See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/225938598

Semi-empirical equations for the systematic decrease in permeability with

depth in porous and fractured media

Article  in  Hydrogeology Journal · June 2010

DOI: 10.1007/s10040-010-0575-3

CITATIONS READS

49 348

3 authors:

Xiao-Wei Jiang Xu-Sheng Wang

China University of Geosciences (Beijing) China University of Geosciences (Beijing)
44 PUBLICATIONS   630 CITATIONS    101 PUBLICATIONS   1,073 CITATIONS   

SEE PROFILE SEE PROFILE

Li Wan
China University of Geosciences (Beijing)
74 PUBLICATIONS   987 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Basin-scale groundwater circulation View project

Selection of papers related to Regional Groundwater Flow View project

All content following this page was uploaded by Xu-Sheng Wang on 03 June 2014.

The user has requested enhancement of the downloaded file.

Semi-empirical equations for the systematic decrease in permeability
with depth in porous and fractured media

Xiao-Wei Jiang & Xu-Sheng Wang & Li Wan

Abstract Permeability loss with depth is a general trend fractured media (e.g., Snow 1968a; Louis 1974;
in geological media and plays an essential role in Carlsson et al. 1983; Zhao 1998; Jiang et al. 2009a).
subsurface fluid flow and solute transport. In the near For porous sedimentary rocks such as carbonate and
surface zone where groundwater movement is active, the sandstone reservoirs, the decrease in porosity with
decrease in permeability with depth is dominated by the depth has been discussed in company with the depth-
mechanical compaction of deformable media caused by dependent permeability (Amthor et al. 1994; Budd
the increase in lithostatic stress with depth. Instead of 2001; Ehrenberg and Nadeau 2005). In fact, a unique
using empirical equations from statistical analysis, by relationship between permeability and depth is hardly
considering the well-defined relationships among perme- observed in the field because permeability measure-
ability, porosity, fracture aperture and effective stress ments at any depth show a remarkable variation.
under lithostatic conditions, new semi-empirical equations However, a systematic trend of decrease in porosity,
for the systematic depth-dependent permeability are hydraulic conductivity or permeability with depth can be
derived, as well as the equations for the depth-dependent commonly identified. Moreover, it is found that these trends
porosity in a porous medium and the depth-dependent are generally nonlinear.
fracture aperture in a fractured medium. The existing The dependence of permeability on depth, which is a
empirical equations can be included in the new equations kind of large-scale permeability heterogeneity, plays an
as special cases under some simplification. These new essential role in subsurface flow (Saar and Manga 2004).
semi-empirical equations perform better than previous The pumping tests in chalk and limestone aquifers
equations to interpret the depth-dependent permeability of usually show an obvious decrease of yield with low-
the Pierre Shale (with a maximum depth of approximately ering of water level, which is dominated by variation
4,500 m) and the granite at Stripa, Sweden (with a of hydraulic conductivity with depth (Rushton and
maximum depth of about 2,500 m). Chan 1976). In the investigation of groundwater
discharge into rock tunnels, Zhang and Franklin
Keywords Permeability . Porosity . Porous media . (1993) found that a constant average hydraulic con-
Fractured rocks . Sweden ductivity was unrealistic and the permeability gradient
(the rate of decrease in permeability with depth) would
significantly affect the discharge patterns. The effect of
gradually decreasing hydraulic conductivity on water-
Introduction table depths for steady-state subsurface drainage is also
discussed by Gallichand (1994). In the study of topography
The decrease in permeability (as well as hydraulic driven groundwater flow, Marklund and Wörman (2007)
conductivity) with depth has been documented for both found that the depth-dependent permeability controls the
porous media (e.g., Neuzil 1986; Whittemore et al. distribution of vertical flux driven by fluctuation of the
1993; Hart and Hammon 2002; Wang et al. 2009) and landscape and would even affect the infiltration at ground
surface. In sedimentary basins, the change in permeability
with depth significantly affects the pore pressure distribu-
Received: 24 August 2009 / Accepted: 5 January 2010 tion, which is significantly different from the results by using
Published online: 16 February 2010 average permeability (Walder and Nur 1984; Bethke and
Corbit 1988; Belitz and Bredehoeft 1988; Rice 1992; Wong
© Springer-Verlag 2010
et al. 1997).
A quantitative description of the permeability-depth
X.-W. Jiang : X.-S. Wang ()) : L. Wan relationship is necessary when the permeability loss is
School of Water Resources & Environment, accounted for in numerical modeling of subsurface flow or
China University of Geosciences,
Xueyuan Load 29, Haidian District, Beijing, 100083, China in other involved research. Traditionally, empirical equations
e-mail: [email protected] based on statistical analysis of permeability measurements
Tel.: +86-10-82320552 are used to establish the permeability-depth correlation. In

Hydrogeology Journal (2010) 18: 839–850 DOI 10.1007/s10040-010-0575-3

840
particular, the (log-permeability)-depth relationship, which ability, porosity, fracture aperture and effective stress under
supposes that permeability decreases exponentially with lithostatic conditions.
depth, is the most frequently used model. The exponential
correlation between permeability and depth was initially
suggested by Louis (1974) and has been widely accepted The equations for permeability-depth in porous
(Zhang and Franklin 1993; Budd 2001; Marklund and media
Wörman 2007). This exponential equation for hydraulic
conductivity is applied in one of the additional programs for For porous media (including the intact rock in fractured
the hydrogeologic-unit flow (HUF) package in MOD- rock masses), the permeability is generally considered to
FLOW-2000 (Anderman and Hill 2003). For porous media, be a result of the microstructure, characterized by the
similar exponential correlation is also applied between porosity, pore geometry, particle size (specific surface
porosity and depth as Athy’s law (Athy 1930). The area) and other factors. These microstructure factors
advantage of a simple exponential function is its depend on the geological processes involved in forming
convenience for analysis. However, in some cases, the the porous media and the depth-dependent stress
simple exponential function fails to match field meas- condition. For the type of porous media formed under
urements and some non-exponential equations were similar geological processes to those that result in
employed. The (log-permeability)-(log-depth) relation- similar distribution of solid particles (in this presenta-
ship was proposed by Snow (1968b) to analyze tion, denoted as a porous medium), porosity can play
permeability measurements of fractured crystalline the role of a dominant factor in controlling the
rocks from several dam sites (with a maximum depth permeability, which is relatively sensitive to buried
of 120 m). Manning and Ingebritsen (1999) also used depth. Compaction and aquathermal pressuring are
the (log-permeability)-(log-depth) function to analyze considered to be two of the most important causes of
the permeability loss with depth using permeability data porosity loss with depth (Luo and Vasseur 1992;
of the continental crust inferred from geothermal data Domenico and Schwartz 1998). Although aquathermal
and progress of metamorphic reactions (with a max- pressuring would be the essential cause of porosity loss
imum depth of 30 km). Another empirical function in for deep buried sediments, mechanical compaction is
the form of log(log-permeability)-depth, which implies the primary factor controlling porosity for sediments
that log-permeability decreases exponentially with near the ground surface. It is reported by Schneider et
depth, was suggested by Belitz and Bredehoeft (1988) al. (1996) that the transition zone would be probably at
for sandstones in the Denver basin. a depth of a few hundred meters for carbonate
Numerous factors such as tectonic activities, difference sediments (Dunnington 1967; Beall and Fisher 1969)
in lithology, thermal or chemical cementation and others and around 1.5 km for sandstones (Fuchtbauer 1967;
would lead to permeability heterogeneity. It is difficult to Schmidt and McDonald 1979; Bloch et al. 1986).
handle the full effect of these geological processes in a Therefore, in the near-surface zone of the earth, where
single explicit equation. However, extracting a unique groundwater movement is most active, porosity-depth
permeability-depth relationship dominated by some well- correlation can be analyzed based on mechanical
defined processes as a background of permeability compaction due to the change of effective stress.
heterogeneity is possible and expected. It is widely In this section, a simplified model of depth-dependent
accepted that the gradual decrease in permeability (porosity) permeability for porous media is developed based on the
with depth can be described using the theory of hydro- inter-relationship among permeability, porosity and
mechanical coupling for both porous media (Neuzil 2003) mechanical properties under lithostatic stress. The objective
and fractured media (Rutqvist and Stephansson 2003). As a is to find an equation that accounts for the most important
result, hydro-mechanical coupling based models might be factors causing the systematic permeability-depth trend in a
able to interpret the permeability-depth correlation in semi- porous medium. This equation should adopt the fundamental
empirical and semi-theoretical approaches. For porous discoveries of relationships among permeability, porosity,
media, theoretical solutions of porosity-depth relationship effective stress and depth that can be described by functions
have been developed by using compaction models (Nagumo as follows:
1965; Ramm 1992; Connolly and Podladchikov 2000).
However, to the authors’ knowledge, few similar studies k ¼ Fk;f ðfÞ ð1Þ
have been done for permeability–depth relationship. Even if
a theoretical solution of porosity-depth relationship can be
0 
applied to produce a permeability-depth relationship given a f ¼ Ff;A m ð2Þ
certain relationship between porosity and permeability for
porous media, such an approach cannot satisfy fractured 0

media, whose permeability is highly dependent on the m ¼ FA;z ðzÞ ð3Þ

special behavior of discontinuities. In this presentation,
semi-empirical equations for the permeability-depth trend where F indicates a function for relationship between two
in porous media and fractured media are developed by factors for a porous media, k is the permeability
0
of porous
considering the well-defined relationships among perme- media, f is the porosity, σ is stress, m is the mean effective

Hydrogeology Journal (2010) 18: 839–850 DOI 10.1007/s10040-010-0575-3

 841
stress (positive for compaction), z is the depth. Then, the range from 4.275 to 8.537 for three ranges of permeability
depth-permeability equation is derived as follows: in sedimentary formations. In Ghabezloo et al. (2009), n is
found to be 11 in a low permeability creeping material.
dk dFk;f dFf;A dFA;z However, some theoretical models suggest that n=2
¼ ð4Þ
dz df d0m dz (Turcotte and Schubert 2001) or n=3 (Zhu et al. 1995;
Walsh and Brace 1984). For a given value of f0, it is
Widely accepted formulae for Eqs. (1)–(3) in the found that the Kozeny-Carman model is equivalent to a
literature are applied in this presentation to find a semi- case of the power-law model with a special n value. As
empirical solution for Eq. (4) which can interpret the shown in Fig. 1a, when f0 =0.4, the curve of k/k0 versus
general nonlinear systematic decrease of permeability with f/f0 derived by Eq. (5a) is almost the same as that given
depth. by Eq. (5b) while n=4. When f0 =0.2, this special n value
is 3.4 as shown in Fig. 1b. It implies that the Kozeny-
Carman model could be included by the power-law
model.
The permeability-porosity relationship
In this study, it is assumed that the parameters, S in
Numerous approaches have been proposed to relate the
Eq. (5a) and n in Eq. (5b), for the permeability-porosity
permeability to porosity in porous media. Civan (2000)
relationship are relatively stable in a porous medium. Thus
grouped these models into three categories: empirical
permeability change in the porous medium can be
equations, hydraulic tube models and network models.
described as a unique response of porosity change. From
Hydraulic tube models are a reasonable compromise
Eq. (5b)) one has
between the empirical and network models (Civan 2001).
Among hydraulic tube models, the Kozeny-Carman model   
dFk;f
has been extensively used in the literature because of its ¼ n k0 fn0 fn1 ¼ nk=f ð6Þ
simplicity (Bear 1972): df

1 f3 This is the first item prepared for Eq. (4). Note that
k ¼ Fk;f ðfÞ ¼ ð5aÞ the constant n value also implies that the ratio of effective
5S 2 ð1  fÞ2 to non-effective porosity (Bernabe et al. 2003; Ghabezloo
et al. 2009) is assumed changeless in this study. Although
where S is the specific surface area of solid grains. Civan Eq. (5b) is used in this paper, other models of perme-
(2001) improved the Kozeny-Carman model by introducing ability-porosity relationship can also be applied by similar
non-integral exponents in Eq. (5a) to account for the steps.
interactive process between pore fluid and porous media.
In some other classic models (e.g., Walsh and Brace 1984),
permeability is proportional to integer powers of pore a
geometry parameters, i.e., porosity, hydraulic radius, tor- 1
tuosity and/or specific surface area. According to these 0.8 Kozeny-Carman model
models, the permeability-porosity relationship can be
power-law model, n=3
simplified and described by the power-law (Bernabe et al. 0.6
power-law model, n=5
k/k0

2003):
0.4 power-law model, n=4
n
k ¼ Fk;f ðfÞ ¼ k0 ðf=f0 Þ ð5bÞ 0.2

where k0 is the reference permeability corresponding to the 0

0 0.2 0.4 0.6 0.8 1
reference porosity, f0, and n is a coefficient which is
dependent on the grain size and the stack pattern of particles. /
Usually, the initial permeability, k0, and initial porosity, f0,
are used for the reference permeability and reference b
1
porosity. It has been demonstrated that the exponent, n, is
also related to the ratio of effective to non-effective porosity 0.8 Kozeny-Carman model
(Bernabe et al. 2003; Ghabezloo et al. 2009). This non- power-law model, n=3
0.6 power-law model, n=4
k/k0

effective porosity is dominated by dead-end pores (Bear

1972) filled by immobile fluid. It indicates a threshold or 0.4 power-law model, n=3.5
critical porosity for percolation in a porous medium.
Among the two permeability-porosity formulas, the 0.2
power-law relationship seems more general with a 0
variable factor, n, for different types of materials. A 0 0.2 0.4 0.6 0.8 1
compilation of laboratory data by David et al. (1994) /
shows that the value of n ranges from 4.6 to 25 in Fig. 1 A comparison of the Kozeny-Carman model and the power
common geologic materials. In Lucia (1995), n is found to law model of porosity-dependent permeability: a f0 =0.4, b f0 =0.2

Hydrogeology Journal (2010) 18: 839–850 DOI 10.1007/s10040-010-0575-3

842
Pore compressibility where M is the ratio of horizontal stress to vertical stress
Compaction of porous materials, which leads to decrease and is assumed to be constant. In addition, change of the
of porosity, is controlled by the stress-strain behaviors. vertical effective stress with depth, h, is determined as
This compaction can be described by stress-dependent
porosity. There are many different approaches to relate 0

porosity change with stress. For example, Zimmerman et dz ¼ g e dz ð11Þ

al. (1986) defined four compressibilities for porous rock,
relating changes in the bulk and pore volumes to changes where g e is the effective unit weight which depends on
in the pore and confining pressures. A similar approach density of the porous media and its saturation. In this
had been employed by Wang et al. (2009) to relate presentation, the fully saturated situation and the equilib-
porosity with stress by defining bulk modulus of porous rium pore pressure are considered, thus g e can be
media and the matrix, respectively. calculated by
In this study, a different approach is used, i.e. the
modified Athy’s law, which had been proposed by Davis
g e ¼ ðs   w Þg ð1  fÞ ð12Þ
and Davis (1999). The equation is written as follows:

0 

0 where ρs is the density of solid grains, ρw is the density of
f ¼ Ff;A m ¼ fr þ ðf0  fr Þ exp am ð7Þ
water, and g is the acceleration of gravity. Both ρs and ρw
are assumed to be constant in this simplified model.
where f0 is the reference porosity for zero effective stress With Eqs. (10)––(12), change of the mean effective
(the initial porosity), fr is the residual porosity at high stress can be described as
stress, α is a coefficient. A more simple equation can be
written as: 0
dFA;z dm

0
 ¼ ¼ g g ð1  fÞ ð13Þ
f ¼ f0 exp am ð8Þ dz dz

where g g ¼ ð1 þ 2MÞðs  w Þg=3 is a constant. This is

where the residual porosity is ignored, as adopted in the third item prepared for Eq. (4).
Rubey and Hubbert (1959). Although Eq. (7) is empirical,
it has been widely accepted (Rutqvist et al. 2002; Neuzil
2003; Liu et al. 2009). The significant advantage of this
formula is inclusion of the residual porosity so that the The solution of depth-dependent permeability
“residual” permeability can be accounted for. Substituting Eqs. (6), (9) and (13) into Eq. (4), the
With the assumption of a constant α value for a porous equation for depth-dependency of permeability can be
medium, Eq. (7) leads to rewritten as:

dFf;A df
¼ 0 ¼ aðf  fr Þ ð9Þ dk ng g aðf  fr Þð1  fÞ
dm0
dm ¼ k ð14Þ
dz f
This is the second item prepared for Eq. (4).
To solve this equation, the function of depth-dependent
porosity, f(h), must be firstly derived. With Eq. (13),
Eq. (9) can be rewritten as:
Effective stress under lithostatic conditions
The action of effective stress on a porous medium is df
subject to the pressure of confining materials, porous fluid ¼ g g aðf  fr Þð1  fÞ ð15Þ
dz
pressure and tectonic deformation. Measurements of the
complete state of geostress have been conducted in many For constant values of g g and α, the solution of Eq. (15) is
rocks at different depths across the world (Hoek and given by
Brown 1980). It has been found that the lithostatic stress is
generally proportional to depth. ðf0  fr Þð1 þ fr Þ
In this simplified model, the relationship between the fðzÞ ¼ fr þ  
0
vertical effective stress, z , the horizontal effective stress, ðf0  fr Þ þ ð1 þ 2fr  f0 Þ exp g g að1 þ fr Þz
0 0 0
H ¼ x ¼ y , and the mean effective stress under a ð16Þ
lithostatic condition are calculated with equations as
follows: where f0 is the initial porosity, which can be considered as
the porosity on the ground surface.
Equation (16) is the simple semi-empirical solution of
0 0 0 1
0 0

depth-dependent porosity for a porous medium. One can
H ¼ M z ; m ¼ z þ 2H ð10Þ
3 easily obtain the semi-empirical solution of the depth-

Hydrogeology Journal (2010) 18: 839–850 DOI 10.1007/s10040-010-0575-3

 843
dependent permeability by substituting Eq. (16) into where G indicates a function for relationship between two
Eq. (5a) or (5b). Using Eq. (5a), one has factors for a fracture,
0
T is the transmissivity of a facture, b

3 is its aperture, n is the effective normal stress (positive
ð1f0 Þ2 ðf0 fr Þð1þfr Þ for compaction). Then, the transmissivity-depth equation
kðzÞ ¼ k0 f 3 fr þ ðf f Þþð1þ2f f Þ exp g að1þf Þz
0 0 r r 0 ½g r  for a fracture is derived as follows:

2 dT dGT;b dGb;A dGA;z
ðf0 fr Þþð1þ2fr f0 Þ exp½g g að1þfr Þz ¼ ð21Þ
ð1fr Þð1þ2fr f0 Þ exp½g g að1þfr Þz2fr ðf0 fr Þ dz db d0n dz

Widely accepted formulae for Eqs. (18)–(20) in the

ð17aÞ literature are adopted in this presentation to find a
theoretical closed-form solution of Eq. (21). Subsequently,
where k0 is the initial permeability, which can be deemed it is combined with considerations of the change in
as the permeability on the ground surface. Using Eq. (5b), fracture frequency with depth to give an approximation
one has of the average nonlinear depth-dependent permeability of
( )n a fractured rock.
fr ðf0  fr Þð1 þ fr Þ=f0
kðzÞ ¼ k0 þ  
f0 ðf0  fr Þ þ ð1 þ 2fr  f0 Þ exp g g að1 þ fr Þz
ð17bÞ
The transmissivity of a fracture
Fluid flow in a deformable rock facture can be described
Equation 17b is the same solution of Eq. 15. with the cubic law (Witherspoon et al. 1980) as follows:

g ðbr þ bd Þ3 gb3
q¼ J¼ J ð22Þ
The equation for permeability-depth in a fractured 12nf 12nf
medium
where q is the volumetric flow rate through a joint per unit
In fractured media, groundwater flow mainly takes place in width, bd is the apparent aperture, br is the residual
discontinuities. Moreover, the occurrence of discontinuities aperture, J is the hydraulic gradient, ν is the fluid viscosity
can be the single most important factor that governs the and f is a friction factor that accounts for the roughness of
mechanical properties as well as permeability of rock the joint surface, b=br +bd, is the effective hydraulic
masses. Unlike porous media, the permeability of fractured aperture. Transmissivity, T, of a fracture is generally
rock is controlled by the aperture of individual fractures and defined as
their combination. Transmissivity of a facture can be
calculated by the famous cubic law with its aperture under q gb3
a certain stress condition. Studies on the change of perme- T ¼ GT;b ðbÞ ¼ ¼ ð23Þ
J 12nf
ability in loading/unloading procedure for fractured rock
masses have attracted numerous researchers. Equations of Therefore
stress-dependent permeability are developed based on exper-
imental studies and theoretical models for hydro-mechanical
behavior of discontinuities (Rutqvist and Stephansson 2003). dGT;b dT gb2
They encourage the attempt of expressing the permeability- ¼ ¼ ð24Þ
db db 4nf
depth trend with a simple equation for fractured rock masses
(Snow 1968b; Oda 1986; Wei et al. 1995).
In this section, a new simplified model of depth-dependent This is the first item prepared for Eq. (21).
transmissivity for a facture is established and applied to give
a systematic permeability trend of fractured media. This
equation is developed on the fundamental discoveries of Discontinuity closure under stress
relationships among the transmissivity, the aperture, the With increasing normal stresses on the faces, the aperture
effective normal stress and the depth of a fracture that can be of a discontinuity would decrease. The apparent aperture
described by functions as follows: can be interpreted as the residual aperture, br, plus an
apparent mechanical opening (Witherspoon et al. 1980)
T ¼ GT;b ðbÞ ð18Þ
bd ¼ umax  un ð25Þ

0
b ¼ Gb;A n ð19Þ Where un and umax are the normal deformation of the
fracture and the maximum closure, respectively.
0 Empirical models have been developed to give a
n ¼ GA;z ðzÞ ð20Þ correlation between the closure of a discontinuity and

Hydrogeology Journal (2010) 18: 839–850 DOI 10.1007/s10040-010-0575-3

844
the effective normal stress. The hyperbolic function Ground surface
proposed by Bandis et al. (1983) is most widely applied as
0

un ¼  n0   ð26Þ
Kn0 þ n umax
H H
where Kn0 is the initial normal stiffness when un equals or n
is approximate to zero. Kn0 belongs to intrinsic mechan-
ical properties of a discontinuity and can be considered as
a constant. In this presentation, the initial condition is
specified to the situation in which bd =umax and un =0 with
zero normal stress on the ground surface. If one defines Fig. 2 Schematic representation of a facture and the stress
b0 = br + umax as the maximum effective aperture of condition
fracture at the initial condition, one has umax = b0 − br
and un = b0 − b. So Eq. (26) can be rewritten into where g h and g e are assumed to be constant for the fractured
media. The solution of Eq. (30) with the boundary condition

0 
0 0
on the ground surface, n ðz ¼ 0Þ ¼ 0, is
b ¼ Gb;A n ¼ b0   0 n  ð27Þ
Kn0 þ n ðb0  br Þ 0
n ¼ g h z ð31Þ
which gives the stress-dependent effective aperture. For
constant Kn0, there is And, there is
0
dGb;A db ðb  br Þ=ðb0  br Þ dGA;z dn
0 ¼ 0 ¼    ð28Þ ¼ ¼ gh ð32Þ
dn dn Kn0 þ 0n ðb0  br Þ dz dz

This is the second item prepared for Eq. (21). This is the last item prepared for Eq. (21).
It is necessary to mention that laboratory investigations
on single rock discontinuities had showed that both
normal closure and shear dilation can change facture
transmissivity (Tsang and Witherspoon 1981; Raven and The solution of depth-dependent transmissivity
Gale 1985; Olsson and Barton 2001). However, when it of a fracture
comes to a rock section that contains a multitude of Substituting Eqs. (24), (28), (31) and (32) into (21), one
discontinuities, the change in transmissivity is mainly has
determined by the normal closure on fractures caused by

normal stress (Barton et al. 1995; Min et al. 2004). 3g h 1  ðTr =T Þ1=3
dT
Accordingly, as a reasonable approximation, the influence ¼ T ð33Þ
of shear stress on closure of a fracture is currently ignored dz ðb0  br ÞKn0 þ g h z
in this model.
where Tr is the transmissivity when b=br. Equation (33)
indicates that the change of transmissivity with depth is
Effective normal stress under lithostatic conditions highly nonlinear.
Suppose that the fracture has a constant dip angle of θ, The solution of depth-dependent aperture can be
subjecting to the horizontal stress σ1 and the vertical stress derived by substituting Eq. (31) into Eq. (27) as follows
σ3 (Fig. 2), the normal stress on the joint can be calculated
by ghz
bðzÞ ¼ b0  ð34Þ
Kn0 þ ½g h z=ðb0  br Þ
1 þ 3 1  3
n ¼ þ cos 2 ð29Þ
2 2
Appling Eq. (34) in the cubic law, the solution of
transmissivity-depth can be obtained as
Under lithostatic conditions, applying Eqs. (10)–(12),
the relationship between the effective normal stress and 3
g zðb0  br Þ=b0
the depth can be described as T ðzÞ ¼ T0 1  h ð35Þ
ðb0  br ÞKn0 þ g h z
0  
dn 1þM 1M
¼ gh ¼  cos 2 g e ð30Þ where T0 is the transmissivity when b=b0. Equation (35)
dz 2 2 is also the solution of Eq. (34).

Hydrogeology Journal (2010) 18: 839–850 DOI 10.1007/s10040-010-0575-3

 845
To simplify the expression, defining a reference depth significant than the porosity loss with depth. For the
zc =(b0 −br)Kn0/g h, Eqs. (34) and (35) can be rewritten into fractured rock, depth-dependent fracture density should be
 accounted for.
1 ðbr =b0 Þðz=zc Þ
bðzÞ ¼ b0 þ ð36Þ
1 þ ðz=zc Þ 1 þ ðz=zc Þ
Discussion
3 Comparison of equations for porous media
1 ðbr =b0 Þðz=zc Þ
T ðzÞ ¼ T0 þ ð37Þ In Connolly and Podladchikov (2000), an equation of
1 þ ðz=zc Þ 1 þ ðz=zc Þ depth-dependent porosity is derived as follows
Eqs. (36) and (37) are the simple analytical equations of f0
the depth-dependent aperture and the depth-dependent fðzÞ ¼ ð41Þ
f0 þ ð1  f0 Þ expðbDgzÞ
transmissivity of a rock fracture, respectively.
where Δρg is defined similar to g g introduced in Eq. (13),
β is the coefficient of pore compressibility. It is the
The solution of depth-dependent permeability pseudoelastic compaction profile as a typical result of a
of a fractured rock mass complex model in dealing with temperature-dependent
In a hydraulic test of a fractured rock, e.g., packer viscoelastic compaction of sediments (Connolly and
injection test, the test segment generally penetrates a Podladchikov 2000). However, Eq. (41) is also closes to
group or several groups of joints. If the length of the the special case of Eq. (16) if the residual porosity is
segment is relatively short so an average depth can be ignored and β is equal to α. In this situation, the depth-
applied and the flow in the hydraulic test is dominated by dependent permeability of porous material can be sim-
a group of parallel joints with the same aperture, the plified to
permeability of the test segment, k, can be calculated by
k0
kðzÞ ¼   n ð42Þ
k¼
v NT v
¼ lT ð38Þ f0 þ ð1  f0 Þ exp g g az
g L g
Further simplification can be made by assuming a
where N is the number of joints across the segment and L constant effective unit weight, g e, in Eq. (12), i.e.,
is the length of the segment, v/g is applied to transform the g e ¼ ðs  w Þg. It leads to
hydraulic conductivity to the permeability where v is the
coefficient of kinematic viscosity, and l=N/L is the depth-
df
dependent density of the joints (fracture frequency). ¼ g g aðf  fr Þ ð43Þ
Assuming constant value of v/g and substituting Eq. (36) dz
into Eq. (38), the depth-dependent permeability can be
described by By solving Eq. (43), the depth-dependent porosity can be
obtained as

l 1 ðbr =b0 Þðz=zc Þ 3
kðzÞ ¼ k0 þ ð39Þ  
l0 1 þ ðz=zc Þ 1 þ ðz=zc Þ fðzÞ ¼ fr þ ðf0  fr Þ exp g g az ð44Þ

where k0 and l0 is the permeability and the density of It is similar to the modified Athy’s law as shown in
joints for the fractured rock on the ground surface, Eq. (7). If zero residual porosity is considered in Eq. (44),
respectively. the widely used empirical exponential function of depth-
According to Eq. 36, the fracture porosity of the rock dependent permeability can be derived as follows:
mass due to opening of the joints can be approximate by
  
l 1 ðbr =b0 Þðz=zc Þ kðzÞ ¼ k0 exp ng g az ð45Þ
ff ðzÞ / lb  ff 0 þ ð40Þ
l0 1 þ ðz=zc Þ 1 þ ðz=zc Þ
In Wang et al. (2009), the exponential function of
where ff0 is the fracture porosity on the ground surface. depth-dependent porosity had also been derived by
Equations. (39) and (40) give the general trends of assuming a constant effective unit weight and a zero
permeability loss and fracture porosity loss of a fractured residual porosity. However, the solution of the depth-
rock. They are quite different from the trends of perme- dependent permeability in Wang et al. (2009), which was
ability and porosity of a porous medium as described by obtained by the Kozeny-Carman model, is different from
Eqs. (16), 17a and 17b. However, a similar pattern can be either Eq. (42) or Eq. (45). Instead, if zero residual
seen, in that the permeability loss with depth is more porosity is assumed in Eq. (17a), which is also obtained

Hydrogeology Journal (2010) 18: 839–850 DOI 10.1007/s10040-010-0575-3

846
by the Kozeny-Carman model, then Eq. (17a) can be uncertainty can not diminish the advantage of the semi-
simplified as empirical equations in showing a similar pattern of
  systematic nonlinear decrease in permeability with depth
exp 2g g az as observed in the Pierre Shale.
kðzÞ ¼ k0   ð46Þ Comparatively, the permeability-depth trend estimated
f0 þ ð1  f0 Þ exp g g az
by using Eqs. (42) and (46), i.e., with zero residual
porosity, fails to approach a non-zero residual perme-
As a typical example, the permeability-depth trend of ability as represented by curve D in Fig. 3. It shows a too
the Pierre Shale is revisited. The data from Neuzil (1986) rapid decrease below the reference depth (450 m) defined
with a maximum depth of about 4,500 m is analyzed. The as the reciprocal of g gα(1 + fr). Curve E in Fig. 3
initial porosity of the shale approximates f0 =0.4 accord- calculated using Eq. (45) has even more serious problem.
ing to its deformation characteristics (Neuzil 1993). It is It approximately coincides with curve C only in the
difficult to determine an accurate n value for the shale if extremely shallow part, where the change in unit weight is
Eq. (17b) is applied. In this presentation, n=4 is chosen negligible.
following Fig. 1a to make an agreement between the
results of Eqs. (17a) and (17b). In this situation, the two
equations lead to the same result. The value of g gα(1 + fr) Comparison of equations for fractured media
is identified by a trial-and-error method to obtain a best The empirically exponential function in describing
match between the estimation and the observation data. As decrease in permeability with depth for fractured rocks,
shown in Fig. 3, the systematic decrease of permeability which is widely used in the literature (Louis 1974; Zhang
of the Pierre Shale is well characterized by curves A and and Franklin 1993), can be interpreted as a special case of
B with the initial permeability varying from 1.0×10−4 the new equation in this study. In Eq. (33), if the depth is
millidarcies (mD) to 2.0×10−2 mD. The confidence of significantly less than the reference depth and the residual
permeability at a depth between curves A and B is 92.9%. aperture is ignored, there is
An average trend is represented by curve C, which
indicates that the residual permeability is 1.3×10−6 mD d ln T 3 3
on average. Note that curves A, B and C have different ¼   ; z << zc ð47Þ
values of fr. There are uncertainties if both the value of dz zc þ z zc
g gα(1 + fr) and the value of n are unknown for the Pierre
Shale. It is found that for any specific n value between 1 It leads to
and 5, one can search a corresponding value of g gα(1 + fr)
to approximately satisfy the observation. The reason is T ðzÞ  T0 expð3z=zc Þ; z << zc ð48Þ
that the permeability-depth trend given by Eq. (17b) is a
coupling of the effects of n and g gα(1 + fr). However, this Consequently, the exponential equation is an approximation
for relatively shallow buried rock masses.
Permeability (mD)
In Wei et al. (1995), hyperbolic equations have been
7 6 5 4 3 2 1
proposed to describe the depth-dependent hydraulic
10 10 10 10 10 10 10 aperture and permeability for a fractured rock mass as
10 follows:
Test methods

Consolidation b z
¼1 ð49Þ
A þ Bz
Steady Flow
Transient Flow
b0
2
10
Depth (m)

3
k z
¼ 1 ð50Þ
103 k0 A þ Bz
Estimation Results

A B
where A and B are two constants. Oda (1986) derived
C D
similar equations with B being equal to 1. One can see the
E
104 relationship between Eqs. (48) and (36) if A and B are
Fig. 3 Estimation of depth-dependent permeability of the Pierre defined as
Shale (data from Neuzil 1986). The hydraulic conductivity is trans-
formed into the permeability with v/g=10−7 m·s according to a normal
zc 1
condition. Measured data were obtained from three types of tests A¼ ;B¼ ð51Þ
(Neuzil 1986). Three estimation results with Eq. (17b) are presented as 1  ðbr =b0 Þ 1  ðbr =b0 Þ
A: fr =0.112, k0 =1.0 × 10−4 mD, g gα(1 + fr)=0.0033 m−1; B: fr =
0.043, k0 =2.0 × 10−2 mD, g gα(1 + fr)=0.0014 m−1; C: fr =0.073, k0 =
1.3 × 10−3 mD, g gα(1 + fr)=0.0022 m−1. The next two estimation
With these definitions, Eq. (36) can be rewritten as
results with k0 =1.3 × 10−3 mD and g gα=0.0022 m−1 are presented as D Eq. (48). Wei et al. (1995) estimated the data from Snow
from Eq. (42) and E from Eq. (45), respectively (1968a) and A=58.0 m and B=1.02 were obtained. This B

Hydrogeology Journal (2010) 18: 839–850 DOI 10.1007/s10040-010-0575-3

 847
value indicates that br is about 2.0% of b0 and zc is found that br/b0 =0.08 satisfies the average permeability
56.86 m in the study area of Snow (1968a). Wei et al. data from Carlsson et al. (1983) best while zc is fixed at
(1995) declared that the two constants can be fixed and 56.86 m. As shown in Fig. 5, most of the data can be
applied to any rock masses. included between two curves of k0 =4 mD and k0 =25 mD.
In Eq. (39), which characterizes the depth-dependent For fractures in granite, the residual aperture obtained by
permeability for fractured rock masses, fracture density is Witherspoon et al. (1980) is 3.2–13.1 μm which is 3–12%
a pre-determined factor. It has been reported that disconti- of the initial aperture. Accordingly, br/b0 =0.02 seems too
nuity frequency would decrease with depth (Snow 1968b; small and br/b0 =0.08 is more confident. However, more
Carlsson et al. 1983; Martin and Christiansson 2009; Jiang work is needed in study of the in situ characteristics of
et al. 2009b). A high density of fractures near surface is fractures at Stripa to check the efficiency of the alternative
generally caused by the effect of weathering and unload- equations. Figure 5 also shows that in the shallow part
ing. In this study, the data of granite at Stripa, Sweden, are (less than 500 m), the curve by Wei et al. (1995) and the
selected for discussion. Decrease in fracture frequency curve generated using Eq. (39) coincides with each other.
with depth for the granite can be identified based on the
data from four boreholes at Stripa (Carlsson et al. 1983) as
shown in Fig. 4. However, a fractured zone exists 800– Remarks on limitations
900 m below surface and results in an abnormally high The new semi-empirical equations of the depth-dependent
fracture frequency. The decreasing trend of fracture permeability, Eqs. (17a) and (17b) for porous media, and
frequency without the disturbed segment at 700–1,000 m Eq. (39) for fractured media, are constrained by the
depth can be approximated by simplifications in setup of the models and limited by the
efficiency of the adopted formulas in describing the hydro-
l=l0 ¼ 0:07 þ 0:93 expð0:0028zÞ ð52Þ mechanical coupling processes. They are new approxima-
tions of the features that control the variation in permeability
where l0 is 5.6 m−1. This approximate trend is applied in with depth.
this study to analyze the decrease in permeability with Firstly, the equations give unique systematic depth-
depth at Stripa. Wei et al. (1995) analyzed this depth- dependent permeability due to hydro-mechanical coupling
dependent permeability by using Eq. (50) without consid- for a single porous medium and fractured medium,
ering the fracture frequency. As previously mentioned, B= without consideration of the change in lithology. They
1.02 indicates that br/b0 =0.02. However, by considering may well satisfy the measured permeability at a site where
the decrease of fracture frequency as described with deep buried sediments or rocks are similar to that near
Eq. (51), different results of the permeability trend can surface. However, different parameters are required if the
be obtained by using the new equation, Eq. (39). It is lithology of the medium significantly changes with depth.
Faults or fractured zones developed by tectonic move-
ments can also cause nonsystematic change of the
Fracture frequency (m 1) permeability for the same fractured rock mass.
0 2 4 6 8 10 The lithostatic stress and hydrostatic fluid pressure
0 conditions assumed in the models would probably be false
if abnormal tectonic stress or fluid pressure is significant.
Abnormal high or low fluid pressure can be induced by
200

Borehole data Permeability (mD)

400 _4 _3 _2 _1
10 10 10 10 1 101 102
Approximate trend
10
600 c = 56.86 m
Depth (m)

br /b0=0.08 k0 =4 mD
800
Fractured zone 102
Depth (m)

k0 =10 mD
1000
Carlsson et al.(1983)

1200 k0 =25 mD 103

Wei et al.(1995)
1400
Fig. 5 Change of permeability with depth at Stripa, Sweden (data
from Carlsson et al. 1983). The hydraulic conductivity is trans-
Fig. 4 The average fracture frequency versus depth at Stripa, formed into the permeability with v/g=10−7 m·s according to a
Sweden (data from Carlsson et al. 1983) normal condition

Hydrogeology Journal (2010) 18: 839–850 DOI 10.1007/s10040-010-0575-3

848
deposition or erosion in a sedimentary basin (Jiao and a significant change of tectonic and geothermal conditions
Zheng 1998). In this situation, additional trends of in this large scale of depth, it is hard to conclude that this
permeability change with depth over the systematic log-log trend can well represent the permeability-depth
permeability-depth relationship are needed. Temporal trend for a porous medium or a fractured rock in near
variations in total stress and fluid pressure can also occur surface conditions. In fact, the Manning and Ingebritsen
in a sedimentary basin that will lead to time-dependent (1999) equation fails to match the observations shown in
change in permeability with depth. It is not handled in the Figs. 3 and 5. However, one can see an approximate log-
new equations in the current paper. However, for a log trend in a certain depth range with the semi-empirical
relatively short period during a geological process, this equations reported here, as characterized by the linear part
time-dependent effect can be generally ignored. of the curves in Fig. 3 (depth range: 200–1,000 m) and
Equations (17a) and (17b) are based on the Kozeny- Fig. 5 (depth range: 100–1,000 m).
Carman model and power-law model, respectively, in
describing the permeability-porosity relationship. In deri-
vation of the equations, the specific surface area of the Conclusions
particles, S, and the exponent, n, are assumed to be
constants. It is possible that change of these items with Nonlinear decrease in permeability with depth is a general
depth can also influence the permeability-depth trend. trend in both porous and fractured media. In the near
Generally, the compressibility of the solid grains in a surface zone where groundwater movement is active, it is
porous medium is smaller than that of the framework. As dominated by the hydro-mechanical coupling of deformable
a result, change of specific surface area would not be a media with increasing stress under gravity loading. New
serious problem. For unconsolidated fine grained sediments, semi-empirical equations that express the systematic depth-
it is possible that the microstructure of the pore such as its dependent permeability of a porous/fractured medium are
connectivity, space tortuosity and the amount of dead-end presented in this study.
pores (non-effective porosity), will significantly depend on Instead of using empirical equations based on statistical
the stress condition and affect the depth-dependent perme- analysis, by considering the well-defined relationships
ability. However, few theoretical or empirical formulas have among permeability, porosity, and effective stress under
been proposed that account for this process. This is a topic the lithostatic condition, a semi-empirical model is
that deserves further investigation. proposed to describe the unique change in permeability
In particular, Eqs. (39) and (40) satisfy an individual with depth for a porous medium. A new equation of
fracture set with similar dip angle and mechanical depth-dependent permeability is derived as well as a new
properties. They are problematic while the reference equation of depth-dependent porosity. The empirical
depth, zc, and the relative residual aperture, br/b0, differ exponential function and the equation developed by
for different groups of fractures. However, if it is Connolly and Podladchikov (2000) can be considered as
necessary, the permeability tensor can be produced by a specially simplified cases of the new equation. In
combination of the individual fracture sets obeying interpreting the permeability trends of the Pierre Shale
Eq. (39). Identification of different parameters for the (Neuzil 1986), with maximum depth approximating
fracture sets is difficult with the normal hydraulic test 4,500 m, the new equation performs better than the
method in boreholes. In this situation, the estimated zc and previous equations.
br/b0, within a single equation, represent average behavior For fractured rock, a model of depth-dependent trans-
of the rock fractures, which dominate the flow in a missivity of a fracture is presented, which considers the
hydraulic test. In Eq. (30), g h is independent of the dip cubic law, the closure of the fracture aperture under
angle when M=1. This is generally approximated under the lithostatic condition. A widely accepted formula for
lithostatic conditions and encourages the use of Eq. (39) in the relationship between aperture and normal effective
describing the average permeability-depth trend of frac- stress is applied to handle the deformation behavior of the
tures with variable direction but similar mechanical fracture. A closed-form semi-analytical solution of the
properties. transmissivity-depth relationship for a fracture is derived
The semi-empirical equations are available only in the as well as the solution of depth-dependent aperture. The
surficial environment where permeability loss is domi- solution is combined with the consideration of fracture
nated by gravitational compaction and the effect of frequency to produce the equation of systematic depth-
aquathermal processes such as chemical cementation, is dependent permeability for a fractured medium. The
insignificant. In this presentation, the new permeability- empirical exponential function and the equation developed
depth trend equations show their efficiency for a porous by Wei et al. (1995) and Oda (1986) can be considered as
medium with depth less than 4,500 m (Fig. 3) and for a specially simplified cases of the new equation. Comparable
fractured medium with depth less than 2,500 m (Fig. 5). application results of the equations are presented in
Manning and Ingebritsen (1999) reports the permeability interpreting the permeability trends of the granite at Stripa,
of the continental crust with a maximum depth of 30 km Sweden, with a deepest depth of approximately 2,500 m.
in which the (log-permeability)-(log-depth) equation
(logk=−14–3.2logz) is applied. Due to complexity in rock Acknowledgements This research is financially supported by the
types and mixing of porous and fractured media as well as Yalongjiang River Joint Fund established by the National Natural

Hydrogeology Journal (2010) 18: 839–850 DOI 10.1007/s10040-010-0575-3

 849
Science Foundation of China and the Ertan Hydropower Develop- and porosity-permeability relationships. AAPG Bull 89:435–
ment Company (Grant No. 50639090). The research is also partially 445
supported by the Program of Outstanding Overseas Chinese Young Fuchtbauer H (1967) Influence of different types of diagenesis on
Scholars established by the National Natural Science Foundation of sandstone porosity. In: Proc. 7th World Petroleum Congr.,
China (Grant No. 40528003). We appreciate valuable comments Mexico City, April 1967, vol 2, pp 353–369
from Dr. D. Hart and two anonymous reviewers as well as constructive Gallichand J (1994) Numerical simulations of steady-state subsur-
suggestions from the editors, Dr. J. Bahr and Dr. J. Jiao. face drainage with vertically decreasing hydraulic conductivity.
Irrig Drain Syst 8:1–12
Ghabezloo S, Sulem J, Saint-Marc J (2009) Evaluation of a
permeability–porosity relationship in a low-permeability creeping
References material using a single transient test. Int J Rock Mech Min Sci
46:761–768
Hart DJ, Hammon III WS (2002) Measurement of hydraulic
Amthor JE, Mountjoy EW, Machel HG (1994) Regional-scale conductivity and specific storage using the shipboard Manheim
porosity and permeability variations in Upper Devonian Leduc squeezer. In: Salisbury MH, Shinohara M, Richter C, et al (eds)
buildups: implications for reservoir development and prediction Proceedings of the Ocean Drilling Program, 2002, Initial
in carbonates. AAPG Bull 78:1541–1559 Reports vol 195, chap 6, Ocean Drilling Program, College
Anderman ER, Hill MC (2003) MODFLOW-2000, the U.S. Station, TX
Geological Survey modular ground-water model—three addi- Hoek E, Brown ET (1980) Underground excavations in rock.
tions to the hydrogeologic-unit flow (HUF) package: alternative Institution of Mining and Metallurgy, London
storage for the uppermost active cells (SYTP parameter type), Jiang XW, Wan L, Wang XS, Liang SH, Bill H (2009a) Estimation
flows in hydrogeologic units, and the hydraulic-conductivity of fracture normal stiffness using a transmissivity-depth correlation.
depth-dependence (KDEP) capability. US Geol Surv Open-File Int J Rock Mech Min Sci 46:51–58
Rep 03-347 Jiang XW, Wan L, Wang XS, Wu X, Zhang X (2009b) Estimation
Athy LF (1930) Density, porosity and compaction of sedimentary of rock mass deformation modulus using variations in trans-
rocks. AAPG Bull 14:1–22 missivity and RQD with depth. Int J Rock Mech Min Sci 46
Bandis S, Lumsden AC, Barton NR (1983) Fundamentals of rock (8):1370–1377. doi:10.1016/j.ijrmms.2009.05.004
joint deformation. Int J Rock Mech Min Sci Geomech Abstr
20:249–268 Jiao JJ, Zheng C (1998) Abnormal fluid pressures caused by
Barton CA, Zoback MD, Moos D (1995) Fluid flow along deposition and erosion of sedimentary basins. J Hydrol
potentially active faults in crystalline rock. Geology 23:683–686 204:124–137
Beall AO, Fisher AG (1969) Sedimentology. Initial Rep Deep Sea Liu HH, Rutqvist J, Berryman JG (2009) On the relationship
Drilling Project 1:521–593 between stress and elastic strain for porous and fractured rock.
Bear J (1972) Dynamics of fluids in porous media. Elsevier, New York Int J Rock Mech Min Sci 46:289–296
Belitz K, Bredehoeft JD (1988) Hydrodynamics of the Denver Louis C (1974) Rock hydraulics. In: Muller L (ed) Rock mechanics.
basin: explanation of subnormal fluid pressure. AAPG Bull Springer, Vienna, pp 299–387
72:1334–1359 Lucia FJ (1995) Rock-fabric/petrophysical classification of carbonate
Bernabe Y, Mok U, Evans B (2003) Permeability-porosity relation- pore space for reservoir characterization. AAPG Bull 79(9):1275–
ships in rocks subjected to various evolution processes. Pure Appl 1300
Geophys 160:937–960 Luo X, Vasseur G (1992) Contributions of compaction and
Bethke CM, Corbit TF (1988) Linear and nonlinear solutions for aquathermal pressuring to geopressure and the influence of
one-dimensional compaction flow in sedimentary basins. Water environmental conditions. AAPG Bull 76:1550–1559
Resour Res 24:461–467 Manning CE, Ingebritsen SE (1999) Permeability of the continental
Bloch S, Suchecki RK, Duncan JR, Bjorlykke K (1986) Porosity crust implications of geothermal data and metamorphic systems.
prediction in quartz-rich sandstones: Middle Jurassic, Halten- Rev Geophys 37:127–150
banken area, offshore central Norway. AAPG Bull Abstr 70:567 Marklund L, Wörman A (2007) The impact of hydraulic con-
Budd DA (2001) Permeability loss with depth in the Cenozoic ductivity on topography driven groundwater flow. Publs Inst
carbonate platform of west-central Florida. AAPG Bull Geophys Pol Acad Sc E-7:159–167
85:1253–1272 Martin CD, Christiansson R (2009) Comparison of the minimum
Carlsson H, Carlsson L, Jamtlid A, Nordlander H, Olsson O, Olsson horizontal stress from overcoring and hydraulic fracturing at
T (1983) Cross-hole techniques in a deep seated rock mass. Bull Forsmark, Sweden. Proceedings of SINOROCK 2009, Hong-
Int Assoc Eng Geol 26–27:377–384 kong, May 2009
Civan F (2000) Reservoir formation damage: fundamentals, model- Min KB, Rutqvist J, Tsang C-F, Jing L (2004) Stress-dependent
ing, assessment, and mitigation. Gulf, Houston, TX permeability of fractured rock masses: a numerical study. Int J
Civan F (2001) Scale effect on porosity and permeability: kinetics, Rock Mech Min Sci 41:1191–1210
model, and correlation. AIChE J 47(2):271–287 Nagumo S (1965) Compaction of sedimentary rocks: a consid-
Connolly JAD, Podladchikov YY (2000) Temperature-dependent eration by the theory of porous media. Bull Earthq Res Inst
viscoelastic compaction and compartmentalization in sedimentary 43:339–348
basins. Tectonophysics 324:137–168 Neuzil CE (1986) Groundwater flow in low-permeability environ-
David C, Wong TF, Zhu W, Zhang J (1994) Laboratory measurement ments. Water Resour Res 22:1163–1195
of compaction-induced permeability change in porous rocks: Neuzil CE (1993) Low fluid pressure within the Pierre Shale: a
implications for the generation and maintenance of pore pressure transient response to erosion. Water Resour Res 29:2007–2020
excess in the crust. PAGEOPH 143:425–456 Neuzil CE (2003) Hydromechanical coupling in geologic processes.
Davis JP, Davis DK (1999) Stress-dependent permeability: charac- Hydrogeol J 11:41–83
terization and modeling. SPE paper No. 56813, Society of Oda M (1986) An equivalent continuum model for coupled stress
Petroleum Engineers, Richardson, TX and fluid flow analysis in jointed rock masses. Water Resour
Domenico PA, Schwartz FW (1998) Physical and chemical hydro- Res 22:1845–1856
geology, 2nd edn. Wiley, New York Olsson R, Barton N (2001) An improved model for hydro-
Dunnington HV (1967) Aspects of diagenesis and shape change in mechanical coupling during shearing of rock joints. Int J Rock
stylolitic limestone reservoirs. In: Proc. 7th World Petroleum Mech Min Sci 38:317–329
Congr., Mexico City, April 1967, vol 2, pp 339–352 Ramm M (1992) Porosity-depth trends in reservoir sandstones:
Ehrenberg SN, Nadeau PH (2005) Sandstone vs. carbonate theoretical models related to Jurassic sandstones offshore
petroleum reservoirs: a global perspective on porosity-depth Norway. Mar Petrol Geol 9:553–567

Hydrogeology Journal (2010) 18: 839–850 DOI 10.1007/s10040-010-0575-3

850
Raven KG, Gale JE (1985) Water flow in a natural rock fracture as a Tsang YW, Witherspoon PA (1981) Hydromechanical behavior of a
function of stress and sample size. Int J Rock Mech Min Sci deformable rock fracture subject to normal stress. J Geophys
Geomech 22:251–261 Res 86:9287–9298
Rice JR (1992) Fault stress states, pore pressure distributions and Turcotte DL, Schubert G (2001) Geodynamics, 2nd edn. Cambridge
the weakness of the San Andreas fault. In: Evans B, Wong T-F University Press, Cambridge, UK
(eds) Fault mechanics and transport properties of rocks. Walder J, Nur A (1984) Porosity reduction and crustal pore pressure
Academic Press. Ltd., San Diego, CA, pp 475–504 development. J Geophys Res 89:11539–11548
Rubey WW, Hubbert MK (1959) Role of fluid pressure in Walsh JB, Brace WF (1984) The effect of pressure on porosity and
mechanics of overthrust faulting: II. overthrust belt in geo- the transport properties of rock. J Geophys Res 89:9425–9431
synclinal area of western Wyoming in light of fluid-pressure Wang XS, Jiang XW, Wan L, Song G, Xia Q (2009) Evaluation of
hypothesis. Geol Soc Am Bull 70:167–206 depth-dependent porosity and bulk modulus of a shear using
Rushton KR, Chan YK (1976) Pumping test analysis when permeability–depth trends. Int J Rock Mech Mining Sci 46
parameters vary with depth. Ground Water 14:82–87 (7):1175−1181. doi:10.1016/j.ijrmms.2009.02.002
Rutqvist J, Stephansson O (2003) The role of hydromechanical Wei ZQ, Egger P, Descoeudres F (1995) Permeability predictions
coupling in fractured rock engineering. Hydrogeol J 11:7– for jointed rock masses. Int J Rock Mech Min Sci Geomech
40 Abstr 32:251–261
Rutqvist J, Wu Y-S, Tsang C-F, Bodvarsson G (2002) A modeling Whittemore DO, Macfarlane PA, Doveton, JH, Butler JAJ, Chu
approach for analysis of coupled multiphase fluid flow, heat TM, Bassler R, Smith M, Mitchell J, Wade A (1993) The
transfer, and deformation in fractured porous rock. Int J Rock Dakota aquifer Program annual report, FY92, KGS Open-File
Mech Min Sci 39:429–442 Rep 93–1, Kansas Geological Survey, Lawrence, KS
Saar MO, Manga M (2004) Depth dependence of permeability in Witherspoon PA, Wang JSY, Iwai K, Gale JE (1980) Validity of
the Oregon Cascades inferred from hydrogeologic, thermal, cubic law for fluid flow in a deformable rock fracture. Water
seismic, and magmatic modeling constraints. J Geophys Res Resour Res 16:1016–1024
109:B01204 Wong TF, Ko SC, Olgaard DL (1997) Generation and maintenance
Schmidt V, McDonald DA (1979) The role of secondary porosity of pore pressure excess in a dehydrating system 2. Theoretical
in the course of sandstone diagenesis. In: Scholle PA, analysis. J Geophys Res 102:841–852
Schluger PR (eds) Aspects of Diagenesis. SEPM Spec Publ Zhang L, Franklin JA (1993) Prediction of water flow into rock
26:175– 207 tunnels: an analytical solution assuming a hydraulic conductivity
Schneider F, Potdevin JL, Wolf S, Faille I (1996) Mechanical and gradient. Int J Rock Mech Min Sci Geomech Abstr 30:37–46
chemical compaction model for sedimentary basin simulators. Zhao J (1998) Rock mass hydraulic conductivity of the Bukit Timah
Tectonophysics 263:307–317 granite, Singapore. Eng Geol 50:211–216
Snow DT (1968a) Rock fracture spacings, openings, and porosities. Zhu W, David C, Wong TF (1995) Network modeling of
J Soil Mech Found Div ASCE 96:73–91 permeability evolution during cementation and hot isostatic
Snow DT (1968b) Hydraulic character of fractured metamorphic pressing. J Geophys Res 100:15451–15464
rocks of the Front Range and implications of the Rocky Zimmerman RW, Somerton WH, King MS (1986) Compressibility
Mountain arsenal well. Colo Sch Mines Q 63:167–199 of porous rocks. J Geophys Res 91:12765–12777

Hydrogeology Journal (2010) 18: 839–850 DOI 10.1007/s10040-010-0575-3

View publication stats