Crustal structures across the western Weihe Graben,North China: Implications for extrusion tectonicsat the northeast margin of Tibetan PlateauYoucai Tang1,2, Shiyong Zhou1, Y. John Chen1, Eric Sandvol3, Xiaofeng Liang4, Yongge Feng1, Ge Jin1,Mingming Jiang4, and Mian Liu3
1Institute of Theoretical and Applied Geophysics, School of Earth and Space Sciences, Peking University, Beijing, China,2State Key Laboratory of Petroleum Resource and Prospecting, and Unconventional Natural Gas Institute, China Universityof Petroleum, Beijing, China, 3Department of Geological Sciences, University of Missouri, Columbia, Missouri, USA, 4Instituteof Geology and Geophysics, Chinese Academy of Sciences, Beijing, China
Abstract The stable Ordos Plateau, extensional Weihe Graben, and Qinling orogenic belt are locatedat the northeast margin of the Tibetan Plateau. They have been thought to play different roles in theeastward expanding of the Tibetan Plateau. Peking University deployed a linear seismic array across thewestern end of the Weihe Graben to investigate the crustal structures of the tectonic provinces of thisstructure. Receiver function analyses revealed low-to-moderate Poisson’s ratios and anticorrelationsbetween Poisson’s ratios and topography beneath the Qinling Orogen. These features may indicate atectonic thickening of the felsic upper crust by folding and thrusting within the Qinling Orogen. Weobserved a strong horizontal negative signal at the midcrust beneath the Ordos Plateau which may indicatea low-velocity zone. This observation would suggest the stable cratonic Ordos Plateau had been modifieddue to the compression between the Tibetan Plateau and the Ordos Plateau. We also observed an abrupt4 km Moho offset across the Weihe Fault, changing from ~44 km beneath the Ordos Plateau to ~40 kmbeneath the Qinling Orogen. We conclude that the Weihe Fault is a lithosphere-scale fault/shear zone, whichextends into the upper mantle beneath theWeihe Graben. It acts as themajor boundary separating the stableOrdos Plateau and the active Qinling Orogen.
1. Introduction
The Ordos Plateau, Weihe Graben, and Qinling Orogen are located at the northeastern margin of the TibetanPlateau (Figure 1). They are close to each other in geography but have substantially different deformationregimes due to the differential eastward extrusion of the Tibetan Plateau and contrasting rheologies[Tapponnier and Molnar, 1977; Yin and Harrison, 2000; Jiang et al., 2011]. Investigations of the differences incrustal structures between the three tectonic provinces could provide better constraints for the mechanismof eastward expansion of the Tibetan Plateau.
The origin of the rift system surrounding the Ordos Plateau and associated seismicity is likely related to thedifferential eastward extrusion of the Tibetan Plateau [Xu and Ma, 1992]. The Ordos Plateau has been atectonically stable block within the North China Craton. It is nearly aseismic and does not have any significantinternal deformation [Zhang et al., 2002], although along the edges of this stable block there is westerncompression and the southeastern extension [C. Wang et al., 2014]. A series of extensional grabens, suchas the Weihe Graben and the Shanxi Graben, have developed around the Ordos Plateau during theCenozoic (Figure 1). To the southwest of the Ordos Plateau, a series of parallel NW-SE trending mountainranges, e.g., Liupanshan Mountain (Figure 1), have been developed due to the compression between theTibetan Plateau, the Ordos Plateau, and the restraining bend associated with the Haiyuan fault zone[Tapponnier et al., 2001].
The Weihe Graben is part of the rift system that surrounds the Ordos Plateau (Figure 1). In contrast to thetectonic quiescence of the Ordos Plateau, the Weihe Graben is characterized by high seismicity and stronginternal deformation (Figure 1). It has been tectonically active for the past ∼50Ma and has experienced anumber of stages of extension, which have resulted in thick sediments with a maximum thickness of 7 km.The latest event was a NNW-SSE extension during the Late Pliocene–Quaternary [Mercier et al., 2013].
TANG ET AL. MOHO OFFSET AT WESTERN WEIHE GRABEN 5070
PUBLICATIONSJournal of Geophysical Research: Solid Earth
Citation:Tang, Y., S. Zhou, Y. J. Chen, E. Sandvol,X. Liang, Y. Feng, G. Jin, M. Jiang, andM. Liu (2015), Crustal structures acrossthe western Weihe Graben, North China:Implications for extrusion tectonics atthe northeast margin of Tibetan Plateau,J. Geophys. Res. Solid Earth, 120, 5070–5081,doi:10.1002/2014JB011210.
Received 25 APR 2014Accepted 10 JUN 2015Accepted article online 12 JUN 2015Published online 14 JUL 2015
The Weihe Fault is the main fault zone in the middle of the Weihe Graben (Figure 1). It has developed in earlyProterozoic and has controlled the initiation and development of the Weihe Graben [Peng, 1992]. The faultseparates the basement of the Weihe Graben with a high dip angle of 65–80° [Feng et al., 2003; Peng,1992]. The basement north of the Weihe Fault is mainly composed of early Paleozoic and Mesozoic rocks,while the basement south of the Weihe Fault consists mainly of Archaeozoic/Proterozoic metamorphic rocksand Mesozoic granites [Peng, 1992]. The Weihe Fault is also an active seismic zone (Figure 1) [Shi et al., 2007]with 70% of the earthquakes in the Weihe Graben associated with the fault [Peng, 1992]. Source mechanismsof three large historical earthquakes occurred on this fault show large extensional components with left-lateralstrike-slip components (Figure 1) [Wan et al., 2006].
To the south of the Weihe Graben is the Qinling Orogen, which has been developed during the Late Triassiccollision between North China Craton and South China Craton [Faure et al., 2008; Li et al., 1993; Ping et al.,2013]. The Qinling Orogen has been reactivated during the Cenozoic [Ratschbacher et al., 2003; Enkelmannet al., 2006]. The apparent eastward decrease in the amount of surface uplift, as well as the onset age andthe amount of exhumation within the Qinling Orogen, has been thought to be related to the eastwardextrusion of the Tibetan Plateau [Enkelmann et al., 2006]. These data can be interpreted in terms oftwo competing models: crustal shortening model [Tapponnier et al., 2001] and lower crust flow and/orextrusion model [Clark and Royden, 2000]. In the crustal shortening model, rapid uplift and exhumationin the western Qinling have resulted from thrusting and folding in crustal shortening and thickening.The Tibetan Plateau may be growing eastward faster in the western Qinling than in the Ordos Plateauand the Sichuan Basin. Alternatively, the lower crustal flow and/or extrusion model predict/predicts that
Figure 1. (a) Topography map of the study region. Light blue dots are earthquakes from China Earthquake Data Centercatalog during the last 30 years. Three big historical earthquakes (Ms ≥ 6.75) are also shown with focus mechanism [Wanet al., 2006]. Red triangles are seismic stations. White lines denote active faults. H. F.: Haiyuan Fault, W. F.: Weihe Fault,Q. F. F.: Qinling Front Fault, Liupanshan M.: Liupanshan Mountain belt. The red line denotes the location of a reflection/refraction profile [Ren et al., 2012]. The green box shows the region of Figure 2. The inset shows a larger geological map thatindicates the study region by a black box. N.C.: North China Craton, S.C.: South China Craton.
Journal of Geophysical Research: Solid Earth 10.1002/2014JB011210
TANG ET AL. MOHO OFFSET AT WESTERN WEIHE GRABEN 5071
the uplift and exhumation in the western Qinling are maintained by the dynamic pressure from the maficlower crust flow and/or extrusion.
The difference between the crustal shortening model and lower crust flow and/or extrusion model lies inwhether the felsic upper crust or the mafic lower crust dominates the uplift of the Qinling Orogen. Sincethe mafic lower crust has a higher Vp/ Vs ratio or Poisson’s ratio than the upper crust, a study of the relationshipof Poisson’s ratio and topography may provide a way to discriminate between the two models for the uplift ofthe Qinling Orogen.
In this study, we investigated the crustal structure across the western part of the Weihe Graben from receiverfunction analyses of earthquake data recorded by a dense linear portable seismic array operated by PekingUniversity. The differences in the crustal structures between the two tectonic provinces surrounding theOrdos Plateau, Weihe Graben, and Qinling Orogen may provide us clues as to the role of the Tibetan plateaugrowth on the surrounding tectonic provinces.
2. Data and Method
Peking University operated a linear seismic array of 15 portable broadband seismic stations (Figure 1) fromSeptember 2005 to August 2006. The array started from the southwestern part of the Ordos Plateau (stationsN001–N005), crossed thewestern end of theWeihe Graben (stations N006–N009), and then extended southwardinto the Qinling Orogen (stations N010–N015) (Figure 2). Stations N013, N014, and N015 were deployed3months later than others and operated only for 9months. The average station spacing was about 10 km.Each station was equipped with a Guralp CMG-3ESP seismometer and a Reftek 130 digital acquisition system.
We used P wave receiver functions to investigate the crustal structure along the seismic array. P wave receiverfunctions are used to isolate the P to S converted arrivals (Ps) at discontinuities beneath seismic stations and arefrequently used to constrain the depths of the discontinuities [Ammon, 1991; Cassidy, 1992; Langston, 1977].Teleseismic events with epicentral distances of 30 to 90° with high signal-to-noise ratios (SNR) (Figure 2) wereused to calculate receiver functions. We first rotated the two horizontal components of seismograms to radialand transverse components. Then the waveforms were filtered using a band-pass filter between 0.05Hz and10Hz. Radial receiver functions were calculated by deconvolving the vertical components from the radialcomponents. The deconvolution was performed using an iterative time-domain method [Ligorria andAmmon, 1999]. A Gaussian low-pass filter was applied to radial receiver function to minimize high-frequencynoise. To ensure self-consistency, we subjected each trace of receiver function to two Gaussian filters at1.2Hz and 2.4Hz, respectively. Then receiver functions with low SNR were eliminated by visual inspection.
Figure 2. (a) Distribution of the earthquakes (green dots) used for receiver function calculations. The red star is the centerof the seismic array. (b) Piercing points at 40 km depth of the raypath of receiver functions are shown as purple circles.The location of the profile is shown as a green box in Figure 1. Red triangles denote the seismic stations.
Journal of Geophysical Research: Solid Earth 10.1002/2014JB011210
TANG ET AL. MOHO OFFSET AT WESTERN WEIHE GRABEN 5072
Finally, a total of 831 receiver functions with high SNR from 153 earthquakes were retained. The number ofhigh-quality receiver functions at each station varies from 22 to 111 (Figure 3).
A 2.5-D Kirchhoff migration method [Dellinger et al., 2000;Wilson and Aster, 2005] was applied to image thesubsurface structures along profile AA′ (Figure 2). The wavefield u(x, z = 0, t) recorded at each station wasback propagated to all potential scatter points. The amplitudes of radial receiver functions were thenscaled and summed onto the scatter points according to the Kirchhoff integral [Dellinger et al., 2000;Wilson and Aster, 2005]:
u x; zð Þ ¼ 12π ∫ Str
cos θvr
∂∂t
� �12
u x; z ¼ 0; tð Þdx (1)
where v is the root-mean-square (RMS) velocity at the imaging point (x, z), r is the distance between thestation and imaging point, t is the lapse time of receiver functions, θ is the incident angle of the raypathat the surface. The weight Str varies as a function of the angle of raypaths of direct wave and scattering wave
recorded at the station. RMS ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiXni¼1
tiV2i =Xni¼1
ti
sis a travel time weighted average velocity from the surface
to the imaging point [Allaby and Allaby, 1999], n is the number of the layers above the point, ti and Vi aretravel time and velocity for S wave in each layer [Wilson and Aster, 2005].
We also used a slant stack method to estimate the local average Vp/Vs ratio (κ) and the crust thickness (H)[Zhu and Kanamori, 2000]. This method searched the maximum of the stacked receiver function amplitudesin the H� κ space. For a homogeneous and isotropic crust, the travel times of Ps, PpPs, and PsPs + PpSs can beexpressed as
Low frequency RF (1.2 Hz) High frequency RF (2.4Hz)Ps1Ps2PsPs1Ps2
Ord
osP
late
auW
eihe
Gra
ben
Qin
ling
Oro
gen
55
76
29
47
24
54
85
62
72
71
22
62
111
34
27
55
76
29
47
24
54
85
62
72
71
22
62
111
34
27
Figure 3. Stacked radial receiver functions at each station. The vertical axis is the station number which increases fromnorth to south (Figure 2b). The Ps phases at Stations N01–N014 are about 0.74 s earlier than other stations (see thevertical reference line t = 5.86 s). Waveforms are (a) low (1.2 Hz) and (b) high (2.4 Hz) frequency-stacked receiver functions,respectively. RF: receiver function. The number above each trace indicates how many RFs were stacked at that station.
Journal of Geophysical Research: Solid Earth 10.1002/2014JB011210
TANG ET AL. MOHO OFFSET AT WESTERN WEIHE GRABEN 5073
where p is ray parameter, VP and VS are P and S waves velocities, respectively. The summation of amplitudesfor the three phases can be expressed as
where R(T ) is the amplitude of receiver functions at time T, ω1 = 0.50, ω2 = 0.35, ω3 = 0.15 are the weight forthe amplitudes of the three phases. The three phases will be stacked in phase if we can get the correct traveltimes, TPs , TPpPs , and TPpSs + PsPs. So the point (H, κ) that has the maximumwill represent the best estimation ofthe crust thickness and the average Vp/Vs ratio.
3. Results
Figure 3 shows the stacked radial receiver functions in two different frequency ranges for each station. Wefirst applied a moveout correction to each receiver function by aligning its Moho converted phase (Ps) to thatof a reference (ray parameter of 0.06). Then we stacked the aligned receiver functions from all azimuths toproduce one-single receiver function for each station. The number of receiver functions that were used toproduce the stacked trace is labeled at the end of each trace. High-frequency receiver functions exhibit morecomplex waveforms in general, and they show similar trends at lower frequencies.
The Ps phases converted at the Moho show three different types, which are associated with the threedifferent tectonic provinces, the Ordos Plateau, the Weihe Graben, and the Qinling Orogen, respectively.The Ps phases at stations within the Qinling Orogen (N010–N015) are about 0.74 s earlier compared tostations within the Ordos Plateau and the Weihe Graben (N001–N009) (Figure 3). Stations N010–N015have relatively narrow and simple Ps phases, while stations N001–N005 have wider or double Ps phases
0
20
40
60
Dep
th (
km)
2 3 4 5 6 7 8
P Velocity (km/s)1
Ordos VelWeihe VelQinling Vel
Ordos RMSWeihe RMSQinling RMS
Ordos refWeihe refQinling ref
Low velocity zone
00.5
1
00.5
1
00.5
1
0
0.6
0
0.6
0
0.6
N003
N007
N011
N003
N007
N011
1.2 Hz 2.4 Hz
Am
plitu
de
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)
(a)
(b) (c)
Figure 4. (a) P wave velocity models used in receiver function analyses. They are modified from a wide-angle reflection/refraction experiment [Ren et al., 2012]. Colored dots indicate the RMS velocities (6.46 km/s, 6.35 km/s, and 6.32 km/s forOrdos, Weihe, and Qinling, respectively) that were used in the slant stack process (Figure 5). RMS: root-mean-square, ref:the reference model from the reflection/refraction line, vel: the modified model from the reference model. (b and c)Comparison between observed receiver functions (solid lines) and synthetics (doted lines) from the modified models atlow- (1.2 Hz) and high (2.4 Hz)-frequency range, respectively.
Journal of Geophysical Research: Solid Earth 10.1002/2014JB011210
TANG ET AL. MOHO OFFSET AT WESTERN WEIHE GRABEN 5074
at lower frequencies. The wide phases split into two or more peaks for high-frequency receiver functions(Ps1 in Figure 3). In addition to Moho conversions, stations N001–N004 also show a broad negative phaseat 3–4 s Stations N006–N009 (located within the Weihe Graben) have anomalous waveforms with longduration of coda, which may be caused by the reverberations within the thick sedimentary layer of theWeihe Graben.
The receiver functions also display significant differences in the first pulses of the time series: stationsN001–N009 have widening and asymmetric first pulses (Ps2 in Figure 3) at the lower frequency range whilestations N010–N015 have thin and simple first pulses. The widening first pulses at stations N001–N009 splitinto two peaks at the high-frequency range.
30 35 40 45 50 55
1.6
1.7
1.8
1.9
30 35 40 45 50 55
1.6
1.7
1.8
1.9
1.6
1.7
1.8
1.9
1.6
1.7
1.8
1.9
H (km) H (km)
kk
kk
Figure 5. Slant stack results at some stations located within the Ordos Plateau and the Qinling Orogen, respectively. The redstar denotes the best estimations of the Moho depths (H) and the average ratios of Vp/Vs (κ). Blue crosses are possibleestimations from bootstrap tests.
Journal of Geophysical Research: Solid Earth 10.1002/2014JB011210
TANG ET AL. MOHO OFFSET AT WESTERN WEIHE GRABEN 5075
To estimate crustal thickness and Vp/Vs ratio from receiver functions, a reference velocity model is required. Inthis study, the reference velocity model was constructed from a Pwavemodel from a nearby NW-SEwide-anglereflection/refraction profile [Ren et al., 2012] (red line in Figure 1). The profile was located relatively close to ourseismic array (the closest distance is ~50 km) and crossed the same tectonic provinces (Figure 1). Therefore, itcould be an appropriate representation of the velocity model beneath our seismic array (profile AA′ inFigure 2). We assembled three different velocity models (Figure 4) for the Ordos Plateau, Weihe Graben, andQinling Orogen, respectively, by modifying the reference model of the reflection/refraction line. Synthetic data(dotted lines) from the modified models match the observed data (solid lines) very well. We used thesemodified models in the later slant stack and migration analyses.
Slant stack results at several typical stations are shown in Figure 5. The RMS velocities at the Moho depths(colored dots in Figure 4) were used as the average P wave velocities for the three tectonic provinces.Stations N001–N004 (located within the Ordos Plateau) have the Moho depth of ~ 45.5 km, while stationsN010–N013 (located within the Qinling Orogen) have the Moho depth of ~42.0 km. The difference is as muchas 3.5 km. Stations N001 and N003 have a double peak in the ~3–6 s time window. We can see a second peakin the slant stacks at stations of N001 and N003. Stations N002 and N004 have widening Ps, and we can seeonly one peak in the slant stacks with “wavy” contours. The second peaks at stations N001 and N003 maycome from the bottom of the low-velocity zone (Figure 4).
The Vp/Vs ratios (κ) within the Ordos Plateau are systemically higher than those within the Qinling Orogen(Figure 5). Poisson’s ratio σ was calculated from Vp/Vs ratio by using the formula σ = 0.5 * [1� 1/(κ2� 1)].The average of Poisson’s ratios within the Ordos Plateau is ~0.277, while it is ~0.263 within the QinlingOrogen. The Poisson’s ratios within the Qinling Orogen show an apparent anticorrelation with topography(Figure 6). Poisson’s ratio decreases northward when the topography increases (green lines in Figure 6).However, the Poisson’s ratio within the Ordos Plateau does not vary significantly.
A migration image of receiver functions along the profile AA′ (Figure 2b) is also shown in Figure 6. This imagewas constructed by using the 2.5-D Kirchhoff migration method [Dellinger et al., 2000;Wilson and Aster, 2005].
Dep
th/k
m
500
1000
1500
2000
2500
Top
ogra
phy/
m
33.5 34 34.5 35
0
10
20
30
40
50
60
70
Figure 6. Migrated image of receiver functions along profile AA′ (Figure 2b) with positive and negative polarity energyshown in red and blue, respectively. Topography and Poisson’s ratios are shown in the top panel. The red circles andblue circles indicate Poisson’s ratios and the Moho depths induced from slant stack of receiver functions, respectively. Theerror bars indicate the standard deviations from bootstrap tests. The two dashed lines indicate the 40 and 44 km depths.The Moho depths are plotted below the sea level (the elevations are subtracted from the depths).
Journal of Geophysical Research: Solid Earth 10.1002/2014JB011210
TANG ET AL. MOHO OFFSET AT WESTERN WEIHE GRABEN 5076
The migration processes were performed withinthe three tectonic provinces separately by usingtheir RMS velocities (Figure 4) and then wereintegrated together to compose one image. Weobserved an abrupt offset (~4 km) in the Mohodepths beneath the Weihe Graben, changing fromabout ~44 km depth beneath the Ordos Plateau toabout ~40 km depth beneath the Qinling Orogen.These migrated Moho depths are consistent withthe slant stack results.
There is a significant difference in crustal structuresbetween the two sides of the Weihe Graben. TheQinling Orogen has a relatively simple crustal struc-ture, while the Ordos Plateau has a large/stronghorizontal negative polarity signal in the middlepart of the crust with a velocity increase abovethe Moho.
4. Discussion
Station N009 has the most complicated receiverfunction (Figure 3) which could make the interpre-tation of an abrupt Moho offset a bit dubious;therefore, we produced two different migrationimages: one with N009 included and one withoutN009. Because N009 has fewer records (only 29high-quality receiver functions) and consequentlyhas less effect on the final result, and the twoimages show trivial differences. To avoid ambigu-ity, the migration image in Figure 6 does notinclude N009.
The migration image in Figure 6 used differentvelocity profiles for the Ordos Plateau, WeiheGraben, and the Qinling Orogen, which are shownin Figure 4a. To test the effect of different velocityprofiles on the Moho offset, we generated a newmigration image with a laterally constant velocitymodel. The new image did not show significantdifference except for the Moho depths. The
Moho offset was ~0.4 km larger than that in Figure 6 when we used a laterally constant velocity profile.Therefore, the Moho offset should not be due to the use of different velocity models.
To test the resolution of the observed Moho offset (Figure 6), we generated migration images by using thereceiver functions at some stations near the offset (Figure 7). Station N008 to the north of the offset showsthat the Moho depth is ~44 km, while Station N011 to the south of the offset shows that the Moho depthis ~40 km. Station N010 located just above the offset shows a difference in Moho depths: receiver functionsfrom azimuths that sample the Ordos Plateau/Weihe Graben show a Moho depth of ~44 km while receiverfunctions sampling the Qinling Orogen show a Moho depth of ~40 km. These receiver functions recordedat station N010 have almost common raypaths in the uppermost crust; therefore, the difference in Mohodepth between the receiver functions that sampled different units should be the real difference in Mohodepths between different units.
The Ordos Plateau and theWeihe Graben have thick sediments (Figure 4), which could have substantial effecton the receiver functions (Figure 3) and may introduce errors in migration images [Schulte-Pelkum andBen-Zion, 2012]. We made synthetic receiver functions from two different models (Figure 8). Model 1 was
30
20
10
0
50
60
70
40
N010
33.5 34 34.5 35
30
20
10
0
50
60
70
40
N011
Latitude
30
20
10
0
50
60
70
40
N008
Dep
th (
km)
Dep
th (
km)
Dep
th (
km)
Figure 7. Migrated images for the three stations around theWeihe Fault. Positive and negative polarity energy of thereceiver functions are shown in red and blue, respectively.The dashed black lines indicate the 40 and 44 km depths,respectively.
Journal of Geophysical Research: Solid Earth 10.1002/2014JB011210
TANG ET AL. MOHO OFFSET AT WESTERN WEIHE GRABEN 5077
modified from the Qinling model in Figure 4. Model 2 was composed by replacing the near-surface layers inModel 1 by the low-velocity layers modified from the Ordos model in Figure 4. We subjected each receiverfunction to low-pass Gaussian filters at 1.2 Hz and 2.4 Hz, respectively. Synthetic tests show that near-surfacelayers with low seismic velocities can produce widening near-zero pulse (low-frequency range) or split pulses(high-frequency range) and can introduce a delay into all subsequent phases in the receiver functions(Figure 8). The time delay of Ps phases introduced by the near-surface low-velocity layers is about 0.13 s,which is much less than the observed 0.74 s time difference of Ps phases between the Ordos Plateau andthe Qinling Orogen (Figure 3). Consequently, we can largely rule out the possibility that the observedMoho offset was caused by the velocity contrast across the Weihe Fault.
The observed Moho offset between the Ordos Plateau and the Qinling Orogen is consistent with the resultsfrom other profiles across the central part of the Weihe Graben [C. Wang et al., 2014; P. Wang et al., 2014; Xuet al., 2014; Ren et al., 2012] where the graben has been well developed. Ambient noise tomography resultsshow that the crustal thickness beneath the Ordos Plateau is much larger than that beneath the QinlingOrogen [Zheng et al., 2010, 2012]. Receiver function studies across the central part of the Weihe Grabenalso show Moho depth changes significantly across the Weihe Fault [C. Wang et al., 2014]. Some of theMoho depths beneath the Weihe Graben [P. Wang et al., 2014] may correspond to a large fault/shear zonecutting through the Moho. A wide-angle reflection/refraction profile across the central part of the WeiheGraben revealed a sharp change in the Moho depths from 45 km to 35 km beneath the Weihe Fault [Renet al., 2012]. Furthermore, the lower crustal velocity structure changes rapidly from the Ordos Plateau tothe Weihe Graben. These features are consistent with our observed Moho offset beneath the WeiheGraben. However, we did not observe a shallowing of the Moho beneath the western end of the WeiheGraben as others observed in the central part of the Weihe Graben [C. Wang et al., 2014; P. Wang et al.,2014] and in Shanxi Graben [Tang et al., 2010]. One possible reason is that the western end of the WeiheGraben is in the early stages of the basin formation, while the central part of the Weihe Graben ismore mature.
-2 6420 8 10
1.0
0.6
0.4
0.2
0.0
0.8
0.6
0.4
0.2
0.0
-2 6420 8 10
Time (s)
Time (s)
Am
plitu
deA
mpl
itude
0
30
20
10
40
Dep
th (
km)
654 7 8P velocity (km/s)
(a) (c)
(b)
Figure 8. Synthetic receiver functions in (a) high-frequency range and (b) low-frequency range generated from (c) twomodels. Ps: conversions at the Moho, Ps_s: conversions at sedimentary layers.
Journal of Geophysical Research: Solid Earth 10.1002/2014JB011210
TANG ET AL. MOHO OFFSET AT WESTERN WEIHE GRABEN 5078
In addition to observed changes in crustal thickness, we observe significant changes in crustal complexityacross the Weihe Fault. The crustal structure beneath the Qinling Orogen is simple, while the structurebeneath the Ordos Plateau is more complicated. We observed a prominent horizontal negative signal atthe midcrust above the Moho beneath the Ordos Plateau.
The modified velocity model within the Ordos Plateau requires a low-velocity zone between the depths of20–30 km (Figure 4a) to match the negative polarity conversions at stations N001–N004 within the OrdosPlateau (shaded region in Figure 3). Though the negative polarity conversions may be multiples for ashallower anomaly, the low-velocity zone is consistent with the results of the reflection/refraction profilewhich had revealed a low-velocity zone between the depths of 15 and –25 km (Figure 4). Synthetic tests showthat sediments may introduce negative pulses after the direct P wave (Figure 8); however, the amplitudes ofthe negative pulses are much smaller than the Ps pulses. The amplitudes of the negative pulses in theobserved receiver functions (Figure 4) are as large as the Ps pulses, so we conclude that these negative pulsesindicate a low-velocity zone rather than multiples from shallow sediments.
We suggest that this strong horizontal negative signal is caused by a low-velocity zone (LVZ) in the middlecrust beneath the Ordos block. The LVZ appears to terminate at 30 km depth in the lower crust where anincrease in velocity was inferred above the Moho. The existence of a LVZ in the crust may indicate the stablecratonic Ordos Plateau had been modified due to the compression between the Tibetan Plateau and theOrdos Plateau [Tapponnier et al., 2001].
Other studies also suggest that there are large changes in crustal properties across the Weihe Fault. A seismicamplitude tomography study concluded that the Qinling Orogen showed high attenuation, while the OrdosPlateau showed little attenuation [Hearn et al., 2008]. Ambient noise tomography revealed large velocitycontrasts in the lower crust across the Weihe Fault [Bao et al., 2013]. Shear wave splitting has shown thatthe anisotropy is weak north to the Weihe Fault but increases sharply southward up to the Weihe Fault[Huang et al., 2008]. GPS measurements have shown a discontinuity in horizontal crustal displacement acrossthe Weihe Fault [Dai et al., 2004], and the vertical deformation across the fault near Baoji (Figure 1) reached13mm [Peng, 1992].
Given these observations, we conclude that Weihe Fault is a lithosphere-scale fault/shear zone, which cutsthrough the Moho producing the ~3 km offset imaged in this study (Figure 6). It may act as a boundary ofthe different responses between the Ordos Plateau and the Qinling Orogen to the eastward extrusion ofthe Tibetan Plateau [Xu and Ma, 1992].
The Poisson’s ratios within the Qinling Orogen range from 0.258 to 0.269, which are relatively low comparedto the value of ~0.277 within the Ordos Plateau. These features are consistent with previous receiver functionresults [Pan and Niu, 2011; C. Wang et al., 2014] and refraction study [Liu et al., 2006] in the central WeiheGraben. In general, mafic lower crustal rocks have higher Poisson’s ratios than felsic upper crustal rocks.Low-to-moderate Poisson’s ratios and the anticorrelation of the Poisson’s ratio and topography within theQinling Orogen (Figure 6) may indicate tectonic thickening of the felsic upper crust by folding and thrusting,which is the dominant mechanism within the Qinling Orogen since the Cenozoic. This is consistent withambient noise tomography results that the Qinling Orogen has very thick upper crust and very thin lowercrust [Zheng et al., 2010]. The observation of thickening of the upper crust prefers the crustal shorteningmodel rather than lower crust flow and/or extrusion model for the uplift of the Qinling Orogen.
5. Conclusions
We investigated crustal structures along a linear seismic array across the western end of the Weihe Grabenand its surrounding areas using receiver functions. The results revealed the following features of the crustbeneath the study area:
1. Low-to-moderate Poisson’s ratios and anticorrelations between Poisson’s ratios and topography withinthe Qinling Orogen indicate significant tectonic thickening of the felsic upper crust by folding andthrusting. This observation is consistent with a model of eastward expansion of the Tibetan crust alongthe Qinling orogenic belt which produced the highest elevation in the western Qinling.
2. We observed a strong horizontal negative signal at the midcrust beneath the Ordos Plateau which weinterpreted as a LVZ in the midcrust and lower crust. The existence of a LVZ in the crust may indicate
Journal of Geophysical Research: Solid Earth 10.1002/2014JB011210
TANG ET AL. MOHO OFFSET AT WESTERN WEIHE GRABEN 5079
the stable cratonic Ordos Plateau had been modified due to the compression between the TibetanPlateau and the Ordos Plateau.
3. We observed a sharp 3 km offset in the Moho depths beneath Weihe Fault across the western part of theWeihe Graben. The receiver functions also revealed different styles of the crustal structures across theWeihe Graben. We conclude that Weihe Fault may be a lithosphere-scale fault/shear zone, which extendsinto the upper mantle beneath theWeihe Graben. It acts as themajor boundary separating the differentialdeformational regimes between the stable Ordos Plateau and the active Qinling Orogen.
ReferencesAllaby, A., and M. Allaby (1999), Root-mean-square velocity, A Dictionary of Earth Sciences. [Available at Encyclopedia.com.]Ammon, C. J. (1991), The isolation of receiver effects from teleseismic P-wave-forms, Bull. Seismol. Soc. Am., 81, 2504–2510.Bao, X. W., X. D. Song, M. J. Xu, L. S. Wang, X. X. Sun, N. Mi, D. Y. Yu, and H. Li (2013), Crust and upper mantle structure of the North China
Craton and the NE Tibetan Plateau and its tectonic implications, Earth Planet. Sci. Lett., 369, 129–137.Cassidy, J. F. (1992), Numerical experiments in broad-band receiver function-analysis, Bull. Seismol. Soc. Am., 82, 1453–1474.Clark, M. K., and L. H. Royden (2000), Topographic ooze: Building the eastern margin of Tibet by lower crustal flow, Geology, 28,
703–706.Dai, W., J. Ren, X. Zhao, X. Shao, and G. Zhu (2004), Characteristics of horizontal crustal movement in Weihe Basin and adjacent zones by GPS
observation, Acta Seismol. Sin., 26, 256–260.Dellinger, J. A., S. H. Gray, G. E. Murphy, and J. T. Etgen (2000), Efficient 2.5-D true-amplitude migration, Geophysics, 65, 943–950.Enkelmann, E., L. Ratschbacher, R. Jonckheere, R. Nestler, M. Fleischer, R. Gloaguen, B. R. Hacker, Y. Q. Zhang, and Y. S. Ma (2006), Cenozoic
exhumation and deformation of northeastern Tibet and the Qinling: Is Tibetan lower crustal flow diverging around the Sichuan Basin?,Geol. Soc. Am. Bull., 118, 651–671.
Faure, M., W. Lin, P. Monie, and S. Meffre (2008), Palaeozoic collision between the North and South China blocks, Triassic intracontinentaltectonics, and the problem of the ultrahigh-pressure metamorphism, C. R. Geosci., 340, 139–150.
Feng, X., Q. Tian, and X. Shen (2003), Analysis of activity difference of the west section of the Weihe fault, Geol. Rev., 49, 233–238.Hearn, T. M., S. Y. Wang, S. P. Pei, Z. H. Xu, J. F. Ni, and Y. X. Yu (2008), Seismic amplitude tomography for crustal attenuation beneath China,
Geophys. J. Int., 174, 223–234.Huang, Z., M. Xu, L. Wang, N. Mi, D. Yu, and H. Li (2008), Shear wave splitting in the southern margin of the Ordos Block, north China, Geophys.
Res. Lett., 35, L19301, doi:10.1029/2008GL035188.Jiang, M. M., S. Zhou, E. Sandvol, X. Chen, X. Liang, Y. J. Chen, and W. Fan (2011), 3-D lithospheric structure beneath southern Tibet from
Rayleigh-wave tomography with a 2-D seismic array, Geophys. J. Int., 185(2), 593–608.Langston, C. A. (1977), The effect of planar dipping structure on source and receiver responses for constant ray parameter, Bull. Seismol. Soc.
Am., 67, 1029–1050.Li, S., et al. (1993), Collision of the North China and Yangtse Blocks and formation of coesite-bearing eclogites: Timing and processes, Chem.
Geol., 109, 89–111.Ligorria, J. P., and C. J. Ammon (1999), Iterative deconvolution and receiver-function estimation, Bull. Seismol. Soc. Am., 89, 1395–1400.Liu, M. J., W. D. Mooney, S. L. Li, N. Okaya, and S. Detweiler (2006), Crustal structure of the northeastern margin of the Tibetan plateau from
the Songpan-Ganzi terrane to the Ordos basin, Tectonophysics, 420, 253–266.Mercier, J. L., P. Vergely, Y. Q. Zhang, M. J. Hou, O. Bellier, and Y. M. Wang (2013), Structural records of the Late Cretaceous-Cenozoic extension
in Eastern China and the kinematics of the Southern Tan-Lu and Qinling Fault Zone (Anhui and Shaanxi provinces, PR China),Tectonophysics, 582, 50–75.
Pan, S. Z., and F. L. Niu (2011), Large contrasts in crustal structure and composition between the Ordos Plateau and the NE Tibetan Plateaufrom receiver function analysis, Earth Planet. Sci. Lett., 303, 291–298.
Peng, J. (1992), Tectonic evolution and seismicity of Weihe fault zone, Seismol. Geol., 14, 113–120.Ping, X. Q., J. P. Zheng, J. H. Zhao, H. Y. Tang, and W. L. Griffin (2013), Heterogeneous sources of the Triassic granitoid plutons in the southern
Qinling orogen: An E-W tectonic division in central China, Tectonics, 32, 396–416, doi:10.1002/tect.20034.Ratschbacher, L., B. R. Hacker, A. Calvert, L. E. Webb, J. C. Grimmer, M. O. McWilliams, T. Ireland, S. W. Dong, and J. M. Hu (2003), Tectonics of
the Qinling (Central China): Tectonostratigraphy, geochronology, and deformation history, Tectonophysics, 366, 1–53.Ren, J., J. B. Peng, F. Y. Wang, C. Liu, X. J. Feng, and W. Q. Dai (2012), The research of deep structural features of Weihe basin and adjacent
areas, Chin. J. Geophys., 55, 2939–2947.Schulte-Pelkum, V., and Y. Ben-Zion (2012), Apparent vertical Moho offsets under continental strike-slip faults from lithology contrasts in the
seismogenic crust, Bull. Seismol. Soc. Am., 102, 2757–2763.Shi, Y., J. Li, X. Feng, W. Dai, J. Ren, X. Li, andM. Dou (2007), The study of paleoearthquake on theWeihe fault zone, Seismol. Geol., 29, 607–616.Tang, Y.-C., Y.-G. Feng, J. Chen Yongshun, S.-Y. Zhou, J.-Y. Ning, S.-Q. Wei, P. Li, C.-Q. Yu, W.-Y. Fan, and H.-Y. Wang (2010), Receiver function
analysis at Shanxi Rift, Chin. J. Geophys., 53, 2102–2109.Tapponnier, P., and P. Molnar (1977), Active faulting and tectonics in China, J. Geophys. Res., 82, 2905–2930, doi:10.1029/JB082i020p02905.Tapponnier, P., Z. Q. Xu, F. Roger, B. Meyer, N. Arnaud, G. Wittlinger, and J. S. Yang (2001), Oblique stepwise rise and growth of the Tibet
plateau, Science, 294, 1671–1677.Wan, Y., Z. Shen, D. Shang, T. Li, and Y. Zeng (2006), Crustal stress evolution over the past 700 years in North China and earthquake
occurrence, Earthquake Res. China, 20, 244–261.Wang, C., E. Sandvol, L. Zhu, H. Lou, Z. Yao, and X. Luo (2014), Lateral variation of crustal structure in the Ordos block and surrounding
regions, North China, and its tectonic implications, Earth Planet. Sci. Lett., 387, 198–211.Wang, P., Z. Huang, N. Mi, M. Xu, L. Wang, H. Li, D. Yu, H. Huang, and X. Mao (2014), Crustal structure beneath the Weihe Graben in central
China: Evidence for the tectonic regime transformation in the Cenozoic, J. Asian Earth Sci., 81, 105–114.Wilson, D., and R. Aster (2005), Seismic imaging of the crust and upper mantle using regularized joint receiver functions, frequency-wave
number filtering, and multimode Kirchhoff migration, J. Geophys. Res., 110, B05305, doi:10.1029/2004JB003430.Xu, S. B., N. Mi, M. J. Xu, L. S. Wang, H. Li, and D. Y. Yu (2014), Crustal structures of the Weihe graben and its surroundings from receiver
functions, Sci. China Earth Sci., 57, 372–378.
AcknowledgmentsContinuous seismic data used in thisstudy are available at Peking University.Please contact Yongshun John Chen([email protected]) if you are inter-ested in the data. We would like tothank Qingliang Wang at the SecondMonitoring and Application Center,CEA, Xi’an, China for providing help inthe field work. Three anonymousreviewers’ comments had improved themanuscript. This study was supportedby NSFC grants (91128210; 90814002),Science Foundation of China Universityof Petroleum, Beijing (2462013YJRC014,PRP/indep-4-1316) and MOST grant(2009AA093401) in China.
Journal of Geophysical Research: Solid Earth 10.1002/2014JB011210
TANG ET AL. MOHO OFFSET AT WESTERN WEIHE GRABEN 5080
Xu, X. W., and X. Y. Ma (1992), Geodynamics of the Shanxi Rift system, China, Tectonophysics, 208, 325–340.Yin, A., and T. M. Harrison (2000), Geologic evolution of the Himalayan-Tibetan orogen, Annu. Rev. Earth Planet. Sci., 28, 211–280.Zhang, P., Q. Wang, and Z. Ma (2002), GPS velocity field and active crustal blocks of contemporary tectonic deformation in continental China.Zheng, X., C. Zhao, L. Zhou, and S. Zheng (2012), Rayleigh wave tomography from ambient noise in central and eastern Chinese mainland,
Chin. J. Geophys., 55(6), 1919–1928.Zheng, Y., Y. Yang, M. H. Ritzwoller, X. Zheng, X. Xiong, and Z. Li (2010), Crustal structure of the northeastern Tibetan plateau, the Ordos block
and the Sichuan basin from ambient noise tomography, Earthquake Sci., 23, 465–476.Zhu, L. P., and H. Kanamori (2000), Moho depth variation in Southern California from teleseismic receiver functions, J. Geophys. Res., 105,
2969–2980, doi:10.1029/1999JB900322.
Journal of Geophysical Research: Solid Earth 10.1002/2014JB011210
TANG ET AL. MOHO OFFSET AT WESTERN WEIHE GRABEN 5081
