This electronic supplement presents further graphical output from testing of the absolute arrival-time recovery method (AARM) on observed (data from southeast Canada of Boyce et al., 2016) and synthetic datasets. A zipped archive containing a copy of the AARM code is also provided. The code, plotting scripts, and a detailed user guide are included.
Figure S1. Comparison of the prearrival noise suppression of (a) linear, (b) nth root, and (c) phase-weight stacking for an example earthquake. (d) Overlay of the three stack types. SNR, Signal-to-noise ratio.
Figure S2. AARM output (using a cross-correlation scheme) for an impulsive earthquake recorded by stations in southeast Canada (Boyce et al., 2016). (a) Output stacks (blue primary and red final) from alignment of traces. The correction time Tcorr is also shown. (b) Absolute arrival-time residual histogram. (c) Distribution of weights from cross correlation of individual traces with the primary stack on the linear weighting scheme. (d) Absolute arrival-time residual distribution against epicentral distance. (e) The autocorrelation-derived estimate for picking error. (f) The difference between calculated absolute arrival times from AARM and the International Seismological Centre (ISC) picks, where available (only one equivalent source–station pair exists for this event).
Figure S3. AARM output (using SNR scheme) for an emergent earthquake recorded by stations in southeast Canada (Boyce et al., 2016); subplots are the same as in Figure S2. Note that no corresponding ISC picks were available for this earthquake.
Figure S4. (a) Nine variably noisy waveforms (STA 1–9) that contribute to the initial stack (stack1 in black) and final stack (stack2 in red). Vertical lines show manually picked arrival times. (b) An expanded view of the stacks shows the difference between the picks (Δ).
Figure S5. Plot showing how increasing levels of noise (Peterson et al., 1993) affect individual traces at (a) station SUP2 and (b) the final stack for the S-wave synthetics (labels are the same as in Fig. 6 in the main article, P-wave data).
Figure S6. Results from S-wave synthetic testing; subplots are identical to Figure 7 in the main article.
Figure S7. (a) Distribution of multichannel cross-correlation (MCCC)-derived arrival times Talign (VanDecar and Crosson, 1990) with great circle arc length for the synthetic P-wave arrivals. (b) Filtered velocity seismograms aligned at 5 s on MCCC-derived arrival times. (c) The distribution of relative-arrival-time residuals for the array of synthetic seismometers at increasing epicentral distance. The removal of the mean arrival time for the array results in approximately zero residuals for central stations in this symmetric synthetic array.
Figure S8. Modeled power spectra for teleseismic background noise models of Peterson et al. (1993) varying between high (1.0) and low (0.0) extremes. These are used to add progressive levels of noise to synthetic seismograms. Black curves represent the upper and lower bound on observed noise.
Download: Absolute_arrival-time_toolkit.zip [Zip Archive; ~2.53 MB]. This file includes the AARM code, plotting scripts, example data, and detailed user guide for AARM.
Boyce, A., I. D. Bastow, F. A. Darbyshire, A. G. Ellwood, A. Gilligan, V. Levin, and W. Menke (2016). Subduction beneath Laurentia modified the eastern North American cratonic edge: Evidence from P wave and S wave tomography, J. Geophys. Res. 121, no. 7, 5013–5030, doi: 10.1002/2016JB012838.
Peterson, J. (1993). Observations and modeling of seismic background noise, U.S. Geol. Surv. Open-File Rept. 93-322.
VanDecar, J., and R. Crosson (1990). Determination of teleseismic relative phase arrival times using multi-channel cross-correlation and least squares, Bull. Seismol. Soc. Am. 80, no. 1, 150–169.
[ Back ]
