Volcanic-like low-frequency earthquakes beneath Osaka Bay in the absence of a volcano
2011; American Geophysical Union; Volume: 38; Issue: 8 Linguagem: Inglês
10.1029/2011gl046935
ISSN1944-8007
AutoresNaofumi Aso, Kazuaki Ohta, Satoshi Ide,
Tópico(s)Seismic Waves and Analysis
ResumoGeophysical Research LettersVolume 38, Issue 8 Solid EarthFree Access Volcanic-like low-frequency earthquakes beneath Osaka Bay in the absence of a volcano Naofumi Aso, Naofumi Aso [email protected] Department of Earth and Planetary Science, University of Tokyo, Tokyo, JapanSearch for more papers by this authorKazuaki Ohta, Kazuaki Ohta Department of Earth and Planetary Science, University of Tokyo, Tokyo, JapanSearch for more papers by this authorSatoshi Ide, Satoshi Ide Department of Earth and Planetary Science, University of Tokyo, Tokyo, JapanSearch for more papers by this author Naofumi Aso, Naofumi Aso [email protected] Department of Earth and Planetary Science, University of Tokyo, Tokyo, JapanSearch for more papers by this authorKazuaki Ohta, Kazuaki Ohta Department of Earth and Planetary Science, University of Tokyo, Tokyo, JapanSearch for more papers by this authorSatoshi Ide, Satoshi Ide Department of Earth and Planetary Science, University of Tokyo, Tokyo, JapanSearch for more papers by this author First published: 20 April 2011 https://doi.org/10.1029/2011GL046935Citations: 21AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat Abstract [1] Among the many deep low-frequency earthquakes (LFEs) recently discovered worldwide, LFEs beneath Osaka Bay, western Japan, are especially unusual. Their waveforms are monochromatic, resembling those of some volcanic LFEs, but there are no active volcanoes around. The area is close to but clearly distinct from a belt of tectonic LFEs, and is near the site of a large inland earthquake (the 1995 Kobe earthquake). To characterize the activity of these LFEs, we present an extensive catalog constructed using a matched filter analysis on continuous seismic records with template LFEs determined by the Japan Meteorological Agency. The relocated catalog of 1378 events over a period of 5 years shows spatially concentrated activity in two volumetric zones, with several active periods including successive tremor-like events. The magnitude–frequency statistics satisfy the Gutenberg–Richter law with a b-value of 2. Unlike tectonic LFEs, which are highly sensitive to tidal stress, the LFEs beneath Osaka Bay show no spectral peak in activity at tidal periods, and the overall pattern of the spectrum is similar to that of volcanic LFEs beneath Sakurajima, Japan. These findings suggest that the Osaka Bay LFEs are almost same as volcanic LFEs in origin, or at least related to fluid upwelling from the mantle. Key Points Many monochromatic LFEs and tremor-like episodes were found beneath Osaka Bay These LFEs are similar to volcanic LFEs as a result of the new rich catalog These LFEs must be related to water upwelling from the mantle in this area 1. Introduction [2] Low-frequency earthquakes (LFEs) are relatively small earthquakes (magnitude ∼1) that radiate seismic waves depleted in high-frequency radiation. Since about 2000, the Japan Meteorological Agency (JMA) has detected and located many LFEs as part of their routine monitoring. The JMA hypocenter catalog contains two major types of LFEs: volcanic and tectonic. The former have been discovered beneath many active volcanoes in Japan [e.g., Ukawa and Ohtake, 1987; Hasegawa and Yamamoto, 1994], whereas the latter are relatively new phenomena and are known as a constituent of episodic tremor and slip [e.g., Obara, 2002; Rogers and Dragert, 2003; Rubinstein et al., 2010]. While tectonic LFEs represent shear slip on the plate interface [Shelly et al., 2006; Ide et al., 2007], the significance of volcanic LFEs in regional tectonics remains poorly understood. [3] Both types of LFEs occur at depths of around 30 km, near the Mohorovicic discontinuity, either beneath volcanoes or on the plate interface; however, there are some exceptions to these modes of occurrence. Katsumata and Kamaya [2004] reported isolated LFEs located away from volcanoes and belts of tectonic LFEs. An example of such LFEs is that detected beneath the focal region of the 2000 Western Tottori earthquake [Ohmi and Obara, 2002]. Although no active volcano exists above these events, they are located close to a dormant Quaternary volcano; consequently, these LFEs may be categorized as volcanic. In fact, our investigations revealed that most of the isolated LFEs reported to date are located near Quaternary volcanoes. A major exception in this regard is the LFEs beneath Osaka Bay, which are the focus of the present study (Figure 1a). These events occur in an area located 100 km from the nearest Quaternary volcano, but close to the 1995 Kobe earthquake (Mw 6.8), which was the most destructive earthquake in Japan during the past 50 years. These observations give rise to a number of questions. Do hidden volcanoes exist beneath Osaka Bay? Do the LFEs behave like tectonic LFEs that occur on a plate boundary, or do they provide a missing link between tectonic and volcanic LFEs? To approach the answer to these questions, we investigated the hypocenter distribution, frequency–size statistics, and tidal sensitivity of the Osaka Bay LFEs. Figure 1Open in figure viewerPowerPoint (a) Distribution of low-frequency earthquakes (LFEs) and volcanoes in western Japan. Blue dots and green dots indicate tectonic LFEs on the subducting plate and inland LFEs (including volcanic LFEs), respectively. Triangles show Quaternary volcanoes [Committee for Catalog of Quaternary Volcanoes in Japan, 1999]. Yellow stars are epicenters of three LFEs represented in Figure 1b, LFEs beneath Izu-Ooshima [Ukawa and Ohtake, 1987], the 1995 Kobe earthquake, and the 2000 Western Tottori earthquake, respectively. (b) Representative (left) waveforms and (right) spectra of LFEs for (top) Osaka Bay LFEs that occurred at 14:17:05 (JST) on 9 November 2007, (middle) volcanic LFEs at Sakurajima at 23:04:49 (JST) on 23 October 2007, and (bottom) tectonic LFEs at Kii Peninsula at 12:16:43 (JST) on 21 March 2005. These events have similar magnitudes and were observed at similar hypocentral distances. In Figure 2b (left), three-component velocity waveforms are shown versus the time elapsed from the event origin time, and vertical lines show P- and S-wave arrivals. In Figure 2b (right), black lines show the power spectrum density for velocity waveforms of the east–west component within 10 seconds after the S-wave arrival, and gray lines show the background spectrum. Distinct peaks are marked by downward arrows. [4] Before studying the hypocenters, we briefly consider the characteristics of their waveforms (Figure 1b). Although they are located close to a belt of tectonic LFEs, their waveforms resemble those of some volcanic LFEs, being monochromatic, and their spectra show several distinct peaks. Therefore, the Osaka Bay LFEs resemble the monochromatic LFEs deep beneath volcanoes, such as Izu–Ooshima [Ukawa and Ohtake, 1987] and Sakurajima, and differ from tectonic LFEs. 2. Construction of an Event Catalog [5] JMA has detected about 150 events beneath Osaka Bay since 2000. We supplemented this catalog to obtain a more complete catalog of events. First, we detected events from continuous records using a matched filter analysis with template events in the JMA catalog. Then, we estimated their magnitudes and relocated their hypocenters. [6] The continuous records are three-component velocity seismograms recorded at 23 stations around Osaka Bay, part of the Hi-net of the National Research Institute for Earth Science and Disaster Prevention, Japan, for the 5-year period beginning in April 2004 (Figure 2). During this period, the JMA catalog contains 107 LFEs, which we used as template events. As preprocessing, original records were band-pass filtered between 2 and 8 Hz, and decimated to 20 samples per second. Figure 2Open in figure viewerPowerPoint (a) Spatial distribution of template low-frequency earthquakes (LFEs) in the Japan Meteorological Agency (JMA) catalog and LFEs detected in this study, shown in map view and in latitudinal and longitudinal cross-sections. Gray ellipses indicate the hypocenters of template LFEs, with error (1σ). Color dots indicate detected LFEs (the color indicates the magnitude). The bottom panel is a map around Osaka Bay. Hi-net stations used in this study are indicated by triangles; template LFEs in the JMA catalog are plotted as black dots. The rectangle indicates the area shown in the top panel. Station names are shown in Figure S2 of the auxiliary material. (b) Space–time plot of detected LFEs. Vertical axis represents latitude, and the colors of the symbols correspond to magnitude. (c) Tremor-like episode of LFEs on 25 November 2005. These waveforms show the velocity amplitude of the east–west component at 10 selected stations. [7] Matched filter analysis has been applied previously in the analysis of small earthquakes and LFEs [Gibbons and Ringdal, 2006; Shelly et al., 2007a]. This method detects similar waveforms to those of a template event as signals from an earthquake with a similar hypocenter and mechanism to those of the template event. The degree of similarity is measured by cross-correlation between template waveforms and a portion of the continuous records. For this purpose, we used the network correlation coefficient (NCC), which is the sum of the correlation coefficients of seismograms for all channels in the network, where the vector ui is a time series consisting of a seismogram in channel i around the arrival time, τ is the event origin time of the template event, and Ti is the travel time for the channel i. We adopted the Pearson product-moment correlation coefficient, as expressed with dot products (brackets), each of which takes a value between −1 and +1. The NCC is normalized using the number of channels n and the data length m. In particular, the summation increases the variance by a factor of n, and the variance of the correlation coefficient is empirically estimated to be 2/m, which is 1/m or 1 if all the data points are completely independent or completely dependent, respectively. [8] The time window employed for correlating waveforms is 8 seconds, beginning from 2 seconds before the theoretical arrival time calculated assuming a horizontally layered 1D structure, following Shelly et al. [2006]. The arrival times of P and S waves are referenced for vertical and horizontal components, respectively. To prevent the repeated detection of the signal from a single event, successive high NCC values with short intervals (< 2 seconds) were counted as a single event. [9] We detected events with NCC values above 8 (i.e., the correlation coefficient is eight times as large as the standard deviation), which in theory is sufficiently high to avoid false detections. The validity of this threshold value was also confirmed by the fact that the calculated distribution of NCC values deviates from a Gaussian distribution for an NCC value of 8 (Figure S1 of the auxiliary material). [10] The magnitude of each detected LFE was measured from the ratio of the maximum velocity amplitudes between the detected LFE and the template LFE. Assuming the self-similarity of the earthquake source [e.g., Aki and Richards, 2002], an increase in the velocity amplitude by a factor of 10 corresponds to the increase in the duration time by a factor of 10, so the seismic moment increases by a factor of 1000, which means a 2-unit increase in magnitude. Using this relationship, we estimated the magnitudes as follows: where Mtemplate and Mdetected are the magnitudes of the template and detected LFEs, respectively, and vtemplatemax and vdetectedmax are the maximum velocity amplitudes of the template and detected LFEs, respectively. It is possible that this magnitude is unrelated to seismic moment or seismic energy if the scaling of these LFEs is completely different from the self-similarity. Moreover, even for ordinary earthquakes, the relation breaks down at high frequencies due to attenuation. However, it appears that attenuation is minor in the low-frequency range considered in the present study. [11] We relocated the hypocenters of the template and detected events using the network correlation relocation method proposed by Ohta and Ide [2008], which utilizes the NCC to constrain the relative hypocenter locations. First, we determined the relative location and time of each event pair by maximizing NCC. Then, a linear inversion technique was employed to determine the hypocenters, consistent with the relative locations of all the event pairs. We preliminarily relocated the template LFEs, and then relocated the detected LFEs for each small subset partitioned using their templates. 3. Characteristics of LFE Activity [12] The matched filter with 107 template events detected more than 6600 new events over the 5-year period of analysis. Among these, 1378 events were detected by more than one template event, which we therefore consider reliable and suitable for further analysis, because any event detected by one template LFE should resemble another nearby template LFE to some extent, if it exists. This procedure excludes events detected by a single isolated template LFE; nevertheless, we expect that we covered nearly all the events above a certain magnitude that occur in dense LFE clusters. Although many events occur outside of these clusters, this catalog is suitable for the analysis of clustered events. [13] Figure 2a shows the spatial distribution of the detected events. Two large clouds of events found in the JMA catalog became some small clusters in our new rich catalog. The distribution of events is three-dimensional rather than two-dimensional (planar), perhaps representing a geological structure—similar three-dimensional distributions of volcanic LFEs are commonly considered to represent magma chambers or channels in the filed of volcanology. The distribution is clearly different from that of tectonic LFEs occurring on a plate boundary [Shelly et al., 2006]. The estimation errors of the hypocenter location are <100 m for 81% of the events, and <300 m for 93% of the events. [14] Figure 2b shows the temporal distribution of the detected events. The activity is higher in the northern zone than in the southern zone. The two zones have distinct active periods, with little overlap between them. Highly active periods are evident in Figure 2b as vertically aligned groups of dots, which indicate the occurrence of many events in a short time and with a wide range of hypocenter locations (within 1–2 km of each other). These episodes continue for several minutes (Figure 2c), resembling tremor, especially the deep tectonic tremor that comprises a swarm of LFEs in western Shikoku, Japan [Shelly et al., 2007a, 2007b]. However, there is no evidence for the migration of hypocenters, which is a widely observed feature of tectonic tremor worldwide. [15] The periodicity of the detected events was examined by evaluating a spectrum of the delta function sequence (Figure 3). We use the Fourier transformation of the Dirac delta functions at each event origin time tj, which is normalized to distribute according to the chi-square distribution. No tidal response was found for Osaka Bay LFEs, whereas tectonic LFEs in western Shikoku are highly sensitive to tide [Shelly et al., 2007a, 2007b; Nakata et al., 2008; Ide, 2010], as revealed in Figure 3, which shows large peaks at tidal periods (e.g., M2 and O1 in Figure 3). The statistical significance can be measured based on the fact that the density function of P(f) has a Chi-square distribution with two degrees of freedom in the case that the sequence is temporally random (i.e., a Poisson process). The two peaks of tectonic LFEs are too large to have been generated by a random process. Another significant spectral peak at a meteorological period of 24 hours may not necessarily indicate the occurrence of a periodic event, because the detection ability depends on the circadian signal-to-noise ratio. Figure 3Open in figure viewerPowerPoint Spectrum of the delta function sequence of low-frequency earthquakes (LFEs). The green line indicates Osaka Bay LFEs, the blue line indicates tectonic LFEs in western Shikoku analyzed by Ide [2010], and the red line indicates volcanic LFEs at Sakurajima. Gray horizontal lines indicate the 0.01% point and 99.99% point of chi-square distribution with 6 degrees of freedom (the values of (f), which is the 3-point summation of P(f) normalized to a common median, have a chi-square distribution with 6 degrees of freedom). Gray vertical lines indicate major tidal and meteorological periods. S2, M2, K1 and O1 indicate the solar semi-diurnal tide, the lunar semi-diurnal tide, the lunisolar diurnal tide, and the lunar diurnal tide, respectively; S1 is the meteorological diurnal period. [16] The spectrum of Osaka Bay LFEs appears to possess gaps at tidal frequencies where tectonic LFEs have large peaks, although we cannot assess the statistical significance of this trend. The overall behavior of the spectrum around the tidal peaks resembles that of volcanic LFEs at Sakurajima, which we calculated similarly using the origin times of 705 events in the JMA catalog over a period of about 10 years. Although more analyses are required before arriving at firm conclusions, the unusual behavior of the Osaka Bay LFEs may provide clues to the background processes that underlie these events. [17] Figure 4 shows the relation between the magnitude and number of LFEs. Although the range is fairly limited, the Osaka Bay LFEs seem to obey the Gutenberg–Richter law (G–R law) with a b-value of 2. Few studies have analyzed the frequency–size statistics of deep volcanic LFEs, probably because of the small number of events, although some works have analyzed B-type events, which is the small events originating at shallow depths (< 3 km) beneath volcanoes [Minakami, 1960]. These studies reported b-values greater than 1.8, indicating the existence of magma or gas [McNutt, 2005]. Although B-type events are shallow earthquakes and they also contain both BL–type (1–3 Hz) and BH–type (5–8 Hz) [Ishihara and Iguchi, 1989], meaning that the reason for high b-values may be different to that for deep LFEs, it is possible that the deep and shallow events represent similar phenomena. An important result of this study is that the b-value for the Osaka Bay LFEs is high (∼2), meaning that a similar analysis could be applied to deep LFEs in other regions. Using the same sequence of analyses, the comparison of b-values in various regional LFEs will be done in the future. Figure 4Open in figure viewerPowerPoint (a) Histogram of LFE number versus magnitude. The number of events is shown on a log scale. Dashed line represents an exponential function with an index of −2. (b) Semi-log plot of cumulative LFE number versus magnitude. Dashed line represents an exponential function with an index of −2. 4. Discussion and Conclusions [18] We analyzed the activity of LFEs beneath Osaka Bay. Their isolated location makes it difficult to judge whether they are volcanic LFEs, tectonic LFEs that occur on a plate boundary, or a new type of event. We detected more than 1000 new LFEs, some of which appear to be tremor sequences. There exist two active zones (north and south) with contrasting active periods. Our analysis of the new rich catalog indicates that the Osaka Bay LFEs have a three-dimensional distribution, obey the G–R Law with a b-value of 2, and show no tidal response. These properties indicate that the events are similar to volcanic LFEs rather than tectonic LFEs, although there are no volcanoes near Osaka Bay. [19] Because volcanoes occur where a mantle diapir is generated by water released by dehydration reactions at a plate boundary, such water (or another fluid of tectonic origin) may currently be generated at the plate boundary beneath Osaka Bay. A split in the subducting plate in this area [Ide et al., 2010] may act as a fluid conduit or source. The two LFE zones of the events, and the hypocenter of the Kobe earthquake, are aligned sub-parallel to the proposed split in the subducting plate (the split is oriented NNW–SSE). If fluid is currently upwelling at this split, the fluid may be short-lived or it may become another medium around the Mohorovicic discontinuity, because the hypocenter distribution has a limited vertical extent. To further examine the physical process of the Osaka Bay LFEs, it would be useful to obtain information on the focal mechanisms of these events. [20] Although our study was restricted to the Osaka Bay LFEs, the sequence of analyses presented in this study is potentially applicable to many similar phenomena; the only requirements are continuous records and a reference catalog. In the present case, we were able to demonstrate behavior according the G–R law and obtain a characteristic spectral profile, mainly because of the rich number of detected events. This approach shows promise in terms of overall understanding and classifying LFEs. Acknowledgments [21] We thank T. Ohminato for helpful discussion on volcanic LFEs. We also thank for constructive comments from Zhigang Peng and Justin Brown. This work was supported by JSPS KAKENHI (20340115) and MEXT KAKENHI (21107007). All the data were obtained from the NIED Hi-net data server. [22] The Editor thanks Zhigang Peng and Justin Brown for their assistance in evaluating this paper. Supporting Information Auxiliary material for this article contains two figures. Auxiliary material files may require downloading to a local drive depending on platform, browser, configuration, and size. To open auxiliary materials in a browser, click on the label. To download, Right-click and select "Save Target As…" (PC) or CTRL-click and select "Download Link to Disk" (Mac). Additional file information is provided in the readme.txt. Filename Description grl27955-sup-0001-readme.txtplain text document, 1.3 KB readme.txt grl27955-sup-0002-fs01.pdfPDF document, 1 MB Figure S1. The distribution of NCC. grl27955-sup-0003-fs02.pdfPDF document, 1 MB Figure S2. Detailed map around Osaka Bay. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article. References Aki, K., and P. G. Richards (2002), Quantitative Seismology, 2nd ed., Univ. Sci., Sausalito, Calif. Committee for Catalog of Quaternary Volcanoes in Japan (1999), Catalog of Quaternary volcanoes in Japan, Bull. Volcanol. Soc. Jpn., 44, 285– 289. Gibbons, S. J., and F. Ringdal (2006), The detection of low magnitude seismic events using array-based waveform correlation, Geophys. J. Int., 165, 149– 166, doi:10.1111/j.1365-246X.2006.02865.x. Hasegawa, A., and A. Yamamoto (1994), Deep, low-frequency microearthquakes in or around seismic low-velocity zones beneath active volcanoes in northeastern Japan, Tectonophysics, 233, 233– 252, doi:10.1016/0040-1951(94)90243-7. Ide, S. (2010), Striations, duration, migration and tidal response in deep tremor, Nature, 466, 356– 359, doi:10.1038/nature09251. Ide, S., G. C. Beroza, D. R. Shelly, and T. Uchide (2007), A scaling law for slow earthquakes, Nature, 447, 76– 79, doi:10.1038/nature05780. Ide, S., K. Shiomi, K. Mochizuki, T. Tonegawa, and G. Kimura (2010), Split Philippine Sea plate beneath Japan, Geophys. Res. Lett., 37, L21304, doi:10.1029/2010GL044585. Ishihara, K., and M. Iguchi (1989), The relationship between micro-earthquake swarms and volcanic activity at Sakurajima Volcano [in Japanese with English abstract], Ann. Disast. Prev. Res. Inst. Kyoto Univ., 32(B1), 1– 11. Katsumata, A., and N. Kamaya (2004), Low-frequency events away from volcanoes in the Japan islands [in Japanese with English abstract], J. Seismol. Soc. Jpn., 57, 11– 28. McNutt, S. R. (2005), Volcanic seismology, Annu. Rev. Earth Planet. Sci., 33, 461– 491, doi:10.1146/annurev.earth.33.092203.122459. Minakami, T. (1960), Fundamental research for predicting volcanic eruption (Part 1), Bull. Earthquake Res. Inst. Univ. Tokyo, 38, 497– 544. Nakata, R., N. Suda, and H. Tsuruoka (2008), Non-volcanic tremor resulting from the combined effect of Earth tides and slow slip events, Nat. Geosci., 1, 676– 678, doi:10.1038/ngeo288. Obara, K. (2002), Nonvolcanic deep tremor associated with subduction in southwest Japan, Science, 296, 1679– 1681, doi:10.1126/science.1070378. Ohmi, S., and K. Obara (2002), Deep low-frequency earthquakes beneath the focal region of the Mw 6.7 2000 western Tottori earthquake, Geophys. Res. Lett., 29(16), 1807, doi:10.1029/2001GL014469. Ohta, K., and S. Ide (2008), A precise hypocenter determination method using network correlation coefficients and its application to deep low frequency earthquakes, Earth Planets Space, 60, 877– 882. Rogers, G., and H. Dragert (2003), Episodic tremor and slip on the Cascadia subduction zone: The chatter of silent slip, Science, 300, 1942– 1943, doi:10.1126/science.1084783. Rubinstein, J. L., D. R. Shelly, and W. L. Ellsworth (2010), Non-volcanic tremor: A window into the roots of fault zones, in New Frontiers in Integrated Solid Earth Sciences, edited by S. Cloetingh, and J. Negendank, pp. 287– 314, Springer, New York, doi:10.1007/978-90-481-2737-5_8. Shelly, D. R., G. C. Beroza, S. Ide, and S. Nakamula (2006), Low-frequency earthquakes in Shikoku, Japan, and their relationship to episodic tremor and slip, Nature, 442, 188– 191, doi:10.1038/nature04931. Shelly, D. R., G. C. Beroza, and S. Ide (2007a), Non-volcanic tremor and low frequency earthquake swarms, Nature, 446, 305– 307, doi:10.1038/nature05666. Shelly, D. R., G. C. Beroza, and S. Ide (2007b), Complex evolution of transient slip derived from precise tremor locations in western Shikoku, Japan, Geochem. Geophys. Geosyst., 8, Q10014, doi:10.1029/2007GC001640. Ukawa, M., and M. Ohtake (1987), A monochromatic earthquake suggesting deep-seated magmatic activity beneath the Izu-Ooshima volcano, Japan, J. Geophys. Res., 92, 12,649– 12,663, doi:10.1029/JB092iB12p12649. Citing Literature Volume38, Issue828 April 2011 FiguresReferencesRelatedInformation
Referência(s)