Seismic Forward Modeling

E. S. Krebes, Department of Geology and Geophysics, University of Calgary, Calgary

Introduction change very much over distances of the order of the domi-
nant wavelength. This rule of thumb also implies that
The technique of forward modeling in seismology begins “high frequency” is a relative term – ray theory could also
with the numerical solution of the equation of motion for be used with frequencies that would be considered “low”
seismic waves, or more specifically, the numerical compu- in certain situations, as long as the heterogeneity is so weak
tation of theoretical or synthetic seismograms, for a given that the rule of thumb applies. Under these conditions of
geological model of the subsurface. The idea is then to high frequency, the travel time T(x) of the wave from the
compare the synthetic seismic traces with real seismic data source to a point x = (x,y,z) in a heterogeneous isotropic
acquired in the field. If the two agree to within an accept- medium obeys the eikonal equation,
able level of accuracy, the given geological model can be
taken to be a reasonably accurate model of the subsurface.
If not, the geological model is altered, and new synthetic (1)
traces are computed and compared with the data. This
process continues iteratively until a satisfactory match is
obtained between the synthetics and the real data. The a non-linear partial differential equation, where υ = υ(x) is
forward modeling approach is, in a sense, the opposite of the seismic wave speed (for either a P or S wave) at the
the inverse modeling approach in which the parameters of point x. The eikonal equation (1) can be obtained by substi-
the geological model are computed from the acquired real tuting a trial solution for u into the equation of motion and
data. Both methods though ultimately have the same goal making the appropriate high-frequency approximations.
– the determination of the geological structure and
Much recent work has been done on computing travel
lithology of the subsurface.
times by solving the eikonal equation, especially using the
As mentioned above, in forward modeling, one attempts to finite difference method (see below), which is computa-
solve the equation of motion for seismic waves, which is tionally efficient (see, e.g., Vidale, 1990; van Trier and
nothing more than the mathematical expression of Symes, 1991; and Kim and Cook, 1999).
Newton’s second law of motion, i.e., force = mass x accelera-
To actually trace a ray through a medium, a set of “ray
tion, for material particles in a solid body set in motion by
equations” must generally be solved. Consider first a single
elastic waves. It is a second-order partial differential equa-
isotropic heterogeneous medium in which υ varies
tion for the vector displacement u experienced by a point
smoothly with x. For a given wave speed υ(x), the geomet-
or particle in a solid medium due to the passage of a wave.
rical ray path can be determined, i.e., the coordinates of any
Once u is known (after solving the equation of motion), the
and all points x on the ray path can be computed, by solving
displacements of geophones can be computed and
the following system of ordinary differential equations:
synthetic seismograms can be produced. There are a
number of computational methods, and variants thereof,
that are commonly used to compute the synthetic seismic (2)
traces, such as ray theory, the finite difference method, and
the reflectivity method. All the methods have their indi-
vidual advantages and disadvantages, and an in-depth
modeling effort for a given data set can involve the use of
several methods. In this article, I will discuss some of the
basic principles and concepts of some of these methods.
Many seismologists have contributed to the development
of these methods, and it is unfortunately not possible to
represent all of them in this article.

Ray theory
Ray theory can be used to compute seismic wave travel
times and amplitudes along ray paths in a heterogeneous
medium when the frequencies present in the wave are high
enough so that the “geometrical optics” approximation can
be used. As a rule of thumb, in order for ray theory to be Figure 1. A ray path in a vertically heterogeneous medium in which the
applicable, the medium parameters (e.g., the Lamé param- wave velocity υ varies with depth z as υ = exp(z2). The horizontal and
vertical directions are offset and depth, respectively. The ray starts at the
eters λ and µ, or the P and S wave speeds) should not origin with a take-off angle of 5.74º and has a travel time of 1.79 s.

Continued on Page 29

28 CSEG RECORDER April 2004

 Article Cont’d
Seismic Forward Modeling
Continued from Page 28

where s is the arc length along the ray path. q ≡ ∇ T is called the error). In some of the simpler cases however, mathematical
slowness vector. It points in the direction of travel of the wave, formulas can be developed which allow one to compute, rela-
i.e., it is tangential to the ray path at any point on the ray path. tively easily, the travel times and amplitudes for the rays at any
Note that the eikonal equation (1) can be written as q•q = q2 = offsets without actually “tracing” or “shooting” any rays. For
1/υ2. The quantity q = 1/υ is the reciprocal of the wave speed example, consider the case of a geometrical ray in a medium
(hence the term “slowness”). For convenience in numerically consisting of a stack of flat homogeneous isotropic horizontal
solving (2), ds is often replaced by υdT in (2), so that the more elastic layers, a model commonly used in land exploration seis-
useful quantity T (the travel time) is used in the equations, mology. Figure 2 shows an example of such a ray. If the source
rather than the arc length s. Figure 1, which was produced by is a spherically symmetric point source located on the surface,
numerically solving (2), shows an example of a ray path in a then the travel time and amplitude of the wave arriving at the
medium in which the velocity υ varies with depth z as υ = receiver can be computed from relatively simple algebraic
exp(z2). formulas as follows.

The standard ray tracing equations (2) can be modified and First, the desired source-receiver offset X is chosen. Then the
approximated to simplify and facilitate ray tracing in complex ray parameter p for the ray going from the source to the receiver
subsurface structures. For example, consider a geometrical ray R is computed by numerically solving the equation
going between two points in a certain region of the subsurface,
and suppose that the paths of the nearby adjacent rays in this (4)
region do not deviate much from the ray R. Such nearby rays are
called paraxial rays. In this case, first-order Taylor expansions can
be used to approximate the wave speed υ = υ(x) for the paraxial for p, where m is the number of ray segments, and hj and υj are
rays. Then, using a coordinate system centred on the ray results the layer thickness and wave speed, resp., for the jth ray
in a coupled system of linear differential equations for the segment (assumed to be known). p is simply the horizontal
components of slowness and the coordinates, which are gener- component of the slowness vector along the ray, i.e., p = qx .
ally easier to solve than (2). Such methods, and modifications Snell’s law states that p is constant along the ray path, i.e., p =
thereof, are generally called paraxial ray tracing methods or sin(θj)/υj , j = 1, …, m, where θj is the angle that the jth ray
dynamic ray tracing methods (see, e.g., âerven˘, 2001). segment makes with the vertical axis. Consequently, once p is
determined, the take-off angle θ1 of the ray can be computed if
Ray amplitudes can also be computed, in the high-frequency needed, as well as all the other angles θj . Equation (4) can be
approximation, by solving the so-called transport equation for solved easily using any root-finding method, e.g., the Newton-
the wave. For example, in the relatively simple case of an Raphson method. Once p is known, the travel time T of the ray
acoustic wave, the transport equation for the amplitude A(x) of can be computed from
the pressure wave at the point x is

(5)
(3)
The amplitude A(ω) of the wave at the receiver for a single
Once T is known (e.g., by solving the eikonal equation 1), A can frequency ω can be computed from
be computed, in principle.

Tracing a ray between known source and receiver points in a (6)

structurally complex geological model of the subsurface is
generally done by a ray shooting method, in which the take-off
angle of a ray (the angle at which the ray leaves the source where Y is the product of displacement reflection and trans-
point) is determined iteratively (essentially through trial and mission coefficients along the ray path, and L is the geometrical
spreading factor which gives the amplitude loss due to the
geometrical spreading of the wavefront, and i is the imaginary
unit. For a ray such as the one in Figure 2, L is given by

(7)

A complex exponential is used in (6) for mathematical conven-

ience – it is the real part that represents the physical sinusoidal
wave. Sometimes in practice, L is roughly estimated simply by
the total length of the ray path. However, this estimate can be
very inaccurate in general, differing from the correct L computed
from (7) by a large amount (e.g., a factor of 2 or 3 or more).
Figure 2. A typical geometrical ray of m segments in a layered medium.

Continued on Page 30

April 2004 CSEG RECORDER 29

Article Cont’d
Seismic Forward Modeling
Continued from Page 29

If the mathematical form of the source pulse is known, then the T from (5). The real part of T, i.e., Re(T), gives the actual travel
amplitude of the waveform (consisting of a superposition of time of the wave, and the imaginary part, Im(T), gives the
frequencies) at the receiver can also be computed. For example, absorption factor, resulting in an exponential amplitude decay
the displacement u of the receiver is given by an inverse Fourier due to absorption (in (6), eiωT becomes eiωT = e-ωIm(T) eiωRe(T) ). As p is
transform, i.e., complex, Y and L in (6) also become complex, as well as b in (8).
These results are then used in (8) to compute the waveform.
For super-critical offsets, care must be taken to correctly choose
(8)
the signs of the square roots (cj ) in the above formulas, to avoid
errors. For example, the waveform discrepancy at 1.5 km
shown in Figure 12 of Hearn and Krebes, 1990, is due to an erro-
where S(ω) is the frequency spectrum of the source pulse, and neous sign choice, and not due to anelastic effects.
b, the polarization vector, is a unit vector in the positive direc-
tion of displacement (defined by convention). For example, for One of the main drawbacks of the ray method is that it is not
an incoming P wave traveling in the xz plane, b = [sin(θm), 0, accurate near critical offsets. For instance, for the ray in Figure 2,
cos(θm)] = [pυm, 0, cm]. If the free surface effect is included, b if the incidence angle at the deepest reflection point were near
would be more complicated – it would have to include the free the critical angle for the P or S transmitted wave (at which the
surface reflection coefficients (the free surface effect takes into transmitted wave becomes evanescent, and propagates along the
account the fact that the displacement of the receiver is affected interface), then the ray amplitude computed from (6) and (8)
not only by the incoming wave, but also by the waves reflected above would not be accurate. Corrections can be made to ray
off the surface). Equation (8) is normally computed using a fast theory to improve the accuracy in these cases, but they often
Fourier transform algorithm. involve elaborate and cumbersome mathematical extensions,
and they are often not generally or easily applicable
The above method can then be used to compute synthetic seis- (e.g.,âerven˘ and Ravindra, 1971; Plumpton and Tindle, 1989;
mograms which can be compared with real data to obtain a Gallop, 1999; Thomson, 1990; Aki and Richards, 2002). Figure 3
geological model of the subsurface. For sub-critical offsets, this shows a set of seismic traces in which basic ray theory ampli-
method can also be easily extended to the case of absorbing tudes agree well with exact calculations for pre-critical offsets,
(dissipative, anelastic) layers – one simply goes through the and reasonably well for post-critical offsets, but not at all for
same calculations (4)-(8) but with complex-valued and near-critical offsets. Including a mathematical correction to the
frequency-dependent wave speeds υj (Hearn and Krebes, 1990). basic ray theory result in Figure 3 to include the head wave (the
For example, if the quality factors Qj (typically frequency-inde- refraction arrival) would have improved the accuracy in the
pendent) for each ray segment are known (Qj = ∞ for no absorp- post-critical zone, and the extent of the zone of inaccuracy
tion), then υj ≈ Vj (ω)[1 - i/(2Qj )], where Vj is the real wave surrounding the critical offset would also have been reduced –
speed. One then obtains, by solving (4), a complex-valued p, but not eliminated. Figure 4 shows a comparison of the ray
which one then uses to compute a complex-valued travel time

Figure 3. Synthetic SH seismograms for two anelastic isotropic half spaces sepa- Figure 4. Amplitude versus offset curves for the SH case for two elastic isotropic
rated by an interface, with the source and receiver 500 m above the interface. The half spaces separated by an interface, with the source and receiver 1000 m above
densities, shear wave speeds, and QS values in the upper and lower half spaces are the interface. The densities and shear wave speeds in the upper and lower half
1.5 and 2.0 g/cm3, 1.0 and 2.0 km/s, and 50 and 100, respectively. The dominant spaces are 2.0 and 2.0 g/cm3, and 1.0 and 2.0 km/s, respectively. The frequency is
frequency of the wavelet is 50 Hz. The critical offset is 0.577 km. The solid line 20 Hz. The graph shows the exact result computed by numerical evaluation of the
traces give the exact response (computed by the ω-k method of Abramovici et al., generalized reflection integral, and the approximate result computed with ray
1990) and the dashed line traces are computed by ray theory. The figure is taken theory.
from Le et al. (2004).
Continued on Page 31

30 CSEG RECORDER April 2004

 Focus Article Cont’d
Seismic Forward Modeling
Continued from Page 30

theory amplitude and the exact amplitude against offset. Again, Generalized Ray Theory
we see that ray theory amplitudes are quite accurate for small
offsets corresponding to sub-critical angles, are reasonably accu- The geometrical ray theory result in (6) and (8) above, an
rate for super-critcal offsets, but not at all accurate for near-crit- approximation to the exact wavefield, makes use of only one
ical offsets. For deep reflectors in horizontally layered media, the value of the ray parameter p, namely, the value (call it p0) which
reflection angles are generally small (well below critical), and ray satisfies (4), the equation for X. A more accurate result would
theory can safely be used to produce synthetic seismograms. involve an appropriate superposition of waves with different
However, for shallow reflectors with large velocity contrasts, ray parameter values.
reflection angles can be large for the larger offsets, in which case A simplistic way to see this is shown in Figure 5, which shows
ray theory could not be used for near-critical angles and offsets. a 4-segment geometrical ray, with ray parameter p0, repre-
Reflections near critical angles could also occur for deeper reflec- senting the reflected wave going from a source point at x = 0 to
tors if they are dipping. a receiver at a known offset x = X0. The layers are homoge-
Ray theory is also inaccurate near caustic zones, where rays neous. The plane wavefronts associated with the geometrical
from a single source converge or focus because of changing ray are also shown in Figure 5. In addition, another ray with ray
velocity gradients, for example. Focusing of wave energy obvi - parameter p and with a larger offset X(p) is shown. This ray
ously produces high amplitudes (e.g., a magnifying lens can does not arrive at the receiver point X0. Nevertheless, note that
focus sunlight enough to burn paper), but in the ray theory the plane wavefronts associated with this ray also strike the
approximation, the computed ray amplitudes are infinite at receiver. Consequently, in this sense, this ray also contributes to
caustics, because the geometrical spreading factor L → 0 there. the displacement at the receiver, and should therefore be
L is a measure of the spreading of the wavefronts, and so when included in the computation of the receiver response if greater
they converge rather than spread, L decreases (and vanishes at accuracy is desired. In fact, one can then surmise that superim-
the caustic in ray theory). posing the contributions of all such rays with different p values
might give the exact response, i.e., the full wavefield, or at least
In addition, as ray theory is a high frequency approximation, or a more accurate response, which is in fact the case.
low wavelength approximation, it cannot be applied accurately
to structures with thin layers, “thin” meaning thinner than a By analyzing this problem mathematically, with the intent of
dominant wavelength, roughly speaking. obtaining the exact response, one obtains a formula expressing
the displacement of the receiver, due to the reflection from the
Ray theory also does not produce the complete or full wave second interface in Figure 5, as an integral over p, with p going
field: ray-synthetic seismograms do not contain the waveforms from 0 to ∞ (the superposition of plane waves mentioned in the
for all the waves arriving at the receiver, but only those for the previous paragraph). In fact, this integral contains within it not
specific rays selected by the user. Of course, one may include as only the displacement due to the reflected wave but also that due
many rays as one wants in a typical computation, but one is to the head wave (the refraction arrival) from the second interface
often not sure in advance whether a particular ray will have a – the integral is said to give the generalized reflection response. In
high enough amplitude to make it worthy of inclusion Figure 5, the geometrical ray path and the ray path associated
(although some progress has been made in this regard in deter- with the head wave, taken together, can be called a generalized ray.
mining threshold amplitudes of rays – see, e.g., Hron, 1971, The basic geometrical ray formula (6) is actually the lowest-order
1972). approximation of the generalized reflection integral.

Ray theory can also be extended and applied to more complex Mathematical formulas for generalized ray responses can also
cases involving other types of sources, laterally heterogeneous
layers, interfaces of any dip and strike, curved interfaces, and
anisotropic media (e.g., Richards et al, 1991; Chapman, 1985,
âerven˘, 2001; Aki and Richards, 2002; P‰enãík et al., 1996).
However, the additional mathematical theory and numerical
computations involved are generally extensive, and the final
results still generally include the inaccuracies mentioned in the
preceding paragraphs.

Nevertheless, in spite of the drawbacks, basic ray theory is still

widely used to compute synthetic seismograms, because of the
relative simplicity of the simpler versions of the method, the
fast computational times, and the fact that the ray paths for all
the events on ray-synthetic seismograms can be identified from
the event travel times (because the rays comprising the
synthetic are chosen by the user). In particular, ray theory is
very useful for computing travel times, if not always signal
Figure 5. A two-layer-over-half-space model in which velocity increases with
amplitudes. However, for more accurate amplitude computa- depth. The geometrical ray and head wave between the source and receiver are
tions, and for synthetics which include all the possible waves shown, as well as a ray with parameter p arriving at the offset X(p). The figure
than can arrive at a receiver, one must resort to exact or nearly- represents a simple way of understanding the concept of a generalized ray.
exact full-wave methods, such as those discussed below. Continued on Page 32

April 2004 CSEG RECORDER 31

Article Cont’d
Seismic Forward Modeling
Continued from Page 31

include multiples and any other wave types than are associated were welded together at the interface. T is also sometimes called
with the ray. Typically, they are applied for high frequencies the stress vector because its components are the normal and
(which in this case means travel times much greater than the shear components of the stress tensor. The boundary conditions,
wave period, or source-receiver offsets much greater than the when expressed as mathematical equations and solved, yield
wavelength). They are appealing because they give the the formulas for the reflection and transmission coefficients. Y
complete (or nearly complete) wavefield at the receiver, and in (6) is a product of such coefficients.
have been used in the literature to calculate exact synthetic seis-
mograms. The fact that they are superpositions of plane waves For plane wave propagation in the xz plane (where x is offset
means that plane wave reflection and transmission coefficients and z is depth) of a vertically heterogeneous solid medium
can be used in the integrands. Note however that in computing (e.g., a stack of horizontal homogeneous layers, or a medium in
synthetic seismograms for a layered geological model, it is still which the medium parameters vary smoothly with depth), the
up to the user to decide which generalized rays to include in displacement u and the traction T can be computed anywhere
the computation. If certain generalized rays are left out of the in the medium using the propagator matrix. More specifically, for
calculation that should have been included, the synthetics will plane harmonic P-SV waves, u and T at any depth z can be
not include the corresponding events, which could hamper computed from given values of u and T at some depth z0 by
interpretation when the synthetics are compared with real data. using the following equation:
Also, a disadvantage of generalized ray theory is that the inte-
grals are notoriously difficult to evaluate – the integrands oscil-
late wildly. However, numerical methods have been developed (9)
to evaluate them as efficiently as possible (see, e.g., Aki and
Richards, 2002). The exact results in Figures 3 and 4 were
computed by evaluating generalized ray integrals. where u and T are 2x1 column vectors whose elements are the
vector components of displacement (ux and uz) and traction (Tx
Matrix Methods and Tz), respectively, and where P(z, z0) is the propagator
When computing synthetic seismograms for the stack of layers matrix, a 4x4 matrix containing the medium parameters, the
in Figure 2 using basic geometrical ray theory or generalized ray frequency, the ray parameter, and the vertical phase factors of
theory, one must select the particular rays to include in the the plane waves. The column vector f is called the displacement-
synthetics. For example, it is common to choose only the stress vector.
primary reflections (rays which reflect only once before For a stack of layers, f can be continued through the layers
returning to the receiver). However, in some cases, one may because f contains precisely the quantities that are continuous
wish to include multiples (rays which reflect more than once), across an interface. In other words, if we know f in the first layer,
such as the ray in Figure 2. Clearly, there are an infinite number we can compute f in the last layer (or any layer). Also, for a stack
of multiples in a stack of layers, but only a finite number of them of layers, the propagator matrix P(z, z0) is itself the product of a
can be included in a synthetic based on the ray methods sequence of matrices, known as layer matrices, and their inverses,
discussed above. The choice of which multiples to include which contain the parameters for the individual layers.
depends on the problem one is attempting to solve. Typically,
one would wish to include the multiples whose amplitudes are The propagator matrix method can be used to solve a number
above a specified threshold, and mathematical formulas have of different types of problems in seismic forward modeling in
been developed to facilitate the choice (Hron, 1971, 1972). Also, vertically heterogeneous media. For example, it can be used to
it is clear that the more rays (primaries or multiples) that are compute the reflection and transmission coefficients for a stack
included, the greater the computation time. of horizontal homogeneous layers (as opposed to the coeffi-
cients for a single interface). More precisely, suppose that a
However, it is in fact possible to include all the multiples (an downgoing plane wave of a given frequency is incident upon
infinite number) by using the so-called propagator matrix method. the top layer of a stack of such layers. The incident wave will be
This is in essence done by replacing an infinite series with a transmitted into the stack and be multiply reflected and trans-
finite expression, in the same sense that the infinite series 1 + x mitted inside all the layers. The upgoing wave emerging from
+ x2 + x3 + … is equivalent to the function 1/(1-x). the top of the stack into the incidence medium will be a super-
The basic ideas involved in the propagator matrix method are position of all the upgoing waves (produced by multiple reflec-
described in the following paragraphs. tion and transmission at all the interfaces of the stack) passing
upwards through the top of the stack. The relative amplitude of
The boundary conditions that seismic waves satisfy at each this upgoing wave can be thought of as the reflection coefficient
interface are that the x, y and z components of displacment u of the stack. It will be a function of the thicknesses, densities and
and traction T (the force per unit area acting across an interface wave speeds of all the layers. Similarly, the downgoing wave in
due to a wave) must be continuous across the interface, i.e., their the medium below the stack will be a superposition of all the
values in the layers above and below the interface must match downgoing waves (due to multiple reflection and tranmission
at the interface. These are called welded-contact boundary condi- in the interior of the stack) passing through the bottom of the
tions, and they express the fact that two layers move as a unit at stack. The relative amplitude of this downgoing wave can
the interface when disturbed by a wave – they move as if they be thought of as the transmission coefficient of the stack. It is

Continued on Page 33

32 CSEG RECORDER April 2004

 Article Cont’d
Seismic Forward Modeling
Continued from Page 32

important to note that the stack reflection and transmission As mentioned above, an integral over the ray parameter p can
responses contain the contributions of all the multiples (there also be performed to obtain the generalized reflection response,
are an infinite number of them). The stack coefficients are also and also an integral over the frequency ω to obtain the wave-
functions of frequency, whereas interface coefficients are not form at the receiver.
(except for dissipative media).
It can be shown that these many-layer stack formulas give the
Such stack reflection and transmission coefficients have been same result as the propagator matrix method. Although they
used in the so-called reflectivity method (e.g., Fuchs and Müller, are more complicated and involve more computation, they
1971) and its variants, which is a popular method for computing offer several advantages over the propagator matrix approach:
synthetic seismograms for a stack of horizontal layers for (a) the reflection and transmission processes of individual wave
forward modeling purposes. As an example of an application of modes (P and S) can be followed more easily, (b) P-S conver-
the reflectivity method, consider a geological model consisting sions can be included or left out (which is useful for deter-
of a single layer overlying a stack of many thin layers, and mining the effect of conversions on the synthetics, (c) useful
suppose one wants to model the reflection response of the thin- approximate formulas can be developed from the exact stack
layer stack due to a point source in the single upper layer. The formulas, e.g., approximate formulas for the case in which only
continuation equation (9) would be used to obtain the ampli- a single conversion is allowed, (d) if the infinite series is used in
tude of the wave incident upon the stack, the stack reflection the formulas (instead of the corresponding finite expression), it
response, and to continue it to a receiver on the surface. Then, an can be truncated to include only as many multiples as desired
integral over the ray parameter p is performed to obtain the (which is useful for determining the effect of multiples), (e) if
generalized reflection response, and also an integral over the the infinite series is used, the individual terms have a very
frequency ω to obtain the waveform at the r eceiver. In practice, simple dependence on frequency, allowing the formulas to be
the integral over p is often limited to sub-critical values, as the used with other frequency-based methods, and (f) the terms in
interest is often just in the small-angle body wave response. The the formulas are directly related to the physical processes
resulting waveform would contain all the multiples as well as occurring in the stack (the multiple reflections and transmis-
the primary reflections. sions), allowing more physical insight into the processes,
whereas the propagator matrix approach gives only the
The reflectivity method can also be combined with other combined cumulative effect. If however, only the total response
methods to enhance the output of the modeling process. For is required, the propagator matrix method is generally easier
example, it can be combined with asymptotic ray theory (Hron, and more straightforward to use.
1971, 1972) in which multiples with sufficiently large ampli-
tudes are ray-traced and included in the synthetics, to give the The matrix methods discussed above can also be used for
so-called ray-reflectivity method (Daley and Hron, 1982). One wavefield extrapolation (e.g., Bale and Margrave, 2003), which
advantage of this combination is that all the events on the is a part of most seismic wave equation migration processes,
synthetics can be identified (in terms of ray paths). and also other inversion methods which contain a forward
modeling stage.
Matrix methods are also used in a technique sometimes called
the recursive reflectivity method (e.g., Kennett, 2001). In this tech- The finite difference method
nique, mathematical formulas for the stack reflection and
transmission coefficients for a single layer are first developed Another popular seismic forward modeling method is the finite
by systematically adding up the contributions of all the indi- difference (FD) method, which is a numerical method for
vidual multiple reflections and transmissions in the layer to solving partial differential equations. It can be applied to the
obtain the total response. The resulting sum is an infinite series, seismic equation of motion to compute the displacement u at
which is replaced by a finite expression using the above- any point in the given geological model, e.g., at the surface (for
mentioned formula, 1 + x + x2 + x3 + … = 1/(1-x) , where x is a generating synthetic seismograms for comparison with
product of the internal reflection coefficients of the layer and a recorded data), or at some depth z (for doing wavefield extrap-
vertical phase factor. These one-layer stack formulas are then olation or downward continuation, a stage of wave equation
used to obtain the stack coefficient formulas for a two-layer migration). There are a number of variants of the method, and
stack, which are in turn used to obtain the formulas for a three- many small improvements have been made to the basic method
layer stack, etc. These formulas all have the same mathematical by many researchers.
form, regardless of the number of layers involved. Although
the stack coefficients for any size stack can be computed in this The main idea behind the FD method is to compute the wave-
way, it can be shown that a more efficient two-step recursive field u(x,y,z,t) at a discrete set of closely-spaced grid points (xl ,
procedure can be developed for computing them. This proce- ym , zn , tq ) , with l, m, n, q = 0, 1, 2, 3, 4, 5, …, by approximating
dure involves applying vertical phase factors to bring the stack the derivatives occurring in the equation of motion with finite
coefficients from the bottom of a given layer to the top of the difference formulas, and recursively solving the resulting
layer, then using the stack coefficient formulas to cross the difference equation.
interface into the bottom of the next layer above the given one. As an example of a finite difference formula for approximating
This process is continued recursively. a derivative, consider the basic definition of a derivative, which
can be found in any elementary calculus text, i.e.,

Continued on Page 34

April 2004 CSEG RECORDER 33

Article Cont’d
Seismic Forward Modeling
Continued from Page 33

(10)

This leads to the forward-difference approximation for dy/dx, given (16)

by which can be used to recursively compute u at the grid point (x,
t+ ∆t) if one knows the value of u at the grid points (x,t) and (x-
∆x,t). Given some starting values, i.e., initial conditions or initial
(11)
values, such as u(x,0) = f(x), where f(x) is an arbitrary function of
x (the waveform at t = 0) and given values of the grid steps ∆x
where ∆x is a small finite-sized interval or grid step in the x and ∆t, i.e, a value for g, one can then compute the values u(x,
direction (adjacent grid points in the x direction are separated ∆t) for all x, which are then in turn used to compute the values
by the amount ∆x). Similarly, the backward-difference approxima- u(x, 2∆t) for all x, etc. In other words, u can be computed at all
tion is given by times t by “marching forward in time” in this way.

It can be shown that if one chooses ∆x and ∆t so that g ≤ 1, then

(16) is a stable FD scheme, i.e., the errors resulting from making
(12)
FD approximations to the derivatives remain finite – they do
not grow without limit as the time-marching computations
and the central-difference approximation is given by progress. In fact, if one chooses g = 1, the resulting simple FD
scheme, u(x, t+∆t) = u(x-∆x,t), gives the exact solution to (14).
For g < 1 however, the scheme in (16) is first-order accurate in
(13) space (x) and time (t), because the errors in the forward- and
backward-difference approximations used for the derivatives
are of first order in ∆x and ∆t.
It can be easily shown, by applying the Taylor series expansion
to these formulas, that the error made in using the forward- and If one naively uses the forward-difference approximation (11)
backward-difference approximations in (11) and (12) is of order to approximate both ∂u/∂x and ∂u/∂t in (14), then one obtains
∆x (i.e., it is linear in the small quantity ∆x), but that the error an unstable FD scheme, in which the error grows without limit
for the central-difference approximation in (13) is of order (∆x)2 during the computations, regardless of the value of g.
(it is quadratic in ∆x). Consequently, the central-difference In general, for any 3D partial differential equation, such as the
formula in (13) is a more accurate approximation to dy/dx. elastic equation of motion, it is necessary to choose the correct
As an example of the application of the FD method, consider FD approximations to the derivatives appearing in the equa-
the 1D partial differential equation tion, and to choose appropriate values for the step sizes ∆x, ∆y,
∆z and ∆t, to ensure numerical stability. Often, a preliminary
mathematical analysis of the partial differential equation under
(14) study is useful, and sometimes necessary, to determine which
type of FD approximations to use.

where u = u(x,t). It is known as the one-way wave equation The 3D seismic equation of motion, a second-order partial
because it admits solutions representing waves travelling with differential equation in the displacement u, can be solved in the
constant speed c in the positive x direction but not the negative same way, although the resulting difference equation is much
x direction. In mathematical terms, the exact solution is u = f(x- more complicated than (16). Normally, the difference equation
ct), where f is an arbitrary function. Since the exact solution of is marched forward in time t, with u being computed at all
(14) is known, one would not of course need to use the FD values of x, y and z for each time level. Because the wavefield is
method to solve (14) in practice – we use it here only as a simple then known everywhere, it is possible to create movies of the
example of how to apply the FD method. wavefronts propagating through the material, which are
helpful in understanding seismic wave propagation.
There are many ways to obtain a FD approximation to this
equation. For example, one possibility is to use a backward- FD computations are extremely time-consuming. 3D FD
difference approximation for ∂u/∂x and a forward-difference computer programs for realistic geological models can take
approximation for ∂u/∂t, resulting in the equation many hours to run. Consequently, it is common practice to do
FD calculations in 2D instead, to reduce the computation time.
Also, various computational techniques have been developed
(15) to improve efficiency. Even in 2D however, computation times
are typically much longer than those of other methods, such as
ray theory. However, as computer technology advances, 3D
where ∆t is a small finite-sized interval or grid step in the t calculations should become more practical.
direction. Upon re-arrangement, one obtains

Continued on Page 35

34 CSEG RECORDER April 2004

 Article Cont’d
Seismic Forward Modeling
Continued from Page 34

Accuracy of the computed solution is another primary concern derivatives are still computed using a FD approximation in the
in FD computations. One way to generally improve the accu- time domain, but the space derivatives are computed in the
racy of the FD solution is to use smaller grid step sizes (while wavenumber domain. More specifically, to obtain a spatial
making sure that the stability condition is always satisfied, e.g., derivative, say ∂ux/∂x, at a given time level, a spatial Fourier
g ≤ 1 for (16) ). However, this increases the computation time. transform is performed on the sequence of wavefield values ux
In addition, if the grid step sizes are made too small, accuracy (known from a previous stage of the FD computation) along the
can decrease, due to the occurrence of significant roundoff x direction to obtain the wavenumber (kx ) spectrum of ux
error. Another way to generally improve accuracy, without (another sequence of numbers). Then this new sequence is
reducing the grid step sizes, is to use higher-order approxima- multiplied by ikx which, in the kx domain, is equivalent to
tions for the derivatives. As indicated above, the forward- and taking the derivative ∂ux /∂x (i is the imaginary unit). Finally, an
backward-difference approximations used in (11) and (12) lead inverse spatial Fourier transform is performed on this altered
to first-order accurate FD schemes, but the central-difference sequence to give the differentiated sequence of values ∂ux /∂x
approximation leads to second-order accurate FD schemes, and in the x domain. The same technique is applied to the other
is commonly used in constructing typical FD schemes. Higher- components and in the other 2 spatial directions. The results are
order approximation formulas, involving a series or combina- then used to compute all the wavefield values at the next time
tion of forward-, backward- or central-difference level. The method gives quite accurate results because the
approximations, can be used as well, but they result in more spatial derivatives are done in the wavenumber domain,
complicated and cumbersome FD schemes and in an increased avoiding the inaccuracies associated with FD approximations
computation time. to derivatives. Another advantage of this method is that it
requires considerably fewer grid points per wavelength than
An unwanted artifact of FD calculations is a degradation of the the standard FD schemes in the space-time domains, making it
accuracy of the solution caused by grid dispersion (e.g., Alford et attractive for 3D computations. With today’s computers,
al., 1974). This occurs if the spatial grid steps are chosen too running at 2-3 GHz or higher, medium-sized 3D model compu-
large, i.e., if the grid is too coarse. To avoid grid dispersion, a tations can be done in several hours or less.
typical rule of thumb is that the spatial grid steps should be no
larger than one-tenth the size of the shortest wavelength in the Another unwanted artifact of the FD method is the generation
wavefield, i.e., there should be at least 10 grid points per wave- of artificial reflections. These are reflections coming from the
length. If higher-order FD schemes are used (i.e., if higher-order edges of the numerical grid, rather than from the bona-fide
derivative approximations are used), this number can be geological interfaces in the model. For example, to compute a
reduced, e.g., 5 grid points per wavelength (e.g., Alford et al., value for ux at the time t+∆t for a node at the very edge of the
1974; Levander, 1988). However, this does not necessarily result grid, one generally needs to know the value of ux at the earlier
in a reduction in computation time, as the higher-order time t at a node just off the grid (many FD schemes work that
schemes are more complicated. way). If one naively sets the off-grid value equal to zero, then
strong artificial reflections from the grid edge are produced,
The main advantage of using FD methods in seismic modelling because by setting the off-grid displacement value equal to
is that they produce the full wavefield – all the different types zero, one is implicitly making the off-grid region into a rigid
of waves that exist (reflections, refractions, etc.) generally will medium, resulting in an elastic-rigid interface. The resulting
appear in the computed solution with correct amplitudes and impedance contrast produces the artificial reflections which
phases. One difficulty of the method is the question of how to propagate back into the grid, and appear in the synthetics.
insert a given type of source of waves. A straightforward way
to do this is to use an exact solution in the vicinity of the source, To suppress such artificial reflections, non-reflecting or absorbing
and then continue the solution with the FD method. Other boundary conditions are used at the grid edges. In the case of
ways involve modifications of this approach (e.g., Alterman non-reflecting boundary conditions, FD schemes which are
and Karal, 1968; Kelly et al., 1976). different from the one used in the interior of the grid (which is
based on the standard equation of motion) are used. These
A number of different variants of the FD method are in use, for alternate schemes are mathematically designed so that the arti -
example, the velocity-stress FD method (e.g., Madariaga, 1976; ficial reflections produced have low (ideally zero) amplitudes.
Virieux, 1986). In this method, the equation of motion, which is a Sometimes they involve solving one-way wave equations near
second-order partial differential equation in the particle- the grid edges – equations whose solutions are waves which
displacement u, is re-expressed as a coupled system of first-order can travel in one direction only, i.e., off the grid but not back on
partial differential equations in the components of the particle- to it. In the case of absorbing boundary conditions, one uses
velocity (∂u/∂t) and the traction T. First-order equations are schemes which absorb the energy of the artificial reflections.
generally easier to solve. In addition, the use of staggered grids, in Typically, the grid is enlarged by a small amount – a few layers
which the particle-velocity components and the traction compo- – and a modified FD scheme which includes the effects of phys-
nents are computed on different grids which are shifted or stag- ical absorption or dissipation is solved in the added grid zone.
gered relative to each other by half a spatial grid step in all If the FD scheme used in the interior of the grid already
directions, makes for efficient and accurate computation. accounts for seismic wave absorption (i.e., if it is based on an
Another popular variant of the FD method is the pseudo-spectral equation of motion containing absorption terms – see, e.g.,
method (Kosloff and Baysal, 1982). In this method, the time Emmerich and Korn, 1987; Krebes and Quiroga-Goode, 1994),

Continued on Page 36

April 2004 CSEG RECORDER 35

Article Cont’d
Seismic Forward Modeling
Continued from Page 35

then it is a simple matter to suppress artificial reflections – one These figures illustrate how various physical effects in seismic
simply makes Q very small near the grid edges. Q is a measure wave propagation can be investigated, because the FD method
of the elastic “quality” of the medium – the lower the Q, the produces the full wavefield.
more absorbing or dissipative the medium is (Q = ∞ for
a perfectly elastic medium in which waves experience no Figure 6 and 7 show 2D synthetic shot records for an explosive
absorption). As with all computational methods, non-reflecting acoustic pressure source buried at depths of 8 m and 18 m,
and absorbing boundary conditions have their limitations, but resp., in a horizontal homogeneous elastic layer overlying a
in general, they work quite well. rigid half-space. The synthetics show the expected events, i.e.,
the direct arrival, the dispersive surface wave and the reflec-
Figures 6-16 show examples of synthetic shot records and tion. They also show how the deeper source results in a
wavefronts computed with the standard 2D FD method. lowering of the resolution and a reduction in the surface wave
Figures 6 and 7 were produced by Peter Manning and Gary amplitude.
Margrave, and Figures 8-16 were produced by Louis Chabot, all
members of the CREWES project at the University of Calgary. Figures 8 and 9 again show 2D synthetic shot records for a
source consisting of a force directed vertically downward at a
point on the surface (e.g., a hammer blow). The non-symmetric
source generates both a P wave and an SV wave. The geological
model is a horizontal homogeneous elastic layer overlying a
homogeneous elastic half-space. The top of the elastic layer is a
free surface. The synthetics show the expected events, i.e., the
direct arrival, the reflected and converted waves, the head
waves (refraction arrivals) and the prominent dispersive
guided waves. They also show the amplitude and phase
(polarity) relationships among the wave types. In addition,
Figure 9, for the horizontal component of particle motion,
shows a polarity reversal in all events as one goes from the left
side of the record to the right side. This happens because the
waves arriving at geophones left and right of centre produce
geophone motions whose x components of displacement have
opposite signs. Some numerical artifacts can also be seen in
Figure 9 on the lower left and lower right edges of the grid.

Figures 10-16 show a series of “snapshots” of the wavefronts

Figure 6. A 2D synthetic shot record produced by the finite difference method for corresponding to the shot records in Figures 8 and 9, at the
a buried pressure source. The horizontal axis is the source-receiver offset in meters
and the vertical axis is the 2-way travel time in milliseconds. The geological model
is a flat elastic layer over a rigid half-space. The upper surface of the elastic layer
is a free surface (e.g., an air/rock interface). The source depth is 8 m. The figure
was produced by Peter Manning and Gary Margrave, CREWES project, U. of
Calgary.

Figure 8. A 2D synthetic shot record for the vertical component of geophone

motion produced by the finite difference method for a surface source. The hori-
zontal axis is the source-receiver offset in meters and the vertical axis is the 2-way
travel time in seconds. The model is 3 km x 1 km. The geological model is a flat
elastic layer with a free upper surface overlying an elastic half-space. The source
is a vertically directed point force located on the free surface. The density, P and
Figure 7. Same as Figure 6, except the source depth is 18 m. The figure was S wave velocities in the elastic layer are 2.1 g/cm3, 1000 m/s and 500 m/s, resp.,
produced by Peter Manning and Gary Margrave, CREWES project, U. of and in the half-space they are 2.3 g/cm3, 1850 m/s and 925 m/s. Figures 8-16 were
Calgary. produced by Louis Chabot.

Continued on Page 37

36 CSEG RECORDER April 2004

 Article Cont’d
Seismic Forward Modeling
Continued from Page 36

Figure 9. Same as Figure 8, except it is for the horizontal component of motion. Figure 11. Same as Figure 10, except the elapsed time is 500 ms.

Figure 10. Fixed-time plot, or "snapshot", of the wavefronts corresponding to Figure 12. Same as Figure 10, except the elapsed time is 600 ms.
the shot records and geological model of Figures 8 and 9, produced by the finite
difference method, at an elapsed time of 300 ms (the source is activated at time nature of the wavefield produced by a non-symmetric source,
t = 0). The horizontal component of particle motion is shown in the upper half
even for a simple one-layer model.
and the vertical component in the lower half. The locations of the free surface
(where the source is) and the interface between the layer and half-space can be
Regarding point (e) in the preceding paragraph, the variation of
inferred from the locations of the incident and reflected wavefronts on this and
the following figures. amplitude observed along the wavefronts is consistent with the
radiation pattern for a vertically directed point force (known
following fixed times: 300 ms, 500 ms, 600 ms, 800 ms, 1000 ms, from theoretical studies of seismic wave propagation). For
1200 ms, and 1400 ms. The upper half of each figure shows the example, this radiation pattern predicts that a downwards
horizontal component of motion and the lower half shows the hammer blow would result in mainly a large vertical (longitu-
vertical component. Several interesting wave propagation dinal) displacement of the medium at points vertically below
features can be seen in the figures – they show (a) that the the source point, with little or no horizontal (transverse)
“hammer blow” source produces both a downgoing P and SV displacement at these points (which one might also predict
wave; (b) that there is a downgoing head wave connecting the using physical intuition). This prediction is confirmed, for
downgoing P and SV wavefronts emerging from the source example, in the transmitted P wavefront in Figure 12 (the
point; (c) the upgoing Pand SV head waves (refraction arrivals) lowermost wavefront in each half, upper and lower, of the
produced by the transmitted P wavefront beyond the critical figure) – for the horizontal component, the centre of the trans-
angle (when it breaks away from the other wavefronts at the mitted P wavefront has zero amplitude, but for the vertical
interface); (d) the expected left-to-right polarity reversal in the component, it has a high amplitude.
horizontal component; (e) the variation of amplitude along
some of the reflected and transmitted wavefronts due to the A good way to study Figures 10-16, with the intent of under-
non-symmetric “hammer blow” source; and (f) the complex standing wave propagation, is to select a certain wave and
Continued on Page 38

April 2004 CSEG RECORDER 37

Article Cont’d
Seismic Forward Modeling
Continued from Page 37

follow what happens to it from one snapshot to the next (as

time progresses). For example, consider the first reflected P
wave – the vertical component. Beginning with Figure 10
(lower half), which shows the downgoing high-amplitude inci-
dent P wave that has not yet reached the interface, we move to
Figure 11 and note the beginnings of the first reflected P wave.
Since the incident P wave is reflecting off a medium (the half-
space) which has a larger acoustic impedance (density x Pwave
velocity), the vertical component of displacement experiences a
polarity reversal (the central peak of the reflected P wavelet is
white, not black as for the incident P wave). This is just like the
polarity reversal in displacement experienced, upon reflection,
by an incident wave on thin string connected to a thick rope.
We then follow the reflected P wave upwards in Figures 12 and
13, where it then reflects off the free surface. This free surface
reflection can be seen in Figure 14 – note that there is now no
polarity reversal in the vertical component of displacement
because the wave is reflecting off a medium with a lower
Figure 13. Same as Figure 10, except the elapsed time is 800 ms. acoustic impedance, i.e., the air layer (although there is a
polarity reversal in the pressure at the free surface – a compres-
sion reflects as a rarefaction, i.e., a dilatation, off a medium with
a lower acoustic impedance). We then follow the P wave
reflected from the free surface back down to the interface in
Figure 15, after which it reflects off the interface with another
polarity reversal (due again to the larger impedance in the half-
space), which can be seen in Figure 16. Note also the decreasing
amplitude of the wave.

FD forward modeling has also been done for solid media

containing interfaces that are not in welded contact with each
other (e.g., Slawinski and Krebes, 2002a, 2002b). Non-welded
contact means that the displacement u is no longer continuous
across an interface, although the traction T still is. Media
containing joints, fractures and faults may in some cases fall into
this category.

Figure 14. Same as Figure 10, except the elapsed time is 1000 ms.

Figure 15. Same as Figure 10, except the elapsed time is 1200 ms. Figure 16. Same as Figure 10, except the elapsed time is 1400 ms.

Continued on Page 39

38 CSEG RECORDER April 2004

 Article Cont’d
Seismic Forward Modeling
Continued from Page 38

Other methods Gallop, J.B., 1999. Asymptotic Descriptions of Seismic Waves. Ph.D. thesis, University
of Alberta.
Forward modeling has also been done with other methods. For Hearn, D. and Krebes, E.S., 1990. On computing ray-synthetic seismograms for
example, some success has been achieved with the finite element anelastic media using complex rays, Geophysics, v. 55, p. 422-432.

method. This method still involves computations on a grid or Hron, F., 1971. Criteria for selection of phases in synthetic seismograms for layered
media. Bull. Seism. Soc. Am., v. 61, p. 765-779.
mesh, but involves representing the components of the solution
Hron, F., 1972. Numerical methods of ray generation in multilayered media. Methods in
u(x,y,z,t) to the equation of motion as linear superpositions of Computational Physics, v. 12, ed. B.A. Bolt, Academic Press, New York.
appropriately selected basis functions. For example, see the arti- Kay, I., and Krebes, E.S., 1999. Applying finite element analysis to the memory variable
cles in Kelly and Marfurt (1990), as well as Moczo et al. (1997), formulation of wave propagation in anelastic media . Geophysics, v. 64, p. 300-307.
Kay and Krebes (1999), and the review article by Carcione et al. Kelly, K.R. and Marfurt, K.J. (eds.), 1990. Numerical Modeling of Seismic Wave
(2002). The finite element method gives reliable results, and it Propagation. Society of Exploration Geophysicists, Tulsa.
can be more flexible and more accurate than FD methods when Kelly, K.R., Ward, R.W., Treitel, S., and Alford, R.M., 1976. Synthetic seismograms: a
finite-difference approach. Geophysics, v.41, p. 2-27.
the geological interfaces in the model are irregular, but it is also
Kennett, B.L.N., 2001. The Seismic Wavefield. Volume 1: Introduction and Theoretical
more difficult to implement and apply than the FD method. A Development. Cambridge University Press, Cambridge, U.K.
variety of other methods have also been used for models with
Kim, S., and Cook, R., 1999. 3-D traveltime computation using second order ENO
irregular interfaces (see, e.g., P‰enãík et al., 1996). scheme. Geophysics, v. 64, p. 1867-1876.
Kosloff, D. and Baysal, E., 1982. Forward modelling by a Fourier method. Geophysics,
Lastly, seismic wave propagation in media containing small- v. 47, p. 1402-1412.
scale heterogeneities such as cracks, fractures and inclusions, Krebes, E.S. and Quiroga-Goode, 1994. A standard finite-difference scheme for the
which scatter the waves, can be usefully modelled by solving time-domain computation of anelastic wavefields. Geophysics, v. 59, p. 290-296.
integral equations which describe their behaviour (e.g., see the Le, L.H.T., Krebes, E.S. and Quiroga-Goode, G.E., 1994. Synthetic seismograms for
references listed by Carcione et al., 2002). SH waves in anelastic transversely isotropic media. Geophys. J. Int., v. 116, p. 598-604.
Levander, A.R. 1988. Fourth-order finite-difference P-SV seismograms. Geophysics, v.
53, p. 1425-1436.
Acknowledgements Madariaga, R., 1976. Dynamics of an expanding circular fault. Bull. Seism. Soc. Am.,
v. 66, p. 639-666.
Thanks go to Louis Chabot, formerly of the CREWES project in
Moczo, P., Bistricky, E., and Bouchon, M., 1997. Hybrid Modeling of P-SV Seismic
the Department of Geology and Geophysics at the University of Motion at Inhomogeneous Viscoelastic Topographic Structures. Bull. Seism. Soc. Am.,
Calgary, and Peter Manning and Gary Margrave of the v. 87, p. 1305-1323.
CREWES project, for the finite difference synthetics. Plumpton, N.G. and Tindle, C.T., 1989. Saddle point analysis of the reflected acoustic
field, J. Acoust. Soc. Am., v. 85, p. 1115-1123.
P‰enãík, I., âerven˘, V., and Klimes, L., 1996. Seismic Waves in Laterally
References Inhomogeneous Media, Part I. Birkhäuser Verlag, Basel, Switzerland.
Abramovici, F., Le, L.H.T., and Kanasewich, E.R., 1990. The solution of Cagniard’s Richards, P.G., Witte, D.C. and Ekstrom, G., 1991. Generalized ray theory for seismic
problem for a SH line source in elastic and anelastic media, calculated using ω-k integrals. waves in structures with planar nonparallel interfaces. Bull. Seism. Soc. Am., v. 81, p
Bull. Seism. Soc. Am., v. 80, p. 1297-1310. 1309-1331.
Aki, K. and Richards, P.G., 2002. Quantitative Seismology. 2nd edition, University Slawinski, R., and Krebes, E.S., 2002a. Finite-difference modeling of SH-wave propa-
Science Books, Sausalito, California. gation in nonwelded contact media. Geophysics, v. 67, p. 1656-1663.
Alford, R.M., Kelly, K.R. and Boore, D.M., 1974. Accuracy of finite-difference Slawinski, R., and Krebes, E.S., 2002b. The homogeneous finite-difference formulation
modeling of the acoustic wave equation. Geophysics, v. 39, p. 834-842. of the P-SV-wave equation of motion. Stud. Geophys. Geod., v. 46, p. 731-751.
Alterman, Z.S. and Karal, F.C. Jr., 1968. Propagation of elastic waves in layered media
Thomson, C.J., 1990. Corrections for critical rays in 2-D seismic modelling. Geophys.
by finite difference methods. Bull. Seism. Soc. Am., v. 58, p. 367-398.
J. Int., v. 103, p. 171-210.
Bale, R.A. and Margrave, G., 2003. Elastic wavefield extrapolation in HTI media. Van Trier, J., and Symes, W.W., 1991. Upwind finite-difference calculations of travel-
CREWES Research Report, v. 15. times. Geophysics, v. 56, p. 812-821.
Carcione, J. M., Herman, G. C. and ten Kroode, A. P. E., 2002, Seismic modelling : Vidale, J., 1990. Finite-difference calculation of traveltimes in three dimensions.
Geophysics 67, 1304-1325. Geophysics, v. 55, p. 521-526.
âerven˘, V., 2001. Seismic Ray Theory. Cambridge University Press., Cambridge, Virieux, J., 1986. P-SV wave propagation in heterogeneous media: velocity-stress finite
UK. difference method. Geophysics, v. 51, p. 889-901.
âerven˘, V. and Ravindra, R., 1971. Theory of Seismic Head Waves. University of
Toronto Press, Toronto, Canada.
Chapman, C.H., 1985. Ray theory and its extensions : WKBJ and Maslov seismograms. Further Reading
J. Geophys., v. 58, p. 27-43. Carcione, J.M., 2001. Wave Fields in Real Media: Wave Propagation in Anisotropic,
Daley, P. and F. Hron, 1982. Ray-reflectivity method for SH-waves in stacks of thin and Anelastic and Porous Media. Pergamon Press, Amsterdam.
thick layers. Geophys. J. R. astr. Soc., v. 69, p. 527-535. Kelly, K.R. and Marfurt, K.J. (eds.), 1990. Numerical Modeling of Seismic Wave
Emmerich, H. and M. Korn, 1987. Incorporation of attenuation into time-domain Propagation. Society of Exploration Geophysicists, Tulsa.
computations of seismic wave fields. Geophysics, v. 52, p. 1252-1264. Slawinski, M.A., 2003. Seismic Waves and Rays in Elastic Media.
Fuchs, K. and Müller, G., 1971. Computation of synthetic seismograms with the Elsevier/Pergamon, Amsterdam. R
reflectivity method and comparison with observations. Geophys. J. R. astr. Soc., v. 23, p.
417-433.

April 2004 CSEG RECORDER 39