- Open Access
Source parameters of the 2011 Yellow Sea earthquake (ML 5.3): Different features from earthquakes on the Korean Peninsula
Earth, Planets and Space volume 64, pages379–388(2012)
A moderate earthquake of magnitude ML 5.3 occurred in the Yellow Sea on January 12, 2011. We estimated the source parameters and found that the quake was a shallow strike-slip fault event with a moment magnitude of 4.6. The stress drop of this event, 1.2–2.0 MPa, is lower than that of moderate earthquakes inland and at the eastern offshore of the Korean Peninsula, and also that of the typical value for shallow intraplate earthquakes. A stronger event (M ∼ 6) in the southern Yellow Sea in 1984 was previously reported to have a low stress drop. Therefore the low stress drop is probably characteristic of earthquakes in the Yellow Sea region. We found that aftershocks of the 2011 Yellow Sea event, with magnitudes greater than 2, occurred for about 5 days, while similar-sized aftershocks of some inland earthquakes of the Korean Peninsula continued for several hours only. The lower stress drop and greater active aftershocks in the Yellow Sea region might reflect a different tectonic setting from that on the Korean Peninsula.
A moderate earthquake of magnitude ML 5.3 occurred in the Yellow Sea on January 12, 2011, according to the Korea Meteorological Administration (KMA) catalog. The epicenter was located over 200 km away from both the Korean Peninsula and China (Fig. 1); however, this event was large enough to be felt in the southwestern region of South Korea as well as in Shanghai, China. Figure 1 shows the epicenters of shallow earthquakes (shallower than 70 km) with magnitudes greater than 4 in the last 30 years (1981–2010) according the International Seismological Centre (ISC) catalog. Many earthquakes are located along plate boundaries around Japan and Taiwan, and in the western part of China, with a cluster related to the 2008 Wenchuan earthquake. Seismic activity around the Korean Peninsula is very low compared to that in the surrounding regions, such as Japan and China. Several events occurred near the eastern coast of the peninsula, and some more events occurred in the Yellow Sea. The region where the 2011 Yellow Sea event occurred does not seem to be very seismically active. Even though there are several events in the Yellow Sea, there does not seem to be a considerable difference in seismic activity on the Korean Peninsula and in the Yellow Sea.
We have assessed seismic activity in the last 30 years around the Korean Peninsula and the Yellow Sea; this activity is illustrated in Fig. 2. The circles indicate shallow earthquakes reported by the ISC; the sizes of the circles reflect the earthquake magnitude. Figure 2 also shows seismicity (denoted by diamonds) determined by the KMA during the same time period. The size of the diamonds depends on the magnitude. Earthquakes recorded as having a magnitude greater than 5 in both catalogs, except those in Japan, are indicated by bold symbols along with the year and magnitude. Earthquakes mentioned in this paper are shown with the date (yyyy/mm/dd) and magnitude, and are listed in Table 1. None of the events on land in South Korea is greater than M 5 (where M refers to Magnitude) and some events in the M 5 class occurred offshore. Many other small events within the seismic network were also observed by the KMA. If we assume that the rate of seismicity is similar between small events (M < 5) and larger events (M > 5) in the two regions, inland of the Korean Peninsula and the Yellow Sea, we may expect more small earthquakes in the Yellow Sea. These earthquakes might be undetected because of their small size and their large distance from seismic networks.
The Korean Peninsula and northeastern China are thought to be located on the Amur Plate (Heki, 1999; Bird, 2003; Petit and Fournier, 2005). The boundaries of the Amur Plate suggested by Bird (2003) and Petit and Fournier (2005) are shown in Fig. 1. As can be seen in the figure, these suggested boundaries are mostly similar but are vastly different around the Yellow Sea and Mongolia. The 2011 Yellow Sea event occurred along the boundary of the Amur Plate suggested by Bird (2003) but was slightly distant from that suggested by Petit and Fournier (2005). Although we cannot judge which proposed boundary is correct, tectonic conditions around the 2011 Yellow Sea event might be different from those on the Korean Peninsula. If so, the earthquake generation mechanism and the seismic features between the two regions, near the 2011 event and on the Korean Peninsula, may be different.
The 2007 Odaesan earthquake, which is one of the largest inland events to have occurred in South Korea in the last 30 years, has been extensively researched (Jo and Baag, 2007; Kim and Park, 2010; Kim et al., 2010; Park and Hahm, 2010). Furthermore, several studies have investigated the source parameters of earthquakes around the Korean Peninsula (Park and Mori, 2005; Park et al., 2007; Choi, 2009, 2010; Jun and Jeon, 2010). In this study, we have analyzed the source parameters of the 2011 Yellow Sea event and compared them to those of Korean earthquakes, including the 2007 Odaesan event, to determine the differences, if any, in the characteristics of earthquake generation in the Yellow Sea and the Korean Peninsula.
2. Source Parameters
To understand the source parameters of the 2011 Yellow Sea event, we first relocated the epicenter. Because the earthquake occurred far from the seismic network of the KMA, we added some F-net stations and two stations in China to the stations of the KMA. In all, thirteen stations were used: the station distribution is given in Table 2 and Fig. 3 (open inverted triangles). To determine the hypocenter, we used HYPOELLIPSE (Lahr, 1980). For using this method, it is necessary to identify the first arrival of the seismic waves, without distinguishing between the direct P or Pn phases. For this particular event, we identified Pn phases at the 13 stations. We also employed the velocity structure proposed by Chang and Baag (2006) for all analyses in this study. This velocity model is driven for the Korean Peninsula and divided into 4 layers without oceanic crust. Since we used this structure for stations in Japan and China, as well as in South Korea, there may be some uncertainties related to calculating the theoretical arrival time, leading to some errors in the determined hypocenter. As the result of hypocenter determination, uncertainties in epicenter and focal depth were calculated to be 1.8 and 33.3 of 68% confidence limit for the horizontal direction (SHE) and for the depth (SEZ), respectively. The root mean square (rms) was 0.4177. Uncertainty, especially in the depth, seems to be large.
The estimated hypocenter is shown in Table 3 and Fig. 3, comparing the results determined by the KMA (cross), the U.S. Geological Survey (USGS; circle) and the China Earthquake Administration (CEA; diamond). The epicenter obtained in this study (star) seems to be closer to that of the USGS and the CEA. The difference in the epicenter locations may be due to differences of the stations used, velocity structures and hypocenter determination methods. We obtained the focal depth as 15 km. The USGS and the CEA estimated the depth as 10 km. Although the depths are a little different, we may consider this earthquake to be a shallow crustal event.
Using the hypocenter estimated in this study, we analyzed moment tensor solutions. A time domain waveform inversion technique (Dreger and Helmberger, 1993; Pasyanos et al., 1996), and three KMA stations (solid inverted triangles in Fig. 3), were used. We tested another combination of stations; however, these three stations provided the best solution. Figure 4 shows the moment tensor solutions. Although variance reduction is somewhat low (54.1%), the determined focal mechanism seemed to explain the polarities of the first P phase. The focal mechanism obtained here is nearly a strike-slip motion, with the compression axis trending NE-SW. This is consistent with other intraplate earthquakes around the Korean Peninsula (Chung and Brantley, 1989; Kim et al., 2004; Jun and Jeon, 2010). Seismic moment and moment magnitude are obtained as 9.94 × 1015 N m and 4.6, respectively. It is difficult to compare the magnitudes (Table 3) directly with each other because all three organizations, and this study, provided different types of magnitude. However, there may be a consistency that this event is a moderate earthquake with a magnitude of about 5.
Next, we estimated the fault radius, rise time, and stress drop of the 2011 Yellow Sea event. We deconvolved the source time function using an empirical Green’s function method (Mori and Frankel, 1990; Park and Mori, 2005; Park and Hahm, 2010). To use this method, a small earthquake is needed for the empirical Green’s function. Since none of the aftershocks or foreshocks was reported, we identified them from continuous waveform data at the HUK station (Fig. 3). Three more stations, JDO, GOS and JJU, at similar epicentral distances (Fig. 3), were also used to search for the small events. However, they had lower signal-to-noise ratios (JDO, GOS), or an ambient noise with similar periods as earthquake signals (JJU). We identified a small earthquake as a foreshock or an aftershock if it was observed at other stations (JDO, GOS, or JJU) as well, or if it had a similar P-S time and shape of waveforms to the mainshock. We looked at waveforms of 5 and 10 days long before, and after, the mainshock, respectively, and found no foreshocks and 16 aftershocks.
Figure 5 shows the velocity waveforms of 16 aftershocks (AF.1–16) recorded at the HUK station. Waveforms in Fig. 5 are bandpass filtered at 0.5–5 Hz and aligned from the arrival of Pn at 10 s, except aftershocks numbered from 2 to 4 (AF.2–4) which are aligned arbitrarily. The first and largest aftershock, AF.1, occurred about three minutes after the mainshock and was followed by three aftershocks, AF.2–4. AF.2–4 occurred sequentially and were contaminated by the waveforms of the previous events. However, we can still identify similar phases in the S-wave train (see the enlargement on the right below). Their S-wave train has mainly four pulses that are identified by dots over the waveforms. Also the correlation coefficients between the AF.1 event and each aftershock of AF.2–AF.5 were larger than about 60%. The correlation coefficient between the mainshock and AF.1 was about 74%. We can also recognize similar later phases in the waveforms of the mainshock and AF.1 that were generated by wave propagation along the same ray path. At the end of each trace, peak-to-peak amplitude of S waves (unit: m/s) and magnitude are shown. Here, magnitude was estimated as M = M (mainshock) – log10[Amp(mainshock)/Amp(aftershock)], where Amp is the amplitude of (peak-to-peak)/2.
We used waveforms of aftershock number 6 (AF.6) as an empirical Green’s function for the deconvolution. The focal depth was assumed to be 15 km, as obtained previously, and the shear wave velocity and rupture speed (80% of S-wave velocity) were assumed to be 3.79 and 3.03 km/s, respectively, using the velocity structure of Chang and Baag (2006). Direct S-waves of the mainshock were deconvolved using those of AF.6, and the result is shown in Fig. 6. Applying the formula from Boatwright (1980), we determined that the rise time was 0.2 s and the fault radius was 1.5–1.8 km. Using Brune’s formula (Brune, 1970), the stress drop was estimated to be 1.2–2.0 MPa. The dynamic stress drop of the 2007 Odaesan event obtained by Jo and Baag (2007) was about 8 MPa, and the static stress drop was estimated to be more than 20 MPa by Kim et al. (2010) and Park and Hahm (2010). The stress drop of the 2011 Yellow Sea event obtained in this study seems to be lower than both that of the 2007 Odaesan event and the standard value for the shallow intraplate earthquakes, which is about 10 MPa (Ruff, 2002).
3. Seismic Activities
According to the KMA catalog, two earthquakes in 1991 and 1993 were observed in the region where the 2011 Yellow Sea event occurred (Table 1). Their epicenters are indicated by crosses with circles in Fig. 7(a). The radius of the circle represents the range of location error estimated by the KMA. The two events seem to occur within about 30 km of the 2011 event. The KMA was operating an analog short-period seismometer, S-13, at the GOS station, when the two events occurred. The waveforms are shown in Fig. 7(b), with those of the 2011 event. The interval of tick marks is 10 s and waveforms are aligned by the first arrival, which is the Pn phase. Although we do not know the absolute amplitudes it is obvious from the relative amplitudes, that the 1993 event is larger than the 1991 event, because the waveforms of the 1993 event are saturated while the full range of amplitudes of the 1991 event are recorded. The waveforms of the 2011 event recorded by a digital short-period seismometer, SS-1, were corrected to the response of S-13. It seems that the 1993 and 2011 events have relatively clearer direct P and S waves, while the amplitudes gradually increases with no distinctly recognizable phases in the waveforms of the 1991 event. This might represent that the 1993 and 2011 events propagated along a similar ray path.
As shown in Fig. 2, it seems that some moderate earthquakes have occurred in the region around the 2011 Yellow Sea event. The 1991 and 1993 events may be examples of a higher seismicity in this region relative to on the Korean Peninsula. The 2007 Odaesan event occurred in the region with the lowest seismicity (dotted ellipse in Fig. 2) on the peninsula. To see if there are some differences in seismic activity between the two regions, we investigated fore- and after-shock activities related to the two events. Figure 8 shows examples of aftershock waveforms of the two events. Seismograms in Fig. 8 show vertical velocity data from the 2007 event at the DGY station (a) and from the 2011 event at the HUK station (b). The data are 2000 s in duration and high-pass filtered at 5 Hz. The amplitude was magnified to see the small aftershocks. Many small phases, corresponding to aftershocks, are observed in Fig. 8(a). Because of the short epicentral distance of the DGY station (7.2 km) for the 2007 event, we were able to observe such small events. Meanwhile, in Fig. 8(b), aftershock waveforms of the 2011 event seem to have longer durations, and there are a small number of aftershocks visible. This is because the P-S time is long, and small aftershocks cannot be detected due to the long epicentral distance of the HUK station (215 km) for the 2011 event.
We measured the Peak Ground Accelerations (PGAs) of the fore-, main- and after-shocks of the two events and show them as a function of time in the top and bottom parts of Fig. 9(a). The triangle, crosses and circles represent fore-, main- and after-shocks, respectively. The horizontal axis indicates the time difference from the origin time of the mainshock in hours. We searched for fore- and after-shocks from continuous waveform data within 5 days before and after the mainshock. We used 100-Hz continuous data to identify earthquakes as small as possible. We found 27 foreshocks and 91 aftershocks of the 2007 event and no foreshocks and 16 aftershocks for the 2011 event.
To generalize the seismic activity on the Korean Peninsula, we looked at other earthquakes. Because 100-Hz continuous data have been available only since 2007, only two inland events could be analyzed. One is the 2008 Gongju event with a local magnitude of 3.4, and the other is the 2009 Andong event with a magnitude of 4.0. The earthquake parameters are listed in Table 1 and the epicenters are shown in Fig. 9(b). For these two events, we investigated fore- and after-shocks in the same way as the 2007 and 2011 events. For the 2008 Gongju event, we used the vertical seismogram at the TEJ station at an epicentral distance of 11.3 km and found 10 aftershocks. Unfortunately, about 60 hours of data were unavailable for this event and we could not detect any fore shocks, even if there might have been some. We used vertical waveform data from the ADO station of a distance of 22.9 km for the 2009 Andong event and found 12 aftershocks. There were no foreshocks for the 2009 event.
PGAs as a function of time for all four events are compared in Fig. 9(a). Foreshocks were observed only for the 2007 Odaesan event. Aftershocks of earthquakes on the Korean Peninsula followed within 3 days, while those of the 2011 Yellow Sea event followed by as long as 5 days. The last aftershock of the 2011 event that we observed occurred about 118 hours after the mainshock. We investigated seismograms of 5 more days for the 2011 event, but could not find any other aftershocks. Because the epicentral distances of the stations are quite different from each other, it is impossible to compare the absolute number of aftershocks. Instead we compare the number of aftershocks that are large enough to be detected at a distance of 215 km. The broken horizontal lines on each graph indicate the relative values of the PGA corresponding to ML 2.0, to the PGA of the mainshock assuming ML reported by the KMA. Considering events larger than ML 2.0, inland earthquakes of the Korean Peninsula have very few aftershocks within several hours. However, aftershocks of the 2011 Yellow Sea event continued for at least 5 days. This may indicate that the seismic activity is relatively higher in the region around the 2011 Yellow Sea event than on the Korean Peninsula.
4. Discussion and Conclusion
We estimated the source parameters of a moderate 2011 Yellow Sea earthquake in order to see if there are any differences in the characteristics of earthquake generation in the Yellow Sea and on the Korean Peninsula. The 2011 Yellow Sea event is a shallow strike-slip fault event and the moment magnitude was 4.6. The fault radius, rise time, and stress drop were determined to be 1.5–1.8 km, about 0.2 s and 1.2–2.0 MPa, respectively.
The stress drop of this event, 1.2–2.0 MPa, is lower than that of many earthquakes on the Korean Peninsula, including the Odaesan earthquake in 2007 at more than 8 MPa (Jo and Baag, 2007; Kim et al., 2010; Park and Hahm, 2010), a moderate earthquake in 2004 (2004 M 5.3 event in Fig. 2) at about 7–14 MPa (Park and Mori, 2005), and the 1996 quake (1996 M 4.8 event in Fig. 2) at about 14 MPa (Choi, 2009). Chung and Brantley (1989) determined the stress drop of the 1984 M 6.0 (Ms 6.3) Southern Yellow Sea event (Fig. 2) to be 4.2 MPa and summarized the values of some large earthquakes (M s ∼ 7) in northern China as 10–20 MPa. They suggested that the southern Yellow Sea region is characterized by short recurrence intervals while the northern China region has very long recurrence intervals. Additionally, they argued that short recurrence intervals and low stress drops reflect a lower material strength in the region of the southern Yellow Sea. The 2011 Yellow Sea event also has a low stress drop, which is lower than the typical value for shallow intraplate earthquakes, similar to the 1984 Southern Yellow Sea event. Meanwhile, earthquakes on the Korean Peninsula and just off the east coast seem to have higher stress drops, indicating that they are typical intraplate events. Therefore, the Yellow Sea region may be in a different geologic, or tectonic, setting from that of the Korean Peninsula or northern China. Choi (2010) estimated stress drops of two moderate Yellow Sea events (2003 M 4.9 and 2003 M 5.0 events in Fig. 2) to be about 10–30 MPa. So the southern part of the Yellow Sea may be characterized by low stress drop.
From examination of continuous waveform data following the 2011 Yellow Sea earthquake, we found 16 aftershocks with magnitudes greater than 2 that occurred over 5 days. However, following three earthquakes on the Korean Peninsula, aftershocks with a magnitude greater than 2 followed within several hours. The low level of aftershock activity and high stress drop on the Korean Peninsula may indicate that earthquakes are not so easily triggered on the Korean Peninsula than in the Yellow Sea. The Korean Peninsula seems to be different from the Yellow Sea, but is, however, similar to northern China.
Bird, P., An updated digital model of plate boundaries, Geochem. Geophys. Geosyst., 4, 1027, doi:10.1029/2001GC000252, 2003.
Boatwright, J. A., A spectral theory for circular seismic sources: simple estimates of source dimension, dynamic stress drops, and radiated energy, Bull. Seismol. Soc. Am., 70, 1–28, 1980.
Brune, J. N., Tectonic stress and the spectra of seismic shearwaves from earthquake, J. Geophys. Res., 75, 4997–5009, 1970.
Chang, S.-J. and C.-E. Baag, Crustal structure in southern Korea from joint analysis of regional broadband waveforms and travel times, Bull. Seismol. Soc. Am., 96, 856–870, 2006.
Choi, H., Source parameters for the December 13 1996 ML 4.5 Earthquake in Yeongwol, South Korea, J. Earthq. Eng. Soc. Korea, 69, 23–29, doi:10.5000/EESK.2009.13.5.023, 2009.
Choi, H., Source parameters of two moderate earthquakes at the Yellow Sea area in the Korean Peninsula on March 22 and 30, 2003, Jigu-Mulli-wa-Mulli-Tamsa, 13, 235–242, 2010.
Chung, W.-Y. and B. J. Brantley, The 1984 Southern Yellow Sea earthquake of eastern China: Source properties and seismotectonic implications for a stable continental area, Bull. Seismol. Soc. Am., 79, 1863–1882, 1989.
Dreger, D. S. and D. V. Helmberger, Determination of source parameters at regional distances with single station or sparse network data, J. Geophys. Res., 98, 8107–8125, 1993.
Heki, K., S. Miyazaki, H. Takahashi, M. Kasahara, F. Kimata, S. Miura, N. F. Vasilenko, A. Ivashchenko, and K.-D. An, The Amurian Plate motion and current plate kinematics in eastern Asia, J. Geophys. Res., 104, 29147–29155, 1999.
Jo, N. and C.-E. Baag, The 20 January 2007, Mw 4.5, Odaesan, Korea, earthquake, Geosci. J., 11, 51–58, 2007.
Jun, M.-S. and J. S. Jeon, Focal mechanism in and around the Korean Peninsula, Jigu-Mulli-wa-Mulli-Tamsa, 13, 198–202, 2010.
Kim, K.-H. and Y. Park, The 20 January 2007 ML 4.8 Odaesan earthquake and its implications for regional tectonics in Korea, Bull. Seismol. Soc. Am., 100, 1395–1405, 2010.
Kim, S. G., E. Lkhasuren, and P.-H. Park, The low seismic activity of the Korean Peninsula surrounded by high earthquake countries, J. Seismol., 8, 91–103, 2004.
Kim, W.-Y., H. Choi, and M. Noh, The 20 January 2007 Odaesan, Korea, earthquake sequence: Reactivation of a buried strike-slip fault?, Bull. Seismol. Soc. Am., 100, 1120–1137, 2010.
Lahr, J. C., HYPOELLIPSE/MULTICS: a computer program for determining local earthquake hypocentral parameters, magnitude, and first motion pattern, U.-S. Geol. Surv. Open File Rep., 59–80, 1980.
Mori, J. and A. Frankel, Source parameters for small events associated with the 1986 North Palm Springs, California, earthquake determined using empirical Green functions, Bull. Seismol. Soc. Am., 80, 278–295, 1990.
Park, J.-C., W. Kim, T. W. Chung, and C.-E. Baag, Focal mechanisms of recent earthquakes in the Southern Korean Peninsula, Geophys. J. Int., 169, 1103–1114, 2007.
Park, S.-C. and I.-K. Hahm, Rupture pattern of a moderate earthquake in Odaesan, Korea, Geosci. J., 14, 191–200, 2010.
Park, S.-C. and J. Mori, Source parameters of the May 29, 2004 South Korea earthquake (ML 5.2), Earth Planets Space, 57, 471–475, 2005.
Pasyanos, M. E., D. S. Dreger, and B. Romanowicz, Toward real time estimation of regional moment tensors, Bull. Seismol. Soc. Am., 86, 1255–1269, 1996.
Petit, C. and M. Fournier, Present-day velocity and stress fields of the Amurian Plate from thin-shell finite-element modeling, Geophys. J. Int., 160, 357–369, 2005.
Ruff, L. J., State of stress within the Earth, in International Handbook of Earthquake & Engineering Seismology (Part A), edited by Lee, W. H. K., et al., 539–558 pp, Academic Press, London, 2002.
We used waveform data observed at seismic stations of the Korea Meteorological Administration, Korea Institute of Geology, Mining and Materials (KIGAM) and Broadband Seismograph Network (F-net) operated by the National Research Institute for Earth Science and Disaster Prevention (NIED). Some of the figures were made using the Generic Mapping Tools (GMT). Moment tensors were computed using the MTpackageV2.1 package developed by Douglas Dreger of the Berkeley Seismological Laboratory, and relating Green’s functions were computed using the FKRPROG software developed by Chandan Saikia of URS. This study was supported by the project of ‘Study on Development and Application of Earthquake Monitoring Techniques’ by the National Institute of Meteorological Research, Korea Meteorological Administration. We thank the EPS reviewers and editor for their helpful comments to improve the manuscript.
About this article
Cite this article
Park, S., Kong, M., Park, E.H. et al. Source parameters of the 2011 Yellow Sea earthquake (ML 5.3): Different features from earthquakes on the Korean Peninsula. Earth Planet Sp 64, 379–388 (2012). https://doi.org/10.5047/eps.2011.12.001
- Moderate earthquake
- source parameter
- Yellow Sea
- Korean Peninsula