On coupling in the Biot model and its influence on the behavior of the

low-frequency P2 wave and distribution of the pore pressure: long wave

asymptotics of the Biot model

Inna Ya. Edelman

Russian Academy of Sciences, Institute of Physics of the Earth, Moscow, Russia

ABSTRACT: This work is devoted to the investigation of the bifurcation behavior of the Biot (P2) wave and
its influence on the pore pressure distribution in a fluid-saturated rock. The bifurcation occurs in a neighborhood
of the critical value of wave number kcr , which depends on the permeability of a medium and the viscosity of
a fluid. The P2 wave is fully attenuated if its wave number is smaller than kcr and it becomes propagatory with
wave numbers bigger than kcr . It is proven that fully attenuated long wavelength P2 waves influence significantly
the pore pressure state in a medium: the initial-boundary value problem to the Biot system of equations is solved
by asymptotic methods and the integral solution (it is an analogy of the Fourier integral) shows that the main
contribution to this integral is due to the P2 wave within the k-domain of diffusive behavior. However, pure
diffusion of the pore pressure does not occur because of the coupling of the Biot equations. The obtained results
are applied to the analysis and interpretation of the fluid injection induced microseismicity.

1 INTRODUCTION permeability of a medium and the viscosity of a fluid.

It is a well-known fact that pore pressure distribu-
tion in a porous fluid-saturated rock is coupled to the
stress state in the medium. M. Biot first derived the
equations for coupled poroelasticity (Biot, 1956; Biot,
1962). The most interesting phenomenon, predicted
by Biot, is the existence of a longitudinal wave of the
second kind (P2), the so-called Biot slow wave. In
the low frequency range the Biot theory assumes that
the slow wave is highly dispersive and strongly atten-
uated below some critical frequency, which depends
on the pore size in the skeleton and the viscosity of
the fluid. This critical frequency is typically around
110 kHz for water saturated porous materials of
around 1 Darcy permeability.
The coupling of the Biot system greatly magnifies
the difficulties of its mathematical analysis and, as a
consequence, there are a very limited number of ana-
lytical solutions for the Biot equations (Detournay&
Cheng, 1993). In this work the Biot system is studied
by asymptotic methods. The focus of this research is on
the investigation of the bifurcation behavior of the Biot
slow wave (see also Edelman, 2003) and its influence
on the pore pressure distribution in a fluid-saturated
rock. The dispersion relation for the P1 and P2 waves
is derived and investigated. It is proven that the P2
wave has a bifurcation behavior depending on its wave
number. The bifurcation occurs in a neighborhood of
the critical value kcr , which indeed depends on the
The P2 wave is fully attenuated if its wave number
is smaller than kcr and it becomes propagatory with
wave numbers bigger than kcr . The bifurcation point
describes the transition from low to high frequency
regimes of propagation of the P2 mode. The estimates
of the critical values for wave number in real rocks
show that the critical wave number is rather large and,
consequently, the Biot slow wave becomes propaga-
tory with a short enough wavelength. Thus, in the low
frequency range of interest in seismology (1100 Hz)
the P2 modes behave like diffusive waves.
Although the long wavelength P2 waves are not
propagatory and have a diffusive behavior, they influ-
ence significantly the pore pressure state in a medium.
Since Terzaghis pioneering works, earth scientists
have recognized that the pore pressure distribution
in a porous fluid-saturated rock plays an important
role in a wide variety of geophysical and geotechni-
cal phenomena. For example, it is a wellknown fact
that an injection of the borehole fluids into surround-
ing rocks often results in low-magnitude earthquakes
controlled by the changes in the pore pressure (e.g.,
Grasso, 1992), which, in turn, is influenced mainly
by the low-frequency Biot slow wave (Shapiro et al.,
1997; Shapiro et al., 1999; Shapiro et al., 2002;
Edelman & Shapiro, 2004). In order to analyze the
pore pressure evolution in a porous rock during the
injection of a borehole fluid the following problem is
considered: the Biot system of equations is formulated

173
Copyright 2005 Taylor & Francis Group plc, London, UK
for the radial symmetry case and is supplemented by where
the relevant boundary conditions at the borehole. The
latter are the given pore pressure and stress distribu-
tions at the borehole. To solve this problem is a very
complicated task because of the fact that the equa-
tions constituting the Biot model are coupled. The
solution is constructed analytically. It has the integral
form (it is an analogy of the Fourier integral) and, to
some degree, resembles the solution of the correspond-
For convenience, the equations (3), (4) can be rewritten
ing boundary value problem to the diffusion equation.
as follows:
The generalized diffusion coefficient is a compli-
cated combination of the parameters of the fluid and
solid phases and frequencies of the waves. It is proven
that the main contribution to this integral is due to the
P2 wave within the k-domain of diffusive behavior.
However, pure diffusion of the pore pressure does
not occur because of the coupling of the P1 and P2
waves.
The constructed solution can be applied to the anal-
ysis of the fluid injection induced micro earthquakes.
Injection of fluids under high pressure appears to have We intend to investigate the pore pressure and stress
induced active seismicity in hydrocarbon and geother- changes in a rock caused by the injection of a bore-
mal reservoirs. The fundamental mechanism behind hole liquid. Let us consider a radial-spherical flow.
fluid injection induced seismicity is the following: the We assume that at the initial time moment t = 0 the
seismicity occurs due to weakening of rocks (effective pressure at the borehole instantly reaches the value p0
normal stress decrease) brought on by a pore pressure and it is maintained constant up to the moment t = t0
change (e.g. Trifu, 2002). The constructed solution (the end of injection), i.e.
and temporal pattern of induced seismicity can be
applied to infer the local state pore pressure suffi-
cient to cause an earthquake and the hydromechanical
properties (permeability, porosity, and coupling coef- where r0 is the radius of a borehole and
ficient) of a fluid-filled rock. Obviously, an injection
of a liquid into the borehole leads to an increase of
the reservoir pressure that causes a growth of the pore
pressure in a fluid-filled rock. Pore pressure changes
may lead to the occurrence of the micro earthquakes We also assume that ij nj = p0 (t), i.e. tre = ( 1)
if the value of the pressure exceeds some threshold. p0 (t)/(0 + 2). After the standard substitution
The critical values of the pore pressure can be evalu- tre = E/r, pf /M = P/r one gets:
ated on the base of the obtained solution. This allows
one to describe the microseismic cloud (spatial-time
distribution of the microearthquakes) and to esti-
mate the threshold values of the pore pressure for a
real rock.

2 CONCEPTUAL MODEL
where
Consider the Biot system of equations for elastic
isotropic fluid-saturated matrix. The momentum bal-
ance equations for a fluid phase and for the mixed
fluid-solid system are (in the notations of Ref.(Biot,
1962)):

Our task is to solve the initial-boundary value prob-

lem (9), (10) and, thereby, to obtain the pore pressure
distribution in the reservoir adjacent to the borehole.

174
Copyright 2005 Taylor & Francis Group plc, London, UK
3 ANALYSIS
Applying the change
the mixed problem (9), (10) is transformed to the
Cauchy problem with a source:
After the Fourier transform w.r.t. , the problem (13),
(14) is rewritten as follows:
The solution to the problem (15), (16) has the following
general form:
where the frequencies j (k), j = 1 . . . 4, are defined
from the dispersion equation
the eigenvectors Rj (k) = (rj1 , rj2 )T correspond to these
frequencies, and the functions Cj (t, k) are defined by
the method of variation of parameters.
The dispersion equation (18) has four roots wj (k),
where k R1 is the wave number. They correspond to
the forward- and backward-directed P1 and P2 waves.
The construction of long wave asymptotic expansions
for the frequencies of P1 waves (i.e. the low frequency
range is considered) is straightforward:
175
Copyright 2005 Taylor & Francis Group plc, London, UK
However, the P2 wave construction of the asymp-
totic solution for the corresponding roots of (18) is
much more complicated. Similar to (4) one can prove
that the solution P2 (k) possesses a bifurcation in a
vicinity of some critical point kcr . Thus, the asymp-
totic expansion for P2 (k) has a different structure
depending on whether the wave number of the P2
wave is smaller or bigger than its critical value kcr . For
forward- and backward-directed P2 waves with wave
numbers |k| kcr one obtains, respectively:
Obviously, the expansions (20) and (21) consist of the
imaginary terms only. The latter means that the phase
velocity of the P2 wave is equal to zero, i.e. the wave
is not propagatory (it is fully attenuated) if its wave
number is smaller than kcr . Otherwise, the Biot wave
begins to emerge with velocity cP2 = 0 (see Fig. 1).
The asymptotic analysis shows that the long wave-
length P2 modes are not only strongly attenuated, as
the Biot theory assumes, but may not be propagatory
at all in the earth rocks. The critical wave number is
derived asymptotically
and, indeed, it depends on the permeability of a
medium and the viscosity of a fluid. The estimates
of the critical values for wave number and wavelength
allow one to suppose that in real geological materials
of low permeability the Biot slow wave does not propa-
gate at seismic frequencies. In Fig. 1 the phase velocity
of the P2 wave and the real part of its frequency are
evaluated for water-saturated sandstone (see example
in (Edelman, 2003)).
The expansions (19)(21) are required to explicitly
derive solution (17). One can prove that the solution
(17) yields
 CP2,m/s where
800

600

400

200

k,1/m
kcr 200 400 600

P2,1/s

300000

100000

k,1/m
kcr 200 400
Here
Figure 1. Phase velocity (upper figure) and real part of
frequency (lower figure) of the P2 wave: kcr 50 m1 .

where matrix T (t s, k) is a complicated combination

of medium parameters, frequencies of the P1 and P2
waves, and corresponding eigenvectors. Applying the
inverse Fourier transform to (23), one obtains: and the matrices T , T P2 B (t s, k) and TP1 (t s, k)

are defined from the construction of the solution. The

first term in (25) UP2F (t, r) reflects the contribution of

the forward-directed P2 wave into the solution (24).

The asymptotic and numerical analysis of the integral
terms in (25) shows that the main input to (24) is due
to integral (26), i.e. due to contribution of the P2 wave
within the interval of those wave numbers where it
The integral (24) is an exact solution of the mixed possesses the diffusive behavior, i.e.
problem (9), (10) that reflects the response of both
the adjoining porous saturated rock and pore fluid, i.e.
stress and pore pressure alternations, to the applied
liquid injection. It is evident the solution (24) is too
complicated for the practitioners. Let us try to sim- Thus, taking into account the boundary conditions
plify it. The integral (24) splits into several terms (10), one concludes:
accordingly to the input of the P2 and P1 waves and
accordingly to the k-domain of the diffusive behavior
of the Biot slow mode, namely:

176
Copyright 2005 Taylor & Francis Group plc, London, UK
Applying the method of stationary phase to the integral
(28), one gets:
where
and A 2 (k)
 is the second coordinate
 of the vec-
( 1)/(0 + 2)
tor A(k) . The formula (29)
1/M
describes the alternations of the pore pressure in the
adjoining borehole rock (t > 0, r > r0 ) caused by the
application at r = r0 loading (liquid injection). Analo-
gously, the first coordinate of the solution (28) gives
the description of strain changes. Consequently, the
stress state in a medium can be recovered from (51 ).
Before propounding a method on application of the
solution (29) to the description of space-time distribu-
tion of microseismic events, it is instructive to compare
the solution (29) with the analogous one for the dif-
fusion equation. Such a comparison is very important
in order to clarify the validity of the widely accepted
idea that, in the low frequency range, the Biot sys-
tem of equations is decoupled and the pore pressure
satisfies the diffusion equation (e.g., (Nur & Booker,
1972; Rice & Cleary, 1976; Norris, 1989)). In reality,
the Biot system can be decoupled only in the case of
irrotational displacement field (3). The latter is a rather
restrictive assumption which is obviously not valid in
most cases. Consider the following initial-boundary
value problem (radial symmetry case):
The solution to this problem has the form (Landau &
Lifshitz, 1987):
Close examination of the formulae (29) and (32)
indeed shows the strong resemblance of the solution
(29) to the solution (32). But then significant differ-
ences are also obvious, namely: in contrast to (32),
Copyright 2005 Taylor & Francis Group plc, London, UK
the coefficient in (29) depends on t and r and, more-
over, on frequencies of all waves existing in a porous
rock. The latter is due to the coupling of the equations
(3), (4) constituting the Biot system. Thus, although
the pore pressure and stress changes are basically con-
trolled by the low-frequency Biot slow wave (see (27)),
there is also a significant contribution from the other
modes. In addition, the asymptotic solution (29) allows
us to draw an inference that the pore pressure does not
satisfy the diffusion equation in spite of the fact that
the structure of the solution of the dispersion equa-
tion corresponding to the P2 wave is identical to that
of the dispersion relation of the diffusion equation. In
other words, even if the P2 mode behaves like a dif-
fusive wave within the interval |k| kcr , the diffusion
equation does not describe the evolution of the pore
pressure because of the fluid-solid coupling.
4 EXAMPLES
Next let us demonstrate how the constructed asymp-
totic solution (29) for the pore pressure can be applied
to the description of the spatial-temporal pattern of
the microseismic events (see also (Edelan & Shapiro,
2004)). Microseismic monitoring has been carried out
during the hydraulic stimulation of experimental reser-
voirs since the mid 1970-s. For example, in 1993 a
massive hydraulic fracturing at the Soultz-sous-Forets
geothermal site, located in the Alsace region of France,
yielded over 116,000 locatable seismic events (Jones,
1995) (see Fig. 2). Obviously, an injection of a liquid
into the borehole leads to an increase of the reser-
voir pressure, which causes an elevation of the pore
pressure above its initial value. Pore pressure growth
may lead to the occurrence of microearthquakes if the
value of the pressure exceeds some threshold. The crit-
ical values of the pore pressure can be evaluated on
the basis of the obtained solution. In Fig. 3 the lines
across the upper and lower portions of data are it ana-
lytically determined from (29) fits to maximum and
minimum distances between injection point and the
majority of the microearthquake locations vs. time.
These curves reflect the critical minimum pmin f and
maximum pmax f values of the pore pressure sufficient
to cause an earthquake. In other words, if a fluid injec-
tion results in such increase of the pore pressure in a
fluid-filled rock at the point (t, r) that the value of the
pore pressure pf (t, r) exceeds a threshold pmin f then a
microearthquake occurs at this point.
However, the amount of pore pressure increase is
not dependent on fluid injection rates alone. Pressure
changes are strongly dependent on the characteristics
of the reservoir into which fluid is being injected. One
of the most important properties is a medium perme-
ability. Because of an explicit dependence of pore pres-
sure (29) on hydraulic diffusivity 1 , which is directly
177
 r, m
800
600
400
200
100 300 t0 500
Figure 2. Distance from center of injection to
microearthquake location as a function of time from
the beginning of injection; t0 denotes the end of injection:
the Soultz-sous-Forets experiment (1993).
r,m
1000
800
600
400
200
t,h
100 200 300 400 500 600 700 800
Figure 3. Upper and lower envelopes of the microearth-
quake cloud constructed for the medium permeability
K = 8 1017 m2 , the Soultz-sous-Forets experiment (1993).
proportional to a medium permeability, the proposed
method represents the possibility of inverse procedure
to discern a permeability of intact and fractured rocks.
Moreover, the direct dependence of the solution (29)
on porosity, coupling coefficient, etc. makes it possible
to backtrack these rock characteristics.
5 CONCLUSIONS
The results presented in the paper concern the prop-
agation of the Biot slow wave in a fluid-saturated
porous medium at low frequencies and its influence
on pore pressure variation. It is proven analytically
that the P2 wave is fully attenuated if its wave num-
ber is lower than some critical value. This critical
wave number is a bifurcation point, above which the
longitudinal wave of the second kind becomes prop-
agatory. The asymptotic formula (22) demonstrates
that the P2 wave behavior is dominated by the perme-
ability of a medium. The theoretical results presented
Copyright 2005 Taylor & Francis Group plc, London, UK
in the paper might be helpful to experimentalists in
the interpretation of field data. The foregoing analysis
shows that the long wavelength P2s modes, generated
for instance as a result of explosions or earthquakes,
are not only strongly attenuated, as the Biot theory
assumes, but may not be propagatory at all in the earth
rocks. However, these modes influence significantly
the pore pressure state in a medium. The constructed
asymptotic solution (29) for the mixed problem to
the Biot coupled equations gives the distribution of
the pore pressure and shows that the main contribu-
tion to this solution is due to the P2 wave within the
t, h k-domain of diffusive behavior. The derived solution
650
has the integral form. It evokes the analogous inte-
gral solution for the initial-boundary value problem to
the diffusion equation. However, pure diffusion of
the pore pressure in a porous medium does not occur
because of the pressure-stress coupling in a porous
rock and, therefore, contribution of the other modes to
(29). The solution (29) allows us to draw an inference
that in the low frequency range the pore pressure does
not satisfy the diffusion equation and the Biot system
of equations is not decoupled. The obtained results
are applied for the description of the liquid injection
induced microseismicity.
REFERENCES
Biot, M.A. 1956. Theory of propagation of elastic
waves in a fluid saturated porous solids I. Low frequency
range & II. High frequency range, J . Acoust. Soc. Amer.
28: 168186.
Biot, M.A. 1962. Mechanics of deformation and acous-
tic propagation in porous media, J. Appl. Phys. 33(4):
14821498.
Detournay, E. & Cheng, A. H.-D. 1993. Fundamentals of
poroelasticity. In: J.A.Hudson (ed), Comprehensive rock
engineering: Principles, practice, and projects: 113171.
Oxford: Pergamon Press.
Edelman, I. 2003. Bifurcation of the Biot slow wave in a
porous medium, J. Acoust. Soc. Am. 14(1): 9097.
Grasso, J-R. 1992. Mechanics of Seismic Instabilities
Induced by the Recovery of Hydrocarbons. In Induced
seismicity: PAGEOPH 139. Basel: Birkhaeuser.
Shapiro, S.A. et al. 1997. Estimating the crust permeability
from fluid-injection-induced seismic emission at the KTB
site. Geophys. J. Internat. 131: F15F18.
Shapiro, S.A. et al. 1999. Large-scale in situ permeability
tensor of rocks from induced microseismicity. Geophys.
J. Internat. 137: 207213.
Shapiro, S.A. et al. 2002. Characterization of fluid transport
properties of reservoirs using induced microseismicity.
Geophysics. 67: 212220.
Edelman, I. & Shapiro, S.A. 2004. An analytical approach
to the description of fluid injection induced microseis-
micity in porous rock. Doklady Earth Sciences. 399(8):
11081112.
Trifu, C.I. (Ed.) 2002. The Mechanism of Induced Seismicity.
PAGEOPH 159(1-3). Basel: Birkhaeuser.
178
Nur, A. & Booker, J.T. 1972. Aftershocks caused by pore fluid
flow? Science. 175: 885887.
Rice, J.R. & Cleary, M.P. 1976. Some basic stress diffusion
solutions for fluid- saturated elastic porous media with
compressible constituents. Rev. Geophys. 14: 227241.
Norris, A.N. 1989. Stoneley-wave attenuation and dispersion
in permeable formations. Geophysics. 54(3): 330341.
Landau, L.D. & Lifshitz, E. 1987. Fluid Mechanics. Oxford:
Pergamon Press.
Jones, R.H. et al. 1995. Imaging induced microseismicity dur-
ing the 1993 injection tests at Soultzsous- Forets, France.
Proc. World Geotherm. Cong., Florence. 26652669.

179
Copyright 2005 Taylor & Francis Group plc, London, UK