Research ArticleGEOPHYSICS

Lower-crustal rheology and thermal gradient in the Taiwan orogenic belt illuminated by the 1999 Chi-Chi earthquake

See allHide authors and affiliations

Science Advances  27 Feb 2019:
Vol. 5, no. 2, eaav3287
DOI: 10.1126/sciadv.aav3287
  • Fig. 1 Fourteen-year postseismic GPS displacements following the 1999 Chi-Chi earthquake.

    (A) Cumulative postseismic displacements with respect to S01R at continuous-recording GPS installed before 2001 are shown as white vectors (see fig. S3 for the whole dataset). The focal mechanism indicates the Chi-Chi mainshock, and contours represent the coseismic slip on the CLPF (red line) projected on the surface. The six physiographic regions in Taiwan include the Coastal Plain (CP), the Western Foothills (WF), the Hsuehshan Range (HR), the Central Range (CR), the Longitudinal Valley (LV), and the Coastal Range (CoR). (B) Horizontal GPS position time series of S167 and FLNM, respectively, projected to the azimuth of N290°, roughly parallel to the direction of plate motion. Regression results are shown as white dashed lines, and the decompositions of basis 1 and 2 are indicated as red and blue curves, respectively.

  • Fig. 2 Modeled 14 years of cumulative strain in the lower crust and afterslip on the CLPF.

    (A) Second invariants of lower-crustal strains and afterslip on the CLPF in map view. Black dashed rectangle marks the area with brittle-ductile transition at 10-km depth. Contours show the coseismic stress change at 20-km depth calculated from the coseismic slip model (12) with a shear modulus of 30 GPa. White triangles denote GPS sites. (B) Strain and coseismic stress change along the AA′ cross section in (A). Inferred principal strain axes (black arrows) are shown as side-projected. (C) Fits to the 14-year GPS postseismic displacements at continuous-recording GPS sites installed before 2001 (circles) and after 2001 (squares), as well as survey-mode sites operating before 2001 (triangles). Observed and predicted horizontal displacements at selected GPS sites are indicated as white and black vectors, respectively. Background color represents predicted vertical displacements overlaid by colored polygons showing observations. (D) Fit to the horizontal GPS time series at the selected sites denoted in (C). Daily solutions with 2σ error bars and predictions (red lines) are projected to the azimuth of N290°.

  • Fig. 3 Lower-crustal rheology and geothermal gradients.

    (A) Strain transect and temperature contours from Simoes et al. (25) along the AA′ in Fig. 2A. Gray dots show aftershocks over a 3-month period, while pink and cyan circles denote the thrust and normal faulting, respectively (±15 km along the cross section). Volumes 1 and 2 beneath the CR and CP are indicated with red and blue boxes, respectively. (B) Inferred ranges of thermal gradients beneath the CR (red) and CP (blue) by incorporating laboratory-derived flow laws for quartz and feldspar (table S1). (C) Stress and strain rate derived from inversions (black lines) and theoretical estimates [red and blue lines for volumes 1 and 2 in (A), respectively] based on a power-law Burgers rheology with a stress exponent of 3. The rheological properties include shear modulus (GM), work-hardening coefficient (GK), steady-state material constant (AM), transient material constant (AK), and stress exponent (n). Black slopes show the benchmarks for steady-state creep with different stress exponents. Vertical and horizontal error bars are 1σ of the inverted strain rate and 0.025 MPa, respectively.

  • Fig. 4 Spatial-temporal pattern of lower-crustal effective viscosities.

    Estimated effective viscosities using a power-law Burgers rheology with a stress exponent of 3 and the best-fit rheological properties in volumes 1 and 2. (A to C) Spatial heterogeneous effective viscosities immediately, 1 year, and 10 years, respectively, after the 1999 Chi-Chi earthquake. The solid and dashed lines indicate the rupture area of the Chi-Chi earthquake and the shallow brittle-ductile transition beneath the CR. GPS sites are shown as white triangles. (D and E) Effective viscosities in volumes 1 and 2 as a function of time. Black and colored curves indicate estimates from inversions and predictions of the best-fit equivalent power-law Burgers model. (F and G) Relation between effective viscosity and strain in volumes 1 and 2, respectively. Black dots show time periods of a day, week, month, year, and decade after the earthquake. The purple dashed lines indicate the fraction of strain rate in Kelvin element.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/5/2/eaav3287/DC1

    Fig. S1. GPS horizontal velocity field from 1993 to 1999 before the Chi-Chi earthquake.

    Fig. S2. Searching for optimal temporal basis functions.

    Fig. S3. The entire GPS dataset including continuous-recording and survey-mode sites.

    Fig. S4. Checkerboard test for 4-by-4 grids without afterslip.

    Fig. S5. Checkerboard test for 4-by-4 grids with afterslip.

    Fig. S6. Checkerboard test for 3-by-3 grids without afterslip.

    Fig. S7. Checkerboard test for 3-by-3 grids with afterslip.

    Fig. S8. Rheology and stress versus strain rate.

    Fig. S9. Temperature-water fugacity relations and flow lows.

    Fig. S10. Effective viscosities derived from a power-law Burgers rheology.

    Table S1. Flow laws for quartz (qtz) and feldspar (fsp) used in this study.

  • Supplementary Materials

    This PDF file includes:

    • Fig. S1. GPS horizontal velocity field from 1993 to 1999 before the Chi-Chi earthquake.
    • Fig. S2. Searching for optimal temporal basis functions.
    • Fig. S3. The entire GPS dataset including continuous-recording and survey-mode sites.
    • Fig. S4. Checkerboard test for 4-by-4 grids without afterslip.
    • Fig. S5. Checkerboard test for 4-by-4 grids with afterslip.
    • Fig. S6. Checkerboard test for 3-by-3 grids without afterslip.
    • Fig. S7. Checkerboard test for 3-by-3 grids with afterslip.
    • Fig. S8. Rheology and stress versus strain rate.
    • Fig. S9. Temperature-water fugacity relations and flow lows.
    • Fig. S10. Effective viscosities derived from a power-law Burgers rheology.
    • Table S1. Flow laws for quartz (qtz) and feldspar (fsp) used in this study.

    Download PDF

    Files in this Data Supplement:

Navigate This Article