Wapenaar 2010

GEOPHYSICS, VOL. 75, NO. 5 共SEPTEMBER-OCTOBER 2010兲; P. 75A195–75A209, 15 FIGS.
10.1190/1.3457445
Downloaded 05/08/13 to 132.203.235.189. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/
Tutorial on seismic interferometry:

Part 1 — Basic principles and applications
Kees Wapenaar1, Deyan Draganov1, Roel Snieder2, Xander Campman3, and Arie Verdel3
ABSTRACT retrieval of seismic surface-wave responses from ambient noise

and the subsequent tomographic determination of the surface-
Seismic interferometry involves the crosscorrelation of re- wave velocity distribution of the subsurface. Seismic interferom-
sponses at different receivers to obtain the Green’s function be- etry is not restricted to retrieving direct waves between receivers.
tween these receivers. For the simple situation of an impulsive In a classic paper, Claerbout shows that the autocorrelation of the
plane wave propagating along the x-axis, the crosscorrelation of transmission response of a layered medium gives the plane-wave
the responses at two receivers along the x-axis gives the Green’s reflection response of that medium. This is essentially 1D reflect-
function of the direct wave between these receivers. When the ed-wave interferometry. Similarly, the crosscorrelation of the
source function of the plane wave is a transient 共as in exploration transmission responses, observed at two receivers, of an arbitrary
seismology兲 or a noise signal 共as in passive seismology兲, then the inhomogeneous medium gives the 3D reflection response of that
crosscorrelation gives the Green’s function, convolved with the medium. One of the main applications of reflected-wave interfer-
autocorrelation of the source function. Direct-wave interferome- ometry is retrieving the seismic reflection response from ambient
try also holds for 2D and 3D situations, assuming the receivers noise and imaging of the reflectors in the subsurface. A common
are surrounded by a uniform distribution of sources. In this case, aspect of direct- and reflected-wave interferometry is that virtual
the main contributions to the retrieved direct wave between the sources are created at positions where there are only receivers
receivers come from sources in Fresnel zones around stationary without requiring knowledge of the subsurface medium parame-
points. The main application of direct-wave interferometry is the ters or of the positions of the actual sources.
INTRODUCTION source positions. Passive seismic interferometry, on the other hand,

is a methodology for turning passive seismic measurements 共ambi-
In this two-part tutorial, we give an overview of the basic princi- ent seismic noise or microearthquake responses兲 into deterministic
ples and the underlying theory of seismic interferometry and discuss seismic responses. Here, we further distinguish between retrieving
applications and new advances. The term seismic interferometry re-
surface-wave transmission responses 共Campillo and Paul, 2003;
fers to the principle of generating new seismic responses of virtual
Shapiro and Campillo, 2004; Sabra, Gerstoft, et al., 2005a兲 and ex-
sources4 by crosscorrelating seismic observations at different re-
ploration reflection responses 共Claerbout, 1968; Scherbaum, 1987b;
ceiver locations. One can distinguish between controlled-source and
passive seismic interferometry. Controlled-source seismic interfer- Draganov et al., 2007, 2009兲. In passive interferometry of ambient
ometry, pioneered by Schuster 共2001兲, Bakulin and Calvert 共2004兲, noise, no explicit summation of correlations over different source
and others, comprises a new processing methodology for seismic ex- positions is required because the correlated responses are a superpo-
ploration data. Apart from crosscorrelation, controlled-source inter- sition of simultaneously acting uncorrelated sources.
ferometry also involves summation of correlations over different In all cases, the response that is retrieved by crosscorrelating two
Manuscript received by the Editor 30 November 2009; published online 14 September 2010.
1
Delft University of Technology, Department of Geotechnology, Delft, The Netherlands. E-mail: [email protected]; [email protected].
2
Colorado School of Mines, Center for Wave Phenomena, Golden, Colorado, U.S.A. E-mail: [email protected].
3
Shell International Exploration and Production, Rijswijk, The Netherlands. E-mail: [email protected]; [email protected].
© 2010 Society of Exploration Geophysicists. All rights reserved.
4
In the literature on seismic interferometry, the term virtual source often refers to the method of Bakulin and Calvert 共2004, 2006兲, which is discussed extensive-
ly in Part 2. However, creating a virtual source is the essence of nearly all seismic interferometry methods 共see e.g., Schuster 共2001兲, who already used this term兲.
In this paper 共Parts 1 and 2兲 we use the term virtual source whenever appropriate. When it refers to Bakulin and Calvert’s method, we will mention this explicitly.
75A195
75A196 Wapenaar et al.
receiver recordings 共and summing over different sources兲 can be in- of 1D direct-wave interferometry and conclude with a discussion of
terpreted as the response that would be measured at one of the re- the principles of 3D reflected-wave interferometry. We present ap-
ceiver locations as if there were a source at the other. Because such a plications in controlled-source as well as passive interferometry and,
point-source response is equal to a Green’s function convolved with where appropriate, review the historical background. To stay fo-
a wavelet, seismic interferometry is also often called Green’s func- cused on seismic applications, we refrain from a further discussion
tion retrieval. Both terms are used in this paper. The term interferom- of the normal-mode approach, nor do we address the many interest-
etry is borrowed from radio astronomy, where it refers to crosscorre- ing applications of Green’s function retrieval in underwater acous-
lation methods applied to radio signals from distant objects 共Thomp- tics 共e.g., Roux and Fink, 2003; Sabra et al., 2005; Brooks and Ger-
son et al., 2001兲. The name Green’s function honors George Green stoft, 2007兲.
who, in a privately published essay, introduced the use of impulse re-
sponses in field representations 共Green, 1828兲. Challis and Sheard
共2003兲 give a brief history of Green’s life and theorem. Ramírez and DIRECT-WAVE INTERFEROMETRY
Weglein 共2009兲 review applications of Green’s theorem in seismic
processing. 1D analysis of direct-wave interferometry
Early successful results of Green’s function retrieval from noise We start our explanation of seismic interferometry by considering
correlations were obtained in the field of ultrasonics 共Weaver and an illustrative 1D analysis of direct-wave interferometry. Figure 1a
Lobkis, 2001, 2002兲. The experiments were done with diffuse fields shows a plane wave, radiated by an impulsive unit source at x ⳱ xS
in a closed system. Here diffuse means that the amplitudes of the nor- and t ⳱ 0, propagating in the rightward direction along the x-axis.
mal modes are uncorrelated but have equal expected energies. We assume that the propagation velocity c is constant and the medi-
Hence, the crosscorrelation of the field at two receiver positions does um is lossless. There are two receivers along the x-axis at xA and xB.
not contain cross-terms of unequal normal modes. The sum of the re- Figure 1b shows the response observed by the first receiver at xA. We
maining terms is proportional to the modal representation of the
denote this response as G共xA,xS,t兲, where G stands for the Green’s
Green’s function of the closed system 共Lobkis and Weaver, 2001兲.
function. Throughout this paper, we use the common convention
This means that the crosscorrelation of a diffuse field in a closed sys-
that the first two arguments in G共xA,xS,t兲 denote the receiver and
tem converges to its impulse response. Later, it was recognized 共e.g.,
source coordinates, respectively 共here, xA and xS兲, whereas the last
Godin, 2007兲 that this theoretical explanation is akin to the fluctua-
argument denotes time t or angular frequency ␻ . In our example, this
tion-dissipation theorem 共Callen and Welton, 1951; Rytov, 1956;
Green’s function consists of an impulse at tA ⳱ 共xA ⳮ xS兲 / c; there-
Rytov et al., 1989; Le Bellac et al., 2004兲.
fore, G共xA,xS,t兲 ⳱ ␦ 共t ⳮ tA兲, where ␦ 共t兲 is the Dirac delta function.
The earth is a closed system; but at the scale of global seismology,
Similarly, the response at xB is given by G共xB,xS,t兲 ⳱ ␦ 共t ⳮ tB兲, with
the wavefield is far from diffuse. At the scale of exploration seismol-
tB ⳱ 共xB ⳮ xS兲 / c 共Figure 1c兲.
ogy, an ambient-noise field may have a diffuse character, but the en-
Seismic interferometry involves the crosscorrelation of responses
compassing system is not closed. Hence, for seismic interferometry,
at two receivers, in this case at xA and xB. Looking at Figure 1a, it ap-
the normal-mode approach breaks down. Throughout this paper, we
pears that the raypaths associated with G共xA,xS,t兲 and G共xB,xS,t兲
consider seismic interferometry 共or Green’s function retrieval兲 in
have the path from xS to xA in common. The traveltime along this
open systems, including half-spaces below a free surface. Instead of
common path cancels in the crosscorrelation process, leaving the
a treatment per field of application or a chronological discussion, we
traveltime along the remaining path from xA to xB, i.e., tB ⳮ tA ⳱ 共xB
have chosen a setup in which we explain the principles of seismic in-
ⳮ xA兲 / c. Hence, the crosscorrelation of the responses in Figure 1b
terferometry step by step. In Part 1, we start with the basic principles
and c is an impulse at tB ⳮ tA 共see Figure 1d兲. This impulse can be in-
terpreted as the response of a source at xA observed by a receiver at
xB, i.e., the Green’s function G共xB,xA,t兲. An interesting observation is
that the propagation velocity c and the position of the actual source
xS need not be known. The traveltimes along the common path from
xS to xA compensate each other, independent of the propagation ve-
locity and the length of this path. Similarly, if the source impulse
would occur at t ⳱ tS instead of at t ⳱ 0, the impulses observed at xA
and xB would be shifted by the same amount of time tS, which would
be canceled in the crosscorrelation. Thus, the absolute time tS at
which the source emits its pulse need not be known.
Let us discuss this example a bit more precisely. We denote the
crosscorrelation of the impulse responses at xA and xB as
G共xB,xS,t兲 ⴱ G共xA,xS,ⳮt兲. The asterisk denotes temporal convolu-
tion, but the time reversal of the second Green’s function turns the
convolution into a correlation, defined as G共xB,xS,t兲 ⴱ G共xA,xS,ⳮt兲
Figure 1. A 1D example of direct-wave interferometry. 共a兲 A plane ⳱ 兰G共xB,xS,t Ⳮ t⬘兲G共xA,xS,t⬘兲dt⬘. Substituting the delta functions
wave traveling rightward along the x-axis, emitted by an impulsive into the right-hand side gives 兰␦ 共t Ⳮ t⬘ ⳮ tB兲␦ 共t⬘ ⳮ tA兲dt⬘ ⳱ ␦ 共t
source at x ⳱ xS and t ⳱ 0. 共b兲 The response observed by a receiver
at xA. This is the Green’s function G共xA,xS,t兲. 共c兲 As in 共b兲 but for a re- ⳮ 共tB ⳮ tA兲兲 ⳱ ␦ 共t ⳮ 共xB ⳮ xA兲 / c兲. This is indeed the Green’s func-
ceiver at xB. 共d兲 Crosscorrelation of the responses at xA and xB. This is tion G共xB,xA,t兲, propagating from xA to xB. Because we started this
interpreted as the response of a source at xA, observed at xB, i.e., derivation with the crosscorrelation of the Green’s functions, we
G共xB,xA,t兲. have obtained the following 1D Green’s function representation:
Tutorial on interferometry: Part 1 75A197
G共xB,xA,t兲 ⳱ G共xB,xS,t兲 ⴱ G共xA,xS,ⳮ t兲. 共1兲 tween xA and xB, convolved with SN共t兲, i.e., the autocorrelation of the
noise N共t兲. The correlation is shown in Figure 2c, which indeed re-
This representation formulates the principle that the crosscorrelation veals a bandlimited impulse centered at t ⳱ 0.6 s 共the traveltime
of observations at two receivers 共xA and xB兲 gives the response at one from xA to xB兲. Note that from registrations at two receivers of a noise
of those receivers 共xB兲 as if there were a source at the other receiver field from an unknown source in a medium with unknown propaga-
共xA兲. It also shows why seismic interferometry is often called tion velocity, we have obtained a bandlimited version of the Green’s
Green’s function retrieval. function. By dividing the distance between the receivers 共1200 m兲
Note that the source is not necessarily an impulse. If the source by the traveltime estimated from the bandlimited Green’s function
function is defined by some wavelet s共t兲, then the responses at xA and 共0.6 s兲, we obtain an estimate of the propagation velocity between
xB can be written as u共xA,xS,t兲 ⳱ G共xA,xS,t兲 ⴱ s共t兲 and u共xB,xS,t兲 the receivers 共2000 m / s兲. This illustrates that direct-wave interfer-
⳱ G共xB,xS,t兲 ⴱ s共t兲, respectively. Let Ss共t兲 be the autocorrelation of ometry can be used for tomographic inversion.
the wavelet, i.e., Ss共t兲 ⳱ s共t兲 ⴱ s共ⳮt兲. Then the crosscorrelation of Until now, we considered a single plane wave propagating in the
u共xA,xS,t兲 and u共xB,xS,t兲 gives the right-hand side of equation 1, con- positive x-direction. In Figure 3a, we consider the same configura-
volved with Ss共t兲. This is equal to the left-hand side of equation 1, tion as in Figure 1a, but now an impulsive unit source at x ⳱ x⬘S radi-
convolved with Ss共t兲. Therefore, ates a leftward-propagating plane wave. Figure 3b is the response at
xA, given by G共xA,x⬘S,t兲 ⳱ ␦ 共t ⳮ t⬘A兲, with t⬘A ⳱ 共x⬘S ⳮ xA兲 / c. Similarly,
G共xB,xA,t兲 ⴱ Ss共t兲 ⳱ u共xB,xS,t兲 ⴱ u共xA,xS,ⳮ t兲. 共2兲 the response at xB is G共xB,x⬘S,t兲 ⳱ ␦ 共t ⳮ t⬘B兲, with t⬘B ⳱ 共x⬘S ⳮ xB兲 / c
共Figure 3c兲. The crosscorrelation of these responses gives ␦ 共t ⳮ 共t⬘B
In words: If the source function is a wavelet instead of an impulse, ⳮ t⬘A兲兲 ⳱ ␦ (t Ⳮ 共xB ⳮ xA兲 / c), which is equal to the time-reversed
then the crosscorrelation of the responses at two receivers gives the Green’s function G共xB,xA,ⳮt兲. So, for the configuration of Figure
Green’s function between these receivers, convolved with the auto- 3a, we obtain the following Green’s function representation:
correlation of the source function.
This principle holds true for any source function, including noise. G共xB,xA,ⳮ t兲 ⳱ G共xB,xS⬘,t兲 ⴱ G共xA,xS⬘,ⳮt兲. 共3兲
Figure 2a and b shows the responses at xA and xB, respectively, of a
bandlimited noise source N共t兲 at xS 共the central frequency of the We can combine equations 1 and 3 as follows:
noise is 30 Hz; the figure shows only 4 s of a total of 160 s of noise兲.
2
In this numerical example, the distance between the receivers is
1200 m and the propagation velocity is 2000 m / s; hence, the travel- G共xB,xA,t兲 Ⳮ G共xB,xA,ⳮt兲 ⳱ 兺 G共xB,x共i兲共i兲
S ,t兲 ⴱ G共xA,xS ,ⳮt兲,
i⳱1
time between these receivers is 0.6 s. As a consequence, the noise re-
sponse at xB in Figure 2b is 0.6 s delayed with respect to the response 共4兲
at xA in Figure 2a 共similar to the impulse in Figure 1c delayed with re-
spect to the impulse in Figure 1b兲. Crosscorrelation of these noise re- where x共i兲
S for i ⳱ 1,2 stands for x S and x ⬘
S , respectively.
sponses gives, analogous to equation 2, the impulse response be- For the 1D situation, this combination may not seem very useful.
We analyze it here, however, because this representation better re-
sembles the 2D and 3D representations we encounter later. Note that
because G共xB,xA,t兲 is the causal response of an impulse at t ⳱ 0
共meaning it is nonzero only for t ⬎ 0兲, it does not overlap with
G共xB,xA,ⳮt兲共which is nonzero only for t ⬍ 0兲. Hence, G共xB,xA,t兲
can be resolved from the left-hand side of equation 4 by extracting
the causal part. If the source function is a wavelet s共t兲 with autocor-
relation Ss共t兲, we obtain, analogous to equation 2,
Figure 2. As in Figure 1 but this time for a noise source N共t兲 at xS. 共a兲
The response observed at xA, i.e., u共xA,xS,t兲 ⳱ G共xA,xS,t兲 ⴱ N共t兲. 共b兲
As in 共a兲 but for a receiver at xB. 共c兲 The crosscorrelation, which is Figure 3. As in Figure 1 but this time for a leftward-traveling impul-
equal to G共xB,xA,t兲 ⴱ SN共t兲, with SN共t兲 the autocorrelation of the sive plane wave. The crosscorrelation in 共d兲 is interpreted as the
noise. time-reversed Green’s function G共xB,xA,ⳮt兲.
兵G共xB,xA,t兲 Ⳮ G共xB,xA,ⳮ t兲其 ⴱ Ss共t兲具u共xB,t兲 ⴱ u共xA,ⳮ t兲典

2
⳱ 兺 u共xB,x共i兲共i兲
S ,t兲 ⴱ u共xA,xS ,ⳮ t兲. 共5兲 ⳱ 冓兺
2
兺
2
G共xB,x共j兲
S ,t兲 ⴱ N j共t兲
冔
i⳱1 j⳱1 i⳱1
Here, G共xB,xA,t兲 ⴱ Ss共t兲 may have some overlap with G共xB,xA, ⴱ G共xA,x共i兲
S ,ⳮ t兲 ⴱ Ni共ⳮ t兲
ⳮt兲 ⴱ Ss共t兲 for small 兩t兩, depending on the length of the autocorrela-
tion function Ss共t兲. Therefore, G共xB,xA,t兲 ⴱ Ss共t兲 can be extracted 2
from the left-hand side of equation 5, except for small distances 兩xB ⳱ 兺 G共xB,x共i兲共i兲
S ,t兲 ⴱ G共xA,xS ,ⳮ t兲 ⴱ SN共t兲. 共6兲
ⳮ xA兩. i⳱1
The right-hand sides of equations 4 and 5 state that the crosscorre-
lation is applied to the responses of each source separately, after Combining equation 6 with equation 4, we finally obtain
which the summation over the sources is carried out. For impulsive
兵G共xB,xA,t兲 Ⳮ G共xB,xA,ⳮ t兲其 ⴱ SN共t兲 ⳱ 具u共xB,t兲 ⴱ u共xA,ⳮt兲典.
sources or transient wavelets s共t兲, these steps should not be inter-
changed. Let us see why. Suppose the sources at xS and x⬘S act simul- 共7兲
taneously, as illustrated in Figure 4a. Then the response at xA would Expression 7 shows that the crosscorrelation of two observed fields
be given by u共xA,t兲 ⳱ 兺2i⳱1G共xA,x共i兲 S ,t兲 ⴱ s共t兲 and the response at x B by at xA and xB, each of which is the superposition of rightward- and left-
u共xB,t兲 ⳱ 兺2j⳱1G共xB,x共j兲 S ,t兲 ⴱ s共t兲. These responses are shown in Fig- ward-propagating noise fields, gives the Green’s function between
ure 4b and c for an impulsive source 共s共t兲 ⳱ ␦ 共t兲兲. The crosscorrela- xA and xB plus its time-reversed version, convolved with the autocor-
tion of these responses, shown in Figure 4d, contains two cross- relation of the noise 共see Figure 5c兲. The cross-terms, unlike Figure
terms at tB ⳮ t⬘A and t⬘B ⳮ tA that have no physical meaning. Hence, for 4d, do not contribute because the noise sources N1共t兲 and N2共t兲 are
impulsive or transient sources, the order of crosscorrelation and uncorrelated.
summation matters. Miyazawa et al. 共2008兲 apply equation 7 with xA and xB at different
The situation is different for noise sources. Consider two simul- depths along a borehole in the presence of industrial noise at Cold
taneously acting noise sources N1共t兲 and N2共t兲 at xS and x⬘S, respec- Lake, Alberta, Canada. By choosing for u different components of
tively. The responses at xA and xB are given by u共xA,t兲 multicomponent sensors in the borehole, they retrieve separate
⳱ 兺2i⳱1G共xA,x共i兲共j兲
S ,t兲 ⴱ N i共t兲 and u共x B,t兲 ⳱ 兺 j⳱1G共x B,x S ,t兲 ⴱ N j共t兲, re-
2 Green’s functions for P- and S-waves, the latter with different polar-
spectively 共see Figure 5a and b兲. Because each of these responses is izations. From the arrival times in the Green’s functions, they derive
the superposition of a rightward- and a leftward-propagating wave, the different propagation velocities and accurately quantify shear-
the response in Figure 5b is not a shifted version of that in Figure 5a wave splitting.
共unlike the responses in Figure 2a and b兲. We assume that the noise Despite the relative simplicity of our 1D analysis of direct-wave
sources are uncorrelated; thus, 具N j共t兲 ⴱ Ni共ⳮt兲典 ⳱ ␦ ijSN共t兲, where interferometry, we can make several observations about seismic in-
␦ ij is the Kronecker delta function and 具·典 denotes ensemble averag- terferometry that also hold true for more general situations. First, we
can distinguish between interferometry for impulsive or transient
ing. In practice, the ensemble averaging is replaced by integrating
over sufficiently long time. In the numerical example the duration of
the noise signals is again 160 s 共only 4 s of noise is shown in Figure
5a and b兲. For the crosscorrelation of the responses at xA and xB, we
can now write
Figure 4. As in Figures 1 and 3 but with simultaneously rightward- Figure 5. As in Figure 4 but this time with simultaneously rightward-
and leftward-traveling impulsive plane waves. The crosscorrelation and leftward-traveling uncorrelated noise fields. The crosscorrela-
in 共d兲 contains cross-terms that have no physical meaning. tion in 共c兲 contains no cross-terms.
sources on the one hand 共equations 4 and 5兲 and interferometry for the traveltimes in Figure 6b and c. This is because in the crosscorre-
noise sources on the other hand 共equation 7兲. In the case of impulsive lation process only the time difference along the paths to xA and xB
or transient sources, the responses of each source must be crosscor- matters.
related separately, after which a summation over the sources takes The source in Figure 6a with ␾ S ⳱ 0° plays the same role as the
place. In the case of uncorrelated noise sources, a single crosscorre- plane-wave source at xS in Figure 1a. For this source, the crosscorre-
lation suffices. lation gives a signal at 兩xB ⳮ xA兩 / c ⳱ 0.6 s, seen in the trace at ␾ S
Second, it appears that an isotropic illumination of the receivers is ⳱ 0° in Figure 6d. Similarly, the source at ␾ S ⳱ 180° plays the same
required to obtain a time-symmetric response between the receivers role as the plane-wave source at x⬘S in Figure 3a and leads to the trace
共of which the causal part is the actual response兲. In one dimension, at ␾ S ⳱ 180° in Figure 6d with a signal at ⳮ0.6 s. Analogous to
isotropic illumination means equal illumination by rightward- and equation 5, we sum the crosscorrelations of all sources, i.e., we sum
leftward-propagating waves. In two and three di-
mensions, it means equal illumination from all di-
rections 共discussed in the next section兲.
Finally, instead of the time-symmetric re-
sponse G共xB,xA,t兲 Ⳮ G共xB,xA,ⳮt兲, in the litera-
ture we often encounter an antisymmetric re-
sponse G共xB,xA,t兲 ⳮ G共xB,xA,ⳮt兲. This is merely
a result of differently defined Green’s functions.
Note that a simple time differentiation of the
Green’s functions would turn the symmetric re-
sponse into an antisymmetric one, and vice versa
共see Wapenaar and Fokkema 关2006兴 for a more
detailed discussion on this aspect兲.
2D and 3D analysis of direct-wave

interferometry
We extend our discussion of direct-wave inter-
ferometry to configurations with more dimen-
sions. In the following discussion, we mainly use
heuristic arguments, illustrated with a numerical
example. For a more precise derivation based on
stationary-phase analysis, we refer to Snieder
共2004兲.
Consider the 2D configuration shown in Figure
6a. The horizontal dashed line corresponds to the
1D configuration of Figure 1a, with two receivers
at xA and xB, 1200 m apart 共x denotes a Cartesian
coordinate vector兲. The propagation velocity c is
2000 m / s, and the medium is again assumed to
be lossless. Instead of plane-wave sources, we
have many point sources denoted by the small
black dots, distributed over a “pineapple slice,”
emitting transient signals with a central frequen-
cy of 30 Hz. In polar coordinates, the positions of
the sources are denoted by 共rS,␾ S兲. The angle ␾ S
is equidistantly sampled 共⌬␾ S ⳱ 0.25° 兲, whereas
the distance rS to the center of the slice is chosen
randomly between 2000 and 3000 m. The re-
sponses at the two receivers at xA and xB are
shown in Figure 6b and c, respectively, as a func-
tion of the 共polar兲 source coordinate ␾ S 共for dis- Figure 6. A 2D example of direct-wave interferometry. 共a兲 Distribution of point sources,
play purposes, only every sixteenth trace is isotropically illuminating the receivers at xA and xB. The thick dashed lines indicate the
shown兲. These responses are crosscorrelated 共for Fresnel zones. 共b兲 Responses at xA as a function of the 共polar兲 source coordinate ␾ S. 共c兲
each source separately兲, and the crosscorrelations Responses at xB. 共d兲 Crosscorrelation of the responses at xA and xB. The dashed lines indi-
cate the Fresnel zones. 共e兲 The sum of the correlations in 共d兲. This is interpreted as
are shown in Figure 6d, again as a function of ␾ S. 兵G共xB,xA,t兲 Ⳮ G共xB,xA,ⳮt兲其 ⴱ Ss共t兲. The main contributions come from sources in the
Such a gather is often called a correlation gather. Fresnel zones indicated in 共a兲 and 共d兲. 共f兲 Single crosscorrelation of the responses at
Note that the traveltimes in this correlation gather xA and xB of simultaneously acting uncorrelated noise sources. The duration of the noise
vary smoothly with ␾ S, despite the randomness of signals was 9600 s.
all traces in Figure 6d, which leads to the time-symmetric response sources 共in two or three dimensions兲 are not necessarily primary
in Figure 6e, with two events at 0.6 and ⳮ0.6 s. These two events sources but can also be secondary sources, i.e., scatterers in a homo-
are again interpreted as the response of a source at xA, observed at xB, geneous embedding. These secondary sources are not independent,
plus its time-reversed version, i.e., 兵G共xB,xA,t兲 Ⳮ G共xB,xA, but the late coda of the multiply scattered response reasonably re-
ⳮt兲其 ⴱ Ss共t兲, where Ss共t兲 is the autocorrelation of the source wavelet. sembles a diffuse wavefield. Thus, in situations with few primary
Because the sources have a finite frequency content, not only do the sources but many secondary sources, only the late coda is used for
sources exactly at ␾ S ⳱ 0° and ␾ S ⳱ 180° contribute to these events Green’s function retrieval 共Campillo and Paul, 2003兲. It is, however,
but also the sources in Fresnel zones around these angles. These unclear how well a scattering medium should be illuminated by dif-
Fresnel zones are denoted by the thick dashed lines in Figure 6a and ferent sources for the scatterers to act as independent secondary
d. In Figure 6d, the centers of these Fresnel zones are the stationary sources. Fan and Snieder 共2009兲 show an example where the scat-
points of the traveltime curve of the crosscorrelations. Note that the tered waves excited by a single source are equipartitioned, in the
events in all traces outside the Fresnel zones in Figure 6d interfere sense that energy propagates equally in all directions, but where the
destructively and give no coherent contribution in Figure 6e. The crosscorrelation of those scattered waves does not resemble the
noise between the two events in Figure 6e results because the travel- Green’s function.
time curve in Figure 6d is not 100% smooth, caused by the random- One of the most widely used applications of direct-wave interfer-
ness of the source positions in Figure 6a. ometry is the retrieval of seismic surface waves between seismome-
The response in Figure 6e is obtained by summing crosscorrela- ters and the subsequent tomographic determination of the surface-
tions of independent transient sources. Using the arguments in the wave velocity distribution of the subsurface. This approach has been
previous section, we can replace the transient sources with simulta- pioneered by Campillo and Paul 共2003兲; Shapiro and Campillo
neously acting noise sources. The cross-terms disappear when the 共2004兲; Sabra et al. 共2005a, 2005b兲; and Shapiro et al. 共2005兲. In lay-
noise sources are uncorrelated; hence, a single crosscorrelation of ered media, surface waves consist of several propagating modes, of
noise observations at xA and xB gives, analogous to equation 7, which the fundamental mode is usually the strongest.As long as only
兵G共xB,xA,t兲 Ⳮ G共xB,xA,ⳮt兲其 ⴱ SN共t兲, where SN共t兲 is the autocorrela- the fundamental mode is considered, surface waves can be seen as an
tion of the noise 共see Figure 6f兲. Note that the symmetry of the re- approximate solution of a 2D wave equation with a frequency-de-
sponses in Figure 6e and f relies again on the isotropic illumination pendent propagation velocity. So by considering the 2D configura-
of the receivers, i.e., on the net power flux of the illuminating wave- tion of Figure 6a as a plan view, the analysis above holds for ambient
field being 共close to兲 zero 共van Tiggelen, 2003; Malcolm et al., 2004; surface-wave noise. The Green’s function of the fundamental mode
Sánchez-Sesma et al., 2006; Snieder et al., 2007; Perton et al., 2009; of the direct surface wave can thus be extracted by crosscorrelating
Weaver et al., 2009; Yao et al., 2009兲. ambient-noise recordings at two seismometers. When many seis-
Of course, what is demonstrated here for a 2D distribution of mometers are available, this procedure can be repeated for any com-
sources also holds for a 3D source distribution. In that case, all bination of two seismometers. In other words, each seismometer can
sources in Fresnel volumes rather than Fresnel zones contribute to be turned into a virtual source, the response of which is observed by
the retrieval of the direct wave between xA and xB. Furthermore, the all other seismometers.
Figure 7, reproduced with permission from Lin
et al. 共2009兲, shows a beautiful example of the
Rayleigh-wave response of a virtual source
southeast of Lake Tahoe, California, U.S.A. The
white triangles represent more than 400 seis-
mometers 共USArray stations兲. Ocean-generated
ambient seismic noise 共Longuet-Higgins, 1950;
Webb, 1998; Stehly et al., 2006兲 was recorded be-
tween October 2004 and November 2007. Be-
cause this noise is coming from the ocean, it is far
from isotropic. This means the crosscorrelation
of the noise between any two stations does not
yield time-symmetric results such as those in Fig-
ure 6. However, as long as one of the Fresnel
zones is sufficiently covered with sources, it is
possible to retrieve either G共xB,xA,t兲 ⴱ SN共t兲 or
G共xB,xA,ⳮt兲 ⴱ SN共t兲共note that the location and
shape of the Fresnel zone is different for each
combination of stations兲. The snapshots shown in
Figure 7a and b were obtained by crosscorrelating
the noise recorded at the station denoted by the
star with noise recorded at all other stations. The
amplitudes exhibit azimuthal variation due to the
Figure 7. Two snapshots of the Rayleigh-wave response of a virtual source 共the white anisotropic illumination. Responses such as this
star兲 southeast of Lake Tahoe, California, U.S.A. 共Lin et al., 2009兲. The white triangles are used for tomographic inversion of the Ray-
represent more than 400 seismometers 共USArray stations兲. The shown response was ob- leigh-wave velocity of the crust and for measur-
tained by crosscorrelating three years of ambient noise, recorded at the station denoted by
the star, with that recorded at all other stations. ing azimuthal anisotropy in the crust.
Bensen et al. 共2007兲 show that it is possible to retrieve the Ray- lossless half-space 共Figure 9a兲. An impulsive unit source in the low-
leigh-wave velocity as a function of frequency. Brenguier et al. er half-space emits a vertically upward-propagating plane wave that
共2007兲 combine these approaches to 3D tomographic inversion. reaches the surface after a time t0. Because it was transmitted by a
From noise measurements at the Piton de la Fournaise volcano, they single interface on its way to the surface, the first arrival is given by
have retrieved the Rayleigh-wave group velocity distribution as a ␶ ␦ 共t ⳮ t0兲, where ␶ is the transmission coefficient of the interface
function of frequency and used this to derive a 3D S-wave velocity 共we use lowercase symbols for local transmission and reflection co-
model of the interior of the volcano. In the past couple of years, the efficients兲. This arrival is represented by the impulse at t ⳱ t0 in Fig-
applications of direct surface-wave interferometry have expanded ure 9b. The wave is reflected downward by the free surface 共reflec-
spectacularly. Without any claim of completeness, we mention tion coefficient ⳮ1兲 and subsequently reflected upward by the inter-
Larose et al. 共2005兲, Gerstoft et al. 共2006兲, Kang and Shin 共2006兲, face 共reflection coefficient r兲. Therefore, the next arrival reaching
Larose et al. 共2006兲, Yao et al. 共2006兲, Bensen et al. 共2008兲, Goué- the surface is ⳮr␶ ␦ 共t ⳮ t0 ⳮ ⌬t兲, with ⌬t ⳱ 2⌬z / c, where ⌬z is the
dard et al. 共2008a, 2008b兲, Liang and Langston 共2008兲, Lin et al. thickness of the first layer and c its propagation velocity. Figure 9b
共2008兲, Ma et al. 共2008兲, Yao et al. 共2008兲, Li et al. 共2009兲, and Pi- shows the total upgoing wavefield reaching the free surface — de-
cozzi et al. 共2009兲. The success of these applications is explained by noted as T共t兲, where T stands for the global transmission response. It
the fact that surface waves are by far the strongest events in ambient consists of an infinite series of impulses with regular intervals ⌬t
seismic noise. In the next section, we show that the retrieval of re- 共starting at t0兲 and amplitudes a0 ⳱ ␶ , a1 ⳱ ⳮr␶ , a2 ⳱ r2␶ , a3 ⳱
flected waves from ambient seismic noise is an order more difficult. ⳮr3␶ , etc.
Direct surface-wave interferometry has an interesting link with Seismic interferometry for a vertically propagating plane wave
early work by Aki 共1957, 1965兲 and Toksöz 共1964兲 on the spatial au- reduces to evaluating the autocorrelation of the global transmission
tocorrelation 共SPAC兲 method. The SPAC method uses a circular ar- response, i.e., T共t兲 ⴱ T共ⳮt兲. We obtain the simplest result if we con-
ray of seismometers plus a seismometer at the center of the circle. sider so-called power-flux normalized up- and downgoing waves
For a distribution of uncorrelated fundamental-mode Rayleigh 共Frasier, 1970; Kennett et al., 1978; Ursin, 1983; Chapman, 1994兲.
waves propagating as plane waves in all directions, the spatial auto- This means we define the local transmission coefficient ␶ as the
correlation function obtained from the circular array reveals the lo- square root of the product of the transmission coefficients for acous-
cal surface-wave velocity as a function of frequency and, subse- tic pressure and particle velocity. Hence, for an upgoing wave, ␶
quently, the local depth-dependent velocity profile. An important ⳱ 冑共1 ⳮ r兲共1 Ⳮ r兲 ⳱ 冑1 ⳮ r2 共which is also the transmission coeffi-
difference with the interferometry approach is that the distances be- cient for a downgoing wave兲. The autocorrelation for zero time lag is
tween the receivers in the SPAC method are usually smaller than half
a wavelength 共Henstridge, 1979兲, making it a local method; whereas 共a20 Ⳮ a21 Ⳮ a22 Ⳮ a23 Ⳮ ¯ 兲␦ 共t兲 ⳱ ␶ 2共1 Ⳮ r2 Ⳮ r4 Ⳮ r6
in direct-wave interferometry, the distances are assumed much larg-
Ⳮ ¯ 兲␦ 共t兲 ⳱ ␶ 2共1 ⳮ r2兲ⳮ1␦ 共t兲 ⳱ ␦ 共t兲.
er than the wavelength because otherwise the stationary-phase argu-
ments would not hold. More recent discussions on the SPAC method This is represented by the impulse at t ⳱ 0 in Figure 9c. The autocor-
are given by Okada 共2003, 2006兲 and Asten 共2006兲. An interesting relation for time lag ⌬t is
discussion on the relation between the SPAC method and seismic in-
terferometry is given by Yokoi and Margaryan 共2008兲. 共a1a0 Ⳮ a2a1 Ⳮ a3a2 Ⳮ ¯ 兲␦ 共t ⳮ ⌬t兲 ⳱ ⳮr␶ 2共1 Ⳮ r2 Ⳮ r4
Ⳮ ¯ 兲␦ 共t ⳮ ⌬t兲 ⳱ ⳮr␦ 共t ⳮ ⌬t兲,
REFLECTED-WAVE INTERFEROMETRY
which is represented by the impulse at ⌬t in Figure 9c. For time lags
1D analysis of reflected-wave interferometry 2⌬t, 3⌬t, etc., we obtain r2␦ 共t ⳮ 2⌬t兲, ⳮr3␦ 共t ⳮ 3⌬t兲, etc. Apart
The figure on the cover of Schuster’s book on seismic interferom- from an overall minus sign, these impulses together 共except the one
etry 共Schuster, 2009兲, reproduced in Figure 8, explains the basic at t ⳱ 0兲 represent the global reflection response R共t兲 of a downgo-
principle of reflected-wave interferometry very well. Figure 8a
shows a source in the subsurface that radiates a transient wave to the
earth’s surface, where it is received by a geophone. The trace con-
tains the delayed source wavelet. Figure 8b shows how the wave is
reflected downward by the surface, reflected upward again by a scat-
terer in the subsurface, and received by a second geophone at the
earth’s surface. The trace contains the wavelet, which is further de-
layed due to the propagation along the additional path from receiver
1 via the scatterer to receiver 2. The propagation paths in Figure 8a
and b have the path from the subsurface source to the first receiver in
common. By crosscorrelating the two traces 共Schuster denotes this
by 丢兲, the propagation along this common path is eliminated, leav-
ing the path from receiver 1 via the scatterer to receiver 2 共Figure 8c兲.
Hence, the result can be interpreted as a reflection experiment with a Figure 8. Basic principle of reflected-wave interferometry 共Schuster,
source at the position of the first geophone, of which the reflection 2001, 2009兲. 共a兲 A subsurface source emits a wave to the surface
response is received by the second geophone. where it is received by a geophone. 共b兲 A second geophone receives a
reflected wave. 共c兲 Crosscorrelation eliminates the propagation
Let us see how this method deals with multiple reflections. To this along the path from the source to the first geophone. The result is in-
end, we consider a configuration consisting of a homogeneous loss- terpreted as the reflection response of a source at the position of the
less layer, sandwiched between a free surface and a homogeneous first geophone, observed by the second geophone.
ing plane wave, illuminating the medium from the free surface. Con- normalized. Hence, the global transmission response of the downgo-
sequently, the causal part of the autocorrelation is equal to ⳮR共t兲. ing plane wave source at the free surface is equal to that of an upgo-
Similarly, the acausal part is ⳮR共ⳮt兲. Taking everything together, ing plane-wave source below the lowest interface 共Frasier, 1970兲.
we have T共t兲 ⴱ T共ⳮt兲 ⳱ ␦ 共t兲 ⳮ R共t兲 ⳮ R共ⳮt兲, or Because we consider a lossless medium, we can use the principle
of power conservation to derive a relation between the wavefields at
R共t兲 Ⳮ R共ⳮt兲 ⳱ ␦ 共t兲 ⳮ T共t兲 ⴱ T共ⳮt兲. 共8兲 the top and the bottom of the configuration. The power flux is most
easily defined in the frequency domain. To this end, we define the
Expression 8 shows that the global reflection response can be ob- Fourier transform of a time-dependent function as
tained from the autocorrelation of the global transmission response.
This concept can be understood intuitively if one bears in mind that
the reflection response, including all its multiples, is implicitly
present in the coda of the transmission response 共see Figure 9b兲.
f̂共␻ 兲 ⳱ 冕
ⳮ⬁
⬁
f共t兲exp共ⳮj␻ t兲dt, 共9兲
Note the analogy of equation 8 with the expression for direct-wave where ␻ is the angular frequency and j the imaginary unit. The net
interferometry 共equation 4兲. In both cases, the left-hand side is a su- power flux just below the free surface is given by
perposition of a causal response and its time-reversed version. The
main difference is that the right-hand side of equation 4 is a superpo- D̂D̂* ⳮ ÛÛ* ⳱ 共1 ⳮ R̂兲共1 ⳮ R̂*兲 ⳮ R̂R̂* ⳱ 1 ⳮ R̂ ⳮ R̂*,
sition of crosscorrelations of rightward- and leftward-propagating
waves, which was necessary to get the time-symmetric response, 共10兲
whereas the right-hand side of equation 8 is a single autocorrelation. where the superscript asterisk denotes complex conjugation. The net
The free surface in Figure 9a acts as a mirror, which removes the re- power flux is independent of depth, so the right-hand side of equa-
quirement of having sources at both sides of the receivers to obtain a tion 10 is equal to the net power flux in the lower half-space T̂T̂*.
time-symmetric response. Thus, 1 ⳮ R̂ ⳮ R̂* ⳱ T̂T̂*, or
It can easily be shown that equation 8 holds for arbitrary horizon-
tally layered media. To this end, consider the configuration in Figure
R̂ Ⳮ R̂* ⳱ 1 ⳮ T̂T̂* . 共11兲
9d. Here, the illuminating wavefield is an impulsive downgoing
plane wave at the free surface 共denoted by ␦ 共t兲 in Figure 9d兲. The Complex conjugation in the frequency domain corresponds to time
upgoing wave arriving at the free surface is the global reflection re- reversal in the time domain, so the inverse Fourier transform of this
sponse R共t兲, which is reflected downward by the free surface with re- equation again gives equation 8, which has now been proven to hold
flection coefficient ⳮ1. Accordingly, the total downgoing wavefield for arbitrarily layered media.
just below the surface is D共t兲 ⳱ ␦ 共t兲 ⳮ R共t兲 and the total upgoing The central assumption in this derivation is the conservation of
wavefield is U共t兲 ⳱ R共t兲. The total downgoing wavefield below the acoustic power, which of course only holds in lossless media. We as-
lowest interface is given by the global transmission response T共t兲. sumed in our discussions of direct-wave interferometry that the me-
We assume again that the downgoing and upgoing waves are flux dium was lossless; but in the present derivation, the essence of this
a) c)
δ (t)
_1 r2 r2
r
r4 r4
r6 r6
τ
t
0
_r 5 _r 3 _r 3 _r 5
_r
_r _r
_ R( _ t) _ R(t)
τ
b) r 2τ d)
r 4τ δ (t) R(t) _ R(t)
T(t) τ
r6
t0 t
0
_r 5τ
_r 3τ
_rτ
T(t)
Figure 9. From transmission to reflection response 共1D兲. 共a兲 Simple layered medium with an upgoing plane wave radiated by a source in the low-
er half-space. 共b兲 The transmission response T共t兲 observed at the free surface. 共c兲 The autocorrelation T共t兲 ⴱ T共ⳮt兲. The causal part is, apart from
a minus sign, the reflection response R共t兲. 共d兲 Configuration used to derive the same relation for an arbitrarily layered medium.
assumption has become manifest. Most approaches to seismic inter- position of the specular reflection point are unknown. However,
ferometry rely on the assumption that the medium is lossless. In Part when there are multiple 共unknown兲 sources in the subsurface, it is
2 of this tutorial, we also encounter approaches that account for loss- again possible to extract the reflection response.
es or that use the essence of this assumption to estimate loss parame- To see this, consider the situation depicted in Figure 10a, in which
ters. there are multiple sources buried in the subsurface. The ray that
We should note here that equation 8 for arbitrarily layered media leaves the source at x1,S ⳱ ⳮ300 m reflects at xA 共the position of the
was derived more than 40 years ago by Jon Claerbout at Stanford first geophone兲 on its way to the scatterer at xD and the second geo-
University 共Claerbout, 1968兲. His expression looks slightly different phone at xB; this is the specular ray. The rays leaving the other sourc-
because he did not use flux normalization. For his derivation, he used es have their specular reflection points left and right from xA 共the sol-
a recursive method introduced by Thomson 共1950兲, Haskell 共1953兲, id rays in Figure 10a兲. The direct arrivals at xA follow the dashed
and others. Later, he proposed the shorter derivation using energy paths and do not coincide with the solid rays, except for the source at
conservation 共Claerbout, 2000兲. Frasier 共1970兲 generalizes Claer- x1,S ⳱ ⳮ300 m. For each of the sources, we crosscorrelate the direct
bout’s result for obliquely propagating plane P- and SV-waves in a arrival at xA with the scattered wave recorded at xB. This gives the
horizontally layered elastic medium. correlation gather shown in Figure 10b, in which the horizontal axis
Analogous to equations 5 and 7, equation 8 can be modified for denotes the source coordinate x1,S. The trace at x1,S ⳱ ⳮ300 m
transient or noise signals. For example, let u共t兲 ⳱ T共t兲 ⴱ N共t兲 be the shows an impulse 共indicated by the vertical arrow兲 at tAB, which is the
upgoing wavefield at the surface, with N共t兲 representing the noise traveltime from xA via the scatterer to xB. The impulses in the sur-
signal emitted by the source in the lower half-space. Then we obtain rounding traces arrive before tAB.
from equation 8 If we sum the traces for all x1,S, the main contribution comes from
an area 共the Fresnel zone, indicated by the dashed lines兲 around the
兵R共t兲 Ⳮ R共ⳮt兲其 ⴱ SN共t兲 ⳱ SN共t兲 ⳮ 具u共t兲 ⴱ u共ⳮt兲典, 共12兲 point x1,S ⳱ ⳮ300 m where the traveltime curve is stationary 共indi-
where SN共t兲 is the autocorrelation of the noise. Equation 12 shows cated by the vertical arrow兲; the other contributions cancel. Hence,
that the autocorrelation of passive noise measurements gives the re- the sum of the correlations 共Figure 10c兲 contains an impulse at tAB
flection response of a transient source at the surface. Quite remark- and can be interpreted as the reflection response that would be mea-
able indeed! Again, the position of the actual source does not need to
be known, but it should lie below the lowest interface. In the next
section, we show that the latter assumption can be relaxed in 2D and
3D configurations.
Early applications of equation 12, some more successful than oth-
ers, are discussed by Baskir and Weller 共1975兲, Scherbaum 共1987a,
1987b兲, Cole 共1995兲, Daneshvar et al. 共1995兲, and Poletto and Petro-
nio 共2003, 2006兲.
2D and 3D analysis of reflected-wave interferometry

Claerbout conjectured for the 2D and 3D situation that “by cross- x
correlating noise traces recorded at two locations on the surface, we
can construct the wavefield that would be recorded at one of the loca-
tions if there was a source at the other” 共citation is from Rickett and
Claerbout 关1999兴, but the conjecture is also mentioned by Cole
关1995兴兲. This statement could be applied literally to direct-wave in-
terferometry, as discussed in a previous section, but Claerbout’s con- t
jecture concerns reflected-wave interferometry. Of course, this ter-
minology was not used by these authors, and the links between di-
rect-wave and reflected-wave interferometry were discovered sever-
al years later. Duvall et al. 共1993兲 and Rickett and Claerbout 共1999兲
applied crosscorrelations to noise observations at the surface of the
sun and were able to retrieve helioseismological shot records.
Claerbout’s 1D relation 共equation 8兲 and his conjecture for the 3D t
situation inspired Jerry Schuster at the University of Utah. During a
sabbatical in 2000 at Stanford University, Schuster analyzed the x
conjecture by the method of stationary phase. Let us briefly review
his line of thought 共Schuster, 2001; Schuster et al., 2004; Schuster Figure 10. Basic principle of reflected-wave interferometry revisit-
and Zhou, 2006兲. First, consider again the configuration shown in ed. 共a兲 Configuration with multiple sources in the subsurface. Only
Figure 8. It was implicitly assumed that the first geophone is located the ray emitted by the source at x1,S ⳱ ⳮ300 m has its specular re-
precisely at the specular reflection point of the drawn ray in Figure flection point at one of the geophone positions. 共b兲 Crosscorrelations
8b. As a consequence, the ray in Figure 8a coincides with the first of the responses at xA and xB as a function of the source coordinate
x1,S. The traveltime curve connecting these events is stationary at
branch of the ray in Figure 8b; so in a 1D crosscorrelation process, x1,S ⳱ ⳮ300 m. The thick dashed lines indicate the Fresnel zone. 共c兲
the traveltime along this ray cancels, which leaves the traveltime of The sum of the correlations in 共b兲. This is interpreted as the reflection
the reflection response. In practice, the source position and hence the response of a source at xA observed by a receiver at xB.
sured at xB if there were a source at xA. In other words, the source has Lu et al., 2008兲. Interferometry can be used to turn multiples in VSP
been repositioned from its unknown position at depth to a known po- data into primaries and in this way enlarge the illuminated area 共Yu
sition xA at the surface. Note that this procedure works for any xA and and Schuster, 2006; He et al., 2007; Jiang et al., 2007兲. Surface mul-
xB as long as the array of sources contains a source that emits a specu- tiples can be turned into primaries at the position of missing traces
lar ray via xA and the scatterer to xB. In Appendix A, we give a simple 共Wang et al., 2009兲. Crosscorrelation of refracted waves gives virtu-
proof that the stationary point of the traveltime curve in a correlation al refractions that can be used for improved estimation of the subsur-
gather corresponds to the source from which the rays to xA and xB face parameters 共Dong et al., 2006b; Mikesell et al., 2009兲. Surface
leave in the same direction. waves can be predicted by interferometry and subsequently sub-
This example shows that it is possible to reposition 共or redatum兲 tracted from exploration seismic data 共Curtis et al., 2006; Dong et
sources without knowing the velocity model and the position of the al., 2006a; Halliday et al., 2007, 2010; Xue et al., 2009兲. In his recent
original sources. In exploration geophysics, redatuming is known as book, Schuster 共2009兲 systematically discusses all possible interfer-
a process that brings sources and/or receivers from the acquisition ometric transformations between surface data, VSP data, single-well
level to another depth level, using extrapolation operators based on a profiles, and crosswell data. Figure 11 shows some examples. An-
macro velocity model 共Berryhill, 1979, 1984兲. In seismic interfer- other approach to interferometric redatuming of controlled-source
ometry, as illustrated in Figure 10, the extrapolation operator comes data, known as the virtual-source method 共Bakulin and Calvert,
directly from the data 共in this example, the observed direct wave at 2004, 2006兲, is discussed in Part 2 of this paper.
xA兲. The example discussed in Figure 10 deals with primary reflec-
In the years following his sabbatical, Schuster showed that the in- tions and therefore confirms Claerbout’s conjecture only partly. The
terferometric redatuming concept, indicated in Figure 10, can be ap- 1D analysis in the previous section showed that not only primary re-
plied to a wide range of configurations 共mostly for controlled-source flections but also all multiples are recovered from the autocorrela-
data兲. His work inspired many other researchers to develop interfer- tion of the transmission response. Claerbout’s conjecture for the 3D
ometric methods for exploration geophysics. For example, VSP data situation can be proven along similar lines. Instead of using the prin-
can be transformed into crosswell data 共Minato et al., 2007兲 or into ciple of power conservation, a so-called power reciprocity theorem
single-well reflection profiles to improve salt-flank delineation and is used as the starting point. In general, an acoustic reciprocity theo-
imaging 共Willis et al., 2006; Xiao et al., 2006; Hornby and Yu, 2007; rem formulates a relation between two acoustic states 共de Hoop,
Figure 11. Some examples of interferometric redatuming 共Schuster, 2009兲. Each diagram shows that crosscorrelation of the trace recorded at A
with the one at B and summing over source locations leads to the response of a source at A, closer to the target than the original sources.
1988; Fokkema and van den Berg, 1993兲. One can distinguish be- related to reflectors below the sources, whereas the physical re-
tween convolution and correlation reciprocity theorems. The theo- sponse of these deeper reflectors shows up correctly in R共xB,xA,t兲.
rems of the correlation type reduce to power-conservation laws This has also been explained with theoretical arguments 共Wapenaar
when the two states are chosen identical, which is why they are also and Fokkema, 2006兲.
called power reciprocity theorems. Because reflection and transmis- Equations 13 and 14 are used by various authors to turn ambient
sion responses are defined for downgoing and upgoing waves, for seismic noise into virtual exploration seismic reflection data 共Draga-
the proof of Claerbout’s conjecture we make use of a correlation rec-
nov et al., 2006, 2007, 2009; Hohl and Mateeva, 2006; Torii et al.,
iprocity theorem for 共flux-normalized兲 one-way wavefields 共Wap-
2007兲. Interestingly, the teleseismic community has recognized in-
enaar and Grimbergen, 1996兲.
Consider the configuration in Figure 12a. An arbitrary inhomoge- dependently that the coda of transmission responses from distant
neous lossless medium is sandwiched between a free surface and a sources contains reflection information which can be used to image
homogeneous lower half-space. Impulsive sources are distributed the earth’s crust 共Bostock et al., 2001; Rondenay et al., 2001;
along a horizontal plane in this lower half-space. For this configura- Shragge et al., 2001, 2006; Mercier et al., 2006兲. The link between
tion, we derive 共Wapenaar et al., 2002, 2004兲 teleseismic coda imaging and seismic interferometry is exploited by
Kumar and Bostock 共2006兲, Nowack et al. 共2006兲, Chaput and Bos-
R共xB,xA,t兲 Ⳮ R共xB,xA,ⳮt兲 tock 共2007兲, and Tonegawa et al. 共2009兲.
⬇ ␦ 共xH,B ⳮ xH,A兲␦ 共t兲 ⳮ 兺 T共xB,x共i兲共i兲 We conclude this section with an example of retrieving virtual ex-
S ,t兲 ⴱ T共xA,xS ,ⳮt兲.
i ploration seismic reflection data from ambient noise, recorded by
Shell in a desert area near Ajdābiya, Libya. Figure 13a shows 10 s of
共13兲 noise, arbitrarily selected from 11 hours of noise, recorded along a
Here, xH,A and xH,B denote the horizontal components of xA and xB, 20-km line. Each receiver channel represents a group of 48 vertical-
respectively; T共xA共B兲,x共i兲 S ,t兲 is the upgoing transmission response of
component geophones, designed to suppress surface waves. Never-
an impulsive point source at x共i兲 S in the subsurface, observed at x A共B兲 at
theless, the main events in Figure 13a are parts of the surface waves
the free surface. Its coda includes all surface-related and internal that fell outside the suppression band of the geophone groups; these
multiple reflections 共only a few rays are shown in Figure 12a兲. The surface waves were caused by traffic on a road intersecting the line at
right-hand side of equation 13 involves a crosscorrelation of trans- x1 ⳱ 14 km. Band-pass and frequency-wavenumber 共 f-k兲 filtering
mission responses at xA and xB for each source x共i兲 S , followed by a were used to suppress the surface waves further 共Figure 13b兲.
summation over all source positions. The time-symmetric response
on the left-hand side is the reflection response that would be record-
ed at xB if there were a source at xA, plus its time-reversed version.
The main approximation is the negligence of evanescent waves.
Apart from that, the retrieved reflection response R共xB,xA,t兲 contains
all primary, surface-related, and internal multiple reflections. They
are unraveled by equation 13 from the coda of the transmission re-
sponses.
When the impulsive sources are replaced by uncorrelated noise
sources, then the responses at xA and xB are given by u共xA,t兲
⳱ 兺iT共xA,x共i兲共j兲
S ,t兲 ⴱ N i共t兲 and u共x B,t兲 ⳱ 兺 jT共x B,x S ,t兲 ⴱ N j共t兲共see Fig-
ure 12b, where each dashed ray represents a complete transmission
response兲. Using a derivation similar to the one that transforms
equation 4 into equation 7, we obtain from equation 13
兵R共xB,xA,t兲 Ⳮ R共xB,xA,ⳮ t兲其 ⴱ SN共t兲

⬇ ␦ 共xH,B ⳮ xH,A兲SN共t兲 ⳮ 具u共xB,t兲 ⴱ u共xA,ⳮ t兲典,
共14兲
where SN共t兲 is the autocorrelation of the noise.
Equation 14 shows that the direct crosscorrelation of passive
noise measurements gives the reflection response of a transient
source at the free surface. Although equations 13 and 14 are derived Figure 12. From transmission to reflection response 共3D兲. 共a兲 Arbi-
for a situation in which the sources at x共i兲
S lie at the same depth 共Figure
trary inhomogeneous lossless medium, with sources in the homoge-
12a兲, these equations remain approximately valid when the depths neous lower half-space and receivers at xA and xB at the free surface.
According to equation 13, the reflection response R共xB,xA,t兲, implic-
are randomly distributed 共as in Figure 12b兲 because in the crosscor- itly present in the coda of the transmission response, is retrieved by
relation process, only the time difference matters 共we used a similar crosscorrelating transmission responses observed at xA and xB and
reasoning for direct-wave interferometry to explain why the travel- summing over the sources. 共b兲 When the sources are simultaneously
time curves in Figure 6d remain smooth兲. Moreover, despite the ini- acting mutually uncorrelated noise sources, the observed responses
at xA and xB are each a superposition of transmission responses.
tial assumption that the medium is homogeneous below the sources, According to equation 14, the reflection response R共xB,xA,t兲 is now
Draganov et al. 共2004兲 show with numerical examples that the ran- retrieved from the direct crosscorrelation of the observations at xA
domness of the source depths helps to suppress nonphysical ghosts and xB.
We use equation 14 to retrieve the reflection response. Strictly CONCLUSIONS

speaking, equation 14 requires decomposition of the filtered geo-
phone data of Figure 13b into the upgoing transmission response. In We have discussed the basic principles of seismic interferometry
the acoustic approximation, decomposition mainly involves the ap- in a heuristic way. We have shown that, whether we consider con-
plication of an angle-dependent amplitude filter. Because it is very trolled-source or passive interferometry, virtual sources are created
difficult to obtain true amplitude responses from ambient noise any- at positions where there are only receivers. Of course, no new infor-
way, the decomposition step is skipped. Using equation 14, with xA mation is generated by interferometry, but information hidden in
fixed 共x1,A ⳱ 1 km兲 and xB chosen variable 共x1,B ⳱ 0,. . .,4 km兲, we noise or in a complex scattering coda is reorganized into easily inter-
retrieve a seismic shot record R共xB,xA,t兲 from the noise, of which the pretable responses that can be further processed by standard tomog-
first 2.5 s are shown in Figure 14a. The red star at x1,B ⳱ x1,A ⳱ 1 km raphic inversion or reflection-imaging methodologies. The main
denotes the position of the virtual source. An active seismic reflec- strength is that this information unraveling requires no knowledge of
tion experiment, carried out with the source at the same position, is the subsurface medium parameters nor of the positions or timing of
shown in Figure 14b. Particularly in the red areas, the reflections re- the actual sources. Moreover, the processing consists of simple
trieved from the ambient noise 共Figure 14a兲 correspond quite well crosscorrelations and is almost entirely data driven.
with those in the active shot gather 共Figure 14b兲. For more details In Part 2, we discuss the relation between interferometry and
about this experiment as well as a pseudo-3D reflection image ob- time-reversed acoustics, review a mathematically sound derivation,
tained from the ambient noise, see Draganov et al. 共2009兲. and indicate recent and new advances.
ACKNOWLEDGMENTS
1 (km) 1 (km)
This work is supported by the Netherlands Re-
search Centre for Integrated Solid Earth Science
共ISES兲, the Dutch Technology Foundation
共STW兲共grant VENI.08115兲, and by the U. S. Na-
tional Science Foundation 共NSF兲共grant
EAS-0609595兲. We thank associate editor Sven
Treitel and reviewers Jerry Schuster, Kasper van
Wijk, and Mark Willis for their valuable com-
ments and suggestions, which improved this pa-
per. We are grateful to Jerry Schuster, Fan-Chi
Lin, Mike Ritzwoller, and Wiley-Blackwell 共pub-
lisher of Geophysical Journal International兲 for
Figures 11 and 7. Last but not least, we thank the
Libyan National Oil Company for permission to
publish the Ajdabeya data results and Shell in
Libya for collecting and making available the
passive data.
APPENDIX A
Figure 13. 共a兲 Ten seconds of ambient noise, arbitrarily selected from 11 hours of noise, STATIONARY-PHASE ANALYSIS
recorded in a desert area near Ajdābiya, Libya. The main events are remnants of surface
waves caused by traffic at x1 ⳱ 14 km. 共b兲 The same noise window after further suppres- We give a simple proof that the stationary
sion of the surface waves.
point of the traveltime curve in a correlation gath-
er corresponds to the source from which the rays
to the receivers at xA and xB leave in the same di-
rection. Consider two rays A and B that propagate from an arbitrary
source point to the two receivers 共Figure A-1兲. This propagation may
be direct, or it may involve bounces off reflectors or scatterers; the
fate of these rays is irrelevant for the argument presented here. The
sources involved in interferometry are located on the surface, indi-
cated by the dashed line in Figure A-1. This surface, which need not
be planar, is in three dimensions parameterized by two orthogonal
coordinates q1 and q2. We first keep q2 fixed and consider only varia-
tions in q1.
The traveltime from a given source to the receiver at xA is denoted
by tA, and the traveltime from that source to the receiver at xB is de-
Figure 14. 共a兲 Reflection response 共shot record兲 obtained by cross-
correlating 11 hours of ambient noise 共Draganov et al., 2009兲. 共b兲 noted by tB. These traveltimes are, in general, functions of the source
For comparison, an active shot record measured at the same location. position q1. In seismic interferometry, the traveltimes of the signals
ent noise surface wave tomography across the United States: Journal of
Geophysical Research — Solid Earth, 113, B05306-1–B05306-21.
Berryhill, J. R., 1979, Wave-equation datuming: Geophysics, 44, 1329–
1344.
——–, 1984, Wave-equation datuming before stack: Geophysics, 49, 2064–
2066.
Bostock, M. G., S. Rondenay, and J. Shragge, 2001, Multiparameter two-di-
mensional inversion of scattered teleseismic body waves. 1. Theory for

oblique incidence: Journal of Geophysical Research, 106, 30771–30782.
Brenguier, F., N. M. Shapiro, M. Campillo, A. Nercessian, and V. Ferrazzini,
2007, 3-D surface wave tomography of the Piton de la Fournaise volcano
using seismic correlations: Geophysical Research Letters, 34, L02305-1–
L02305-5.
Brooks, L. A., and P. Gerstoft, 2007, Ocean acoustic interferometry: Journal
of the Acoustical Society of America, 121, 3377–3385.
Figure A-1. Two rays, A and B, that propagate from a common point Callen, H. B., and T. A. Welton, 1951, Irreversibility and generalized noise:
on the surface with sources 共dashed line兲 and their take-off angles at Physical Review, 83, 34–40.
this source surface. Campillo, M., and A. Paul, 2003, Long-range correlations in the diffuse seis-
mic coda: Science, 299, 547–549.
Challis, L., and F. Sheard, 2003, The Green of Green functions: Physics To-
that are crosscorrelated are subtracted. This means that the travel- day, 56, 41–46.
time tcorr of the crosscorrelation for a given source position is Chapman, C. H., 1994, Reflection/transmission coefficients reciprocities in
anisotropic media: Geophysical Journal International, 116, 498–501.
Chaput, J. A., and M. G. Bostock, 2007, Seismic interferometry using non-
tcorr共q1兲 ⳱ tB共q1兲 ⳮ tA共q1兲. 共A-1兲 volcanic tremor in Cascadia: Geophysical Research Letters, 34, L07304-
1–L07304-5.
The condition that the traveltime is stationary means that Claerbout, J. F., 1968, Synthesis of a layered medium from its acoustic trans-
mission response: Geophysics, 33, 264–269.
⳵tcorr共q1兲 ⳵tB共q1兲 ⳵tA共q1兲 ——–, 2000, Introduction to acoustic daylight imaging, http://
⳱ ⳮ ⳱ 0. 共A-2兲 sepwww.stanford.edu/data/media/public/sep//jon/eqcor/index.html, ac-
⳵q1 ⳵q1 ⳵q1 cessed 21 May 2010.
Cole, S., 1995, Passive seismic and drill-bit experiments using 2-D arrays:
A standard derivation 共Aki and Richards, 1980兲 relates the slowness Ph.D. dissertation, Stanford University.
Curtis, A., P. Gerstoft, H. Sato, R. Snieder, and K. Wapenaar, 2006, Seismic
along the surface to the take-off angle interferometry — Turning noise into signal: The Leading Edge, 25,
1082–1092.
⳵tA共q1兲 sin iA Daneshvar, M. R., C. S. Clay, and M. K. Savage, 1995, Passive seismic imag-
⳱ , 共A-3兲 ing using microearthquakes: Geophysics, 60, 1178–1186.
⳵q1 c de Hoop, A. T., 1988, Time-domain reciprocity theorems for acoustic wave
fields in fluids with relaxation: Journal of the Acoustical Society of Ameri-
with c the propagation velocity. A similar expression holds for tB. In- ca, 84, 1877–1882.
Dong, S., R. He, and G. T. Schuster, 2006a, Interferometric prediction and
serting this in equation A-2 implies that, at the stationary point, least squares subtraction of surface waves: 76th Annual International
Meeting, SEG, Expanded Abstracts, 2783–2786.
i A ⳱ i B, 共A-4兲 Dong, S., J. Sheng, and G. T. Schuster, 2006b, Theory and practice of refrac-
tion interferometry: 76th Annual International Meeting, SEG, Expanded
which means the rays take off in the same direction. Abstracts, 3021–3025.
This reasoning is applicable to variations in the source coordi- Draganov, D., X. Campman, J. Thorbecke, A. Verdel, and K. Wapenaar,
2009, Reflection images from ambient seismic noise: Geophysics, 74, no.
nate q1. The same reasoning applies to variations with the orthogonal 5, A63–A67.
source coordinate q2. The rays take off in the same direction as mea- Draganov, D., K. Wapenaar, W. Mulder, and J. Singer, 2006, Seismic inter-
ferometry on background-noise field data: 76th Annual International
sured in two orthogonal planes; hence, the rays have the same direc- Meeting, SEG, Expanded Abstracts, 590–593.
tion in three dimensions. Therefore, the rays radiating from the sta- Draganov, D., K. Wapenaar, W. Mulder, J. Singer, and A. Verdel, 2007, Re-
tionary source position are parallel. trieval of reflections from seismic background-noise measurements: Geo-
physical Research Letters, 34, L04305-1–L04305-4.
Draganov, D., K. Wapenaar, and J. Thorbecke, 2004, Passive seismic imag-
REFERENCES ing in the presence of white noise sources: The Leading Edge, 23, no. 9,
889–892.
Duvall, T. L., S. M. Jefferies, J. W. Harvey, and M. A. Pomerantz, 1993,
Aki, K., 1957, Space and time spectra of stationary stochastic waves, with Time-distance helioseismology: Nature, 362, 430–432.
special reference to micro-tremors: Bulletin of the Earthquake Research Fan, Y., and R. Snieder, 2009, Required source distribution for interferome-
Institute, 35, 415–457. try of waves and diffusive fields: Geophysical Journal International, 179,
——–, 1965, A note on the use of microseisms in determining the shallow 1232–1244.
structures of the earth’s crust: Geophysics, 30, 665–666. Fokkema, J. T., and P. M. van den Berg, 1993, Seismic applications of acous-
Aki, K., and P. G. Richards, 1980, Quantitative seismology, vol. 1: W. H. tic reciprocity: Elsevier Scientific Publ. Co., Inc.
Freeman & Co. Frasier, C. W., 1970, Discrete time solution of plane P-SV waves in a plane
Asten M. W., 2006, On bias and noise in passive seismic data from finite cir- layered medium: Geophysics, 35, 197–219.
cular array data processed using SPAC methods: Geophysics, 71, no. 6, Gerstoft, P., K. G. Sabra, P. Roux, W. A. Kuperman, and M. C. Fehler, 2006,
V153–V162. Green’s functions extraction and surface-wave tomography from mi-
Bakulin, A., and R. Calvert, 2004, Virtual source: New method for imaging croseisms in southern California: Geophysics, 71, no. 4, SI23–SI31.
and 4D below complex overburden: 74th Annual International Meeting, Godin, O. A., 2007, Emergence of the acoustic Green’s function from ther-
SEG, Expanded Abstracts, 2477–2480. mal noise: Journal of the Acoustical Society of America, 121, no. 2, EL96–
——–, 2006, The virtual source method: Theory and case study: Geophysics, EL102.
71, no. 4, SI139–SI150. Gouédard, P., P. Roux, M. Campillo, and A. Verdel, 2008a, Convergence of
Baskir, E., and C. E. Weller, 1975, Sourceless reflection seismic exploration the two-point correlation function toward the Green’s function in the con-
共abstract兲: Geophysics, 40, 158–159. text of a seismic-prospecting data set: Geophysics, 73, no. 6, V47–V53.
Bensen, G. D., M. H. Ritzwoller, M. P. Barmin, A. L. Levshin, F. Lin, M. P. Gouédard, P., L. Stehly, F. Brenguier, M. Campillo, Y. Colin de Verdière, E.
Moschetti, N. M. Shapiro, and Y. Yang, 2007, Processing seismic ambient Larose, L. Margerin, P. Roux, F. J. Sánchez-Sesma, N. M. Shapiro, and R.
noise data to obtain reliable broad-band surface wave dispersion measure- L. Weaver, 2008b, Cross-correlation of random fields: Mathematical ap-
ments: Geophysical Journal International, 169, 1239–1260. proach and applications: Geophysical Prospecting, 56, 375–393.
Bensen, G. D., M. H. Ritzwoller, and N. M. Shapiro, 2008, Broadband ambi- Green, G., 1828, An essay on the application of mathematical analysis to the
theories of electricity and magnetism: Privately published. Nowack, R. L., S. Dasgupta, G. T. Schuster, and J.-M. Sheng, 2006, Correla-
Halliday, D. F., A. Curtis, J. O.A. Robertsson, and D.-J. van Manen, 2007, In- tion migration using Gaussian beams of scattered teleseismic body waves:
terferometric surface-wave isolation and removal: Geophysics, 72, no. 5, Bulletin of the Seismological Society of America, 96, 1–10.
A69–A73. Okada, H., 2003, The microtremor survey method: SEG.
Halliday, D. F., A. Curtis, P. Vermeer, C. Strobbia, A. Glushchenko, D.-J. van ——–, 2006, Theory of efficient array observations of microtremors with
Manen, and J. O. A. Robertsson, 2010, Interferometric ground-roll remov- special reference to the SPAC method: Exploration Geophysics, 37,
al: Attenuation of scattered surface waves in single-sensor data: Geophys- 73–85.
ics, 75, no. 2, SA15–SA25 Perton, M., F. J. Sánchez-Sesma, A. Rodríguez-Castellanos, M. Campillo,

Haskell, N. A., 1953, The dispersion of surface waves on multilayered me- and R. L. Weaver, 2009, Two perspectives on equipartition in diffuse elas-
dia: Bulletin of the Seismological Society of America, 43, 17–34. tic fields in three dimensions: Journal of the Acoustical Society of Ameri-
He, R., B. Hornby, and G. Schuster, 2007, 3D wave-equation interferometric ca, 126, 1125–1130.
migration of VSP free-surface multiples: Geophysics, 72, no. 5, S195– Picozzi, M., S. Parolai, D. Bindi, and A. Strollo, 2009, Characterization of
S203. shallow geology by high-frequency seismic noise tomography: Geophysi-
Henstridge, J. D., 1979, A signal processing method for circular arrays: Geo- cal Journal International, 176, 164–174.
physics, 44, 179–184. Poletto, F., and L. Petronio, 2003, Transmitted and reflected waves in tunnel
Hohl, D., and A. Mateeva, 2006, Passive seismic reflectivity imaging with SWD: 73rd Annual International Meeting, SEG, Expanded Abstracts,
ocean-bottom cable data: 76th Annual International Meeting, SEG, Ex- 1211–1214.
panded Abstracts, 1560–1563. ——–, 2006, Seismic interferometry with a TBM source of transmitted and
Hornby, B. E., and J. Yu, 2007, Interferometric imaging of a salt flank using reflected waves: Geophysics, 71, no. 4, SI85–SI93.
walkaway VSP data: The Leading Edge, 26, 760–763. Ramírez, A. C., and A. B. Weglein, 2009, Green’s theorem as a comprehen-
Jiang, Z., J. Sheng, J. Yu, G. T. Schuster, and B. E. Hornby, 2007, Migration sive framework for data reconstruction, regularization, wavefield separa-
methods for imaging different-order multiples: Geophysical Prospecting tion, seismic interferometry, and wavelet estimation: A tutorial: Geophys-
55, 1–19. ics, 74, no. 6, W35–W62.
Kang, T.-S., and J. S. Shin, 2006, Surface-wave tomography from ambient Rickett, J., and J. Claerbout, 1999, Acoustic daylight imaging via spectral
seismic noise of accelerograph networks in southern Korea: Geophysical factorization: Helioseismology and reservoir monitoring: The Leading
Research Letters, 33, L17303-1–L17303-5. Edge, 18, 957–960.
Kennett, B. L. N., N. J. Kerry, and J. H. Woodhouse, 1978, Symmetries in the Rondenay, S., M. G. Bostock, and J. Shragge, 2001, Multiparameter two-di-
reflection and transmission of elastic waves: Geophysical Journal of the mensional inversion of scattered teleseismic body waves. 3. Application to
Royal Astronomical Society, 52, 215–230. the Cascadia 1993 data set: Journal of Geophysical Research, 106, 30795–
Kumar, M. R., and M. G. Bostock, 2006, Transmission to reflection transfor- 30807.
mation of teleseismic wavefields: Journal of Geophysical Research — Roux, P., and M. Fink, 2003, Green’s function estimation using secondary
Solid Earth, 111, B08306-1–B08306-9. sources in a shallow water environment: Journal of the Acoustical Society
Larose, E., A. Khan, Y. Nakamura, and M. Campillo, 2005, Lunar subsurface of America, 113, 1406–1416.
investigated from correlation of seismic noise: Geophysical Research Let- Rytov, S. M., 1956, On thermal fluctuations in distributed systems: Doklady
ters, 32, L16201-1–L16201-4. Akademii Nauk SSSR, 110, 371.
Larose, E., L. Margerin, A. Derode, B. van Tiggelen, M. Campillo, N. Sha- Rytov, S. M., Y. A. Kravtsov, and V. I. Tatarskii, 1989, Principles of statistical
piro, A. Paul, L. Stehly, and M. Tanter, 2006, Correlation of random wave radiophysics. 3: Elements of random fields: Springer-Verlag Berlin.
fields: An interdisciplinary review: Geophysics, 71, no. 4, SI11–SI21. Sabra, K. G., P. Gerstoft, P. Roux, W. A. Kuperman, and M. C. Fehler, 2005a,
Le Bellac, M., F. Mortessagne, and G. G. Batrouni, 2004, Equilibrium and Extracting time-domain Green’s function estimates from ambient seismic
non-equilibrium statistical thermodynamics: Cambridge University Press. noise: Geophysical Research Letters, 32, L03310-1–L03310-5.
Li, H., W. Su, C.-Y. Wang, and Z. Huang, 2009, Ambient noise Rayleigh ——–, 2005b, Surface wave tomography from microseisms in southern Cali-
wave tomography in western Sichuan and eastern Tibet: Earth and Plane- fornia: Geophysical Research Letters, 32, L14311-1–L14311-4.
tary Science Letters, 282, 201–211. Sabra, K. G., P. Roux, and W. A. Kuperman, 2005c, Arrival-time structure of
Liang, C., and C. A. Langston, 2008, Ambient seismic noise tomography and the time-averaged ambient noise cross-correlation function in an oceanic
structure of eastern North America: Journal of Geophysical Research — waveguide: Journal of the Acoustical Society of America, 117, no. 1, 164–
Solid Earth, 113, B03309-1–B03309-18. 174.
Lin, F.-C., M. P. Moschetti, and M. H. Ritzwoller, 2008, Surface wave to- Sánchez-Sesma, F. J., J. A. Pérez-Ruiz, M. Campillo, and F. Luzón, 2006,
mography of the western United States from ambient seismic noise: Ray- Elastodynamic 2D Green function retrieval from cross-correlation: Ca-
leigh and Love wave phase velocity maps: Geophysical Journal Interna- nonical inclusion problem: Geophysical Research Letters, 33, L13305-1–
tional, 173, 281–298. L13305-6.
Lin, F.-C., M. H. Ritzwoller, and R. Snieder, 2009, Eikonal tomography: Sur- Scherbaum, F., 1987a, Seismic imaging of the site response using micro-
face wave tomography by phase front tracking across a regional broad- earthquake recordings. Part I. Method: Bulletin of the Seismological Soci-
band seismic array: Geophysical Journal International, 177, 1091–1110. ety of America, 77, 1905–1923.
Lobkis, O. I., and R. L. Weaver, 2001, On the emergence of the Green’s func- ——–, 1987b, Seismic imaging of the site response using microearthquake
tion in the correlations of a diffuse field: Journal of the Acoustical Society recordings. Part II. Application to the Swabian Jura, southwest Germany,
of America, 110, 3011–3017. seismic network: Bulletin of the Seismological Society of America, 77,
Longuet-Higgins, M. S., 1950, A theory for the generation of microseisms: 1924–1944.
Philosophical Transactions of the Royal Society of London Series A, 243, Schuster, G. T., 2001, Theory of daylight/interferometric imaging: Tutorial:
1–35. 63rd Conference & Technical Exhibition, EAGE, Extended Abstracts,
Lu, R., M. Willis, X. Campman, J. Ajo-Franklin, and M. N. Toksöz, 2008, A32.
Redatuming through a salt canopy and target-oriented salt-flank imaging: ——–, 2009, Seismic interferometry: Cambridge University Press.
Geophysics 73, no. 3, S63–S71. Schuster, G. T., J. Yu, J. Sheng, and J. Rickett, 2004, Interferometric/daylight
Ma, S., G. A. Prieto, and G. C. Beroza, 2008, Testing community velocity seismic imaging: Geophysical Journal International, 157, 838–852.
models for southern California using the ambient seismic field: Bulletin of Schuster, G. T., and M. Zhou, 2006, A theoretical overview of model-based
the Seismological Society of America, 98, 2694–2714. and correlation-based redatuming methods: Geophysics, 71, no. 4, SI103–
Malcolm, A. E., J. A. Scales, and B. A. van Tiggelen, 2004, Extracting the SI110.
Green function from diffuse, equipartitioned waves: Physical Review E, Shapiro, N. M., and M. Campillo, 2004, Emergence of broadband Rayleigh
70, 015601共R兲-1–015601共R兲-4. waves from correlations of the ambient seismic noise: Geophysical Re-
Mercier, J.-P., M. G. Bostock, and A. M. Baig, 2006, Improved Green’s func- search Letters, 31, L07614-1–L07614-4.
tions for passive-source structural studies: Geophysics, 71, no. 4, SI95– Shapiro, N. M., M. Campillo, L. Stehly, and M. H. Ritzwoller, 2005, High-
SI102. resolution surface-wave tomography from ambient seismic noise: Sci-
Mikesell, D., K. van Wijk, A. Calvert, and M. Haney, 2009, The virtual re- ence, 307, 1615–1618.
fraction: Useful spurious energy in seismic interferometry: Geophysics, Shragge, J., B. Artman, and C. Wilson, 2006, Teleseismic shot-profile migra-
74, no. 3, A13–A17. tion: Geophysics, 71, no. 4, SI221–SI229.
Minato, S., K. Onishi, T. Matsuoka, Y. Okajima, J. Tsuchiyama, D. Nobuoka, Shragge, J., M. G. Bostock, and S. Rondenay, 2001, Multiparameter two-di-
H. Azuma, and T. Iwamoto, 2007, Cross-well seismic survey without mensional inversion of scattered teleseismic body waves. 2. Numerical
borehole source: 77th Annual International Meeting, SEG, Expanded Ab- examples: Journal of Geophysical Research, 106, 30783–30793.
stracts, 1357–1361. Snieder, R., 2004, Extracting the Green’s function from the correlation of
Miyazawa, M., R. Snieder, and A. Venkataraman, 2008, Application of seis- coda waves: A derivation based on stationary phase: Physical Review E,
mic interferometry to extract P- and S-wave propagation and observation 69, 046610-1–046610-8.
of shear-wave splitting from noise data at Cold Lake, Alberta, Canada: Snieder, R., K. Wapenaar, and U. Wegler, 2007, Unified Green’s function re-
Geophysics, 73, no. 4, D35–D40. trieval by cross-correlation: Connection with energy principles: Physical
Review E, 75, 036103-1–036103-14. Acoustical Society of America, 126, 1817–1826.

Stehly, L., M. Campillo, and N. M. Shapiro, 2006, A study of the seismic Weaver, R. L., and O. I. Lobkis, 2001, Ultrasonics without a source: Thermal
noise from its long-range correlation properties: Journal of Geophysical fluctuation correlations at MHz frequencies: Physical Review Letters, 87,
Research, 111, B10306-1–B10306-12. 134301-1–134301-4.
Thompson, A. R., J. M. Moran, and G. W. Swenson Jr., 2001, Interferometry ——–, 2002, On the emergence of the Green’s function in the correlations of
and synthesis in radio astronomy: Wiley Interscience. a diffuse field: Pulse-echo using thermal phonons: Ultrasonics, 40,
Thomson, W. T., 1950, Transmission of elastic waves through a stratified sol- 435–439.
id medium: Journal of Applied Physics, 21, 89–93. Webb, S. C., 1998, Broadband seismology and noise under the ocean: Re-
Toksöz, M. N., 1964, Microseisms and an attempted application to explora- views of Geophysics, 36, 105–142.
tion: Geophysics, 39, 154–177. Willis, M. E., R. Lu, X. Campman, M. N. Toksöz, Y. Zhang, and M. V. de
Tonegawa, T., K. Nishida, T. Watanabe, and K. Shiomi, 2009, Seismic inter- Hoop, 2006, A novel application of time-reversed acoustics: Salt-dome
ferometry of teleseismic S-wave coda for retrieval of body waves: An ap- flank imaging using walkaway VSP surveys: Geophysics, 71, no. 2, A7–
plication to the Philippine Sea slab underneath the Japanese Islands: Geo- A11.
physical Journal International, 178, 1574–1586. Xiao, X., M. Zhou, and G. T. Schuster, 2006, Salt-flank delineation by inter-
Torii, K., T. Matsuoka, K. Onishi, K. Shirashi, T. Aizawa, Y. Yamanaka, S. ferometric imaging of transmitted P- and S-waves: Geophysics, 71, no. 4,
Ito, T. Kimura, Y. Asano, and T. Takeda, 2007, Application of seismic in- SI197–SI207.
terferometry to natural earthquakes measured by small-scale array: 77th Xue, Y., S. Dong, and G. T. Schuster, 2009, Interferometric prediction and
Annual International Meeting, SEG, Expanded Abstracts, 1362–1366. subtraction of surface waves with a nonlinear local filter: Geophysics, 74,
Ursin, B., 1983, Review of elastic and electromagnetic wave propagation in no. 1, SI1–SI8.
horizontally layered media: Geophysics, 48, 1063–1081. Yao, H., X. Campman, M. V. de Hoop, and R. D. van der Hilst, 2009, Estima-
van Tiggelen, B. A., 2003, Green function retrieval and time reversal in a dis- tion of surface wave Green’s functions from correlation of direct waves,
ordered world: Physical Review Letters, 91, 243904-1–243904-4. coda waves, and ambient noise in SE Tibet: Physics of the Earth and Plane-
Wang, Y., Y. Luo, and G. T. Schuster, 2009, Interferometric interpolation of tary Interiors, 177, 1–11.
missing seismic data: Geophysics, 74, no. 3, SI37–SI45. Yao, H., R. D. van der Hilst, and M. V. de Hoop, 2006, Surface-wave array to-
Wapenaar, K., D. Draganov, J. Thorbecke, and J. Fokkema, 2002, Theory of mography in SE Tibet from ambient seismic noise and two-station analysis
acoustic daylight imaging revisited: 72nd Annual International Meeting, — I. Phase velocity maps: Geophysical Journal International, 166,
SEG, Expanded Abstracts, 2269–2272. 732–744.
Wapenaar, K., and J. Fokkema, 2006, Green’s function representations for Yao, H., C. Beghein, and R. D. van der Hilst, 2008, Surface-wave array to-
seismic interferometry: Geophysics, 71, no. 4, SI33–SI46. mography in SE Tibet from ambient seismic noise and two-station analysis
Wapenaar, C. P. A., and J. L. T. Grimbergen, 1996, Reciprocity theorems for — II. Crustal and upper-mantle structure: Geophysical Journal Interna-
one-way wave fields: Geophysical Journal International, 127, 169–177. tional, 173, 205–219.
Wapenaar, K., J. Thorbecke, and D. Draganov, 2004, Relations between re- Yokoi, T., and S. Margaryan, 2008, Consistency of the spatial autocorrelation
flection and transmission responses of three-dimensional inhomogeneous method with seismic interferometry and its consequence: Geophysical
media: Geophysical Journal International, 156, 179–194. Prospecting, 56, 435–451.
Weaver, R., B. Froment, and M. Campillo, 2009, On the correlation of non- Yu, J., and G. T. Schuster, 2006, Crosscorrelogram migration of inverse verti-
isotropically distributed ballistic scalar diffuse waves: Journal of the cal seismic profile data: Geophysics, 71, no. 1, S1–S11.

Wapenaar 2010

Uploaded by

Copyright:

Available Formats

Wapenaar 2010

Uploaded by

Document Information

Original Title

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

Wapenaar 2010

Uploaded by

Copyright:

Available Formats

GEOPHYSICS, VOL. 75, NO. 5 共SEPTEMBER-OCTOBER 2010兲; P. 75A195–75A209, 15 FIGS.

Tutorial on seismic interferometry:

ABSTRACT retrieval of seismic surface-wave responses from ambient noise

INTRODUCTION source positions. Passive seismic interferometry, on the other hand,

兵G共xB,xA,t兲 Ⳮ G共xB,xA,ⳮ t兲其 ⴱ Ss共t兲具u共xB,t兲 ⴱ u共xA,ⳮ t兲典

2D and 3D analysis of direct-wave

2D and 3D analysis of reflected-wave interferometry

兵R共xB,xA,t兲 Ⳮ R共xB,xA,ⳮ t兲其 ⴱ SN共t兲

We use equation 14 to retrieve the reflection response. Strictly CONCLUSIONS

mensional inversion of scattered teleseismic body waves. 1. Theory for

ics, 75, no. 2, SA15–SA25 Perton, M., F. J. Sánchez-Sesma, A. Rodríguez-Castellanos, M. Campillo,

Review E, 75, 036103-1–036103-14. Acoustical Society of America, 126, 1817–1826.

You might also like

Wapenaar 2010

Uploaded by

Copyright:

Available Formats

Wapenaar 2010

Uploaded by

Document Information

Original Title

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

Wapenaar 2010

Uploaded by

Copyright:

Available Formats

GEOPHYSICS, VOL. 75, NO. 5 共SEPTEMBER-OCTOBER 2010兲; P. 75A195–75A209, 15 FIGS.

Tutorial on seismic interferometry:

ABSTRACT retrieval of seismic surface-wave responses from ambient noise

INTRODUCTION source positions. Passive seismic interferometry, on the other hand,

兵G共xB,xA,t兲 Ⳮ G共xB,xA,ⳮ t兲其 ⴱ Ss共t兲 具u共xB,t兲 ⴱ u共xA,ⳮ t兲典

2D and 3D analysis of direct-wave

2D and 3D analysis of reflected-wave interferometry

兵R共xB,xA,t兲 Ⳮ R共xB,xA,ⳮ t兲其 ⴱ SN共t兲

We use equation 14 to retrieve the reflection response. Strictly CONCLUSIONS

mensional inversion of scattered teleseismic body waves. 1. Theory for

ics, 75, no. 2, SA15–SA25 Perton, M., F. J. Sánchez-Sesma, A. Rodríguez-Castellanos, M. Campillo,

Review E, 75, 036103-1–036103-14. Acoustical Society of America, 126, 1817–1826.

You might also like

兵G共xB,xA,t兲 Ⳮ G共xB,xA,ⳮ t兲其 ⴱ Ss共t兲具u共xB,t兲 ⴱ u共xA,ⳮ t兲典