiment setup.
    • fig. S2. Low-magnification SEM images showing cross sections of the entire samples of the experiments.
    • fig. S3. Example waveforms of AE events in groups 1 and 2.
    • fig. S4. An example of grid search result.
    • fig. S5. Histograms of ISO parameter ζ, CLVD parameter χ, and strike ϕ of 593 moment tensor solutions (light gray).
    • fig. S6. An example of cross-correlation template searching results for smaller AE events.
    • fig. S7. Waveforms of events with different moment magnitudes.
    • fig. S8. Orientation relations between olivine grains and NSBs.
    • fig. S9. BSE micrographs of recovered samples.
    • Legends for movies S1 to S6

    Download PDF

    Other Supplementary Material for this manuscript includes the following:

    • movie S1 (.mov format). XMT image of the recovered sample D1247.
    • movie S2 (.mov format). XMT image of the recovered sample D1247.
    • movie S3 (.avi format). Relocated AE events in group 1 superimposed with the digitized faults, as shown in movie S2.
    • movie S4 (.avi format). Relocated AE events in group 2 superimposed with the digitized faults, as shown in movie S2.
    • movie S5 (.avi format). Group 1 events viewed along the strike of fault plane 1, throughout the deformation/transformation history.
    • movie S6 (.avi format). Group 2 events viewed along the strike of fault plane 2, throughout the deformation/transformation history.

    Files in this Data Supplement: