Small scale hot upwelling near the North Yellow Sea of eastern China

Transcription

1 GEOPHYSICAL RESEARCH LETTERS, VOL. 35, L20305, doi: /2008gl035269, 2008 Small scale hot upwelling near the North Yellow Sea of eastern China Yinshuang Ai, 1 Tianyu Zheng, 1 Weiwei Xu, 1 and Qiang Li 2 Received 11 July 2008; revised 30 August 2008; accepted 23 September 2008; published 22 October [1] In order to image the structure of the upper-mantle discontinuity beneath Eastern China, we have applied a common conversion point (CCP) stacking method of receiver function (RF). Both the 410-km and the 660-km discontinuities (hereafter called the 410 and the 660) clearly show continuous positive phases along the selected profile. The 410 shows depression, whereas the 660 shows uplift, in the eastern section of the study profile. The transition zone (TZ) to the west of longitude 122 is thicker than the global average, though the TZ is thinner to the east of that longitude. The thinnest part of the TZ, with 1015 km of thinning (an increase in temperature of up to 100 C), is located at the North Yellow Sea. We suggest that either the small-scale convection associated with the deep penetration of the sinking slab into the lower mantle, or a small plume from lower mantle, has generated hot upwelling in this region. Citation: Ai, Y., T. Zheng, W. Xu, and Q. Li (2008), Small scale hot upwelling near the North Yellow Sea of eastern China, Geophys. Res. Lett., 35, L20305, doi: / 2008GL Introduction [2] Located near the Western Pacific subduction zone, Eastern China is an ideal region to investigate the present status of the Pacific slab. Late Mesozoic and Cenozoic volcanic rocks, as well as many Cenozoic intraplate volcanoes, are widely distributed in this region (Figure 1). However, the sources of these volcanoes, which were thought to be formed by hotspots or back-arc extension, are a topic of debate [Lei and Zhao, 2005]. Recent global body wave tomography suggests that Pacific-subducted slabs tend to deflect or flatten near the bottom of the upper mantle, without direct penetration into the lower mantle beneath Eastern China [Fukao et al., 2001]. Regional P-wave tomography also demonstrates that the Pacific slab has become stagnant in the mantle transition zone and its western boundary is close to the south-north gravity lineament, which underlies Eastern China [Huang and Zhao, 2006]. In contrast, RF studies reveal a slightly different structure, showing evidence of slab accumulation in the TZ and local penetration into the lower mantle beneath Northeast China [Ai et al., 2003; Li and Yuan, 2003]. Thus, the status and dynamics of the section of the Pacific slab in the TZ beneath Eastern China remains controversial. 1 Key Laboratory of the Earth s Deep Interior, Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing, China. 2 Earthquake Administration of Liaoning Province, Shenyang, China. Copyright 2008 by the American Geophysical Union /08/2008GL [3] Knowledge of the topography of the upper-mantle discontinuities and the transition zone thickness (TZT) can help establish temperature constraints in the TZ. These temperature constraints may provide some data to facilitate the study of the dynamics between the subducting slab and the surrounding upper mantle. In subduction regions the 410 will be uplifted and the 660 depressed, while in hotspot regions the 410 will be depressed and the 660 uplifted, because of the differences in their phase transformations and Clapeyron slopes [Ito and Takahashi, 1989]. It has been found from both SS precursors surveys and RF studies, that in many subduction regions the depression of the 660 supports cold slab stagnant in the TZ [Houser et al., 2008; Niu et al., 2005]. [4] In this study, we use dense seismic data from the Liaoning experiment (Figure 1). We employ the common converted point (CCP) stacking of RFs [Zhu, 2002] to study the topography of the 410 and the 660, as well as the TZT, with high lateral resolution. Through detailed study of the upper mantle discontinuity structure beneath Eastern China, we can gain insight into the relationship between mantle discontinuities and thermal anomalies. Furthermore, the related studies may provide further evidence for the fate of the slab in the TZ beneath Eastern China. 2. Data Acquisition and Methodology [5] Two continuous seismic data sets, Liaoning Seismic Network (LSN) and Bohaiwan Seismic Network (BSN), were used in this study (Figure 1). LSN, deployed in 1999 by the earthquake administration of Liaoning Province, consists of four broadband and 12 short-period permanent stations (Figure 1). The data from LSN that have been used in this study span a period from July 1999 through December BSN, operated by the Institute of Geology and Geophysics, Chinese Academy of Sciences (IGGCAS) from May 2005 to May 2006, consists of 20 portable broadband stations with Guralp CMG-3ESP sensors and Reftek 72A digitizers. Teleseismic events with a magnitude greater than 5.8 in the 30 through 90 distance range, from all stations mentioned above, were selected for the RF analysis. [6] RF was isolated for each event-station pair in the following manner. First, each event was carefully checked for a clear first P wave arrival in three components. Second, the RFs were calculated from the selected data by the maximum entropy deconvolution method [Wu et al., 2003; Ai et al., 2003]. Then, a second-order zero-phase Butterworth bandpass filter with corner frequencies of Hz was applied to all RFs, in order to suppress high-frequency noise contamination. Finally, the RFs with clear P-to-S converted phase from the Moho and its multiple phases were selected. As a whole, we obtained 2248 reliable RFs that included 664 from BSN and 1584 from LSN. L of5

2 Figure 1. Map showing the study region in Eastern China and the seismic stations used in this study (see also the inset maps). Locations of seismic stations are shown by red squares for broadband stations of BSN, blue squares for (LSNB) broadband stations of LSN, and blue circles for (LSNS) short-period stations of LSN. The direction of the subducting Pacific slab is shown by the arrow. Black dots represent the piercing points of converted S waves at depths of 660 km. The locations of profile AA 0 is shown by the red line, which traverses the North Yellow Sea and the western portion of the Korean Peninsula. Three letter station codes are shown near each station. In the left inset map, lines, red stars, and red triangles denote the depth contours of the subducting Pacific slab along the Japan trench (the maximum depth of the most westward slab contour is 600 km), Mesozoic and Cenozoic volcanic rocks, and the intraplate volcanoes (CB, Changbai; XJD, Xianjindao; CUR, Ch Uga-Ryong, respectively [after Lei and Zhao, 2005]). Epicenter distribution of teleseismic events used in this study is shown as red dots in the right inset map. These RFs were calculated from 315 teleseismic events, with most events distributed to the southeast of the study region (Figure 1). [7] The modified CCP stacking method of RFs that takes into account the lateral heterogeneities of the velocity structure was used to constrain the lateral variations of the 410 and the 660, as well as the TZT beneath the study region. The CCP stacking method was performed as described in the following. To begin, the depth domain raypath of each RF was calculated on the basis of a known velocity model. Then, the amplitude at each point on the radial RF was distributed to its ray-path position, where the P-to-S conversion was assumed to exist. Finally, the CCP stacking was done along a 2-D profile, which was divided into bins. All amplitudes in the same bin were stacked and smoothed to the specific depth [Zhu, 2002]. For the purposes of this experiment, a Gaussian window along the stacking profile was used to smooth the CCP stacking results, where the smoothing weight for each RF depended on the distance to the center of the bin. 3. The Upper-Mantle Discontinuities and TZT [8] The laterally varying P-wave velocity model [Huang and Zhao, 2006] and the topography of the 410 and the 660 for the profile AA 0 are illustrated in Figure 2. The profile AA 0 is located in the south of the study region, with the eastern section nearer to the subducting slab than the western section. Because the CCP stacking in depth domain is constructed by time-to-depth conversion based on a defined velocity model, it is necessary to carefully construct the velocity model. We used different velocity models for CCP stacking. At depths from the surface to 50 km, we modified the 1-D IASP91 [Kennett and Engdahl, 1991] velocity structure based on the H-K stacking result [Zhu and Kanamori, 2000] for each station. At depths of 50 km to 2of5

3 Figure 2. (a) P-wave velocity perturbations relative to IASP91 model [Huang and Zhao, 2006] considered in the CCP stacking of RFs and (b, c) CCP stacked depth-domain images between the depths of 300 and 800 km for profile AA 0. Figure 2b is constructed using the P-wave velocity model shown in Figure 2a, the IASP91 S-wave velocity model, and a modified crustal structure for each station. Figure 2c is obtained from the same P-wave velocity model as Figure 2b and the S-wave velocity model of Fukao et al. [2001] within the TZ. For Figures 2b and 2c, the observed depths of discontinuities are denoted with blue dots. From west to east, the profile crosses the Bo Sea, the North Yellow Sea, and the West Korean Peninsula. Red squares and blue dots denote the number of RFs (NRF) in each bin at depths of 410 km and 660 km, respectively. 800 km, we used the 3-D P-wave velocity perturbation model of Huang and Zhao [2006] (with a 2-D slice shown in Figure 2a) to generate the P-wave velocity model, used the 1-D IASP91 S-wave velocity model, for constructing the image in Figure 2b. In Figure 2c, the P-wave velocity model was the same as that used in Figure 2b, and the S-wave velocity model was constructed by considering the results of Fukao et al. [2001], in which S-wave velocity has a 2% perturbation relative to the IASP91 model within the TZ in the whole study region. [9] We considered the Fresnel zone sizes of rays along the stacking profile and used different smoothing parameters to carry out the CCP stacking. The smoothing parameters used in Figures 2b and 2c are the minimum variability results. For instance, at a depth of 410 km, the bin size is a cuboid with 120 km in half width, approximately 110 km in half-length along the profile, and 0.5 km in thickness. At a depth of 660 km, the half-length of a cuboid is about 180 km. All RFs, whose piercing points at depths of 410 km and 660 km were within the above cuboids, were stacked by using a Gaussian window to smooth. The 410 and the 660 (the maximum amplitude part of the converted phase shown in Figure 2b and 2c, with blue dots) can be traced clearly as continuous positive phases along profile AA 0. The CCP images revealed that the depths of the 410 range gradually from 395 km to 413 km, at longitudes between and 126.0, and the depths of the 660 range gradually from 655 km to 643 km, at longitudes between and as shown in Figure 2b. In Figure 2c, the depths of the 410 range gradually from 396 km to 414 km, at longitudes between and 126.0, and the depths of the 660 range gradually from 664 km to 654 km, at longitudes between and In both cases, the 410 becomes deeper and the 660 becomes shallower near the North Yellow Sea. Besides the 410 and the 660, we also find a positive phase along the profile at depth of about 510 km which may indicate the 520-km discontinuity in this region (Figures 2b and 2c). [10] The TZT variation, which was less affected by the velocity structure both above 410 km and below 660 km, was considered to be another powerful indicator of the temperature anomaly in the TZ [Owens et al., 2000]. Hence, twelve CCP stacking lines parallel to profile AA 0 were selected, at latitudes from 38.0 to 43.5 with 0.5 spacing intervals, to calculate the TZT variation. The depths of the 410 and the 660 were determined by picking up the maximum positive amplitude in the bands of km and km in each bin along the stacking profiles. The TZT was calculated by subtracting the depth of the 660 from that of the 410, for each point. During the calculation, only bins with the number of receiver functions over 100 were selected. [11] Figures 3a and 3b were calculated on the basis of the same parameters used in Figures 2b and 2c, respectively. For both Figure 3a and Figure 3b, the TZT variations are a bit thicker than that of the global average of 250 km to the west of longitude 122, though they are thinner than that of the global average to the east of longitude 122. The thinnest part of the TZT was found from 230 km to 240 km in Figure 3a, from 240 km to 250 km in Figure 3b, which was located at the North Yellow Sea near the top of the subducted slab in the TZ (Figure 3). [12] Our study demonstrated that the TZT is thinning beneath the North Yellow Sea, which was near the tip of the subducting slab. Since the RF method is quite sensitive to sharp velocity contrasts, large S- and small P-wave velocity anomalies (or small S- and large P-wave velocity anomalies) in the TZ, the 660, and the TZT are probably calculated with some bias. Thus, the influence of unsuitable velocity models on the TZT calculation should be discussed in detail. With the 3-D P-wave velocity perturbation model used in this study, possible bias in our results could mainly come 3of5

4 et al., 2001]. In this case, the TZT in each bin becomes nine to eleven kilometres thicker than that in Figure 3a. However, the TZT are still about 10 km of thinning relative to the global average beneath the North Yellow Sea. [13] According to the relationship between tomographic images and slab age [Widiyantoro et al., 1999; Gorbatov and Kennett, 2003], the slab beneath the North Yellow Sea is younger than that beneath the western section of longitude 122. Therefore, compared with the S-wave anomaly in the TZ, the positive P-wave anomaly will dominate beneath the North Yellow Sea. Thus, if we consider the real 3-D S- wave velocity structure (if this model is available) in this region, the TZT should be in between Figures 3a and 3b. Further evidence that supports our results comes from a previous PASSCAL experiment in Northeastern China [Li and Yuan, 2003], which showed about 20 km of thinning in the TZ beneath the same region. Hence, we believe that the thinning TZ beneath the North Yellow Sea (Figure 3) is a reliable finding. Figure 3. TZT variations based on two different velocity models calculated on the basis of the same parameters used in (a) Figure 2b and (b) Figure 2c. The TZT is determined by subtracting the depth of the 660 from that of the 410 for each bin in the study profile. Contour lines denote the TZT variations. Crosses represent the piercing points of converted S waves at 660-km depth. See the captions in Figures 1 and 2 for further information. from the uncertainty in the S-wave structure. By only considering the lateral P-wave velocity model in Figure 2a, the TZT are about 20 km of thinning beneath the North Yellow Sea (Figure 3a). Because no detailed 3-D S-wave velocity model is available for this region, in Figure 2c we chose an extreme case of model in which the S-wave velocity relative to IASP91 has a 2.0% velocity anomaly in the TZ based on the previous tomographic images [Fukao 4. Discussion and Conclusions [14] West of longitude 122 beneath this study region (Figures 2 and 3), the 660 is relatively depressed and the TZT is greater than the global average of 250 km. This feature was confirmed by recent RF results [Chen et al., 2006]. In addition, south of latitude 37 in Eastern China the depressed 660 and thicker TZ were found to support the finding that the slab is cold and flat-lying in the TZ [Ai and Zheng, 2003]. In contrast, the uplifted 660 and the thinning TZ (by 15 km), which may represent 100 C higher temperature anomalies from mineral physics experiments [Bina and Helffrich, 1994], were found beneath the North Yellow Sea in this study (Figure 3). Due to the 410 did not depress too many kilometers in this study, it was unlikely that chemical heterogeneity within the TZ produced the thinning TZ beneath the North Yellow Sea [Schmerr and Garnero, 2007]. However, to explain the phenomenon that the TZ is hotter near the present subducting slab yet colder far from this same slab, not only previous seismic tomographic results from this region but also the dynamic behavior of the slab should be considered in detail. [15] Although seismic tomographic results may have a lower lateral resolution compared with RF studies, they can provide another constraint to explain the thinning TZ beneath the North Yellow Sea. Some tomographic studies show the stagnant subducting slabs above the 660 [Fukao et al., 2001; Huang and Zhao, 2006], while other tomographic studies show part of the subducting slab reaching a depth of 800 km or more [Bijwaard et al., 1998] beneath Eastern China. However, the slab s age and rigidity may influence the tomographic images and their interpretations [Miller and Kennett, 2006]. [16] On the basis of this study and previous seismic works in this region, we propose two possible models to explain the thinning TZ beneath the North Yellow Sea. One possible explanation is that the cold slab sinking into the lower mantle might generate local hot upwelling perturbation near the sinking slab. First, the Japan subducting slabs were subhorizontally deflecting or flattening in the TZ, which spread over a large area and caused a broad depressed 660 and thicker TZ in eastern China. Second, as the sufficient 4of5

5 trapped slab material increased at the top of the subduction above the 660 (around longitude 130 ), part of the rollback slab would sink into the lower mantle beneath the east of Korea Peninsula. Finally, with the increasing sinking of the slab to the lower mantle, small-scale convection, which may be caused by decomposition of hydrous ringwoodite [Ohtani et al., 2004], would occur, producing hot upwelling in the upper mantle and causing the uplifting of the 660 and the thinning TZ beneath the North Yellow Sea. [17] Another possible explanation for the thinning TZ beneath the North Yellow Sea is a small mantle plume from the lower mantle. Beneath the study region Late Mesozoic and Cenozoic volcanic rocks are widely distributed (Figure 1). The North Yellow Sea is located at the center of these intraplate volcanoes. Recent P wave tomographic images near our study region showed slow anomalies in the upper and the lower mantle near the subducting Pacific slab [Zhao et al., 2007]. If a small mantle plume originates from the lower mantle, the hot upwelling will cause a thinning TZ locally, rather than widely. Thus, we can see the thinning TZ beneath the North Yellow Sea as well as observe the thickening TZ beneath the western part of the study region. [18] Dense data sets and detailed RF analyses, which we obtained around Eastern China, clearly indicated the thinning TZ beneath the North Yellow Sea. This could be explained by either the small-scale mantle convection derived from the sinking slab in the lower mantle, or a small mantle plume, which would have generated the thinning TZ beneath this region. However, there are several problems still to be solved if these models are to be accepted. First, if the convection-relative to the sinking slab in the lower mantle produced the hot upwelling beneath this region, a depressed 660 should exist in the eastern section of this study region (east of longitude 128 ). This would support the theory of the subducting slab sinking into the lower mantle. Unfortunately, our data do not cover this area. Second, if a small mantle plume originates from the lower mantle, a low velocity anomaly image from the lower mantle should be presented by tomographic pictures. At present, we cannot find this small, fine structure among the available tomographic images. More data and much finer tomographic methods are required, in order to support this hypothesis. Fortunately, with the appearance of more permanent and portable seismic stations in Eastern China, more refined tomographic methods, and data from the latest mineral physics experiments, the mechanism of a thinning TZ beneath the North Yellow Sea can be studied in detail in the future. [19] Acknowledgments. We gratefully acknowledge the participation of the Seismic Array Laboratory of IGGCAS. Jinli Huang provided 3-D P-wave velocity model in this study. Ling Chen and two anonymous referees gave valuable comments on the manuscript. This work was supported by the National Natural Science Foundation of China (grant ). References Ai, Y., and T. Zheng (2003), The upper mantle discontinuity structure beneath eastern China, Geophys. Res. Lett., 30(21), 2089, doi: / 2003GL Ai, Y., T. Zheng, W. Xu, Y. He, and D. Dong (2003), A complex 660 km discontinuity beneath northeast China, Earth Planet. Sci. Lett., 212, Bijwaard, H., W. Spakman, and E. R. Engdahl (1998), Closing the gap between regional and global travel time tomography, J. Geophys. Res., 103, 30,055 30,078. Bina, C. R., and G. R. Helffrich (1994), Phase transition Clapeyron slopes and transition zone seismic discontinuity topography, J. Geophys. Res., 99, 15,853 15,860. Chen, L., T. Zheng, and W. Xu (2006), Receiver function migration image of the deep structure in the Bohai Bay Basin, eastern China, Geophys. Res. Lett., 33, L20307, doi: /2006gl Fukao, Y., S. Widiyantoro, and M. Obayashi (2001), Stagnant slabs in the upper and lower mantle transition region, Rev. Geophys., 39, Gorbatov, A., and B. L. N. Kennett (2003), Joint bulk-sound and shear tomography for western Pacific subduction zones, Earth Planet. Sci. Lett., 210, Houser, C., G. Masters, M. Flanagan, and P. Shearer (2008), Determination and analysis of long-wavelength transition zone structure using SS precursors, Geophys. J. Int., 174, Huang, J., and D. Zhao (2006), High-resolution mantle tomography of China and surrounding regions, J. Geophys. Res., 111, B09305, doi: /2005jb Ito, E., and E. Takahashi (1989), Postspinel transformations in the system Mg 2 SiO 4 -Fe 2 SiO 4 and some geophysical implications, J. Geophys. Res., 94, 10,637 10,646. Kennett, B. L. N., and E. R. Engdahl (1991), Traveltimes for global earthquake location and phase identification, Geophys. J. Int., 105, Lei, J., and D. Zhao (2005), P-wave tomography and origin of the Changbai intraplate volcano in northeast Asia, Tectonophysics, 397, Li, X., and X. Yuan (2003), Receiver functions in northeast China Implications for slab penetration into the lower mantle in northwest Pacific subduction zone, Earth Planet. Sci. Lett., 216, Miller, M. S., and B. L. N. Kennett (2006), Evolution of mantle structure beneath the northwest Pacific: Evidence from seismic tomography and paleogeographic reconstructions, Tectonics, 25, TC4002, doi: / 2005TC Niu, F., A. Levander, S. Ham, and M. Obayashi (2005), Mapping the subducting Pacific slab beneath southwest Japan with Hi-net receiver functions, Earth Planet. Sci. Lett., 239, Ohtani, E., K. Litasov, T. Hosoya, T. Kubo, and T. Kondo (2004), Water transport into the deep mantle and formation of a hydrous transition zone, Phys. Earth Planet. Inter., 143, Owens, T. J., A. A. Nyblade, H. Gurrola, and C. A. Langston (2000), Mantle transition zone structure beneath Tanzania, Geophys. Res. Lett., 27, Schmerr, N., and E. J. Garnero (2007), Upper mantle discontinuity topography from thermal and chemical heterogeneity, Science, 318, Widiyantoro, S., B. L. N. Kennett, and R. D. van der Hilst (1999), Seismic tomography with P and S data reveals lateral variations in the rigidity of deep slabs, Earth Planet. Sci. Lett., 173, Wu, Q., X. Tian, N. Zhang, W. Li, and R. Zeng (2003), Receiver function estimated by maximum entropy deconvolution, Acta Seismol. Sin., 25(4), Zhao, D., S. Maruyama, and S. Omori (2007), Mantle dynamics of western Pacific and east Asia: Insight from seismic tomography and mineral physics, Gondwana Res., 11, Zhu, L. (2002), Deformation in the lower crust and downward extent of the San Andreas Fault as revealed by teleseismic waveforms, Earth Planets Space, 54, Zhu, L., and H. Kanamori (2000), Moho depth variation in southern California from teleseismic receiver functions, J. Geophys. Res., 105, Y. Ai, W. Xu, and T. Zheng, Key Laboratory of the Earth s Deep Interior, Institute of Geology and Geophysics, Chinese Academy of Sciences, Beituchengxilu 19, Chaoyang District, Beijing , China. (ysai@mail. iggcas.ac.cn) Q. Li, Earthquake Administration of Liaoning Province, Shenyang , China. 5of5