Evidence for a large strike-slip component during the 1960 Chilean earthquake

Copyright and License

© The Author(s) 2019. Published by Oxford University Press on behalf of The Royal Astronomical Society.

Acknowledgement

We thank Stewart Smith who provided us with the detailed backup materials used for his 1966 study which played a key role for understanding the instrument characteristics of the Isabella strain record. Frank Press provided us with key information on the Isabella strain record used for the 1960 Chilean earthquake study. Duncan Agnew provided us with the strainmeter record of the 2010 Maule, Chile, earthquake recorded at the Piñon Flat Observatory. Guy Masters made available to us hand-digitized data of the Isabella strain seismograms for the 1960 Chilean and the 1964 Alaskan earthquakes archived at the Scripps Institution of Oceanography. Walter Zürn read a preliminary version of the manuscript and provided useful feedback to improve it. We appreciate the effort of Emile Okal and an anonymous reviewer who carefully read the manuscript and raised important questions on some of the key issues. Comments from Shingo Watada were helpful for clarifying some key points.

This work would not have been possible without help from colleagues who sent us, upon our request, the copies of the seismograms of the Chilean earthquake. Following is a partial list.
Peggy Hellweg: Berkeley Seismographic Stations, University of California, Berkeley. Shingo Watada: Earthquake Research Institute, the University of Tokyo.
Brian Ferris: Geonet, New Zealand.
Nobuo Hamada, Noriko Kamaya, and Jumpei Shimizu: Japan Meteorological Agency.
Jim Mori: Kyoto University.
Norihito Umino, Toru Matsuzawa, Tomotsugu Demachi, and Satoshi Hirahara: Tohoku University.

Software References

The Data Management System of the Incorporated Research Institutions for Seismology (http://www.iris.edu/hq/) was used to access the seismic data from the Global Seismic Network and Federation of Digital Seismic Network stations. We used Syngine web service at the Data Management Center of Incorporate Research Institutions for Seismology (Syngine, IRISURL https://service.iris.edu/irisws/syngine/, last accessed 2017 July 20) for computing some of the synthetic seismograms. The Green's functions are pre-computed with the AxiSEM (Nissen-Meyer et al.2014) for several 1-D reference models.

Supplemental Material

Table S1. Details of Computation and Processing of Data and Synthetics.
Table S2. The amplitudes a₀, a_b and a¯ in nanostrain.
Table S3. Peak-to-peak amplitude at three stations.
Figure S1. Top: Benioff ratio Γ. Bottom: T/S amplitude ratio as a function of l for the λ = 90° case for the 1960 Chilean earthquake at ISA station.
Figure S2. Top: Benioff ratio Γ. Bottom: T/S amplitude ratio as a function of l for the λ = 140° case for the 1960 Chilean earthquake at ISA station.
Figure S3. Normal-mode spectrum for (a) λ = 90° and (b) λ = 140° cases. Top: without splitting and coupling; middle: with splitting only; bottom: with splitting and coupling for the 1960 Chilean earthquake at ISA station.
Figure S4. The amplitude decay of each normal mode as a function of time (1 × 10⁵ s = 1.157 d).
Figure S5. Normal-mode strain spectrum (NW–SE component) for the 2010 Maule, Chile, earthquake recorded at PFO, California.
Figure S6. Power spectrum of the Pasadena ultralong period seismogram for three different time windows (duration = 1250 min) for the 1960 Chilean earthquake (λ = 140°). (a) From 21:01 May 22 (1 hr and 50 min after the origin time). (b) From 05:11, May 23 (10 hr after the origin time). (c) From 20:11 May 23 (25 hr after the origin time).
igure S7. The power spectral amplitudes of toroidal (red) and spheroidal (blue) modes for the λ = 140° and λ = 90° models of the 1960 Chilean earthquake at Pasadena. (a) From 21:01 May 22 (1 hr and 50 min after the origin time). (b) From 20:11 May 23 (25 hr after the origin time).
Figure S8. Power spectrum of the strain record at NNA for the 1960 Chilean earthquake (from Fig. 1 of BPS-1961). Blue: Spheroidal; Red: Toroidal.
Figure S9. Comparison of the observed and synthetic seismograms of the 1960 Chilean earthquake. R4 at PAL and G6 at RES. The response of the instrument is assumed to be T_s = 15 s, T_g = 75 s, h_s = 1., h_g = 1., σ = 0.25, V_max = 1000.
Figure S10. Comparison of the observed and synthetic seismograms of the 1960 Chilean earthquake. G4 at SFA. The seismograph is Wood-Anderson with T_s = 1 s, h_s = 0.8, gain = 2500. Synthetics computed for (a) the 10 sources model, (b) point source with a 90 s duration.
Figure S11. Benioff long-period seismograms of the 1960 Chilean earthquake recorded at Matsushiro, Japan. (a) EW component (pendulum period/galvanometer period/peak gain = 1 s/62 s/9640). The bottom three lines contain R3 and R4. (b) NS component (pendulum period/galvanometer period/peak gain = 1 s/70 s/7330). The bottom three lines contain G3 and G4. (c) Magnified figure of the boxed portion shown in (b). (d) Radiation pattern of G and Rayleigh waves for λ = 90° model. The location of the stations ISA and MAT are indicated. (e) Radiation pattern of G and Rayleigh waves for λ = 140° model.
Figure S12. Comparison of the acceleration spectra at Grotta Gigante computed for the two models, λ = 90° and λ = 140°, of the 1960 Chilean earthquakes (record length: 2580 min; beginning: 300 min after the origin time).
Figure S13. Normal-mode strain spectrum at BFO (N60°E component) for the 2004 Sumatra earthquake.
Figure S14. Comparison of 1-D and 3-D synthetic seismograms of the 1960 Chilean earthquake (λ = 140°) for the station Palisades (PAL). The 1-D and 3-D synthetics are computed for PREM and S40RTS (Ritsema et al.2011), respectively.
Figure S15. Comparison of the observed and synthetic seismograms of the 2011 Tohoku-Oki earthquake for the station Harvard (HRV).
Figure S16. Comparison of the observed and synthetic seismograms of the 2010 Maule, Chile, earthquake for the station PFO.
Figure S17. Static displacement field for a dip-slip (δ = 17°, λ = 90°) fault (left column) and an oblique slip (λ = 140°) fault (right column) for a 1000 × 200 km² fault. The left edge of the fault is placed along the trench at a depth of 5 km.
Figure S18. Static displacement field for a dip-slip (δ = 17°, λ = 90°) fault (left column) and an oblique slip (δ = 17°, λ = 140°) fault (right column) for a 1000 x 350 km² fault. The left edge of the fault is placed along the trench at a depth of 5 km.
Figure S19. Normal-mode spectrum of the 1960 Chilean earthquake. (a) Spectrum of the Pasadena Press–Ewing seismogram (vertical component) and the UCLA gravity meter record (Benioff et al.1961; Ness et al.1961; Kanamori & Anderson 1975). The spectral holes at 0S10 (blue dot) and 0S21 (red dot) are indicated. (b) Spectrum of the Pasadena Press–Ewing seismogram (vertical component) (Cifuentes & Silver 1989).
Figure S20. Spectrum of Pasadena Press–Ewing seismogram (vertical component). (a) Spectrum of the observed seismogram. (b) Spectrum of the synthetic seismogram for a point source with λ = 90°. (c) Spectrum of the synthetic seismogram for a finite source with λ = 90°. (d) Spectrum of the synthetic seismogram for a finite source with λ = 140°.