Prediction of maximum P- and S-wave amplitude distributions incorporating frequency- and distance-dependent characteristics of the observed apparent radiation patterns
- 1.5k Downloads
KeywordsBody waves Wave propagation Earthquake ground motions
It is well known that, as frequency increases over 1 Hz, the spatial distributions of observed maximum P- and S-wave amplitudes during local earthquakes (hereafter, this is called the “apparent radiation pattern”) are gradually distorted from the expected four-lobe amplitude pattern of a double-couple point source (e.g., Liu and Helmberger 1985; Satoh 2002a; Takenaka et al. 2003; Takemura et al. 2009; Sawazaki et al. 2011; Kobayashi et al. 2015). The frequency-dependent characteristics of the observed apparent radiation patterns have been incorporated into various applications such as the predictions of strong ground motions (e.g., Pitarka et al. 2000; Pulido and Kubo 2004), the estimation of high-frequency seismic energy radiation during large earthquakes (e.g., Nakahara 2013), nonvolcanic/volcanic tremors (e.g., Maeda and Obara 2009; Kumagai et al. 2010; Cannata et al. 2013; Yabe and Ide 2014), and landslides (e.g., Ogiso and Yomogida 2015), and the earthquake early warning systems (e.g., Okamoto and Tsuno 2015). Although the frequency–distance change model for S-wave radiation pattern proposed by Satoh (2002b) has been used in some applications, to achieve more accurate estimation and prediction of high-frequency seismic radiation, a precise frequency- and distance-dependent model for the apparent radiation pattern for both P and S waves is required. Relationship of the apparent radiation patterns between P and S waves has been important due to recent development real-time systems, such as urgent earthquake detection and earthquake early warning (e.g., Okamoto and Tsuno 2015).
High-quality seismograms recorded by dense regional seismic networks for various distances and wide dynamic ranges enable us to investigate frequency- and distance-dependent characteristics of the apparent radiation pattern. Takemura et al. (2009, 2015) and Kobayashi et al. (2015) reported that the apparent P- and S-wave radiation patterns are distorted with increasing distance but still preserving the original four-lobe pattern at hypocentral distances less than 40 km even for high frequencies. In this study, we firstly investigated the frequency- and distance-dependent characteristics of the apparent P- and S-wave radiation patterns using dense and large number seismograms. On the basis of observed characteristics, then we propose a frequency- and distance-dependent model of the apparent radiation pattern to predict the spatial distributions of maximum P- and S-wave amplitudes of local earthquakes.
Apparent radiation pattern for crustal earthquakes
In some previous studies, energy partition of S wave in each horizontal component was analyzed in order to eliminate the effects of differences in site amplification and source size. However, Sawazaki et al. (2011) pointed out that spatial distribution of maximum amplitudes and energy partitioning in each component show different frequency-dependent properties. Therefore, on the basis of the method by Kobayashi et al. (2015), we measured coda-normalized maximum P- and S-wave amplitudes (hereafter, these are referred to as the “P-wave amplitude” and “S-wave amplitude,” respectively) from three-component root-mean-square (RMS) envelopes for the following different frequency bands: 0.5–1, 1–2, 2–4, 4–8, and 8–16 Hz.
Additional file 1: Figure S1 shows examples of filtered velocity seismograms and RMS vector envelopes normalized by averaged coda amplitudes to eliminate the effects of differences in site amplification and source size (e.g., Yoshimoto et al. 1993). Since coda normalization technique is applicable in the seismograms with hypocentral distance less than approximately 150 km (e.g., Sato et al. 2012 Ch. 3; Takemoto et al. 2012), we employ the lapse times of 60–70 s for calculating averaged coda amplitudes. The time windows of τ-seconds, which represent the averaged pulse durations of P and S waves measured from the displacement waveforms at four F-net stations (filled triangles in Fig. 1a), were used to measure P- and S-wave amplitudes.
Figure 1c, d shows the measured P- and S-wave amplitudes and master attenuation curves as a function of the hypocentral distance for frequencies of 0.5–1 and 4–8 Hz. The color scale represents the magnitude of the P- and S-wave radiation pattern coefficients (|F P | and |F S |; Aki and Richards 2002) estimated from MT solutions in the one-dimensional (1D) crustal velocity structure (Ukawa et al. 1984), which is used in the Hi-net routine hypocenter analysis. The S-wave radiation pattern coefficient |F S | was calculated by RMS of SV- and SH-wave radiation pattern coefficients. Wavelengths in Fig. 1c, d (λ P and λ S , respectively) were calculated by using the central frequencies of each band and seismic velocities in the crust. Observed amplitudes are scattered around master attenuation curves, reflecting the effects of non-isotropic source radiation and fluctuation of amplitude due to small-scale velocity inhomogeneity along propagation path (e.g., Hoshiba 2000; Yoshimoto et al. 2015).
The scatter due to non-isotropic source radiation is most evident in the P-wave amplitudes for the lowest frequency (0.5–1 Hz; left side of Fig. 1c). We confirmed that P-wave amplitudes with larger/smaller |F P | values tend to distribute above/below the master attenuation curve, respectively. As the frequency increased (4–8 Hz; right of Fig. 1c), this tendency become unclear, implying that P-wave amplitudes at higher frequencies do not show a clear four-lobe apparent radiation pattern. Although similar behaviors appeared in the S-wave amplitudes (Fig. 1d), the four-lobe patterns become unclear more rapidly compared to the P-wave ones (Fig. 1c).
Frequency and distance dependences in the apparent radiation pattern
To quantify distortion of the apparent radiation pattern from double-couple point source predictions, we simply calculated the cross-correlation coefficient (CCC) between the observation and theoretical amplitude fluctuations using moving hypocentral distance windows (40–70, 50–80, 60–90, 70–100, 80–110, 90–120, and 100–130 km).
The values of r and CCC0 were determined as −0.38 ± 0.023 and 1.35 ± 0.053, respectively, by a least squares estimation for the range of log(kL) < 2.85. The observed CCC could be described by using resultant Eq. (3) (blue line in Fig. 3). We here introduce a set of values k 1 L 1, k 2 L 2, and k 3 L 3 from log(k i L i ) = 0.92, 2.85, and 3.55 (i = 1, 2, 3), respectively, from Fig. 3, for the following discussions.
Apparent radiation pattern modeling and amplitude predictions
Figure 4a, b shows the spatial distributions of modeled apparent radiation pattern coefficients for frequencies of 0.5–1 and 4–8 Hz, respectively. We also show the spatial distribution of the radiation pattern coefficient for a double-couple point source in a homogeneous medium as a reference (Fig. 4c), where amplitude nodes (R j = 0.00) clearly exist. The azimuthal difference of modeled apparent radiation pattern coefficients became unclear with increasing distances and wave numbers.
Satoh (2014) employed frequency- and distance-dependent S-wave radiation pattern coefficient for the stochastic Green’s function method based on empirical model of Satoh (2002a, b), which showed very weak distance dependency for frequencies of 2–5 Hz and an isotropic radiation pattern for higher (>6 Hz) frequencies irrespective of distance. Satoh (2002b) constructed this model via observed energy partitioning of S waves in each horizontal component to reduce source and site amplification effects, rather than spatial distribution of maximum amplitude. Although our model does not include directivity effects, our model practically succeeds in reproducing observed spatial distribution of maximum amplitude of small-to-moderate local crustal earthquakes compared to Satoh (2002b)’s model (Additional file 1: Figure S2). This difference may be caused by difference in the method for model construction.
We investigated the frequency and distance dependences in the apparent radiation pattern for both P and S waves during local crustal earthquakes. We demonstrated how the four-lobe apparent radiation pattern, which is expected from a double-couple point source, is gradually distorted with increasing frequency and distance. The observed distortions have common decay pattern for P and S waves and could be characterized by the normalized hypocentral distance kL. These results suggest that major causes of frequency- and distance-dependent distortion of the apparent radiation pattern are seismic wave scattering and diffraction in the heterogeneous crust.
The observed frequency and distance dependences in the apparent radiation patterns for both P and S waves could be simply modeled by using a linear function of log(kL). On the basis of this, we proposed a method for prediction of the spatial distributions of maximum P- and S-wave amplitudes. Our method, which incorporates frequency- and distance-dependent characteristics of the observed apparent radiation pattern, successfully reproduced the observed spatial distributions of P- and S-wave amplitudes during small-to-moderate local crustal earthquakes.
Our method could also provide better insights into source rupture process and practical correction for the effects of the apparent radiation pattern. In future study, this would enable us to estimate the radiated source energy precisely and to obtain better insights into high-frequency seismic sources, such as small earthquakes and non-volcanic/volcanic tremors and the other effects, especially rupture directivity and fluctuation of maximum amplitudes, will be taken into consideration within our method.
ST developed basic idea of this work, proposed a new method for amplitude prediction, and drafted this manuscript. MK analyzed Hi-net/F-net waveform data and examined the characteristics of the apparent radiation pattern. KY made comments on seismic wave scattering and helped drafting. All authors read and approved the final manuscript.
Hi-net and F-net waveform data and the MT solutions from the F-net are available via the Web site of the National Research Institute for Earth Science and Disaster Resilience, Japan. We used the unified hypocentral catalog provided by the Japan Meteorological Agency (JMA), which is available via the JMA server. The frequency response of the short-period Hi-net sensors with a natural frequency of 1 Hz was corrected to make a broadband record using the program of Maeda et al. (2011). Generic Mapping Tools (Wessel and Smith 1998) and Seismic Analysis Code (SAC) were used for making the figures and conducting the signal processing work, respectively. We also thank two anonymous reviewers and the editor Prof. H. Takenaka for careful reading and constructive comments that improved the manuscript.
The authors declare that they have no competing interests.
- Aki K, Richards P (1980) Quantitative seismology. University Science Books, Freeman, San FranciscoGoogle Scholar
- Aki K, Richards P (2002) Quantitative seismology, 2nd edn. University Science Books, Freeman, San FranciscoGoogle Scholar
- Boore DM, Boatwright J (1984) Average body-wave radiation coefficients. Bull Seismol Soc Am 74:1615–1621Google Scholar
- Fukuyama E, Ishida M, Dreger DS, Kawai H (1998) Automated seismic moment tensor determination by using on-line broadband seismic waveforms. Zisin 51:149–156 (in Japanese with English abstract) Google Scholar
- Kobayashi M, Takemura S, Yoshimoto K (2015) Frequency and distance changes in the apparent P-wave radiation pattern: effects of seismic wave scattering in the crust inferred from dense seismic observations and numerical simulations. Geophys J Int 202:1895–1907. doi: 10.1093/gji/ggv263 CrossRefGoogle Scholar
- Liu H, Helmberger DV (1985) The 23:19 aftershock of the 15 October 1979 Imperial Valley earthquake: more evidence for an asperity. Bull Seismol Soc Am 75:689–708Google Scholar
- Maeda T, Obara K (2009) Spatiotemporal distribution of seismic energy radiation from low-frequency tremor in western Shikoku, Japan. J Geophys Res 114, B00A09, doi: 10.1029/2008JB006043
- Okada Y, Kasahara K, Hori S, Obara K, Sekiguchi S, Fujiwara H, Yamamoto A (2004) Recent progress of seismic observation networks in Japan—Hi-net, F-net, K-NET and KiK-net. Earth Planets Space, 56, xv–xxviii, doi: 10.1186/BF03353076
- Satoh T (2002b) Radiation pattern and f max of the Tottori-ken Seibu earthquake and the aftershocks inferred from KiK-net strong motion records. J. Struct Constr Eng AIJ 556:25–34 (in Japanese with English abstract) Google Scholar
- Si H, Midorikawa S (1999) New attenuation relationships for peak ground acceleration and velocity considering effects of fault type and site condition. J Struct Constr Eng AIJ 523:63–70 (in Japanese with English abstract) Google Scholar
- Ukawa M, Ishida M, Matsumura S, Kasahara K (1984) Hypocenter determination method of the Kanto-Tokai observational network for microearthquakes. Res Notes Natl Res Cent Disaster Prev 53:1–88 (in Japanese with English abstract) Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.