Tectonophysics 241 (1995) 55-66
Strike-slip faulting on the northern margin of the South China Sea: evidence from two earthquakes offshore of Hainan Island, China, in December 1969 B a o - Z h u Wei, W a i - Y i n g C h u n g * ,1 Center for Earthquake Research and Information, The University of Memphis, Memphis, TN 38152, USA
Received 2 December 1993; revised version accepted 25 August 1994
Abstract In December 1969, two earthquakes (m b = 4.7 and 4.8) occurred in a rift basin offshore of Hainan Island, China, in a passive continental margin environment. By modeling short-period P and long-period SH waveforms recorded at WWSSN stations, we have determined the focal mechanisms and source parameters of these two events to be: strike 45 _+ 7°, dip 78 _+4°, rake - 6 _+ 7°, focal depth 8 _+ 3 km, seismic moment (1.73 _+ 0.4) x 1023 dyne cm and stress drop 56 _+6 bar for the 17 December event; and strike 36 _+6°, dip 60 _+4°, rake - 2 _+ 7°, focal depth 8 + 3 km, seismic moment (1.43 _+0.13) × 1023 dyne cm and stress drop 47 _+ 6 bar for the 20 December event. Results of relative location computed from differential arrival times indicate that the 20 December event is slightly deeper than the 17 December event but within the resolution limit of the waveform constraint. Analysis of the local tectonic setting, geometry of isoseismals, aftershock distribution and fault-plane solutions indicate that the 17 December event probably ruptured along a NE-SW-striking fault on the northern margin of the Songtao uplift in the Qiongdongnan basin. The 20 December event probably ruptured along another major NE-SW-striking fault on the northern margin of the Qiongdongnan basin. The fault motions of both earthquakes were almost pure left-lateral strike slip. P and T axes show that the epicentral area is subject to nearly horizontal N-S-directed compression and E-W-directed extension, consistent with previously observed stress field in the area. The two earthquakes represent reactivation of pre-existing zones of weakness in the basin by the present-day tectonic stress field. A review of earthquakes in the region indicates that strike-slip motion is characteristic of the seismotectonics of the northern margin of the South China Sea.
1. Introduction L o c a t e d o n the n o r t h e r n m a r g i n of the S o u t h C h i n a Sea, H a i n a n I s l a n d is a part of South
* Corresponding author. 1Also with the Department of Geological Sciences, the University of Memphis, Memphis, TN 38152, USA.
China, a stable c o n t i n e n t a l region as d e f i n e d by J o h n s t o n a n d M e t z g e r (1986) a n d J o h n s t o n a n d K a n t e r (1990). I n D e c e m b e r 1969, however, H a i n a n I s l a n d was s h a k e n by two offshore earthquakes (Fig. 1). I n spite of the fact that H a i n a n I s l a n d a n d its a d j a c e n t region are k n o w n as a n area of relatively low seismicity c o m p a r e d to m a n y parts of China, the largest k n o w n e a r t h q u a k e o n the island, the Q i o n g s h a n e a r t h q u a k e of 1605,
0040-1951/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0040-1951(94)00163-4
B.-Z. Wei, W.-Y. Chung / Tectonophysics 241 (1995)55-66
"~ The Earthquakes of December 1969
~r 12/17/19 Event :ff 12/20/69
," Bathymetric Line O M > _ 6 . 0 0 4.5
Fig. 1. Epicentral locations, geological structure sketch and seismicity on H a i n a n lsland and adjacent area. T h e bathymetric data are from C h e n et al. (1993). 14/WF and JWF stand for the W a n g w u - W e n j i a o fault and the Jianfengling-Wanling fault, respectively.
had an estimated magnitude of M s 7.5. Another major earthquake in the offshore region was the 1931 M, 6.8 earthquake, which occurred more
than 200 km east of the island in the South China Sea, a major marginal sea in the Western Pacific. Because of the low seismicity off the coast of
[~ 12/20/69 12117169 FAULT DEPRESSION 11, SEEPS "0" DRY
0 KM 40 I
t I p' i
Fig, 2. Geological structures and some related features in the epicentral area of the two earthquakes (redrawn from Chen et al., 1993), The two epicenters shown here are from relative location,
B.-Z. WeL W.-Y. Chung / Tectonophysics 241 (1995)55-66
Because of poor station coverage and lack of high-quality data, the two largest events of this earthquake sequence were not well studied. To our knowledge, no single event fault plane solutions and seismic moments were determined for the two events. Liu et al. (1983) divided some of the larger events in the sequence into two groups. The first group included the 17 D e c e m b e r event and the second included the 20 D e c e m b e r event. Two composite fault plane solutions were determined for the two groups of events (Figs. 3a and 5a). In this paper, we will first use data from the World Wide Standardized Seismograph Network (WWSSN) and a synthetic seismogram technique to investigate the source mechanisms, focal depths, seismic moments and stress drops of these two earthquakes. We will also determine the relative location of the 20 D e c e m b e r event with respect to the 17 D e c e m b e r event as well as discuss seismotectonics in the epicentral region after an integration of regional seismicity, tectonic setting and marine geology and geophysics.
South China and along the northern margin of the South China Sea, the seismotectonics of the region are not well known. Since 1979, extensive offshore oil exploration work has included a number of wells drilled near the epicentral area of the D e c e m b e r 1969 events (Fig. 2). A study of the two offshore Hainan earthquakes will improve our understanding of the seismotectonics of stable continental earthquakes as well as the seismic hazards for oil exploration in the area. According to the International Seismological Center (ISC), the first event occurred at 08:00:1.1 h 17 D e c e m b e r 1969; the epicenter was 18.11 + 0.069°N, 110.55 _+0.068°E, focal depth 33 km, and magnitude m b 4.7. An M s of 5.1 was reported by the Chinese seismologists (Gu et al., 1983). The second event took place at 02:09 : 13.2 h 20 D e c e m b e r 1969, epicenter at 18.39 + 0.088°N, 110.54 + 0.067°E, focal depth 33 km, and m b 4.8. An M s of 5.2 was determined by Gu et al. (1983). The 17 D e c e m b e r event was felt throughout the island. The 20 D e c e m b e r event, however, was only felt in the southeastern part of the island. According to data from a temporary local seismic network, 138 earthquakes with M >~ 1.5 occurred in the epicentral area from 17 December 1969 to 31 January 1970 (Work T e a m of Hainan Earthquake, 1990). (c)
2. Tectonics and seismicity South China is located in the southeastern part of the Eurasian continent and adjoins the
Short Period P-Waves, 12/17/69, mb=4.7 I
I0 S e e
¢=45 °, 6=78 °, ),=354 °, h = 8 k m Fig. 3. (a) Composite solution determined from a group of events including the 17 December event by Liu et al. (1983). (b) Single event solution from this study using first motions and body wave waveforms. (c) Comparison of observed (upper trace) and synthetic (lower trace) short-period P waves of the 17 December 1969 event. Fault strike (~), dip angle (g), slip angle (A) and source depth (h) are also indicated. The mechanisms are plotted using a lower hemisphere equal-area projection.
B.-Z. Wei, W.-Y Chung / Tectonophysics 241 (1995) 55-66
South China Sea. To the east of South China is the Philippine Sea plate, to the far northeast is the Pacific plate, and to the southwest is the I n d i a n - A u s t r a l i a n plate (Fig. 1). South China is subjected to the applied forces from these surrounding lithospheric plates simultaneously. The observed principal compressive stress axis from earthquake focal mechanisms shows an E - W to E S E - W N W trend in the northeastern part of South China, rotates to more or less n o r t h w e s t southeast in the southeastern part of South China and then to approximately n o r t h - s o u t h in the southern part of South China (Lin et al., 1980). The crust of Hainan Island experienced a number of tectonic movements in the geological past. Both vertical and horizontal differential movements can be observed easily in the field. The island can be divided into three parts by two E - W - t r e n d i n g m a j o r faults, the W a n g w u Wenjiao fault and the Jianfengling-Wanling fault (Fig. 1). Northwest-southeast-trending faults have been developed extensively in the northern part. But in the southern part and the adjacent offshore area, N E - S W - t r e n d i n g faults are widespread. The South China Sea has undergone at least three stages of rifting and two intervening stages of sea-floor spreading since the Early Cretaceous. The episodes of rifting and associated thermal activities began during the Late Cretaceous, the Late Eocene, and the late Early Miocene. The
rift system corresponding to the first episode trends northeast-southwest, whereas those of the second and third trend east-west. These two trends match the orientations of the major tectonic lineations in the basin (Ru and Pigott, 1986). The northern continental shelf of the South China Sea is underlain by four major marginal basins, the Beibuwan basin, the Yinggehai basin, the Qiondongnan basin and the Pearl River Mouth basin (Fig. 1). These basins are rift basins developed in the continental crust and are filled predominantly with Tertiary sediments. The two earthquakes of D e c e m b e r 1969 occurred in the Qiongdongnan basin. Structural development of this basin is closely associated with the spreading of the South China Sea. It is a typical passive marginal basin developed from rifting and regional subsidence. In the Late Miocene, sediment entered the basin from two source regions, from the north off Hainan Island and from the southwest. The southwestern source was abruptly terminated probably by regional tilting, whereas the northern source has continued supplying sediments (Chen et al., 1993). In more detail, Fig. 2 shows two major fault systems in the region, the N W - S E - t r e n d i n g No. 1 Fault system and the N E - S W - t r e n d i n g No. 2 Fault system. The latter passes the epicentral area of the two earthquakes. The grabens and half-grabens in the Qiongdongnan basin are mostly E N E - W S W trending, parallel to the No. 2 Fault system. Off-
Table 1 Seismicity in the northern part of the South China Sea ~ (10-22°N, 108-118°E; M >/4.5) Date
5 Apr. 1524 July 1600 b 13 July 1605 15 Dec. 1605 9 Sept. 1611 1618 b 12 Aug. 1653 13 June 1926 5 Jan. 1927 24 Oct. 1929 21 Sept. 1931
19.2°N, 21.6°N, 19.9°N. 19.9°N, 21.5°N. 20.0ON. 21.7°N, 20.0°N, 17.0ON. 22.0°N, 19.8°N,
5.0 4.8 7.5 6.0 5.5 5.1) 4.8 5.5 5.0 6.5 6.8
21 June 1932 25 Mar. 1950 7 Oct. 1965 4 Feb. 1966 27 Jan. 1969 25 July 1969 17 Dec. 1969 21/Dec. 1969 7 Dec. 1982 27 Jan. 1986
16.5°N 15.0°N 12.5°N 12.4°N 12.5°N 21.6°N 18.0°N 18.4°N 12.4°N 21.7°N
5.6 5.3 5.7 4.9 4.8 6.4 5.0 5.0 4.7 5.0
110.5°E ll0.3°E 110.5°E 110.5°E lll.3°E 110.1°E 110.2°E 116.5°E 118.0°E 118.0°E lI3.1°E
112.0°E 115.0°E 114.5°E 114.3°E 114.4°E 111.8°E 110.6°E 110.6°E 114.6°E 111.9°E
~ The seismicity data listed here are from Gu et al. (1983). (For earthquakes before 1900, all the epicenters are estimated on the basis of intensity data. After 1900, the epicenters are determi ne d from instrumental data.) b This earth quake was reported in Chinese historical literature but the exact date cannot he determined.
B.-Z. Wei, W.-Y. Chung /Tectonophysics 241 (1995) 55-66
BAG--~ ~'t/ 1144--`-/
Mo=l.73XlO23 dyne-cm Fig. 4. Comparison of observed (upper trace) and synthetic (lower trace) long-period SH waves of the 17 December 1969 event. Average seimic moment (M 0) is also indicated. The number by the side of synthetic trace times 1023 dyn cm is the seismic moment at that station.
four M s >/6.0 earthquakes since 1524 (Table 1). The largest earthquake in this region is the Qiongshan earthquake of 13 July 1605, M s = 7.5, on the northern part of the island. Magnitudes for historical events are estimated from intensity data. Both historical data and local seismic data (which are not presented in Fig. 1) show that the northern part of the island is characterized by a much higher seismicity compared to the southern part. Most earthquakes north of the 19°N parallel occurred around 1605.
3. Source parameter study shore drilling shows pre-Upper Oligocene sediments filling these grabens. The 17 December event took place at the northern edge of the N E - S W - t r e n d i n g Songtao uplift. The 20 December event occurred close to the No. 2 Fault system. Compared with other areas in South China, the historical seismicity of Hainan Island and adjacent area is relatively high. If we consider the region between 10-22°N and 108-118°E (Fig. 1), there have been seventeen 6.0 > M s/> 4.5 and
To determine the focal mechanisms and other source parameters of the two earthquakes, we use P and SH waveforms recorded by the WWSSN stations and the waveform modeling technique described by Helmberger and Engen (1980), Helmberger (1983) and Xu and Yao (1988). Six P-wave first motions, five short-period teleseismic (distance >~ 30 °) P waveforms, and two long-period SH waveforms with distance within 30 ° were used for the 17 December 1969 event, and nine P-wave
Short Period P-Waves, 12/20/69, mb=4.8
K B L ~
¢=36 °, 6=60°, ,k=358", h=8km Fig. 5. (a) Composite solution determined from a group of events including the 20 December event by Liu et al. (1983). (b) Single event solution from this study using first motions and body wave waveforms. (c) Comparison of observed (upper trace) and synthetic (lower trace) short-period P waves of the 20 December 1969 event.
B.-Z. Wei, VV.-~ Chung / Tectonophysics 241 (1995) 55-66
first motions, nine short-period teleseismic P waveforms and two long-period SH waveforms with distance within 30 ° for the 20 December 1969 event. Long-period P waveforms were not available because the two earthquakes were not large enough. Because the epicentral distances of stations C H G and B A G were within 12°, a model of a single crustal layer over a mantle half-space was utilized to compute synthetic seismograms of SH waves for each of the stations. The crustal and upper mantle models used for computing synthetic seismograms are the same as those used by Wei and Chung (1993). The parameters of the crustal layer are P wave velocity 6.2 k m / s , S wave velocity 3.5 k m / s and density 2.78 g / c m 3. The mantle half-space has a P wave velocity of 8.2 k m / s , an S wave velocity of 4.5 k m / s and a density of 3.4 g / c m 3. The average crustal thicknesses from the epicenter to C H G and BAG are 33 and 17 km, respectively. First we needed to find a starting model for computing synthetic seismograms. From the Pwave first motions available to us, we could not obtain well-constrained fault-plane solutions for the two earthquakes. Using focal mechanisms from previous studies consistent with our first motions as our starting mechanisms, we then estimated the focal depths and source time functions using short-period waveforms. The observed P and SH waveforms were fitted by forward modeling, searching through different ranges of strike, dip and rake. The best-matching waveforms consistent with P-wave first motions are presented in Figs. 3, 4, 5 and 6. The final focal mechanisms are shown in Figs. 3b, 5b and Table 2. Because short-period waveforms can be significantly affected by local structures, fitting them is often
30 See /~Io--1.43X102a dyne-era Fig. 6. Comparison of observed (upper trace) and synthetic (lower trace) SH waves of the 20 December 1969 event. The number by the side of synthetic trace times 1023 dyn cm is the seismic moment at that station.
difficult. From Figs. 3 and 5, however, the observed and synthetic waveforms fit reasonably well. The SH waveforms in Figs. 4 and 6 also fit very well. The azimuthal coverages of stations as shown in Figs. 3 and 4 are also adequate. Hence, our focal mechanisms should be reliable. The mechanisms of the two events are both strike-slip faulting (Figs. 3 and 5). Comparing our mechanisms with those from Liu et al. (1983), we find that, for the 17 December event, the differences in strike, dip angle and rake are all about 20 °. For the 20 December event, our solution is significantly different from their solution; the differences in strike, dip and rake are about 30, 10 and 180 °, respectively. The P and T axes determined from this study are found to be more consistent with previously observed stress field in the area. The ISC assigned focal depths of 33 km for the two earthquakes. The focal depths of earthquakes in South China are generally less than 25 km (Lin et al., 1980). For the two events, we obtain a focal depth of 8 _+ 3 km using teleseismic short-period waveforms. Because these wave-
Table 2 Focal mechanisms of the two offshore Hainan Island events Event
P axis Az.
17 D e c e m b e r 1969 2 0 D e c e m b e r 1969
45 136 36 127
78 84 60 88
354 192 358 210
B.-Z. We4 W.-Y. Chung / Tectonophysics 241 (1995) 55-66
forms are very sensitive to focal depths, our results should be more reliable. From the amplitudes of long-period SH waves, we obtain seismic moments (Aki, 1966), M 0, of (1.73 _+ 0.4) × 10 23 dyne cm for the 17 December 1969 event and (1.43 _+ 0.13)× 1023 dyne cm for the 20 December 1969 event. Although we used more shortperiod P waveforms than long-period SH waveforms, we computed the seismic moments using the long-period waves, which are usually considered to be more stable, and less affected by complex structure than short-period waves. From the seismic moments, we obtain moment magnitudes M w (Kanamori, 1977) of 4.8 and 4.7 for the 17 December and the 20 December events, respectively. Using the long- and short-period waveforms, we determined a trapezoidal source time function with a rise time ~- of 0.3 s, a rupture time t r of 0.5 s and a total duration T of 0.8 + 0.2 s for both earthquakes. Applying the method described in Geller (1976), Ebel et al. (1978) and Chung and Cipar (1983), we obtain fault radii a of 1.11 and 1.10 km for a circular fault and fault
110030 ' E
I / ~ \ \ ~ ' ~ ~ s ° ~ °
~Mk5.0 O 4.0~2.0). The larger stars are the epicenters after relative location; the smaller stars are the ISC locations. The magnitudes used in this figure are in the Chinese magnitude scale.
61 110 °
i (~ For 12/17/69 ~,~ For 12/20/69
I Event Event
Fig. 8. Isoseismals of the two offshore Hainan Island earthquakes (redrawn from Xie and Cai, 1983). The Chinese intensity scale consists of twelve levels, corresponding approximately to the modified Mercalli scale.
areas A of 3.85 and 3.82 km 2 for the 17 and 20 D e c e m b e r 1969 events, respectively. These yielded respective average dislocations D of 13.22 and 11.02 cm from the relation D =Mo/I~A if rigidity/z is 3.4 × 10 u d y n / c m 2. Respective average dislocation velocities of 44.1 and 36.7 c m / s are given by / ) = D / r for the two earthquakes. Average static stress drops are, respectively, 56 _+ 6 and 47 _+ 6 bar, derived from the formula Act = 7Mo/16a 3 (Keiles-Borok, 1959; Kanamori and Anderson, 1975). Strain drops of 1.64 × 10 - 4 and 1.38 × 10 - 4 w e r e derived from Ae = A~r//x. The respective strain energies released are 1.419 × 1019 and 0.983 × 1019 erg, computed from W = (AcrM0)/2/z, assuming that the residual stress after faulting is zero.
4. Relative hypocentral locations The errors in epicentral location of these two offshore Hainan Island events from ISC, as given in Section 1, are of the order of 8-10 km. Because of the inaccuracy of the earth model used for location and other nonrandom errors, the actual errors in epicentral location can be much larger than the nominal errors mentioned above.
B.-Z. Wei, W-Y. Chung/ Tectonophysics 241 (1995)55-66
In view of the significance of epicentral locations in terms of tectonic interpretations, we applied a relative hypocentral location technique (Chung and Kanamori, 1976; Chung and Liu, 1992) to the 20 D e c e m b e r 1969 event (or the second event) with respect to the 17 D e c e m b e r 1969 event (or the first event). We used arrival time data reported in the ISC Bulletin for 21 stations at teleseismic and regional distances. The epicenter of the second event is located N12°W with respect to that of the first event and is 24 km away. In other words, the 20 D e c e m b e r event is almost to the due north of the 17 D e c e m b e r event. This is consistent with the relative orientation from the ISC epicenters. The hypocenter of the second event is found to be approximately 6 km deeper than that of the first event. This difference in focal depth is within the uncertainty of the depth range determined from waveform modeling. The spatial separation of the ISC epicenters of these two events is approximately 38 km, which is larger than that obtained from relative location. Since the uncertainties of the ISC epicentral locations in the N - S direction are 8 and 10 km for the first and second events, respectively, we can move the first event 7 km toward the second event and
move the second event 7 km toward the first event to make the spatial separation of the two events consistent with the result from relative location. The epicentral locations of the two events after such an adjustment are shown in Figs. 2, 7 and 8.
Because the two earthquakes occurred in an offshore area, the local information available is relatively limited. Isoseismals of the two events are shown in Fig. 8. Even though the m b of the 17 D e c e m b e r event is smaller than that of the 20 D e c e m b e r event, the felt area of the first shock is larger than that of the second one. From the present study, we found that the 17 D e c e m b e r event actually has a larger M 0, M w and stress drop compared to the 20 D e c e m b e r event. Our results are more consistent with the observed difference in felt area between these two shocks. For the 20 D e c e m b e r event, the isoseismals elongate in the N E - S W direction, which is sub-parallel to one of the nodal planes of our mechanism. For the 17 D e c e m b e r event, however, an elonga-
Table 3 Focal mechanisms of seven major earthquakes on Hainan Island and adjacent area Event
Az. 1605 Qiongshan a (M~ = 7.5) 1918 Nanao a (M~ = 7.3) 1936 Lingshan a (M~ = 6.8) 1962 Heyuan ( M s = 5.3) 1965 Nanhai (M~ = 5.7) 1966 NE of Dongsha ( M s = 5.3) 1969 Yangjiang (M~ = 6.4)
313 51 310 216 19 111 158 67 240 45 145 243 170 263
66 72 70 80 86 65 84 77 50 41 79 55 76 79
160 25 11 160 335 184 13 174 100 78 324 194 349 194
References: BC = Brantley and Chung (1991); L L Z X = Lin et al. (1980); WGS = Wang et al. (1979). a The mechanisms of historical earthquakes were estimated from isoseismals and macroscopic data.
B.-Z. Wei, W.-Y. Chung/ Tectonophysics 241 (1995)55-66
tion direction is not obvious. The aftershock distribution from 17 D e c e m b e r 1969 to 23 February 1970 is drawn in Fig. 7 from the data in a Chinese report (Work T e a m of H a i n a n Earthquake, 1990). The aftershock distribution mainly in the Songtao depression and its vicinity is N E - S W trending. In Figs. 2 and 7 is shown that the 17 D e c e m b e r event occurred at the northern edge of the Songtao uplift, which is N E - S W trending, sub-parallel to the strike of one of the nodal planes in our focal mechanism and the elongation direction of the aftershock distribution. Therefore, we infer that a N E - S W - s t r i k i n g fault located at the northern edge of the Songtao uplift, probably is the seismogenic fault. For the 20 D e c e m b e r event, the N E - S W - s t r i k i n g nodal plane of our mechanism is parallel to the trend of the aftershock zone and the strike of the No. 2 Fault in the epicentral area (Figs. 2 and 7). Hence, we infer that the No. 2 Fault probably was the seismogenic fault of the earthquake. A special or odd feature associated with the the aftershocks in Fig. 7 is that it forms two trends which are offset by about 10 km and have the same shape (dashed line). While this could indicate two different faults, it could also represent some type of systematic mislocation. However, the original aftershock arrival time data are not available to us at this point; no further investigation about the details of the aftershocks can be m a d e in the present paper. According to local field mapping, N E - S W - s t r i k i n g faults are fairly c o m m o n on the southern part of Hainan Island (Wu, 1985). Hence, the strikes of the fault planes of the two events inferred from this study are consistent with the geology observed both on land and offshore. The fault motions of these two offshore Hainan events are almost pure left-lateral strike Slip. The 1605 Qiongshan, 1918 Nanao, 1936 Lingshan, 1962 Heyuan, 1966 northeast of the Dongsha Islands and 1969 Yangjiang earthquakes were also found to be strike slip (Table 3; Fig. 9; Wang et al., 1979; L i n e t al., 1980; Brantley and Chung, 1991). The source mechanisms of the 1605, 1918 and 1936 events are, of course, less reliable than those of other events in the table. However, their strike-slip natures a p p e a r to be consistent with the regional tectonics at each
411/19~ ~ I t A N
10/07/196~ N ANI'IAI
9/26/19¢~ NE OF 12ONGSItA
Fig. 9. Locations and focal mechanisms of several major earthquakes on the southeast coast of China. See Table 3 for more information about these events.
individual regions. Strike-slip faulting appears to be characteristic of the seismotectonic m o v e m e n t in the region. The Qiongdongnan basin was formed by rifting, which is usually associated with horizontal extenion and normal faulting. However, the two present left-lateral strike-slip ruptures can be interpreted as a result of a change in the tectonic stress field after the formation of the rift and subsequent reactivation of the preexisting faults in the old rifted basin by the present-day tectonic stress (Chung, 1983). In Figs. 2 and 7, the barbed lines represent thrust faults, an interpretation from prior investigators (Chen et al., 1993). Thrust faulting can occur as a result of compression at a later stage after rifting (Chung, 1993). However, strike slips can happen if the present-day compressional stresses are not perpendicular to the strike of the fault. Our results from the two offshore Hainan earthquakes indicate that the fault motions are predominantly strike slip. No significant thrust faulting have been revealed from our data. The P and T axes of earthquake mechanisms can deviate from the diretions of the maximum and minimum compressive stresses (~r1 and ~r3) (McKenzie, 1969; G e p h a r t and Forsyth, 1984). For midplate areas under compression, the aver-
B.-Z. Wei, P~-Y Chung / Tectonophysics 241 (1995) 55-66
age of P directions from various independent earthquakes has been found to be close to the o-1 direction of the regional stress field (Zoback, 1992). Hence, despite some uncertainties, P and T axes have been used to approximate the maxim u m and minimum compressive stresses. The P and T axes of the two events are oriented almost horizontally, trending approximately n o r t h - s o u t h and east-west. North-south-directed compression is consistent with other observations of the tectonic stress in the region (Lin et al., 1980), and also with the former N - S - o r i e n t e d spreading centered in the South China Sea (Fig. 1; Ru and Pigott, 1986). From Fig. 9, we can see that the P axes of the earthquakes which occurred near l l 0 ° E are N - S to N W - S E trending, such as the 1605 Qiongshan, 1936 Lingshan, 1969 Yangjiang and the two 1969 Hainan earthquakes. The 7 October 1965 Nanhai earthquake ( M s = 5.6) in the South China Sea is a thrust event (Wang et al., 1979), but the P axis of this earthquake is also N W - S E trending and is consistent with our P-axis orientation. The earthquake sequence associated with the two offshore Hainan Island events of 1969 did not occur as a mainshock-aftershock sequence and should be considered as a double shock sequence. A magnitude difference of 0.1 is obtained between the two largest shocks, much lower than the B,~th's law value of 1.2 (Richter, 1958; B~th, 1965). The stress drops of the events are approximately 50 bar. Kanamori and Anderson (1975), however, observed an average stress drop of about 100 bar for intraplate earthquakes near plate boundaries. Hence the stress drops of the two events a p p e a r to be lower than that of the average intraplate events. An earthquake sequence in South China with characteristics very different from the 1969 offshore Hainan Island sequence is the 1969 Yangjiang earthquake sequence (Table 3; Fig. 9; Brantley and Chung, 1991). This sequence is characterized by a single powerful mainshock with very few and small aftershocks. A magnitude difference of 2.7 is obtained between the mainshock and the largest aftershock, much higher than 1.2. The mainshock has a high stress drop of 380 bar. Even though these two earthquake sequences both occurred
along the continental margin, they show remarkable regional variations in earthquake characteristics. The temporal distribution of seismicity on Hainan Island and the northern part of the South China Sea appears to be rather uneven and is characterized by episodic activity. From Table 1 we can see that there have been clusters of activity separated by quiescent periods of variable length. During the period 1600-1618 five major earthquakes occurred in the northern part of the region, from 1926 to 1932 another five earthquakes took place, and between 1965 and 1969 there occurred another six earthquakes. The two offshore Hainan events occurred during the third active episode. The active periods total only 28 years or 6% of a time span of about 470 years.
The offshore Hainan Island earthquake sequence of D e c e m b e r 1969 represents a double shock sequence which occurred in the Qiongdongnan basin, an early Cenozoic rift basin on the passive continental margin in the northern part of the South China Sea. The 17 D e c e m b e r event ruptured a NE-SW-striking fault at the northern edge of the Songtao uplift. Results from relative location indicate that the 20 D e c e m b e r event is located approximately 24 km to the north (N12°W) of the 17 D e c e m b e r event, very close the NE-SW-striking No. 2 Fault at the northern boundary of the basin (Fig. 2). The fault motions of these two earthquakes are almost pure leftlateral strike slip. Strike-slip motion is found to be a major mode of tectonic movement in the region. Consistent with other regional observations, the epicentral area is subject to a nearly horizontal N-S-directed maximum compressional stress and an E - W - d i r e c t e d minimum compressional stress. Even though the 17 D e c e m b e r event has a smaller m b than the 20 D e c e m b e r event, the seismic moment, m o m e n t magnitude and stress drop of the first shock are determined to be larger than those of the second shock. Our results
B.-Z. Wei, W.-Y. Chung/ Tectonophysics 241 (1995)55-66
are consistent with the observations that the isoseismal areas of the first event are larger than those of the second event.
Acknowledgements We would like to thank Arch C. Johnston, James H. Dorman, Kazuya Fujita, Ann G. Metzger and B. Clark Burchfiel for critically reading the manuscript and providing suggestions for improvements. This research was supported by the Division of Earth Science, National Science Foundation, NSF Grant EAR-9105970 and also in part by a Faculty Research Grant of the University of Memphis. CERI contribution number 184.
References Aki, K., 1966. Generation and propagation of G waves from the Niigata earthquake of June 16, 1964. Bull. Earthquake Res. Inst., Tokyo Univ., 44: 23-88. B~th, M., 1965. Lateral inhomogeneities of the upper mantle. Tectonophysics, 2: 483-514. Brantley, B.J. and Chung, W.-Y., 1991. Body-wave waveform constraints on the source parameters of the Yangiiang, China, earthquake of July 25, 1969: a devastating earthquake in a stable continental region. Pure Appl. Geophys., 135: 529-543. Chen, P., Chen, Z.-Y. and Zhang, Q.-M., 1993. Sequence stratigraphy and continental margin development of the northwestern shelf of the South China Sea. Am. Assoc. Pet. Geol. Bull., 77: 842-862. Chung, W.-Y., 1983. Mode of strain release of three grabenassociated earthquakes. Tectonics, 2: 379-398. Chung, W.-Y., 1993. Source parameters of two rift-associated intraplate earthquakes in peninsular India: the Bhadrachalam earthquake of April 13, 1969 and the Broach earthquake of Mach 23, 1970. Tectonophysics, 255: 219-230. Chung, W.-Y. and Cipar, J.J., 1983. Source modeling of the Hsingtai, China earthquakes of March 1966. Phys. Earth Planet. Inter., 33: 111-125. Chung, W.-Y. and Kanamori, H., 1976. Source process and tectonic implications of the Spanish deep-focus earthquake of March 29, 1954. Phys. Earth Planet. Inter., 13: 85-96. Chung, W.-Y. and Liu, C., 1992. The reservoir-associated earthquakes of April 1983 in western Thailand: source modeling and implications for induced seismicity. Pure Appl. Geophys., 138: 17-41.
Ebel, J.E., Burdick, L.J. and Stewart, G.S., 1978. The source mechanism of the August 7, 1966 El Golfo earthquake. Bull. Seismol. Soc. Am., 68: 1281-1292. Geller, R.J., 1976. Scaling relations for earthquake source parameters and magnitudes. Bull. Seismol. Soc. Am., 66: 1501-1523. Gephart, J.W. and Forsyth, D.W., 1984. An improved method for determining the regional stress tensor using earthquake focal mechanism data: Application to the San Fernando earthquake sequence. J. Geophys. Res., 89: 93059320. Gu, G., Lin, T. and Shi, Z., 1983. Catalogue of Earthquakes in China (1831 B.C.-1969). Science Press, Beijing. Helmberger, D.V., 1983. Theory and application of synthetic seismograms. In: H. Kanamori and E. Boschi (Editors), Earthquakes: Observation, Theory and Interpretation. North-Holland Publishing Company, pp. 174-222. Helmberger, D.V. and Engen, G.R., 1980. Modeling the long-period body waves from shallow earthquakes at regional distances. Bull. Seismol. Soc. Am., 70: 1699-1714. Johnston, A.C. and Kanter, L.R., 1990. Stable continental earthquakes. Sci. Am., 262: 68-75. Johnston, A.C. and Metzger, A.G., 1986. The seismicity of stable continental interiors. Tenn. Earthquake Inf. Cent. Spec. Rep. 86-1. Kanamori, H., 1977. The energy release in great earthquakes. J. Geophys. Res., 82: 2981-2987. Kanamori, H. and Anderson, D.L, 1915. Theoretical basis of some empirical relations in seismology. Bull. Seismol. Soc. Am., 65: 1073-1095. Keiles-Borok, V., 1959. An estimation of the displacement in an earthquake source and of source dimensions. Ann. Geofis. (Rome), 12: 205-214. Lin, J.-Z., Liang, G.-Z., Zhao, Y. and Xie, M.-F., 1980. Focal mechanisms and tectonic stress field of coastal southeast China. Acta Seismol. Sin., 3:245-257 (in Chinese). Liu, P.-X., Huang, D.-Y., Yang M.-Y. and Zhang, L.-M., 1983. Types of earthquakes and focal mechanisms. Acta Geophys. Sin., 26:237-248 (in Chinese). McKenzie, D.P., 1969. The relation between fault plane solutions for earthquakes and the directions of principal stresses. Bull. Seismol. Soc. Am., 59: 591-601. Richter, C.F., 1958. Elementary Seismology. Freeman, San Francisco, CA, 763 pp. Ru, K. and Pigott, J.D., 1986. Episodic rifting and subsidence in the South China Sea. Am. Assoc. Pet. Geol. Bull., 70: 1136-1155. Wang, S.-C., Geller, R.J., Stein, S. and Taylor, B., 1979. An intraplate thrust earthquake in the South China Sea. J. Geophys. Res., 84: 5627-5631. Wei, B.-Z. and Chung, W.-Y., 1993. Regional waveform constraints on the source parameters of the Xunwu, China, earthquake of 2 August 1987, with implications for midplate seismotectonics. Phys. Earth Planet. Inter., 78: 57-68. Work Team of Hainan Earthquake, 1990. The Qiongdonghai, Hainan, Guangdong province earthquakes of December 1969: An investigation report of the December 1969
B.-Z. Wei, W.-Y. Chung / Tectonophysics 241 (1995) 55-66
Qiongdonghai, Hainan, Guangdong earthquakes. In: Z.-L. Shi, K.-W. Xu and Y.-F. Chen (Editors), China's Macroseismic Survey, Vol. II. Seismological Press, Beijing, pp. 288-299 (in Chinese). Wu, Z., 1985. Crustal movement and seismic characteristics of the Hainan island. Seismol. Geol., 7:71-76 (in Chinese). Xie, Y. and Cai, M., 1983. Compilation of Historical Materials of Chinese Earthquakes, Vol. 5. Science Press, Beijing (in Chinese).
Xu, D.-F. and Yao, Z.-X., 1988. Low-frequency response of the crustal media at regional ranges. Acta Geophys. Sin., 31:176-183 (in Chinese). Zoback, M.L., 1992. Stress field constraints on intraplate seismicity in eastern North America. J. Geophys. Res., 97: 11,761 - 11,782.