Surface Wave Paper

TECTO-125734; No of Pages 18
Tectonophysics xxx (2013) xxxxxx
Contents lists available at SciVerse ScienceDirect
Tectonophysics
journal homepage: www.elsevier.com/locate/tecto
Mapping the Moho with seismic surface waves: A review, resolution analysis, and
recommended inversion strategies
Sergei Lebedev a,, Joanne M.-C. Adam a, b, Thomas Meier c
a
Dublin Institute for Advanced Studies, School of Cosmic Physics, Geophysics Section, 5 Merrion Square, Dublin 2, Ireland
b
Trinity College Dublin, Department of Geology, Dublin 2, Ireland
c
Christian-Albrechts University Kiel, Institute of Geophysics, Kiel, Germany
a r t i c l e i n f o a b s t r a c t
Article history: The strong sensitivity of seismic surface waves to the Moho is evident from a mere visual inspection of their dis-
Received 30 June 2012 persion curves or waveforms. Rayleigh and Love waves have been used to study the Earth's crust since the early
Received in revised form 21 December 2012 days of modern seismology. Yet, strong trade-offs between the Moho depth and crustal and mantle structure in
Accepted 28 December 2012
surface-wave inversions prompted doubts regarding their capacity to resolve the Moho. Here, we review
Available online xxxx
surface-wave studies of the Moho, with a focus on early work, and then use model-space mapping to establish
Keywords:
the waves' sensitivity to the Moho depth and the resolution of their inversion for it. If seismic wavespeeds within
Rayleigh wave the crust and upper mantle are known, then Moho-depth variations of a few kilometres produce large (>1%) per-
Love wave turbations in phase velocities. However, in inversions of surface-wave data with no a priori information
Mohorovii discontinuity (wavespeeds not known), strong Moho-depth/shear-speed trade-offs will mask ~90% of the Moho-depth signal,
Model space with remaining phase-velocity perturbations ~0.1% only. In order to resolve the Moho with surface waves alone,
Inversion errors in the data must thus be small (up to ~0.2% for resolving continental Moho). With larger errors, Moho-
Tomography depth resolution is not warranted and depends on error distribution with period. An effective strategy for the in-
version of surface-wave data alone for the Moho depth is to, rst, constrain the crustal and upper-mantle structure
by inversion in a broad period range and then determine the Moho depth in inversion in a narrow period range
most sensitive to it, with the rst-step results used as reference. Prior information on crustal and mantle structure
reduces the trade-offs and thus enables resolving the Moho depth with noisier data; such information should be
used whenever available. Joint analysis or inversion of surface-wave and other data (receiver functions, topogra-
phy, gravity) can reduce uncertainties further and facilitate Moho mapping.
2013 Elsevier B.V. All rights reserved.
Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
2. Surface-wave studies of the crust and the Moho . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
3. Sensitivity of surface waves to the Moho. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
4. Trade-offs between the Moho depth and other model parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
5. Inversion of surface-wave measurements for the Moho depth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
5.1. Mapping the model space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
5.2. Resolution and trade-offs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
5.3. Inversion of measured data: Northern Kaapvaal Craton. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
6. Noise in the data: how much is too much for the Moho to be resolved? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
7. Recommended inversion strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
7.1. Inversion of surface-wave data only, with no a priori information. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
7.2. A priori information: include whenever available! . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
7.3. Joint analysis and inversion of surface-wave and other data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
8. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
9. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
Corresponding author. Tel.: +353 1 653 5147x240.

E-mail address: sergei@cp.dias.ie (S. Lebedev).
0040-1951/$ see front matter 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.tecto.2012.12.030
Please cite this article as: Lebedev, S., et al., Mapping the Moho with seismic surface waves: A review, resolution analysis, and recommended
inversion strategies, Tectonophysics (2013), http://dx.doi.org/10.1016/j.tecto.2012.12.030
2 S. Lebedev et al. / Tectonophysics xxx (2013) xxxxxx
1. Introduction Phase velocity Group velocity
The Mohorovii discontinuity, often referred to as the Moho, sep- dC/dVs dU/dVs
arates the Earth's crust from the underlying mantle. Compositional 0 0
differences between the lighter crust and the denser upper mantle 10 10
give rise to an increase in seismic velocities across the Moho, from
Depth, km
Depth, km
the crust to the mantle. The discontinuity can thus be identied seis- 20 20
mically as the location of the seismic-velocity increase (Mohorovii, 30 30
1910).
40 40
During the century since the discovery of the Moho (Mohorovii, Period: 6s Period: 6s
1910), the discontinuity, which can be either sharp or gradational,
0 0
has been detected and imaged in numerous locations around the
10 10
world, at various length-scales and with different seismic techniques.
Depth, km
Depth, km
Controlled-source seismic surveys yield high resolution of the entire 20 20
crust and the Moho by sampling them densely with rays of reected 30 30
or refracted seismic body waves, propagating between local sources 40 40
and receivers (Prodehl and Mooney, 2011, and references therein). 50 50
Relatively expensive and labour-intensive, controlled-source experi- 60 Period: 15s Period: 15s 60
ments can be complemented by passive seismic studies that use 0 0
natural seismic sources (local or teleseismic earthquakes or ambient
50 50
seismic noise). The passive imaging approaches include the analysis
Depth, km
Depth, km
of P-to-S wave conversions at the Moho (e.g., Bostock et al., 2002; 100 100
Kind et al., 2002; Nabelek et al., 2009; Stankiewicz et al., 2002; Zhu 150 150
and Kanamori, 2000), surface-wave imaging, including inversions
200 200
of surface-wave dispersion curves or waveforms and surface-wave
Period: 40s Period: 40s
tomography (e.g., Das and Nolet, 1995; Endrun et al., 2004; Yang 250 250
et al., 2008), joint inversions of the P-to-S conversions (receiver func- 0 0
tions) and surface-wave data (e.g., Juli et al., 2000; Tkali et al., 100 100
Depth, km
Depth, km
2012), local body-wave tomography (e.g., Koulakov and Sobolev, 200 200
2006), and even SS waveform stacking (Rychert and Shearer, 2010).
300 300
Regional crustal models and Moho maps have also been constructed
using combinations of both active-source and passive seismic data, 400 400
as well as other geophysical and geological data (e.g., Grad et al., 500 Period: 100s Period: 100s 500
2009; Kissling, 1993; Molinari and Morelli, 2011; Tesauro et al., 0 0
2008; Thybo, 2001). dC/dVs dU/dVs
Seismic surface waves are particularly sensitive to the structure of
the crust and uppermost mantle and, thus, to the depth of the : Rayleigh : Love
Moho. Because these waves propagate along the Earth's surface, mea-
surements of their speeds characterise average elastic properties of Fig. 1. Depth sensitivity of surface waves. The sensitivity curves are the Frchet deriv-
the crust and upper mantle between seismic sources and stations or atives of the phase and group velocities of the fundamental-mode Rayleigh and Love
waves with respect to S-wave velocities at different depths. The derivatives were com-
between different stations. The Moho can thus be imaged even in lo-
puted for a continental, 1-D Earth model with a 37-km thick crust, at 4 different pe-
cations with no stations or sources. riods. Each graph is scaled independently.
The two main types of surface waves are Rayleigh waves and Love
waves (Aki and Richards, 1980; Dahlen and Tromp, 1998; Kennett,
1983, 2001; Levshin et al., 1989; Nolet, 2008). The speeds of Rayleigh seismology. It also became apparent early that the crustal models in-
waves depend primarily on the speeds of the vertically polarised S ferred from the dispersion data can be highly non-unique. Although
waves in the crust and mantle and, also, on P-wave speeds and densi- the Moho depth has been an inversion parameter in numerous surface-
ty; the particle motion associated with Rayleigh waves in an isotropic, wave studies, the data's sensitivity to the Moho and, in particular, the
laterally homogeneous Earth model is within the great circle plane resolution of the Moho properties given by inversions of surface-wave
containing the source and the receiver. The speeds of Love waves de- data with measurement errors are still uncertain and not agreed upon.
pend primarily on the speeds of the horizontally polarised S waves In this paper we overview the classic surface-wave studies since the
and, also, on density; the associated particle motion is approximately late 19thearly 20th century, as well as some of the more recent work
perpendicular to the great circle plane. focussing on the Moho. We then investigate in detail the sensitivity of
The depth sensitivity of surface waves depends on their period: surface-wave phase velocities to the Moho depth and the trade-offs be-
the longer the period, the deeper within the Earth the waves sample tween Moho-depth and crustal and mantle shear-velocity parameters
(Fig. 1). This makes surface waves strongly dispersive. in inversions of surface-wave dispersion. Exploring the model spaces
Dispersion curves of surface waves (their phase or group velocities in inversions of synthetic and real data, we examine the resolution of
plotted as a function of period or frequency) show a characteristic the Moho by surface-wave measurements as a data type. Finally, we
sharp increase with period associated with the Moho (Figs. 2, 3). discuss strategies for an accurate estimation of the Moho depth using
This increase reects the S-wave velocity increase across the discon- surface-wave data and illustrate some of them with applications to
tinuity, and its period range depends on the depth of the Moho: it oc- phase-velocity measurements from southern Africa.
curs at longer periods if the Moho is deeper. The depth of the Moho
can thus be estimated roughly by a mere visual inspection of a 2. Surface-wave studies of the crust and the Moho
surface-wave dispersion curve (Figs. 2, 3).
Inferences on the crustal structure and thickness have been Rayleigh waves were identied on seismic recordings by Oldham
drawn from surface-wave observations since the early days of modern (1899), and already at that time Wiechert (1899) speculated that
S. Lebedev et al. / Tectonophysics xxx (2013) xxxxxx 3
Rayleigh waves Love waves Rayleigh waves Love waves

5 5 5 5
phase velocity, km/s
group velocity, km/s

MOHO
MOHO
MOHO
4 4
4
3
3 4
2
2
ocean ocean 3
MOHO 1
1 3
5 10 20 50 100 5 10 20 50 100 5 10 20 50 100 5 10 20 50 100
5 5 5 5

MOHO
MOHO
4 4
4 4
normal normal
continent 3
MOHO continent MOHO 3
3 3
5 10 20 50 100 5 10 20 50 100 5 10 20 50 100 5 10 20 50 100
5 5 5 5

MOHO
Tibet Tibet MOHO

4 4
4 4
3 3
3 MOHO 3 MOHO
5 10 20 50 100 5 10 20 50 100 5 10 20 50 100 5 10 20 50 100
period, s period, s period, s period, s
Fig. 2. The signature of the Moho in phase-velocity curves of surface waves. The phase Fig. 3. The signature of the Moho in group-velocity curves of surface waves. The group
velocities of the fundamental-mode Rayleigh and Love waves were computed for an velocities of the fundamental-mode Rayleigh and Love waves were computed for the
oceanic model with a 5-km water layer and a 6-km thick crust (top), a continental same oceanic and continental models as in Fig. 2.
model with a 37-km thick crust (middle), and a model with a 65-km thick crust that
ts surface-wave data from NE Tibet (Agius and Lebedev, 2010). The period ranges
with the characteristic phase-velocity increase with period due to the S-velocity in- study the properties of the Earth's crust and attempted to measure
creases at the Moho are marked with grey shading. both the velocities and amplitudes of Love and Rayleigh waves. He
also pointed out the differences of surface-wave propagation along
the velocities of surface waves which he called main waves oceanic and continental paths and reported, correctly, that recordings
could be used to study the properties of the outer shells of the at shorter epicentral distances are dominated by shorter period
Earth, by means of measuring phase differences between signals waves compared to those at longer distances. The relation between
recorded at nearby stations. In the early 20th century, velocities of dominant periods and crustal thickness, however, was not handled
surface waves have been estimated, at rst, without taking their dis- accurately, leading to erroneous estimates of crustal thicknesses.
persion into account. Angenheister (1906) gave a velocity estimate of In order to describe Love wave propagation in realistic models of
3.1 km/s for long waves, also citing similar, earlier estimates by the crust, Meissner (1921) gave an expression for Love waves in
Omori. Reid (1910) called surface waves regular waves, while also a crust with a linear increase of seismic velocities with depth.
estimating their velocities. Stoneley (1925) claried the differences between group and phase
Golitsyn (cited here from his selected-works compilation: Golitsyn, velocities. He then gave a quantitative expression for Rayleigh-wave
1960) used minor and major arc recordings of the 1908 Messina earth- dispersion in an Earth model with a compressible uid over an elastic
quake made at Pulkovo observatory and computed a global-average, half-space (Stoneley, 1926). This was particularly useful because the
surface-wave velocity of 3.53 km/s; dispersion, again, was not consid- majority of early surface wave observations were performed for
ered. This value, interestingly, is very similar to the group velocities of paths that traversed oceans. Furthermore, the expressions were im-
Love waves in a typical continent at periods below 25 s, well known mediately applicable at the time because the problem could be solved
today (Fig. 3). Golitsyn also argued that the velocity of surface waves analytically.
should depend on the physical properties of the upper layers of the Surface-wave dispersion in an arbitrarily layered elastic half-space
Earth and be different beneath continents and oceans. was determined by Meissner (1926) for Love waves and by Jeffreys
Love (1911) demonstrated the existence of transversely polarised (1935) for Rayleigh waves. Meissner (1926) also noted the non-
and dispersive surface waves in layered media (the Love waves). The uniqueness of dispersion-curve inversions and gave examples of
observation of Love waves was a direct indication for the layering different one-dimensional (1-D) Earth models that produced very
within the Earth. similar dispersion curves. He concluded that highly accurate mea-
Tams (1921) compared Rayleigh waves propagating along oceanic surements using dense networks would be required in order to deter-
and continental paths and proposed that they had different velocities mine the structure of the outer layers of the Earth.
because the crust beneath oceans, unlike the crust beneath conti- In the early 1920s Gutenberg undertook the rst systematic
nents, did not comprise a granitic layer with relatively low seismic studies of surface-wave dispersion for both Love and Rayleigh waves
velocities within it. He deliberately did not account for dispersion, (e.g., Gutenberg, 1924), also including measurements by Macelwane
considering the accuracy of available measurements insufcient. (1923). He identied the now well known normal surface-wave disper-
Angenheister (1921) argued that surface waves are well suited to sion, characterised by a general increase of surface-wave speeds with
period and indicative of the increase of elastic velocities with depth. Algeria and South Africa. They noted the similarity of their measure-
He also inferred different crustal thicknesses for Eurasia, America ments to those by Brilliant and Ewing (1954) for North America
and the Atlantic and Pacic Oceans. Testing crustal thicknesses of 30, and, also, to a theoretical curve corresponding to a 35-km-thick, ho-
60 and 120 km, he estimated the crustal thickness for Eurasia to be mogeneous crust overlying the mantle. They concluded, as well,
around 50 km. This was a remarkable result, even though he com- that a gradual velocity increase in the crust and the mantle might
pared, incorrectly, measured group velocities with theoretical phase be needed to explain the measured dispersion curves.
velocities. Press (1956) measured phase velocities by examining phase dif-
Neumann (1929) found evidence for lateral heterogeneity of the ferences across arrays of stations, each array comprising only three
Pacic plate by analysing Love and Rayleigh waves. Carder (1934) stations. The reading algorithm he applied can be described as a visu-
summarised Love and Rayleigh wave characteristics known at that al f-k analysis in the time domain. Based on the measurements, he
time; the theoretical understanding of surface waves and the inver- presented a two-dimensional (2-D) cross-section for the crustal
sion tools that were available, however, were not sufcient to draw structure in California, with (over-estimated) Moho-depth variations
accurate conclusions on crustal structure. from ~ 15 km near the coast to ~ 50 km beneath the Sierra Nevada.
Ewing and Press (1950) gave a remarkable synthesis of surface- In the 1960s, the emergence of computer programs for surface-
wave observations, using their own as well as previous measure- wave analysis presented unprecedented new opportunities for accu-
ments (Bullen, 1939; DeLisle, 1941; Wilson and Baykal, 1948). The rate analysis and inversion of surface-wave data. Brune et al. (1960)
very title of their paper, Crustal structure and surface-wave disper- analysed the phase of a dispersed waveform in the time domain and
sion, emphasised the inherent link of the early surface observations proposed improved reading schemes for phase-velocity determina-
to the properties of crust and the Moho. The analysis was based on tion. Alterman et al. (1961) calculated Rayleigh-wave, phase and
sophisticated manual readings of group-velocity dispersion, using group velocities in a 10700 s period range and discussed the effects
time-domain measurements of the arrival times of the dominant of gravity and the Earth's sphericity on the waves' propagation, as
periods (about 1530 s) in the dispersed waveform. The Airy phase well as the relation between surface waves and the Earth's free oscil-
described by Pekeris (1948) was identied correctly, based on lations. Using the measurements of Ewing and Press (1956) and Nafe
Stoneley's equation, and theoretical group-velocity curves were tted and Brune (1960), Alterman et al. (1961) also presented evidence
to the observations. The strong inuence of water and sediments on supporting the Gutenberg's model of the Earth with a low velocity
Rayleigh-wave velocities was clearly established. Interpreting the re- asthenosphere.
sults, Ewing and Press (1950, 1952) implicitly applied ray theory and Dorman and Ewing (1962) developed a linearised scheme for the
derived estimates for path-average, sub-crustal velocities and the inversion of surface-wave measurements and computed a Moho
Moho depths by estimating the continental portion of the paths. depth of about 39 km for the New YorkPennsylvania area. Brune
The limited bandwidth of their observations, however, and their use and Dorman (1963) compared Love and Rayleigh waveforms at sta-
of a simplied Earth model with one layer (water and sediments) tions within the Canadian Shield in the time domain. They deter-
overlying a half-space, implied that their results were most meaning- mined phase velocities in a 540 s period range and inverted them
ful for sub-crustal velocities, and less so for the properties of the crust. for a 1-D, S-wave velocity model with a multilayered crust and
Citing Love-wave, group-velocity observations by Wilson (1940), upper mantle, detecting high S-wave velocities within the mantle
Ewing and Press (1950) also noted that Love waves show higher lithosphere of the craton. They also calculated synthetic dispersed
group velocities for oceanic paths compared to continental paths in waveforms for their model.
the period range of 20100 s and conrmed that, in contrast to Rayleigh Toksz and Ben-Menahem (1963) followed an earlier suggestion
waves, Love waves are insensitive to the water layer. by Sato (1955) and measured phase velocities in the frequency do-
Brilliant and Ewing (1954) measured, in the time domain, main, using successive passages of surface waves at a single station.
Rayleigh-wave phase differences between stations in the US, elimi- McEvilly (1964) measured the phase difference between two stations
nating the phase shifts due to the source and the oceanic portions in the frequency domain, for both Love and Rayleigh waves. Inverting
of the paths. They determined the rst phase-velocity curve for the resulting dispersion curves, he established that different 1-D
North America between 18 s and 32 s, a period range where phase models were needed for the horizontally and vertically polarised S
velocities are sensitive mainly to the crust and the Moho. waves (Vsh and Vsv, respectively). This observation became known
Evernden (1954) analysed group velocities of Love waves be- as the LoveRayleigh discrepancy.
tween 7 and 45 s for the Pacic Basin. Using Stoneley's analytical ex- Santo and Sato (1966) developed a regionalization technique that
pressions for the dispersion of Love waves in a three-layered model, may be seen as the rst attempted group-velocity tomography.
he concluded that a sedimentary layer, a high-velocity crust and a Knopoff et al. (1967) described lter and triangulation techniques for
seismic-velocity increase from the crust to the mantle were all neces- the determination of phase velocities. The determination of group ve-
sary to explain the measurements. He also noted that the results of locities using spectrograms was then suggested by Landisman et al.
the surface-wave analysis were compatible with those of seismic re- (1969).
fraction surveys. These general conclusions still stand today, although Since the 1970s, the number of surface-wave studies has grown
the crustal and mantle models have since been improved substantial- steadily. Compilations and reviews of surface-wave analyses in the
ly in their details. beginning of this period are given by Dziewonski (1970), Knopoff
The solution for Rayleigh-wave phase velocities in a model with a (1972), Seidl and Mller (1977), Kovach (1978) and Levshin et al.
water layer overlying two solid layers made it possible to interpret (1989). With long-period surface-wave measurements increasingly
Rayleigh-wave group velocity curves measured along oceanic paths. accurate and abundant, surface waves were now used extensively
Oliver et al. (1955) discussed available dispersion measurements for the study of the upper mantle. In the course of inversions of the
for Rayleigh and Love waves. They concluded that a high velocity long-period data, the substantial sensitivity of surface-wave speeds
crust is present under the oceans and were able to rule out the to crustal structure was often accounted for by means of crustal
hypothesised existence of a large continent submerged beneath the corrections: the effect of the crustal structure on surface-wave
Pacic Ocean. They also showed that at periods longer than about measurements was evaluated using a priori crustal models, usually
25 s the dispersion curves for the Atlantic and Pacic basins were constrained by other seismic methods (e.g., Bassin et al., 2000;
similar. Boschi and Ekstrom, 2002; Bozdag and Trampert, 2008; Ferreira
Press et al. (1956) made the rst single-station measurements et al., 2010; Kustowski et al., 2007; Lekic et al., 2010; Marone and
over a 1070 s period range for a pure continental path, between Romanowicz, 2007; Montagner and Jobert, 1988; Mooney et al.,
1998; Nataf and Ricard, 1996; Nataf et al., 1986; Nolet, 1990; Panning constructive interference of surface waves within the ambient noise
et al., 2010; Woodhouse and Dziewonski, 1984). wave eld that arrive to a pair of stations at and near the station-
Thanks to the rapid growth of broadband seismic networks since station azimuth. Interestingly, the wave elds used for these measure-
the 1990s, increasingly large surface-wave datasets were used in re- ments cannot at all be described by ray theory, but the phase and
gional and global imaging. Many tomographic inversions included group velocities extracted from the cross-correlation functions can
the crustal structure and thickness as inversion parameters (e.g., Das dene surface-wave propagation along inter-station paths in a ray-
and Nolet, 1995; Lebedev and Nolet, 2003; Lebedev et al., 1997; Li theoretical framework.
and Romanowicz, 1996; Pasyanos and Walter, 2002; Shapiro and Over the last few years, the new methods have been applied to
Ritzwoller, 2002; Van der Lee and Nolet, 1997) (Fig. 4), and some broadband array data from around the world. The newly abundant
surface-wave studies targeted primarily the Moho itself (Das and short-period and broad-band surface-wave measurements are, once
Nolet, 1995, 1998; Marone et al., 2003; Meier et al., 2007a, 2007b). again, bringing the crust into the focus of surface-wave seismology
The main difculty in resolving the Moho with surface waves remained (e.g., Adam and Lebedev, 2012; Bensen et al., 2007; Deschamps et
the non-uniqueness of seismic-velocity and Moho depth models consis- al., 2008b; Endrun et al., 2008, 2011; Lin et al., 2011; Moschetti et
tent with surface-wave observations. Resolving the trade-offs between al., 2010; Pawlak et al., 2012; Polat et al., 2012; Shapiro et al., 2005;
the Moho depth and seismic velocities required highly accurate mea- Yang et al., 2008, 2011, 2012; Yao et al., 2008; Zhang et al., 2007,
surements at intermediate and relatively short periods (Fig. 2). Phase 2009) (Fig. 5). It is thus particularly appropriate at this time to exam-
velocities of surface waves, however, were difcult to measure at short ine in detail the sensitivity of surface waves to the Moho and the res-
periods, with the waveforms of teleseismic surface waves distorted by olution of the Moho properties that they can provide.
diffraction at periods below 1520 s, and with regional source-station
measurements biased substantially even by small errors in earthquake 3. Sensitivity of surface waves to the Moho
locations.
The surface-wave crustal imaging has been rejuvenated in the Characteristic signatures of the crustal thickness are clearly seen
2000s by the emergence of new, array techniques for surface-wave in various surface-wave observables, including phase-velocity curves
measurements. Phase velocities of short-period surface waves are (Fig. 2), group-velocity curves (Fig. 3), and waveforms of surface-
now measured routinely using pairs or arrays of broadband stations. wave trains on broad-band seismograms. The wave forms are closely
The measurements are mainly by means of cross-correlation of either related to the frequency-dependent phase velocities. In a weakly
diffracted surface waves from teleseismic earthquakes (Meier et al., heterogeneous Earth, a complete seismogram can be computed as a
2004) or of surface waves within the ambient seismic noise (Shapiro superposition of the fundamental and higher surface-wave modes
and Campillo, 2004). using the JWKB (JeffreysWentzelKramersBrillouin) approxima-
The cross-correlation of surface-wave recordings from nearby station as:
tions is, essentially, the classical two-station method (Brilliant and

Ewing, 1954; McEvilly, 1964; Press, 1956; Sato, 1955; Toksz and s Am exp iC m ; 1
Ben-Menahem, 1963). The difference of the modern and traditional m
applications is in the types of the signal they use. The classical
two-station method was applied only to teleseismic surface waves where the sum is over modes m, is the circular frequency, is the
that obeyed surface-wave ray theory, i.e., were not distorted by diffrac- sourcestation distance, C m are the average phase velocities of
tion. Of the new techniques, teleseismic cross-correlations (Meier et al., the modes along the source-station path, and Am() are the complex
2004) can extract inter-station phase-velocity measurements even amplitudes of the modes, depending on the source mechanism and
from wave elds diffracted at teleseismic distances, and the ambient the Earth structure in the source region, as well as on geometrical
noise cross-correlations (Shapiro and Campillo, 2004) make use of the spreading and attenuation (Dahlen and Tromp, 1998).
Topography Moho, from waveform tomography Moho, from CRUST2. 0
40N 40N
20N 20N
0 0
100E 120E 140E 100E 120E 140E 100E 120E 140E
-12 -6 -5 -4 -0 1 2 3 5 8 5 15 18 21 30 40 50 60 75
topography, km Moho depth, km
Fig. 4. Resolving the Moho in East Asia-Western Pacic with surface-wave, waveform tomography (Lebedev and Nolet, 2003). Centre: Moho depths resulting from a 3D tomograph-
ic inversion of surface-wave forms for crustal and mantle shear-speed structure. The 1D background model had a 25-km Moho depth. The large system of linear equations solved in
the inversion was assembled from results of multi-mode waveform inversions of around 4000 seismograms with source-station paths within the region. Even with a 1D reference
model, the 3D inversion reproduces reasonably correct Moho depths, demonstrating the sensitivity of surface-wave waveforms to the Moho.
Receiver functions Ambient noise Ambient noise

& teleseismic tomograpy
-20 A B C -20
-24 -24
-28 -28
-32 -32
20 24 28 32 20 24 28 32 20 24 28 32
32 34 36 37 38 39 40 41 42 43 45 50
crustal thickness km
Fig. 5. Results of the Moho mapping in southern Africa using receiver functions (left, Nair et al., 2006) and surface-wave phase velocities, measured with ambient noise (centre) and
both ambient noise and teleseismic signals (right) (Yang et al., 2008). (Figure courtesy of Yingjie Yang.)
Group velocity U is the velocity of propagation of the surface depth variations of only a few kilometres will correspond to easily de-
wave's energy; it depends on the phase velocity and its frequency de- tectable (>1%) perturbations in phase velocities.
rivative as Group velocities (Fig. 7) show an even stronger sensitivity to the
Moho depth. For a typical continental crustal thickness (37 km), a
C Moho-depth change of only 1 km for Rayleigh or 2 km for Love
U : 2
1=C dC=d waves results in a group-velocity perturbations up to almost 1%.
The sensitivity of surface waves to the Moho beneath oceans
While exploring the surface waves' sensitivity to the properties of (Figs. 8 and 9) is different from that beneath continents, both because
the Moho, we shall focus on the phase and group velocities only, of the small thickness of the oceanic crust and because of the pres-
while noting that different surface-wave observables may have difference of the water layer, which has a strong effect on the propagation
ent useful properties (for example, local minima in the group velocity of Rayleigh waves (Figs. 13). Until recently, it has been difcult to
curve, causing an Airy phase, have some sensitivity to the sharpness measure phase or group velocities in oceans at periods sufciently
of the Moho). short to resolve the shallow oceanic Moho. In the last few years,
Fig. 6 illustrates the sensitivity of the Rayleigh and Love phase- deployments of arrays of Ocean-Bottom Seismometers (OBS) have -
velocity curves to the Moho depth in a typical continental model nally provided the data for such measurements (e.g., Harmon et al.,
with a 37-km thick crust. If seismic velocities in the crust and upper 2012; Yao et al., 2011). Although measurement errors in the
mantle can be xed (i.e., assumed to be known), then small Moho- surface-wave data from OBS arrays are relatively large, the signal of
Phase velocity
0 5.0 2
A Rayleigh
4.5 1
C, km/s
C, %
0
4.0
50 -1
Depth, km
3.5 B D
-2
2
5.0 Love
100 1
C, km/s
C, %
4.5
0
Normal
Continent 4.0 -1
C E
150 3.5 -2
3.5 4.0 4.5 5 10 20 50 100 200 5 10 20 50 100 200
S wave velocity, km/s Period, s Period, s
Fig. 6. Sensitivity of surface-wave phase velocities to the depth of the Moho in a typical continental model. Vs and other model parameters are xed in the crust and the mantle, and
the Moho is shifted up and down, at 1 km increments, from its 37-km reference depth. Grey and black lines show the 1-D models tested (A), the corresponding Rayleigh- and
Love-wave phase velocity curves computed for these models (B and C, respectively), and the relative changes in phase velocities (D, E), with respect to the curve for the reference
model that has a 37-km thick crust (A). Black lines correspond to the models with the Moho depth within 3 km of the reference value.
Group velocity
0 6
A Rayleigh 4
4.0
2
U, km/s
U, %
3.5 0
50 -2
3.0 -4
Depth, km
B D
-6
6
4.5
Love 4
100 2
U, km/s
U, %
4.0 0
Normal
-2
Continent
3.5 -4
C E
150 -6
3.5 4.0 4.5 5 10 20 50 100 200 5 10 20 50 100 200
Fig. 7. Sensitivity of surface-wave group velocities to the depth of the Moho in a typical continental model. Denitions of the proles and curves are as in Fig. 6.
the crustal structure and thickness in these data is also large. A 1-km smaller phase-velocity changes compared to those in the depth of a
perturbation in the depth of an oceanic Moho corresponds to a per- shallower Moho, because the sensitivity kernels of surface waves
turbation of 0.75% in phase velocity and over 2% in group velocity of sampling the deeper Moho will be broader (Fig. 1). The thickest
Rayleigh waves (Figs. 8 and 9). crust beneath high plateaux, however, is also characterised by low
Fig. 10 summarises the sensitivity of phase and group velocities to seismic velocities within it (e.g., Agius and Lebedev, 2010; Yang et
the Moho depth in different tectonic settings. The cumulative mists al., 2012), which enhances the crust-mantle, seismic-velocity contrast
between the perturbed and reference phase- and group-velocity and, thus, the visibility of the Moho.
curves (top row) are computed over the entire length of the broad-
band curves, with sample spacing increasing logarithmically with in-
creasing period so as to equalize, roughly, the weight of the structural 4. Trade-offs between the Moho depth and other
information given by different parts of the phase-velocity curve, sen- model parameters
sitive to different depth intervals within the Earth (Bartzsch et al.,
2011). The mists do not have a physical meaning; comparisons of The effect of the Moho on surface wave speeds reects primarily
the mists in the analysis and inversion of the same phase-velocity the shear-wave speed increase from the crust to the mantle. Varia-
curves, however, are consistent and meaningful. The mists show tions in shear speeds in the lower crust or uppermost mantle give
steep valleys with clear minima at the correct (reference) Moho rise to perturbations in surface-wave speeds similar to those due to
depth values. For continents with either normal or thickened crust, Moho-depth variations. If seismic velocities in the crust and mantle
Rayleigh and Love waves show a similar sensitivity to the Moho, are not known a priori as is the case most often then an inversion
with the periods of maximum sensitivity increasing with an increas- of surface-wave data will suffer from a trade-off between the param-
ing Moho depth, and with the period ranges of sensitivity broader eters for the Moho depth and the crustal and mantle shear speeds.
for Love waves compared to Rayleigh waves. Generally, perturbations The resulting model non-uniqueness translates into uncertainty in
in the depth of a deeper Moho can be expected to translate into the Moho depth.
Phase velocity
0 5 2
A Rayleigh
4 1
C, km/s
C, %
10
3 0
-1
Depth, km
20 2
B D
-2
2
30 5.0 Love
1
C, km/s
C, %
4.5
0
40
Ocean 4.0 -1
C E
50 3.5 -2
0 1 2 3 4 5 10 20 50 100 200 5 10 20 50 100 200
Fig. 8. Sensitivity of surface-wave phase velocities to the depth of the Moho in a typical oceanic model. Denitions of the proles and curves are as in Fig. 6.
Group velocity
0 5 6
A Rayleigh 4
4
2
U, km/s
U, %
10 3 0
2 -2
Depth, km
20 -4
1 B D
-6
5.0 6
30 Love 4
4.5
2
U, km/s
U, %
4.0 0
40
3.5 -2
Ocean
-4
3.0 C E
50 -6
0 1 2 3 4 5 10 20 50 100 200 5 10 20 50 100 200
Fig. 9. Sensitivity of surface-wave group velocities to the depth of the Moho in a typical oceanic model. Denitions of the proles and curves are as in Fig. 6.
The trade-offs are quantied in Fig. 11. In each of the two-parameter the mist between the synthetic and measured phase-velocity curves.
planes, the parameter on the horizontal axis is for the Moho depth, and Perturbations to the background shear-speed proles (Fig. 12A) are
the parameters on the vertical axes are for shear-speed perturbations in parameterised using 1520 boxcar (crust) and triangle (mantle)
the lower crust and in the uppermost mantle and for the thickness of the basis functions, with the width of the basis functions increasing
Moho. For both Rayleigh and Love waves, the Moho depth and shear with depth (see Bartzsch et al., 2011, for details). It is important
speeds above and below the Moho show the expected trade-offs: a mis- that the crustal and mantle structure is over-parameterised, i.e. that
t due to an increase (decrease) of the Moho depth can be compensat- the number of basis functions is large enough so that the choice of a
ed, to a large extent, by an increase (decrease) of the wavespeeds above particular number does not affect the minimum mist achievable
or below the Moho. This is fundamentally due to the broad depth range with various Moho depths. At the same time, the shear-speed proles
of surface-wave depth sensitivity functions (Fig. 1). are constrained to be relatively smooth, both implicitly, by the nite
The trade-off between the Moho depth and its thickness is weak, widths of the basis functions (10 km or greater depth ranges in
and the sensitivity of surface waves to the Moho thickness in general the crust; a few tens of km in the mantle) and by the slight norm
is low. Surface waves alone are thus insufcient to determine wheth- damping applied to the inversion parameters. Given reasonably
er the crustmantle transition is a sharp discontinuity or a gradient accurate phase-velocity measurements, small damping is sufcient
over a depth range. The ne structure of a discontinuity can, however, to rule out exotic models with unrealistic shear-speed values.
be investigated by means of joint analysis of surface-wave data or Compressional-wave speed perturbations are coupled to shear-wave
models and other data, such as receiver functions (e.g., Endrun speed ones (VP (m/s) = VS (m/s)). (This assumption is reasonable
et al., 2004; Juli et al., 2000; Lebedev et al., 2002a, 2002b; Shen et for the upper mantle but not always for the crust, particularly in
al., 2013; Tkali et al., 2012). The incorporation of such additional sedimentary layers. In the examples below, sedimentary layers are
data can also reduce the trade-offs between the Moho depth and absent or insignicant, but in general the variations in crustal
shear speeds (Fig. 11). Poisson's ratios will add to the uncertainties of the inversion for the
Moho depth; a priori information on the structure of the sediments
5. Inversion of surface-wave measurements for the Moho depth is thus particularly valuable (Section 7.2).) The non-linear gradient
search is performed with the LevenbergMarquardt algorithm. Syn-
We now set up an inversion procedure that will help us to not only thetic phase velocities are computed directly from one-dimensional
determine the best-tting Moho-depth values but also explore the (1-D) Earth models at every step during the gradient search, using a
properties of the multi-parameter model space that are most relevant fast version of the MINEOS modes code (Masters, http://igppweb.
to the Moho depth and its uncertainty. The procedure is similar to ucsd.edu/gabi/rem.dir/surface/minos.html), which we modied
that described by Bartzsch et al. (2011), who projected the from the version of Nolet (1990). The gradient search is not linearised
smallest-mist surface in a multi-dimensional parameter space onto a and converges to true best-tting solutions (Erduran et al., 2008).
two-parameter plane (the two parameters in that study being the The inversion procedure is thus a grid search (the grid, in this case,
depth and the thickness of the lithosphere-asthenosphere boundary). being one-dimensional) that comprises numerous non-linear gradi-
Here, we use a one-parameter axis instead of a two-parameter plane ent searches, one at each knot along the Moho axis. The gradient
and focus on the Moho depth only (the Moho thickness being difcult searches determine best-tting shear-speed proles that minimise
to constrain with surface waves with useful accuracy (Fig. 11)). Our the mist as much as possible with the Moho depth xed at the
goal is to investigate the general properties of the inversion of value that denes the point on the axis. Any trade-offs between the
surface-wave data for the Moho depth. Moho depth and shear speeds above and below it will contribute to
minimizing the impact of the Moho on the mist function (that is,
5.1. Mapping the model space the gradient-search inversion will compensate, as much as possible,
the impact of changes in the Moho depth with changes in shear
For every point along the Moho-depth axis, we perform a speeds above and below it). If the Moho depth, however, is not con-
non-linear gradient search inversion in which the Moho depth is sistent with the data, then the best possible t will still be relatively
xed and the crustal and mantle structure is varied, so as to minimise poor.
Phase velocity Group velocity

2.5 2.5
Ocean Normal : Rayleigh Ocean Normal : Rayleigh
continent : Love continent : Love
2.0 2.0
Tibet Tibet
Misfit x103
Misfit x103
1.5 1.5
1.0 1.0
0.5 0.5
0.0 0.0
10 20 30 40 50 60 70 80 10 20 30 40 50 60 70 80
Depth, km Depth, km
Rayleigh waves Rayleigh waves

200 Ocean Normal Tibet Ocean Normal Tibet 200
continent continent
100 100
Period, s
Period, s
50 50
20 20
10 10
5 5
10 20 30 40 50 60 70 80 10 20 30 40 50 60 70 80
Love waves Love waves

200 Ocean Normal Tibet Ocean Normal Tibet 200
continent continent
100 100
Period, s
Period, s
50 50
20 20
10 10
5 5
10 20 30 40 50 60 70 80 10 20 30 40 50 60 70 80
Depth, km Depth, km
-3 -2 -1 0 1 2 3
Phase and group velocity perturbation, %
Fig. 10. Sensitivity of the phase-velocity (left) and group-velocity (right) curves of fundamental-mode surface waves to the depth of the Moho in different tectonic settings. Top:
mists computed over the length of the broad-band curves. The mists are due to deviations of the Moho depths from their reference values (ocean: 11 km relative to the sea
surface; normal continent: 37 km from the surface; Tibet: 65 km from the surface). Seismic velocities in the crust and mantle are xed. Middle and bottom: Phase- and
group-velocity changes at each period due to changes of the Moho depth from its reference value. The 1-D models and corresponding phase- and group-velocity perturbations
for a normal continent (37-km Moho depth) are the same as in Figs. 6 and 7; for an ocean same as in Figs. 8 and 9. The Tibet reference model has a 65-km thick crust and
a relatively low, 3.6 km/s S-wave velocity in the lower crust (Agius and Lebedev, 2010).
5.2. Resolution and trade-offs Although the V shapes of the mist curves (Fig. 12B) look similar
to those in the sensitivity tests where only the Moho depth was var-
We rst apply the model space mapping procedure to synthetic ied (Fig. 10, top), the curves are, in fact, quite different: the mists are
phase-velocity curves, computed for a reference model with a now around 10 times smaller. The best-tting, phase-velocity curves
37-km Moho depth (Fig. 12A). The results show that both the Ray- for all the Moho depths tested in the inversion are much closer to the
leigh and Love wave data have the capacity to resolve the Moho reference curve (synthetic data) and to each other (Fig. 12C, D) than
depth accurately: the V-shaped mist curves have a clear minimum the different curves in the sensitivity tests (Fig. 6B, C). This
at the correct Moho depth (Fig. 12B). order-of-magnitude reduction in the mists is due to the trade-offs
Phase velocity Group velocity
Rayleigh waves Love waves Rayleigh waves Love waves

30 35 40 45 30 35 40 45 30 35 40 45 30 35 40 45
15 15 30 30
velocity variation, km/s

0 15 0 15 0 30 0 30
0.2 0.2 0.2 0.2
Lower-crust
Lower-crust
0.1 0.1 0.1 0.1
0.0 0.0 0.0 0.0
-0.1 -0.1 -0.1 -0.1
-0.2 -0.2 -0.2 -0.2

30 35 40 45 30 35 40 45 30 35 40 45 30 35 40 45
30 35 40 45 30 35 40 45 30 35 40 45 30 35 40 45
15 15 30 30
0 15 0 15 0 30 0 30

0.2 0.2 0.2 0.2
Upper-mantle
Upper-mantle
0.1 0.1 0.1 0.1
0.0 0.0 0.0 0.0
-0.1 -0.1 -0.1 -0.1
-0.2 -0.2 -0.2 -0.2

30 35 40 45 30 35 40 45 30 35 40 45 30 35 40 45
30 35 40 45 30 35 40 45 30 35 40 45 30 35 40 45
15 15 30 30
0 15 0 15 0 30 0 30
Moho thickness, km
Moho thickness, km
30 30 30 30
20 20 20 20
10 10 10 10
0 0 0 0
30 35 40 45 30 35 40 45 30 35 40 45 30 35 40 45
Moho depth, km Moho depth, km Moho depth, km Moho depth, km
0 2 4 6 8 10
Misfit x10 4
Fig. 11. Trade-offs between the Moho depth and other seismic model parameters. The reference model (a cross in each frame) is as in Fig. 6(A), with a 37-km deep Moho. In each of
the tests, the Moho depth and one other parameter were perturbed within the ranges shown, with the rest of the model unchanged, and the mist was computed at each point
within the 2-parameter planes. The mist is the average relative difference between the perturbed and reference broad-band phase-velocity curves computed over their entire
length (5250 s). Top: the trade-off between the Moho depth and S-wave velocities in the lower crust (between 15-km depth and the Moho). Middle: the trade-off between
the Moho depth and S-wave velocities in the uppermost mantle (between the Moho and a 100-km depth). Bottom: the (weak) trade-offs between the depth and thickness of
the Moho. Variations in the Moho thickness were parameterised using a layer with a linear seismic-velocity increase within it, centred at the value of the Moho depth.
between the Moho depth and the crustal and mantle structure. The 5.3. Inversion of measured data: Northern Kaapvaal Craton
adjustments in the crustal and mantle structure determined in the
course of an inversion can compensate for an incorrect Moho well Applying the inversion to real data, we now invert phase-velocity
enough to mask around 90% of the signal. curves measured in northern Kaapvaal Craton (24-26S, 26-32E),
The effects of the trade-offs can be claried further by a comparison southern Africa (see the map in Fig. 5). Adam and Lebedev (2012)
of the relative differences of phase-velocity curves in the inversion computed the average curves for this region by averaging thousands
(where both the Moho depth and the crustal and mantle structure of inter-station measurements, obtained by both cross-correlation
were varied, Fig. 12E-H) and in the sensitivity tests (where only the and multimode waveform inversion (Lebedev et al., 2006, 2009;
Moho depth was varied, Figs. 6D, E and 10, middle and bottom left). If Meier et al., 2004). (The region-average measurements and inver-
only the Moho depth is perturbed, then its change by a few kilometres sions are meaningful because the Moho depth shows variations of
results in phase-velocity perturbations that vary gradually with period only a few kilometres across the northern Kaapvaal Craton, and
and reach a maximum on the order of 1% (Figs. 5, 6). In the inversion, shear-velocity heterogeneity is also limited, according to published
where the effect of the Moho-depth perturbations is partly compensat- receiver-function studies and tomography (e.g., Kgaswane et al.,
ed by perturbations in crustal and mantle seismic-velocity structure, the 2009; Nair et al., 2006; Yang et al., 2008).) The highly accurate
same Moho-depth changes result in oscillatory phase-velocity pertur- phase-velocity curves span very broad period ranges, particularly
bations up to a maximum on the order of 0.1% only. for Rayleigh waves (up to 5400 s).
S wave velocity, km/s Period, s Period, s Moho depth, km

3.5 4.0 4.5 5 10 20 50 100 200 5 10 20 50 100 200 30 36 42 48
0 5.0
200
A Rayleigh 0.4
4.5 100
Depth, km
Period, s
50
C, km/s
C, %
50
0.0
4.0
100 20
Normal
continent 3.5 -0.4 10
C E G
150 5
3
5.0 200
Love 0.4
100
Misfit x104
Period, s
2
C, km/s
4.5
C, %
50
0.0
1 4.0 20
-0.4 10
B D F H
0 3.5 5
30 36 42 48 5 10 20 50 100 200 5 10 20 50 100 200
Moho depth, km Period, s Period, s -0.2 -0.1 0.0 0.1 0.2
Phase velocity perturbation C, %
Fig. 12. Resolution tests: model-space-map inversions of synthetic phase-velocity curves for the Moho depth. The inversions of Rayleigh (C, E, G) and Love (D, F, H) waves were
performed separately. The model spaces were explored by means of a uniform sampling of the Moho-depth axis, with a non-linear, gradient-search inversion at each point. In
each of the gradient searches, the Moho depth is xed but the crustal and mantle shear-speed structure is allowed to vary, so that the trade-offs between the Moho depth and
shear velocities are taken into account. A: the best-tting shear-speed proles for each Moho depth in the 2949 km range. B: the minimum data-synthetic mists given by the
gradient-search inversions at each of the 1-km-spaced points on the Moho-depth axis (crosses). C, D: Rayleigh and Love phase-velocity curves, coloured for the reference model
and grey and black (very close to or behind the coloured curves) for all the proles in (A). E, F: differences between the best-tting phase-velocity curves determined for various
Moho depths and the reference curves computed for the reference model with a 37-km Moho depth. Black lines in A, E, F indicate globally best-tting models and curves, with
cumulative mists below the threshold indicated by the dashed line in (B). G, H: Period-dependent differences between the best-tting phase-velocity curves at each (xed)
Moho depth and the reference curves. The characteristic patterns of alternating positive and negative differences in the Moho depth-period plane reect the trade-offs of the
Moho depth and shear-speed structure in the crust and the mantle.
In Fig. 13 we show the results of three different inversions of the this reason, the inversions of smoothed data are best used for testing
Rayleigh-wave phase velocities. In the rst (Fig. 13, top row: C, F, I), and validation, and not as the primary way to determine the Moho
we inverted the measured dispersion curve in a relatively broad peri- depth. Ideally, the results of the smoothed-data and original-data
od range (570 s), in which it had substantial sensitivity to the upper inversions should be consistent, indicating their robustness (e.g.,
and lower crust, to the Moho, and to the lithospheric mantle. The Deschamps et al., 2008a; Endrun et al., 2011). The cumulative mist
measured curve can be t with synthetic curves closely (within a curves in our inversions for the Moho depth, however, do not show
line thickness in Fig. 13C). The mists can be seen more clearly such consistency and are substantially different for the two inver-
when relative phase-velocity differences are plotted (Fig. 13F); they sions (Fig. 13B): the larger mists given by original-data inversions
suggest that the noise in the measurements is up to 0.10.2%, varying form a broader smallest-mist valley, centred at Moho depths that
with period. All the shear-velocity proles corresponding to the syn- are 78 km greater, compared to the mists in the smoothed-curve
thetic phase-velocity curves in Fig. 13C and F (as well as in Fig. 13D, E, inversion. This implies that the accuracy of the original-data inver-
G and H) are plotted in Fig. 13A. The pattern of frequency-dependent sion has suffered from the errors in the data (which we estimated
phase-velocity perturbations due to Moho-depth changes (Fig. 13I) is to be up to ~ 0.2%).
similar to that in inversions of noise-free, synthetic data (Fig. 12G), In order to reduce the effect of measurement errors, we set up a
but with distortions due to the errors in the measurements. third inversion (Fig. 13, bottom row: E, H, K). We now invert a
In the second inversion (Fig. 13, middle row: D, G, J), we attempt narrow-band curve, in a period range most sensitive to the Moho
to remove the effects of the noise and invert the dispersion data in (1532 s). Because this narrow-band curve has limited sensitivity to
the same period range as in the rst inversion but smoothed before- the crustal and mantle structure, an accurate reference prole of
hand. The smoothing was by means of an over-parameterised and crustal and mantle shear-velocity must be used. Such prole is pro-
under-damped, gradient-search inversion of the phase-velocity vided by the results of the original, broad-band inversion.
curve for a 1-D shear-velocity prole (the prole itself being of no im- The narrow-band inversion shows a steep mist valley, with
portance). While this inversion can t structural information in the best-tting Moho depths in the 3741 km range (Fig. 13B). These
data, it cannot t random errors with a strong period dependence values are roughly consistent with the Moho depths of 4045 km deter-
(high-frequency noise), inconsistent with any plausible Earth mined in the region using receiver functions (Kgaswane et al., 2009;
models. Random errors thus get smoothed out to a large extent. Com- Nair et al., 2006) and the Moho depths of 4043 km constrained by
pared to the original-data inversion, the inversion of the smoothed Rayleigh-wave measurements in a 640 s period range, made with
dispersion curve reaches smaller mists for best-tting Moho depths cross-correlations of ambient seismic noise and inverted with starting
(Fig. 13G). It also shows a more regular pattern of perturbations of models similar to those given by receiver-function analysis (Yang et
best-tting phase velocities as a function of the Moho depth (Fig. 13J). al., 2008) (Fig. 5). Receiver-function measurements have their own un-
The smoothed-curve inversion (Fig. 13D, G, J) conrms that when certainties due to trade-offs of the Moho depth and the crustal Vp/Vs ra-
random errors in the data are reduced, frequency-dependent mists tios; in surface-wave inversions, uncertainties result from trade-offs of
display patterns that are more similar to those in synthetic-data in- the Moho depth and crustal and uppermost-mantle shear-velocity
versions, and the Moho depth can probably be resolved. The smooth- structure. These uncertainties are the most likely reason for the appar-
ing, however, may by itself introduce new biases into the data. For ent small discrepancy between the different measurements in northern
S wave velocity, km/s Period, s Period, s Moho depth, km

3.5 4.0 4.5 5.0 5 10 20 50 5 10 20 50 30 40 50 60
0 4.2
50
A Rayleigh 0.4 50
4.0
Period, s
C, km/s
C, %
100
3.8 0.0 20
Depth, km
150
3.6
200 10
3.4 -0.4
250
C F I
5
300 4.2
Smoothed 0.4 50
N. Kaapvaal
350 4.0 Rayleigh
Period, s
Craton
C, km/s
C, %
400 3.8 0.0 20
3.6
2.0 10
3.4 -0.4
D G J 5
1.5
Misfit x104
4.2
Narrow-band 0.4 50
4.0 Rayleigh
Period, s
1.0
C, km/s
C, %
3.8 0.0 20
0.5 3.6
10
3.4 -0.4
B E H K
0.0 5
30 40 50 60 5 10 20 50 5 10 20 50
Moho depth, km Period, s Period, s -0.2 -0.1 0.0 0.1 0.2
Phase velocity perturbation C,
Fig. 13. Inversion of the Rayleigh-wave phase-velocity curve from the northern Kaapvaal Craton. A: best-tting, shear-speed proles computed for Moho depths xed at values
between 28 and 60 km. The black proles correspond to the black phase-velocity curves in (H). B: minimum mists given by the gradient searches with the Moho depth xed
at various values (crosses) and the crustal and mantle structure allowed to vary. Blue, red and green curves correspond to the inversion of the measured broad-band curve (top
row: C, F, I), smoothed broad-band curve (middle row: D, G, J), and a narrow-band curve over periods most sensitive to the Moho (bottom row: E, H, K), respectively. When
the measured, broad-band curve is inverted, the modest noise at the shortest and longest periods contributes to mists sufciently to make the Moho depth very uncertain. Inver-
sion of the narrow-band curve, with an accurate background shear-speed model pre-computed in a preliminary broad-band inversion, yields the most robust results.
Kaapvaal Craton. Joint analysis of surface-wave and receiver-function Love-wave data did not contribute usefully to constraining the Moho
data could help to reduce some of these uncertainties and, also, to con- depth, due to the higher levels of noise in them.
strain the ne structure of the Moho. Small discrepancies notwithstand-
ing, the close agreement between the results of the receiver-function
analysis and those yielded by the surface-wave inversion with no a 6. Noise in the data: how much is too much for the Moho
priori information (shear speeds in the crust and upper mantle were to be resolved?
allowed to vary in unlimited, very broad ranges, and the trial Moho
depth values spanned a very broad, 2860 km range) validates the in- As we saw in Section 5.3, errors in surface-wave measurements
version set-up and conrms the resolving power of surface waves. (noise in the data) can bias the Moho-depth values yielded by the in-
The inversion procedure that is optimal thus has two steps: in the version of the data. This will occur regardless of what inversion ap-
rst step, we use a broad-band dispersion curve to determine the proach is used. The trade-offs between the Moho depth and crustal
mantle and crustal structure with a reasonable accuracy (Fig. 13A); and mantle seismic velocities make the signal of the Moho depth in
in the second step, we use that as a reference model (which can still the data very subtle: ~ 0.10.2% of the phase velocity values. If the in-
be perturbed) while inverting only the part of the curve in the narrow version of the data accounts for the trade-offs correctly and if no a
period range where the signal of the Moho depth is the strongest and priori information on seismic velocities is available, then an amount
most likely to be well above the noise level. (This assumes that the of noise that is similar to or higher in amplitude than the signal of
Moho is associated with a seismic-velocity contrast. This contrast is the Moho may bias the results of the inversion.
seen, empirically, in the steep increase in the phase velocity at pe- The resolvability of the depth of a continental Moho with surface-
riods sampling primarily the depth range around the Moho. It is this wave, phase-velocity data alone (and with no a priori information) is
period range that is used in the second-step inversion.) not warranted if the noise level exceeds ~ 0.2%. With stronger noise,
Love-wave phase-velocity curves show sensitivity to the Moho the Moho may or may not be resolved correctly, depending on the
depth similar to that of the Rayleigh-wave ones. Unfortunately, there character of the noise.
is usually more noise in Love-wave measurements. For the northern We illustrate the effects of different noise patterns in Fig. 14. (For
Kaapvaal Craton, errors in the Love-wave phase-velocity curve of completeness, mists as a function of periods are presented in
Adam and Lebedev (2012) appear to be up to 0.20.3% at 550 s and Supplementary Fig. 1). Five different synthetic phase-velocity curves
up to 0.5% at 6070 s. Inversions of Love-wave phase velocities, were inverted in different model-space-map inversion tests. One of
performed in the same way as those for Rayleigh waves (Fig. 13) did the ve curves was computed for a cratonic seismic-velocity prole
not provide robust solutions for the Moho depth. We also attempted (Fig. 14A, dashed line), and the other four were obtained by an addi-
joint Love and Rayleigh inversions, allowing for radial anisotropy, but tion of different patterns of noise to this curve.
S wave velocity, km/s Period, s

3.5 4.0 4.5 5.0 5 10 20 50
0 0.8
A B "High-frequency" noise, <0.2-0.4%
50 0.0
100 -0.8
0.8
C Complete estimated noise, <0.3-0.6%
150
0.0
Depth, km
Noise, %
200 -0.8
0.8
250 D "Ramp" noise, -0.3%
0.0
300
-0.8
350 0.8
E "Ramp" noise, -0.5%
Global reference model 0.0
400 True reference model
-0.8
Moho depth, km Moho depth, km
32 36 40 44 48 52 32 36 40 44 48 52
1.2 1.4
1.0
1.3
0.8
Misfit x104
Misfit x104
1.2
0.6
0.4 1.1
0.2
1.0
F G
0.0
True Global True Global
reference reference reference reference
model model "High-frequency" model model
No noise noise, <0.3%
"Ramp" noise, -0.3%
Complete estimated
"Ramp" noise, -0.5% noise, 0.3-0.6%
Fig. 14. The effect of measurement errors on the results of surface-wave inversions for the Moho depth. A: synthetic phase-velocity curves were computed for the true reference
model (dashed line); in the different tests, both this model and AK135 (solid line) were used as the reference. BE: different patterns of noise added to the synthetic curves before
their inversion. F, G: results of model-space-map inversions for the Moho depth. The minimum of every mist curve shows the best tting Moho value. While the effect of the ref-
erence model is small, errors in the data of only 0.30.6% can cause large errors (up to 10 km) in the retrieved Moho depth, depending on their distribution with period. Complete
presentation of mists as a function of period for each of the tests is given in Supplementary Data (SFig. 1).
In order to isolate the effect of the assumed reference model on the they are reasonable estimates and are well suited for the purposes
model-space-map inversion (this effect is due to the damping applied of our tests as examples of noise distribution with period.
in the gradient searches), each of the ve dispersion curves was The high-frequency noise pattern is characterised by random er-
inverted twice, rst with the correct, true reference model, and then rors oscillating and changing sign every few seconds along the period
with a substantially different, global reference model (Fig. 14A, solid axis. This noise pattern increases the mists but has little effect on the
line). These tests conrmed that the inuence of the reference model Moho-depth values yielded by the inversions (mist-curve minima in
is limited, much smaller than that of the noise in each case (Fig. 14F, G). Fig. 14G). The inuence of such random-noise patterns was investi-
The rst two noise patterns were estimated from real data, mea- gated previously by Bartzsch et al. (2011) in their surface-wave inver-
sured across the Limpopo Belt (Adam and Lebedev, 2012). The com- sions for the depth of the lithosphereasthenosphere boundary. If the
plete estimated noise (Fig. 14C) is the difference between the errors were random (did not persist over broad period ranges), then
measured and synthetic phase velocities, the latter computed for a their effect on the best-tting parameter values was small, even if
preferred, smooth 1-D prole obtained in a damped, gradient- the noise amplitude was as high as 1.0%. Here, we use noise estimates
search inversion of the data. The high-frequency noise is estimated not from a random-number generator, as Bartzsch et al. (2011), but
in the same way but with the gradient-search inversion under- from real data. We arrive, however, to the same conclusion: random,
damped, and the 1-D prole showing some unrealistic roughness, high-frequency noise will be unlikely to bias the results of the in-
most likely introduced by the inversion so as to t the smoother version, in our case for the Moho depth.
components of the noise. Regardless of how accurately these noise In contrast, the complete estimated noise (Fig. 14C) contains errors
patterns represent the actual errors in the Limpopo measurements, that persist over relatively broad period ranges. This noise pattern
causes large errors in the Moho depth values. Although the noise is only Moho depth is not optimal. Because the Moho depth is constrained
up to 0.3% in the 1180 s period range and up to 0.6% in the 511 s pe- by very subtle signal in surface-wave data, even small errors at
riod range, it causes a 10-km error in the Moho depth (Fig. 14G). One short periods (sensitive primarily to the upper crust) or longer pe-
component of this noise pattern is a sharp phase-velocity increase at riods (sensitive primarily to the deep mantle lithosphere) can bias
1011 s. Approximating this component using ramp functions with the inversion results, because of the trade-offs between seismic veloc-
0.3% and 0.5% amplitudes (Fig. 14D, E), we nd that it alone gives rise ities at different depths. In order to determine the Moho depth using
to Moho-depth errors of a few kilometres (Fig. 14F). surface-wave data only, with no a priori information on crustal or
It is clear from these tests that errors in phase-velocity data of only mantle structure, the most effective inversion strategy is to rst con-
0.30.5% can prevent the determination of the Moho depth with use- strain the crustal and mantle structure by an inversion of the data in a
ful accuracy, particularly if the errors vary slowly along the period broad period range and then, as a second step, to nd best-tting
axis. This result is general and not specic to an inversion method. Moho depths in an inversion in a narrow period range with the
Given that there is always noise in the data, the question now is: most sensitivity to the Moho, using the results of the rst step as a
how can we reduce the effect of noise, so as to resolve the Moho? Un- reference model (see Section 5.3 for details).
fortunately, the most damaging errors are the ones that vary smooth-
ly with period, and such errors can map, to a large extent, into 7.2. A priori information: include whenever available!
articial structure in seismic-velocity models. If period ranges with
relatively large suspected errors can be identied for example, by A successful inversion of surface waves alone for the Moho depth re-
an observed increase in data-synthetic, dispersion-curve mist, by quires high accuracy of the measurements, because the strong sensitiv-
an increase of standard deviations or standard errors, or by an in- ity of surface waves to the Moho is reduced, in inversions, by severe
crease in frequency-dependent waveform mist excluding such pe- trade-offs of the depth of the Moho and the shear-speed structure
riod ranges from the inversion would be most effective (Fig. 13). above and below it. Accurate a priori constraints on crustal and mantle
Another way to improve the resolvability of the Moho is to include structure will reduce the trade-offs and should be sought when possi-
a priori information in the inversions, as is done often. The errors test- ble. Such constraints come, for example, from controlled-source crustal
ed in this section are relatively small. Most surface-wave studies of imaging in the region, from receiver-function studies, or from crustal
the Moho to date have used data with errors larger than these. Does xenoliths (e.g., Christensen and Mooney, 1995).
this mean that the results of all these studies should be in doubt?
Not if they incorporated accurate additional, a priori information, ex- 7.3. Joint analysis and inversion of surface-wave and other data
plicitly or implicitly. For example, in our inversions shown in Fig. 13
we included no a priori information and allowed the lower-crustal Joint analysis or joint inversion of surface-wave and other data can
and uppermost-mantle sheer speeds to vary in very broad ranges of reduce model uncertainties, making Moho mapping possible even
3.54.2 km/s and 4.24.6 km/s, respectively. Often, geological or geo- when the accuracy of surface-wave measurements alone is insuf-
physical constraints will be available to make these ranges much cient for the purpose. Surface-wave and receiver-function data com-
narrower. This will steepen the mist valleys and allow the Moho to plement each other especially well: both types of data are yielded
be resolved even with noise higher than a few tenths of a percent. by passive imaging methods and can be obtained from a small array
The Moho can thus be resolved even with relatively noisy data if of broadband stations situated virtually anywhere in the world.
accurate a priori information on the crustal and mantle seismic veloc- And, whereas surface waves with their broad depth sensitivity kernels
ities (or on the difference between the two) is available. Such con- (Fig. 1) provide strong constraints on shear speeds within depth
straints have been used extensively in surface-wave crustal studies, ranges (i.e., on smooth variations of shear speed with depth), receiver
either explicitly, with a clear analysis of the ranges of values consis- functions have particular sensitivity to sharp discontinuities (e.g.,
tent with existing data, or implicitly, through the choice of a reference Endrun et al., 2004; Juli et al., 2000; Shen et al., 2013; Tkali et al.,
or starting model. It is important to keep in mind that the accuracy of 2012).
the Moho depth yielded by such a constrained inversion will depend While the joint inversion of seismic data of different types is well
directly on the accuracy of the a priori constraints. established, recent developments in computational petrological and
geophysical modelling now also make increasingly feasible joint in-
7. Recommended inversion strategies versions of surface-wave and other geophysical and geological data.
Fullea et al. (2012) developed a joint inversion of surface-wave dis-
Basic strategies for the surface-wave inversion for the Moho depth persion curves and topography and applied it to data from central
are the same for the different types of surface-wave observables (phase Mongolia, also incorporating constraints from surface heat ow mea-
velocities, group velocities or waveforms) and are dictated by the sensi- surements, controlled-source seismic data, and crustal xenolith anal-
tivity of surface waves and the trade-offs between the discontinuity ysis. Designed for determination of the thermal structure of the
depth and seismic-velocity structure. Resolving the Moho depth with lithosphere with a high vertical resolution, such petro-physical inver-
surface waves alone is possible but is guaranteed to work only if the er- sion also constrains the depth of the Moho, using the complementary
rors of the measurements are very small (e.g., up to ~0.2% of phase veloc- sensitivities of the surface-wave and other data.
ities for mapping the Moho beneath continents). Data with larger errors
can still be used effectively if accurate a priori information is available 8. Discussion
that can reduce the possible ranges of seismic velocities in the crust
and uppermost mantle or of the velocity contrast across the Moho. In our investigation of the resolution of surface-wave inversions
(More information translates into better resolution in the presence of for the Moho depth, we considered only one type of surface-wave ob-
the same errors.) Similarly, relatively noisy surface-wave data can be servables, their phase velocities. Our results, however, characterise
successfully inverted for the Moho jointly with data of other types that general data-model relationships and apply, with obvious adjust-
provide complementary sensitivity, such as receiver functions. ments, to inversions of other surface-wave observables, including
group velocities and waveforms.
7.1. Inversion of surface-wave data only, with no a priori information Group velocities have a higher sensitivity to the Moho depth, com-
pared to phase velocities, in the absence of errors in the measure-
Contrary to the conventional wisdom, inverting surface-wave ments or with the same error levels (Figs. 611). This advantage of
measurements in a period range as broad as possible directly for the group velocities is offset, normally, by the larger actual errors in
their measurements. In group-velocity tomography, further errors increasingly abundant data from arrays of seismic stations. Thanks
may come from the group delays' off-ray sensitivity to local phase- to the progress in the precision and bandwidth of the measurements,
velocity perturbations and their dispersion (Eq. (2)), in addition surface-wave imaging of the crust and the Moho is now reaching a
to the expected sensitivity to local group velocity perturbations new level of accuracy.
(Dahlen and Zhou, 2006). In spite of this, the high sensitivity of Both Rayleigh and Love waves have strong sensitivity to the Moho
group velocities to the Moho depth has been exploited successfully depth. Tests with synthetic data show that if seismic wavespeeds
in Moho mapping, including in joint inversions of group and phase within the crust and upper mantle can be xed (assumed to be
velocities (e.g., Shapiro and Ritzwoller, 2002). known), then Moho-depth variations of a few kilometres produce
Although we have focussed so far on the fundamental-mode sur- large (>1%) perturbations in phase velocities, varying gradually
face waves only, higher surface-wave modes also sample the crust with period.
and the Moho. The growing global data sets of higher-mode, In inversions of surface-wave data with no a priori information,
phase-velocity measurements in increasingly broad frequency ranges the Moho depth shows strong trade-offs with shear-wave speeds in
(Schaeffer and Lebedev, submitted for publication; Van Heijst and the lower crust and uppermost mantle. Adjustments in the crustal
Woodhouse, 1999; Visser et al., 2007) will offer new types of con- and mantle structure can compensate for as much as 90% of the
straints on the global crustal structure. At a regional scale, the utility Moho-depth signal in surface-wave data. In the inversion, changes
of higher modes has already been clearly established: Levshin et al. of a few kilometres in the depth of a continental Moho result in oscil-
(2005) measured group velocities of Rayleigh- and Love-wave rst latory phase-velocity perturbations that reach a maximum on the
crustal overtones in the 7 18 s period range and showed that the order of 0.1% only. For the inversion to be guaranteed to resolve the
joint inversion of the fundamental and higher mode measurements Moho depth with useful precision, very high accuracy of the measure-
increased the resolution of crustal structure. ments is thus required, with errors up to 0.10.2% at most. Errors that
Lateral variations in the Moho depth (Moho topography) can be persist over broad period ranges are particularly harmful; random er-
resolved using tomography based on surface-wave ray theory if rors that vary rapidly with period have a smaller effect.
they are smooth enough so that the validity of surface-wave ray the- An effective strategy for the inversion of surface-wave data alone
ory is warranted (Dahlen and Tromp, 1998) or, at least, so that suf- for the Moho depth, with no a priori information, is to rst constrain
cient amount of surface-wave signals in the time-frequency planes the crustal and upper-mantle structure by an inversion in a broad pe-
can be selected empirically, such that they can be modelled using riod range, and then, as a second step, to nd best-tting Moho
ray theory (Das and Nolet, 1995; Lebedev et al., 2005; Schaeffer and depths in an inversion of the data in a narrow period range that has
Lebedev, submitted for publication). Strong lateral changes in the the most sensitivity to the Moho, with the results of the rst step
Moho depth result in strong lateral seismic-velocity changes and used as a reference model.
give rise to scattering and multipathing (e.g., Levshin et al., 1992; A priori constraints on crustal and mantle structure (from controlled-
Meier and Malischewsky, 2000; Meier et al., 1997). The Moho topog- source imaging, receiver-functions, xenoliths or regional geology) will
raphy itself, if not accounted for accurately, can also cause biases in reduce the trade-offs. Joint analysis or inversion of surface-wave and
the interpretation of surface-wave measurements that average over other data (receiver functions, topography, gravity) can reduce model
paths or areas (Levshin and Ratnikova, 1984). In order to resolve uncertainties and make Moho mapping possible even when the accuracy
small-scale Moho topography these effects have to be taken into ac- of surface-wave measurements alone would be insufcient. Alone or as a
count in the inversion for local phase or shear velocities (Wielandt, part of multi-disciplinary datasets, surface-wave data offer unique sensi-
1993). tivity to the crustal and upper-mantle structure and are becoming in-
The strong trade-offs between the depth of the Moho and the creasingly important in the seismic imaging of the crust and the Moho.
shear speeds just above and just below it discussed throughout Supplementary data to this article can be found online at http://
this paper also have implications for large-scale mantle tomography dx.doi.org/10.1016/j.tecto.2012.12.030.
that uses intermediate- and long-period surface waves. Here, the
trade-offs play a positive role, reducing the effect of uncertainties in Acknowledgements
the Moho depth on mantle models. If the inversion set-up includes
a few parameters for wavespeeds within the crust and a reasonably We thank the editors, I. Artemieva, L. Brown, B.L.N. Kennett and
dense parameterisation in the lithospheric mantle, and if the refer- H. Thybo for their work on this volume. Insightful comments and sug-
ence model for the tomography is 3-D and includes a realistic crust, gestions by Anatoli Levshin, an anonymous reviewer and one of the
such as CRUST2.0 (Bassin et al., 2000) or a more detailed regional editors have helped us to improve the clarity of the paper. Most
model (e.g., Grad et al., 2009; Molinari and Morelli, 2011; Tesauro gures were created with the Generic Mapping Tools (Wessel and
et al., 2008), then the Moho depth may not need to be perturbed in Smith, 1998). This work was funded by the Dublin Institute for
the inversion at all (e.g., Lebedev and van der Hilst, 2008; Legendre Advanced Studies and Science Foundation Ireland (grants 08/RFP/
et al., 2012; Schaeffer and Lebedev, submitted for publication). The ef- GEO1704 and 09/RFP/GEO2550).
fects of differences between the true and 3-D-reference Moho depths
will be compensated by perturbations within the crust almost entirely, References
with an accuracy most likely better than that of the source-station,
Adam, J.M.C., Lebedev, S., 2012. Azimuthal anisotropy beneath southern Africa, from
surface-wave measurements themselves, affected by uncertainties in very-broadband, surface-wave dispersion measurements. Geophysical Journal
source parameters. International 191, 155174.
Agius, M.R., Lebedev, S., 2010. Shear-velocity proles across the Tibetan Plateau from
broadband, surface-wave, phase-velocity measurements. In: Leech, Mary L.,
9. Conclusions Klemperer, Simon L., Mooney, Walter D. (Eds.), Proceedings of the 25th
HimalayaKarakoramTibet Workshop: U.S. Geological Survey Open-File Report
Surface waves have been used to study the outer layers of the 2010-1099.
Aki, K., Richards, P.G., 1980. Quantitative Seismology, Theory and Methods, Volume 1.
Earth since the early days of modern seismology. Their sensitivity to
W. H. Freeman and Company, San Francisco.
the crustal thickness (the depth to the Moho) has been established Alterman, Z., Jarosch, H., Pekeris, C.L., 1961. Propagation of Rayleigh waves in the Earth.
and applied in structural studies already in the rst half of the 20th Geophysical Journal of the Royal Astronomical Society 4, 219241.
century. The beginning of the 21st century has seen the crustal Angenheister, G., 1906. Bestimmung der Fortpanzungsgeschwindigkeit und Absorp-
tion von Erdbebenwellen, die durch den Gegenpunkt des Herdes gegangen sind.
surface-wave seismology re-energised by the emergence of new Nachrichten der Kniglichen Gesellschaft der Wissenschaften zu Gttingen,
techniques for broad-band surface-wave measurements, using the Mathematisch-physikalische Klasse.
Angenheister, G., 1921. Beobachtungen an pazischen Beben Ein Beitrag zum Gutenberg, B., 1924. Dispersion und Extinktion von seismischen Oberchenwellen
Studium der obersten Erdkruste. Nachrichten der Kniglichen Gesellschaft der und der Aufbau der obersten Erdschichten. Physikalische Zeitschrift XXV, 377381.
Wissenschaften zu Gttingen, Mathematisch-physikalische Klasse. Harmon, N., Henstock, T., Tilmann, F., Rietbrock, A., Barton, P., 2012. Shear velocity
Bartzsch, S., Lebedev, S., Meier, T., 2011. Resolving the lithosphereasthenosphere structure across the Sumatran forearc-arc. Geophysical Journal International 189,
boundary with seismic Rayleigh waves. Geophysical Journal International 186, 13061314.
11521164. Jeffreys, H., 1935. The surface waves of earthquakes. Monthly Notices, Royal Astronom-
Bassin, C., Laske, G., Masters, G., 2000. The current limits of resolution for surface wave ical Society, Geophysical Supplement 3, 253261.
tomography in North America. EOS, Fall Meeting Supplement F897. Juli, J., Ammon, C.J., Herrmann, R.B., Correig, A.M., 2000. Joint inversion of receiver
Bensen, G.D., Ritzwoller, M.H., Barmin, M.P., Levshin, A.L., Lin, F., Moschetti, M.P., function and surface wave dispersion observations. Geophysical Journal Interna-
Shapiro, N.M., Yang, Y., 2007. Processing seismic ambient noise data to obtain reli- tional 143, 99112.
able broad-band surface wave dispersion measurements. Geophysical Journal Kennett, B.L.N., 1983. Seismic Wave Propagation in Stratied Media. Cambridge
International 169, 12391260. University Press, Cambridge.
Boschi, L., Ekstrom, G., 2002. New images of the Earth's upper mantle from measure- Kennett, B.L.N., 2001. The Seismic Waveeld I. Introduction and Theoretical Development.
ments of surface wave phase velocity anomalies. Journal of Geophysical Research Cambridge University Press, Cambridge.
107 (B4), 2059. http://dx.doi.org/10.1029/2000JB000059. Kgaswane, E.M., Nyblade, A.A., Juli, J., Dirks, P.H.G.M., Durrheim, R.J., Pasyanos, M.E.,
Bostock, M.G., Hyndman, R.D., Rondenay, S., Peacock, S.M., 2002. An inverted continen- 2009. Shear wave velocity structure of the lower crust in Southern Africa: evidence
tal Moho and serpentinization of the forearc mantle. Nature 417, 536538. for compositional heterogeneity within Archaean and Proterozoic terrains. Journal
Bozdag, E., Trampert, J., 2008. On crustal corrections in surface wave tomography. of Geophysical Research 114, B12304. http://dx.doi.org/10.1029/2008JB006217.
Geophysical Journal International 172, 10661082. Kind, R., Yuan, X., Saul, J., Nelson, D., Sobolev, S.V., Mechie, J., Zhao, W., Kosarev, G., Ni, J.,
Brilliant, R.M., Ewing, M., 1954. Dispersion of Rayleigh waves across the U.S. Bulletin of Achauer, U., Jiang, M., 2002. Seismic images of crust and upper mantle beneath
the Seismological Society of America 44, 149158. Tibet: evidence for Eurasian plate subduction. Science 298, 12191221.
Brune, J., Dorman, J., 1963. Seismic waves and Earth structure in the Canadian Shield. Kissling, E., 1993. Deep structure of the Alps what do we really know? Physics of the
Bulletin of the Seismological Society of America 53, 167210. Earth and Planetary Interiors 79, 87112.
Brune, J.N., Nafe, J.E., Oliver, J.E., 1960. A simplied method for the analysis and synthe- Knopoff, L., 1972. Observation and inversion of surface-wave dispersion. Tectonophysics
sis of dispersed wave trains. Journal of Geophysical Research 65, 287304. 13, 497519.
Bullen, K.E., 1939. On Rayleigh waves across the Pacic Ocean. Monthly Notices, Royal Knopoff, L., Berry, J., Schwab, F.A., 1967. Tripartite phase velocity observation in later-
Astronomical Society, Geophysical Supplement 4, 579582. ally heterogeneous regions. Journal of Geophysical Research 72, 25952601.
Carder, D.S., 1934. Seismic surface waves and the crustal structure of the Pacic region. Koulakov, I., Sobolev, S.V., 2006. Moho depth and three-dimensional P and S structure
Bulletin of the Seismological Society of America 24, 231301. of the crust and uppermost mantle in the Eastern Mediterranean and Middle East
Christensen, N.I., Mooney, W.D., 1995. Seismic velocity structure and composition of the derived from tomographic inversion of local ISC data. Geophysical Journal Interna-
continental crust: a global view. Journal of Geophysical Research 100, 97619788. tional 164, 218235.
Dahlen, F.A., Tromp, J., 1998. Theoretical Global Seismology. Princeton University Press, Kovach, R.L., 1978. Seismic Surface waves and crustal and upper mantle structure.
Princeton, NJ. Reviews of Geophysics and Space Physics 16, 113.
Dahlen, F.A., Zhou, Y., 2006. Surface-wave group-delay and attenuation kernels. Kustowski, B., Dziewonski, A.M., Ekstrom, G., 2007. Nonlinear crustal corrections for
Geophysical Journal International 165, 545554. normal-mode seismograms. Bulletin of the Seismological Society of America 97,
Das, T., Nolet, G., 1995. Crustal thickness estimation using high frequency Rayleigh 17561762.
waves. Geophysical Research Letters 22 (5), 539542. Landisman, M., Dziewonski, A., Sato, Y., 1969. Recent improvements in the analysis of
Das, T., Nolet, G., 1998. Crustal thickness map of the western United States by surface wave observations. Geophysical Journal of the Royal Astronomical Society
partitioned waveform inversion. Journal of Geophysical Research 103 (B12), 17, 369403.
30,02130,038. Lebedev, S., Nolet, G., 2003. Upper mantle beneath Southeast Asia from S velocity tomog-
DeLisle, J.F., 1941. On dispersion of Rayleigh waves from the North Pacic earthquake raphy. Journal of Geophysical Research 108. http://dx.doi.org/10.1029/2000JB000073.
of November 10, 1938. Bulletin of the Seismological Society of America 31, Lebedev, S., van der Hilst, R.D., 2008. Global upper-mantle tomography with the auto-
303307. mated multimode inversion of surface and S-wave forms. Geophysical Journal
Deschamps, F., Lebedev, S., Meier, T., Trampert, J., 2008a. Azimuthal anisotropy of International 173, 505518.
Rayleigh-wave phase velocities in the east-central United States. Geophysical Journal Lebedev, S., Nolet, G., van der Hilst, R.D., 1997. The upper mantle beneath the
International 173, 827843. Philippine Sea region from waveform inversions. Geophysical Research Letters
Deschamps, F., Lebedev, S., Meier, T., Trampert, J., 2008b. Stratied seismic anisotropy 25, 18511854.
reveals past and present deformation beneath the East-central United States. Lebedev, S., Chevrot, S., van der Hilst, R.D., 2002a. Seismic evidence for olivine phase
Earth and Planetary Science Letters 274, 489498. changes at the 410- and 660-kilometer discontinuities. Science 296, 13001302.
Dorman, J., Ewing, M., 1962. Numerical inversion of seismic surface wave dispersion Lebedev, S., Chevrot, S., van der Hilst, R.D., 2002b. The 660-km discontinuity within the
data and crust-mantle structure in the New York-Pennsylvania Area. Journal of subducting NW-Pacic lithospheric slab. Earth and Planetary Science Letters 205,
Geophysical Research 67, 52275241. 2535.
Dziewonski, A.M., 1970. On regional differences in dispersion of mantle Rayleigh Lebedev, S., Nolet, G., Meier, T., van der Hilst, R.D., 2005. Automated multimode inver-
waves. Geophysical Journal of the Royal Astronomical Society 22, 289325. sion of surface and S waveforms. Geophysical Journal International 162, 951964.
Endrun, B., Meier, T., Bischoff, M., Harjes, H.P., 2004. Lithospheric structure in the area Lebedev, S., Meier, T., van der Hilst, R.D., 2006. Asthenospheric ow and origin of vol-
of Crete constrained by receiver functions and dispersion analysis of Rayleigh canism in the Baikal Rift area. Earth and Planetary Science Letters 249, 415424.
phase velocities. Geophysical Journal International 158, 592608. Lebedev, S., Boonen, J., Trampert, J., 2009. Seismic structure of Precambrian litho-
Endrun, B., Meier, T., Lebedev, S., Bohnhoff, M., Stavrakakis, G., Harjes, H.-P., 2008. S ve- sphere: new constraints from broad-band surface-wave dispersion. Lithos 109,
locity structure and radial anisotropy in the Aegean region from surface wave dis- 96111.
persion. Geophysical Journal International 174, 593616. Legendre, C., Meier, T., Lebedev, S., Friederich, W., Viereck-Goette, L., 2012. A shear-
Endrun, B., Lebedev, S., Meier, T., Tirel, C., Friederich, W., 2011. Complex layered defor- wave velocity model of the European upper mantle from automated inversion
mation within the Aegean crust and mantle revealed by seismic anisotropy. Nature of seismic shear and surface waveforms. Geophysical Journal International 191,
Geoscience 4, 203207. 282304.
Erduran, M., Endrun, B., Meier, T., 2008. Continental vs. oceanic lithosphere beneath the Lekic, V., Panning, M., Romanowicz, B., 2010. A simple method for improving crustal
eastern Mediterranean Sea implications from Rayleigh wave dispersion mea- corrections in waveform tomography. Geophysical Journal International 182,
surements. Tectonophysics 457, 4252. 265278.
Evernden, J.F., 1954. Love-wave dispersion and the structure of the Pacic basin. Bulletin Levshin, A., Ratnikova, L., 1984. Apparent anisotropy in inhomogeneous media.
of the Seismological Society of America 44, 15. Geophysical Journal of the Royal Astronomical Society 74, 6569.
Ewing, M., Press, F., 1950. Crustal structure and surface-wave dispersion. Bulletin of the Levshin, A.L., Yanovskaya, T.B., Lander, A.V., Bukchin, B.G., Barmin, M.P., Ratnikova, L.I.,
Seismological Society of America 40, 271280. Its, E.N., 1989. In: Keilis-Borok, V.I. (Ed.), Seismic Surface Waves in a Laterally
Ewing, M., Press, F., 1952. Crustal structure and surface-wave dispersion, Part II: Solomon Inhomogeneous Earth. Kluwer, Norwell, Mass.
Islands earthquake of July 29, 1950. Bulletin of the Seismological Society of America Levshin, A., Ratnikova, L., Berger, J., 1992. Peculiarities of surface-wave propagation
42, 315325. across central Eurasia. Bulletin of the Seismological Society of America 82,
Ewing, M., Press, F., 1956. Rayleigh wave dispersion in the period range 10 to 500 seconds. 24642493.
Transactions of the American Geophysical Union 37, 213215. Levshin, A.L., Ritzwoller, M.H., Shapiro, N.M., 2005. The use of crustal higher modes to
Ferreira, A.M.G., Woodhouse, J.H., Visser, K., Trampert, J., 2010. On the robustness constrain crustal structure across Central Asia. Geophysical Journal International
of global radially anisotropic surface wave tomography. Journal of Geophysical 160, 961972.
Research 115, B04313. http://dx.doi.org/10.1029/2009JB006716. Li, X., Romanowicz, B., 1996. Global mantle shear velocity model developed using
Fullea, J., Lebedev, S., Agius, M.R., Jones, A.G., Afonso, J.C., 2012. Lithospheric structure nonlinear asymptotic coupling theory. Journal of Geophysical Research 101,
in the Baikalcentral Mongolia region from integrated geophysical-petrological inver- 2224522272.
sion of surface-wave data and topographic elevation. Geochemistry, Geophysics, Lin, F.C., Ritzwoller, M.H., Yang, Y., Moschetti, M.P., Fouch, M.J., 2011. Complex and var-
Geosystems 13, Q0AK09. http://dx.doi.org/10.1029/2012GC004138. iable crustal and uppermost mantle seismic anisotropy in the western United
Golitsyn, B.B., 1960. Selected works. Press of the Academy of Sciences of the USSR, States. Nature Geoscience 4, 5561.
Moscow: Seismology, 2. Love, A.E.H., 1911. Some Problems of Geodynamics. Cambridge University Press, Cambridge.
Grad, M., Tiira, T., ESC Working Group, 2009. The Moho depth map of the European Macelwane, J.B., 1923. A study of the relation between the periods of elastic waves and
Plate. Geophysical Journal International 176, 279292. the distance traveled by them, based upon the seismographic records of the
California earthquake, January 31, 1922. Bulletin of the Seismological Society of Press, F., Ewing, M., Oliver, J., 1956. Crustal structure and surface-wave dispersion in
America 13, 1369. Africa. Bulletin of the Seismological Society of America 46, 97103.
Marone, F., Romanowicz, B., 2007. Non-linear crustal corrections in high-resolution region- Prodehl, C., Mooney, W.D., 2011. Exploring the Earth's crust: history and results of
al waveform seismic tomography. Geophysical Journal International 170, 460467. controlled-source seismology. The Geological Society of America, Memoir 208.
Marone, F., van der Meijde, M., van der Lee, S., Giardini, D., 2003. Joint inversion of local, Reid, H.F., 1910. The California Earthquake of April 18, 1906. Report of the State Earth-
regional, and teleseismic data for crustal thickness in the Eurasia-Africa plate quake Investigation Commission.
boundary region. Geophysical Journal International 154, 499514. Rychert, C.A., Shearer, P.M., 2010. Resolving crustal thickness using SS waveform
McEvilly, T.V., 1964. Central U.S. curst-upper mantle structure from Love and Rayleigh stacks. Geophysical Journal International 180, 11281137.
wave phase velocity inversion. Bulletin of the Seismological Society of America 54, Santo, T., Sato, Y., 1966. World-wide survey of the regional characteristics of group ve-
19972015. locity dispersion of Rayleigh waves. Bulletin of the Earthquake Research Institute,
Meier, T., Malischewsky, P.G., 2000. Approximation of surface wave mode conversion University of Tokyo 44, 939964.
at a passive continental margin by a mode-matching technique. Geophysical Journal Sato, Y., 1955. Analysis of dispersed surface waves by means of Fourier transform: Part
International 141, 1224. 1. Bulletin of the Earthquake Research Institute, University of Tokyo 33, 3347.
Meier, T., Lebedev, S., Nolet, G., Dahlen, F.A., 1997. Diffraction tomography using Schaeffer, A. J., Lebedev, S., submitted for publication. Global shear speed structure of
multimode surface waves. Journal of Geophysical Research 102, 82558267. the upper mantle. Geophysical Journal International.
Meier, T., Dietrich, K., Stockhert, B., Harjes, H.P., 2004. One-dimensional models of Seidl, D., Mller, S., 1977. Seismische Oberchenwellen. Zeitschrift fr Geophysik 42,
shear wave velocity for the eastern Mediterranean obtained from the inversion 283328.
of Rayleigh wave phase velocities and tectonic implications. Geophysical Journal Shapiro, N.M., Campillo, M., 2004. Emergence of broadband Rayleigh waves from cor-
International 156, 4558. relations of the ambient seismic noise. Geophysical Research Letters 31, L07614.
Meier, U., Curtis, A., Trampert, J., 2007a. Fully nonlinear inversion of fundamental mode http://dx.doi.org/10.1029/2004GL019491.
surface waves for a global crustal model. Geophysical Research Letters 34, L16304. Shapiro, N.M., Ritzwoller, M.H., 2002. Monte-Carlo inversion for a global shear-velocity
http://dx.doi.org/10.1029/2007GL030989. model of the crust and upper mantle. Geophysical Journal International 151,
Meier, U., Curtis, A., Trampert, J., 2007b. Global crustal thickness from neural network 88105.
inversion of surface wave data. Geophysical Journal International 169, 706722. Shapiro, N.M., Campillo, M., Stehly, L., Ritzwoller, M.H., 2005. High-resolution surface
Meissner, E., 1921. Elastische Oberchenwellen mit Dispersion in einem inhomogenen wave tomography from ambient seismic noise. Science 307, 16151618.
Medium. Vierteljahrsschrift der Naturforschenden Gesellschaft in Zrich 66, 181195. Shen, W., Ritzwoller, M.H., Schulte-Pelkum, V., Lin, F.-C., 2013. Joint inversion of surface
Meissner, E., 1926. Elastische Oberchen-Querwellen. Proceedings of the 2nd Interna- wave dispersion and receiver functions: a Bayesian Monte-Carlo approach. Geo-
tional Congress for Applied Mechanics, Zrich, pp. 311. physical Journal International 192, 807836.
Mohorovii, A., 1910. Potres od 8.X 1909 (Das Beben vom 8.X. 1909). Jahrbuch des Stankiewicz, J., Chevrot, S., van der Hilst, R.D., de Wit, M.J., 2002. Crustal thickness, dis-
meteorologischen Observatoriums in Zagreb (Agram) fr das Jahr 1909, pp. 156 continuity depth, and upper mantle structure beneath southern Africa: constraints
(English translation: 1992, Earthquake of 8 October 1909, Geozika, 9, 355). form body wave conversions. Physics of the Earth and Planetary Interiors 130,
Molinari, I., Morelli, A., 2011. EPcrust: a reference crustal model for the European plate. 235251.
Geophysical Journal International 185, 352364. Stoneley, R., 1925. Dispersion of seismic waves. Geophysical Journal International 1,
Montagner, J.P., Jobert, N., 1988. Vectorial tomography II. Application to the Indian 280282. http://dx.doi.org/10.1111/j.1365-246X.1925.tb05375.x.
Ocean. Geophysical Journal International 94, 309344. Stoneley, R., 1926. The effect of the ocean on Rayleigh waves. Monthly Notices, Royal
Mooney, W.D., Laske, G., Masters, T.G., 1998. Crust5.1: a global crustal model at 5 5. Astronomical Society, Geophysical Supplement 1, 349356.
Journal of Geophysical Research 103, 727747. Tams, E., 1921. ber die Fortpanzungsgeschwindigkeit der seismischen Oberchenwellen
Moschetti, M.P., Ritzwoller, M.H., Lin, F.-C., Yang, Y., 2010. Crustal shear wave velocity lngs kontinentaler und ozeanischer Wege. Centralblatt der Mineralogie, Geologie und
structure of the western United States inferred from ambient seismic noise and Palontologie 2-3, 4452 (75-83).
earthquake data. Journal of Geophysical Research 115, B10306. http://dx.doi.org/ Tesauro, M., Kaban, M.K., Cloetingh, S., 2008. EuCRUST-07: a new reference model for
10.1029/2010JB007448. the European crust. Geophysical Research Letters 35, L05313. http://dx.doi.org/
Nabelek, J., Hetenyi, G., Vergne, J., Sapkota, S., Kae, B., Jiang, M., Su, H.P., Chen, J., 10.1029/2007GL032244.
Huang, B.S., Team, Hi-CLIMB, 2009. Underplating in the HimalayaTibet collision Thybo, H., 2001. Crustal structure along the EGT prole across the Tornquist Fan
zone revealed by the Hi-CLIMB experiment. Science 325, 13711374. interpreted from seismic, gravity and magnetic data. Tectonophysics 334, 155190.
Nafe, J.E., Brune, J.N., 1960. Observations of phase velocity for Rayleigh waves in the pe- Tkali, H., Rawlinson, N., Arroucau, P., Kumar, A., Kennett, B.L.N., 2012. Multistep
riod range 100 to 400 seconds. Bulletin of the Seismological Society of America 50, modelling of receiver-based seismic and ambient noise data from WOMBAT
427439. array: crustal structure beneath southeast Australia. Geophysical Journal Interna-
Nair, S.K., Gao, S.S., Liu, K.H., Silver, P.G., 2006. Southern African crustal evolution and tional 189, 16801700.
composition: constraints from receiver function studies. Journal of Geophysical Toksz, M.N., Ben-Menahem, A., 1963. Velocities of mantle Love and Rayleigh waves
Research 111. http://dx.doi.org/10.1029/2005JB003802. over multiple paths. Bulletin of the Seismological Society of America 53, 741764.
Nataf, H.C., Ricard, Y., 1996. 3SMAC: an a priori tomographic model of the upper mantle Van der Lee, S., Nolet, G., 1997. Upper-mantle S-velocity structure of North America.
based on geophysical modelling. Physics of the Earth and Planetary Interiors 95, Journal of Geophysical Research 102, 22,81522,838.
101122. Van Heijst, H.J., Woodhouse, J., 1999. Global high-resolution phase velocity distribu-
Nataf, H.C., Nakanishi, I., Anderson, D.L., 1986. Measurements of mantle wave velocities tions of overtone and fundamental-mode surface waves determined by mode
and inversion for lateral heterogeneities and anisotropy. Part III: inversion. Journal branch stripping. Geophysical Journal International 137, 601620.
of Geophysical Research 91, 72617307. Visser, K., Lebedev, S., Trampert, J., Kennett, B.L.N., 2007. Global Love wave overtone
Neumann, F., 1929. The velocity of seismic surface waves over pacic paths. Bulletin of measurements. Geophysical Research Letters 34, L03302. http://dx.doi.org/
the Seismological Society of America 19, 6376. 10.1029/2006GL028671.
Nolet, G., 1990. Partitioned waveform inversion and two-dimensional structure under Wessel, P., Smith, W.H.F., 1998. New, improved version of generic mapping tools re-
the Network of Autonomously Recording Seismographs. Journal of Geophysical leased. Eos Transactions AGU 79 (47), 579. http://dx.doi.org/10.1029/98EO00426.
Research 95, 84998512. Wiechert, E., 1899. Seismometrische Beobachtungen im Gttinger Geophysikalischen
Nolet, G., 2008. A Breviary of Seismic Tomography: Imaging the Interior of the Earth Institut. Nachrichten der Kniglichen Gesellschaft der Wissenschaften zu
and the Sun. Cambridge University Press, Cambridge. Gttingen, Mathematisch-physikalische Klasse.
Oldham, R.D., 1899. Report on the great earthquake of 12th June 1897. Memoirs of the Wielandt, E., 1993. Propagation and structural interpretation of non-plane waves.
Geological Survey of India 29, 1379. Geophysical Journal International 113, 4553.
Oliver, J.E., Ewing, M., Press, F., 1955. Crustal structure and surface-wave dispersion. Wilson, J.T., 1940. The Love waves of the South Atlantic earthquakes of August 28,
Bulletin of the Seismological Society of America 66, 913946. 1933. Bulletin of the Seismological Society of America 30, 273301.
Panning, M., Lekic, V., Romanowicz, B., 2010. Importance of crustal corrections in the Wilson, J.T., Baykal, O., 1948. Crustal Structure of the North Atlantic Basin as deter-
development of a new global model of radial anisotropy. Journal of Geophysical mined from Rayleigh wave dispersion. Bulletin of the Seismological Society of
Research 115. http://dx.doi.org/10.1029/2010JB007520. America 38, 4153.
Pasyanos, M.E., Walter, W.R., 2002. Crust and upper-mantle structure of North Africa, Woodhouse, J.H., Dziewonski, A.M., 1984. Mapping the upper mantle: three-
Europe and the Middle East from inversion of surface waves. Geophysical Journal dimensional modeling of earth structure by inversion of seismic waveforms. Journal
International 149, 463481. of Geophysical Research 89, 59535986.
Pawlak, A.E., Eaton, D., Darbyshire, F., Lebedev, S., Bastow, I.D., 2012. Crustal anisotropy Yang, Y., Li, A., Ritzwoller, M.H., 2008. Crustal and uppermost mantle structure
beneath Hudson Bay from ambient-noise tomography: evidence for post-orogenic in southern Africa revealed from ambient noise and teleseismic tomography.
lower-crustal ow? Journal of Geophysical Research 117, B08301. http://dx.doi.org/ Geophysical Journal International 174, 235248.
10.1029/2011JB009066. Yang, Y., Ritzwoller, M.H., Jones, C.H., 2011. Crustal structure determined from ambient
Pekeris, C.L., 1948. Theory of propagation of explosive sounds in shallow water. noise tomography near the magmatic centers of the Coso region, southeastern
Geological Society of America Memoir 27, 117. California. Geochemistry, Geophysics, Geosystems 12, Q02009. http://dx.doi.org/
Polat, G., Lebedev, S., Readman, P.W., O'Reilly, B.M., Hauser, F., 2012. Anisotropic 10.1029/2010GC003362.
Rayleigh-wave tomography of Ireland's crust: implications for crustal accretion Yang, Y., Ritzwoller, M.H., Zheng, Y., Levshin, A.L., Xie, Z., 2012. A synoptic view of the dis-
and evolution within the Caledonian Orogen. Geophysical Research Letters 39, tribution and connectivity of the mid-crustal low velocity zone beneath Tibet. Journal
L04302. http://dx.doi.org/10.1029/2012GL051014. of Geophysical Research 117, B04303. http://dx.doi.org/10.1029/2011JB008810.
Press, F., 1956. Determination of crustal structure from phase velocity of Rayleigh Yao, H., Beghein, C., van der Hilst, R.D., 2008. Surface-wave array tomography in SE
waves, Part I: Southern California. Geological Society of America Bulletin 67, Tibet from ambient seismic noise and two-station analysis: II Crustal and
16471658. upper mantle structure. Geophysical Journal International 173, 205219.
Yao, H., Goudard, P., Collins, J.A., McGuire, J.J., van der Hilst, R.D., 2011. Structure of Zhang, X., Paulssen, H., Lebedev, S., Meier, T., 2009. 3D shear velocity structure beneath
young East Pacic rise lithosphere from ambient noise correlation analysis of the Gulf of California from Rayleigh wave dispersion. Earth and Planetary Science
fundamental- and higher-mode ScholteRayleigh waves. Comptes Rendus Geoscience Letters 279, 255262.
343, 571583. Zhu, L., Kanamori, H., 2000. Moho depth variation in southern California from
Zhang, X., Paulssen, H., Lebedev, S., Meier, T., 2007. Surface wave tomography of the teleseismic receiver functions. Journal of Geophysical Research 105, 29692980.
Gulf of California. Geophysical Research Letters 34. http://dx.doi.org/10.1029/
2007GL030631.

Surface Wave Paper

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Surface Wave Paper

Uploaded by

Copyright:

Available Formats

TECTO-125734; No of Pages 18

Tectonophysics xxx (2013) xxxxxx

Contents lists available at SciVerse ScienceDirect

Corresponding author. Tel.: +353 1 653 5147x240.

1. Introduction Phase velocity Group velocity

Rayleigh waves Love waves Rayleigh waves Love waves

phase velocity, km/s

group velocity, km/s

group velocity, km/s

phase velocity, km/s

group velocity, km/s

group velocity, km/s

phase velocity, km/s

group velocity, km/s

group velocity, km/s

Tibet Tibet MOHO

Topography Moho, from waveform tomography Moho, from CRUST2. 0

100E 120E 140E 100E 120E 140E 100E 120E 140E

Receiver functions Ambient noise Ambient noise

Phase velocity Group velocity

Rayleigh waves Rayleigh waves

Love waves Love waves

Phase velocity Group velocity

Rayleigh waves Love waves Rayleigh waves Love waves

velocity variation, km/s

0.2 0.2 0.2 0.2

0.1 0.1 0.1 0.1

0.0 0.0 0.0 0.0

-0.1 -0.1 -0.1 -0.1

-0.2 -0.2 -0.2 -0.2

velocity variation, km/s

0.2 0.2 0.2 0.2

0.1 0.1 0.1 0.1

0.0 0.0 0.0 0.0

-0.1 -0.1 -0.1 -0.1

-0.2 -0.2 -0.2 -0.2

S wave velocity, km/s Period, s Period, s Moho depth, km

S wave velocity, km/s Period, s Period, s Moho depth, km

S wave velocity, km/s Period, s

You might also like