Supplementary Materials for Daniel A Frost, Barbara Romanowicz Correspondence to: [email protected]Contents of this file Figures S1 to S14 Tables S1 to S3 Introduction The supporting information contains additional information about the data and method (Figures S1 and S3, and Table S1), observations of lower qualities that those shown in the main paper (Figures S2, and S4-S7), and the results of additional tests (Figures S8, and S9, and Tables S2 and S3). All measurements shown in the paper are listed in table S4 (available separately).
33
Embed
danielafrost.files.wordpress.com · Web viewSupplementary Materials for. aatched region marks time and distance region investagted for in Frost et al., [2017]. itude Constraints on
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
The supporting information contains additional information about the data and method (Figures S1 and S3, and Table S1), observations of lower qualities that those shown in the main paper (Figures S2, and S4-S7), and the results of additional tests (Figures S8, and S9, and Tables S2 and S3). All measurements shown in the paper are listed in table S4 (available separately).
Supplementary Figures
Supplementary Figure 1. Relation beween the angles ξ1 and ξ2 of the two legs of the inner core P′ path with respect to the rotation axis for the Eilson (blue) and Yellowknife (red) arrays. All data are shown, both with and without clear P′P′ arrivals.
Supplementary Figure 2. Beams and array processing of a M7.1 event which occurred on 17/03/2003 at a depth of 33 km, recorded at YK, identified as poor quality (Supplementary Table 1). (a) Beams and F-traces formed on P′P′ab and P′P′df slownesses and back-azimuths. F-vespagrams formed for a range of (b) back-azimuths and (c) slownesses. Predicted times and slownesses from ak135 are shown by green triangles and lines. Observations are marked by blue lines representing onset times and circles showing the maximum F-amplitude, accounting for the time separation between the picks and peaks.
Supplementary Figure 3. Beams and array processing of a M6.7 event which occurred on 05/12/2003 at a depth of 10 km, recorded at IL, identified as poor quality (Supplementary Table 1).
Supplementary Figure 4. Beams and array processing of a M6.3 event which occurred on 06/03/2005 at a depth of 10 km, recorded at YK, identified as good quality (Supplementary Table 1).
Supplementary Figure 5. Beams and array processing of a M6.4 event which occurred on 08/10/2010 at a depth of 19 km, recorded at IL, identified as good quality (Supplementary Table 1).
Supplementary Figure 6. Beams and array processing of a M6.5 event which occurred on 12/03/2011 at a depth of 20 km, recorded at IL, identified as poor quality (Supplementary Table 1).
Supplementary Figure 7. Residual travel times as a function of average angle to the rotation axis for (a) ab-df differential measurements, (b) bc-df differential measurements, and (c) df absolute measurements. Individual measurements are shown by points, while moving averages (calculated in 3° bins with no overlap, calculated only for bins containing a minimum of 2 observations) are shown by purple circles with one standard deviation error bars. Residual travel times plotted along the inner core path (line color), for only the most polar of the source- and receiver-side paths (angle to pole shown by circle color), for (d) ab-df differential measurements, (e) bc-df differential measurements, and (f) df absolute measurements. Data shown have SNR greater than 1.
Supplementary Figure 8. Regions of possible scattering of P′P′ at the surface. Scattering from each location would generate waves with a characteristic (a) travel time shown relative to direct P′P′, (b) back-azimuth, and (c) slowness. The slowness and back-azimuth of scattering from other locations are distinct from reflection from the mid-point (circle) between the source (star) and array (triangle). Dashed and solid lines denote the path of the main phase and an example scattered path, respectively.
Supplementary Figure 9. Observed velocity anomalies from ab-df and bc-df differential residual travel times, as a function of average angle to the rotation axis, with corrections applied for tomographic velocities in (a) and (b) SEMUCB-WM1 scaled to P-waves, (c) and (d) S40RTS scaled to P-waves, (e) and (f) MIT_P08, (g) and (h) GAP_P4. Symbols as in Figure 4.
Supplementary Figure 10. Observed velocity anomalies from ab-df and bc-df differential residual travel times, and df absolute residual travel times, as a function of average angle to the rotation axis, with corrections applied for tomographic
velocities in (a) and (b) SEMUCB-WM1 scaled to P-waves, (c) and (d) S40RTS scaled to P-waves, (e) and (f) MIT_P08, (g) and (h) GAP_P4. Symbols as in Figure 4.
Supplementary Figure 11. Observed velocity anomalies from ab-df and bc-df differential residual travel times, as a function of average angle to the rotation axis, with events reflecting under the South Sandwich Islands removed. (a) Raw measurements for all observations are displayed as crosses and triangles for ab-df and bc-df, respectively, for df signals with SNR greater than 2. (b) Moving averages of observed velocity anomalies (purple) with one standard deviation error bars, and predicted velocity anomalies calculated from several anisotropy models for the observed events. Moving averages are calculated in 3° bins with no overlap, calculated only for bins containing a minimum of 2 observations.
Supplementary Figure 12. Observed velocity anomalies from ab-df and bc-df differential residual travel times, and df absolute differential travel times, as a function of average angle to the rotation axis. (a) Raw measurements for all observations. Velocity anomalies from paths reflecting under the South Sandwich Islands source region are shown in red. (b) Moving averages of observed velocity
anomalies (purple) with one standard deviation error bars, and predicted velocity anomalies calculated from several anisotropy models for the observed events. Moving averages are calculated in 3° bins with no overlap, calculated only for bins containing a minimum of 2 observations.
Supplementary Figure 13. Grid search over eastern and western hemisphere anisotropy. (a) Anisotropy curve (from Irving and Deuss [2011]) that is scaled to
model data. Root mean square misfit for east and west hemisphere anisotropy for: (b) all data with ab-df, bc-df, and absolute df observations, (c) excluding data reflecting under the South Sandwich Islands with ab-df, bc-df, and absolute df observations, (d) all data with only ab-df and bc-df observations, and (e) excluding data reflecting under the South Sandwich Islands with only ab-df and bc-df observations. Star shows minimum misfit with best-fitting anisotropy magnitudes shown above figure. Tested grid points shown in background.
Supplementary Figure 14. Curves to fit velocity anomalies as a function of average angle made with the polar axis, with bootstrap resampling. (a) Original data (circles) fit by a curve of equation 2 (green line). Bootstrap samples (grey) are summarized as a mean line (purple) and one standard deviation range (purple shading).
Coefficients of the line a, b, and c are shown above the figure. Curves to fit data corrected for travel time anomalies in tomographic models (b) SEMUCB-WM1 scaled to P-waves, (c) S40RTS scaled to P-waves, (d) MIT_P08, and (e) GAP_P4.
Supplementary Table 1. Details of events used. Events with discernable phases are classified as either good or poor, while events without phases are left blank.Year
Supplementary Table 2. Slowness and back-azimuth residuals from resolution tests for Elison and Yellowknife arrays. For each slowness, synthetic signals were generated from -180 to 135° back-azimuth, in 45° increments.
Supplementary Table 3: coefficients to fit velocity anomalies with equation 1 Anisotropy magnitude (%)
a b c RMS Misfit
Original 0.494 0.039 0.502 -0.008 0.755Bootstrap sampled
0.483 0.037 0.576 -0.093 0.787
No anisotropy
0 0 0 0 0.829
Supplementary Table 4: details of events used, including relative P′P′ab-df and P′P′bc-df, and P′P′df absolute travel time anomalies, pick qualities, polar angles ξ1 and ξ2, and event-receiver distances, sufficient to recreate figures shown in the paper.