GEOPHYSICS,

FORMATION

VELOCITY

BASICS

STRATIGRAPHIC

G.

H.

FOR

F.

GARDNER,*

AND

L.

W.

VOL.

39, NO.

6 (DECEMBER

DENSITY-THE

19741, P. 770-780,

3 TABLES

DIAGNOSTIC

TRAPS

GARDNER,1

A multiplicity of factors influence seismic reflection coefficients and the observed gravity of
typical sedimentary rocks. Rock velocity and
density depend upon the mineral composition and
the granular nature of the rock matrix, ccmentation, porosity, fluid content, and environmental
pressure. Depth of burial and geologic age also
have an effect.
Lithology and porosity can be related empirically to velocity by the time-average equation.
This equation is most reliable when the rock is
under substantial pressure, is saturated with
brine, and contains well-cemented grains. For
very low porosity rocks under large pressures, the
mineral composition can be related to velocity
by the theories of Voigt and Reuss.
One effect of pressure variation on velocity
results from the opening or closing of microcracks.
For porous sedimentary rocks, only the difference
between overburden and fluid pressure affects the

INTRODUCTION

The purpose of this paper is to set forth certain

relationships between rock physical properties,
rock composition, and environmental conditions
which have been established through extensive
laboratory and field experimentation together
with theoretical considerations. The literature on
the subject is vast. We are concerned primarily
with seismic P-wave velocity and density of
different types of sedimentary rocks in different
environments. These properties govern occur-

AND

A.

R.

GREGORY8

microcrack system. Existing theory does not take

into account the effect of microcrack closure on
the elastic behavior of rocks under pressure or
the chemical interaction between water and clay
particles.
The theory of Gassmann can be used to calculate the effect of different saturating fluids on the
P-wave velocity- of porous rocks. The effect may
be large enough in shallow, recent sediments to
permit gas sands to be distinguished from water
sands on seismic records. At depths greater than
about 6000 it, however, the reflection coefficient
becomes essentially independent of the nature of
the fluid.
Data show the systematic relationship between
velocity and density in sedimentary rocks. As a
result, reflection coefficients can often be estimated satisfactorily from velocity information
alone.

rences of seismic reflections and variations of

observed gravity. They thus have significant
bearing upon the manner of use of these geophysical methods and their effectiveness in finding or
delineating stratigraphic traps.
A stratigraphic trap connotes a porous and
permeable reservoir rock which alters laterally
on one or more sides into a nonpcrmcablc rock by
facics changes or a pinch-out. A particularly inrportant example is a reef surrounded by rock with
different properties. Reflection seismic represen-

Paper presented at the 38th Annual International

Annual International
1974.

9 FIGS.

SEG Meeting, October 3, 1968, Denver, Cola. and the 43rd

SEG Meeting, October 24, 1973, Mexico City. Manuscript recetved by the Editor January 30,

* Gulf Research & Development Co., Pittsburgh, Penn. 15230.

$. Retired, Austin, Tex. 78703; formerly Gulf Research & Development Co.
$ University of Texas, Austin, Tex. 78712; formerly Gulf Research & Development Co.
@ 1974 Society of Exploration Geophysicists. All rights reserved.
770

Formation

Velocity

tation of the convergence of seismic horizons

bracketing pinch-outs may comprise very useful
information. Also, relatively small lateral convergences in seismic transit times may reveal
differential compaction, and lateral variation in
transit time may indicate lateral change in the
physical characteristics of the bracketed rock.
Direct seismic reflection evidence, however,
typically involves lateral variation of interference patterns of reflections from the top and
bottom of a reservoir rock. Evidence furnished
by gravity observations is even more subtle but
can be applicable in some cases.
Many laboratory studies have been made to
show how P-wave velocity in rocks is affected by
pressure and fluid saturation; these studies
principally USCultrasonic techniques. When the
samples are cores recovered from wells, conditions that exist in the earth can be reproduced
with a fair degree of realism. As a result, a number of significant relationships for 1-wave
velocities of rocks under different conditions of
stress and fluid containment have been established.
A change in rock lithology or composition however cannot be simulated very satisfactorily in the
laboratory. Consequently, no rclevant relationships between different specific parameters are
well established. Commonly then, we must resort
to empirical correlations based on field data. Such
correlations generally entail some unknowns, so
they are satisfactorily applicable only for particular formations and environments.
An illustration of the wide range of P-wave
velocities and lesser range oi bulk densities for the
more prevalent sedimentary rock types through
a wide range of basins, geologic ages, and depths
(to 25,000 ft) is given in Figure 1.
An additional consideration is a general tendency for velocity and density to increase with
increase in depth of burial and with increase in
age of formations as verified by Faust (1953).
Successively deeper layers may differ materially
in composition and porosity with accompany,ing
marked local departures in velocity and density
from progressive increase with depth.
An understanding of interrelationships between
different rock properties and environmental conditions, however, requires recognition and consideration of the nature of rocks in general as
being granular with interconnected fluid-filled

and

Density

771

4.5 -

4.4 -

4.3 -

..?

;
0
:
>

..1-

E
z

1.0 -

:
;
;

3.9 -

_a

3.8 -

FIG. 1. Velocity-density relationships in rocks of

different lithology.
interstices. As a consequence, porosity, mineral
composition, intergranular elastic behavior, and
fluid properties are primary factors. These factors
are dependent upon overburden pressure, fluid
pressure, microcracks, age, and depth of burial.
ROCK

COMPOSITION

AND

GASSMANNS

THEORY

The three components which characterize the

composition of rocks and the superscripts which
indicate symbols of their properties arc:
(1)
frame
(2)
(3)

the solid matter of which the skeleton or

is built (index *);
the frame or skeleton (index-); and
the fluid filling of the pores (index I).

Properties of the whole rock are indicated by

symbols without superscripts.
In a classic paper, Gassmann (1951) showed
that when a rock with its fluid is a closed system,
grossly isotropic and homogeneous, the use of elementary elastic theory yields the following interrelationship between the rock parameters:

k = @ + @A~+ Q>,
where

Gardner,

772

Gardner,

Definitions of the symbols are shown below.

r!KmaNCLA~TTJRE

Quantity

Symbol
M
c1
k
V
G
F

Space modulus (or P-wave modulus)

Rigidity modulus (or S-wave modulus)
Bulk modulus (or reciprocal of the
compressibility)
Poissons ratio
Bulk density
Fractional porosity
Pressureon skeleton=Total external
pressure less the internal fluid pore
pressure= Net overburden pressure

Superscripts
Properties of the frame or skeleton
_
(empty porous rock)
Properties of the solid matter (grains
or crystals) of which the skeleton is
built
Properties of fluid occupying the
pore space of the skeleton
Parumeters of Time-Average Equation:
V
VF

Velocity of fluid-saturated rock

Velocity of fluid occupying the pore
space of the rock
l_+l Velocity of solid mineral of which
the matrix of rock is built

and

Gregory

Gassmanns theory was extended by Biot

(1956) to include the dynamic effects of relative
motion between the fluid and the frame and also
to take into account, viscoelastic effects in part
associated with the presence of microcracks.
These refinements are unimportant for waves at
seismic prospecting frequencies, and even at welllogging frequencies Gassmanns simple formula
often is adequate.
Neither Gassmanns nor Biots theory treats
the effect of microcrack closure on the elastic
behavior of rocks under pressure or the chemical
effects such as the interaction between water and
clay particles.
The parameters of the solid matter of the frame
that enter equations (l), (la), and (lb) are $ and
6. Some typical values are listed in Table 1; additional data are available in Clark (1966).
The parameters of typical fluids are also known,
and some are listed in Table 2.
Thus, characteristics of the solid matter and the
pore fluid of many rocks can be assigned numerical values without much difficulty
If values of k/ri are 0.5 or greater, with r$= 0.2
and with i/K= 18, the magnitude of the second
term on the right of equation (la) becomes about
.06 6 or less. This means that if the frame or skeleton has relatively high elastic constants, its
characteristics essentially govern the properties
of the whole rock, regardless of the fluid filling.
Those characteristics depend upon the elastic
interactions between the grains, their bonding,
and the presence of microcracks.
If, however, R and p are zero, p and ,V also
become zero, and equation (la) becomes:

He also noted that p =p and that I&= k+4//3 p.

White (1965) gives these relationships in the form:
M = 7i.Z+ (1 -

n/&/(+/K
(la)
+

(1 -

4)/b

K/L).

It is also known that:

P = (1 -

4M + 4P,

(lb)

and
P-wave

velocity

= yM/p.

(14

There~is generallagreement. among experimenters that equations (la) to (lc) satisfactorily

predict the effect of different saturating fluids on
P-wave velocity in most porous rocks, in spite of
the obvious simplifications of the theory.

Table 1. Bulk modulus and density of some minerals

,.
.
Density,
P
Bulk modulus,
k
dynes/cm 2 x 10 O gm/cm3
Solid
cc-Quartz

38

2.65

Calcite

67

2.71

Anhydrite

54

2 .~96

Dolomite

32

2.a7

294

3.99

Halite

23

2.16

Gypsum

40

2.32

Corundum

Formation

Velocity

Fluid

modulus,

dynes/cm2

Density,

1O1

773

Table 3. Upper and lower theoretical bounds for

velocity in aggregates of quartz crystals, with no
porosity and no microcracks

Table 2. Bulk modulus and density of some fluids

Bulk

and Density

gm/cm3

Volgtm
Water,

25C

2.402

1.025

2.752

1.0686

(1)

0.862

0.85

(2)

1.740

0.80

0.000142

0.001293

0.001325

0.007168

Sea

water,

25OC

(100,000

mg/L)

Crude

20,300

19,300

S-wave

it/set

13,900

12,900

eloclty,

EFFECT

0C

(dry,

76

cm

cm

Hg)

0 C

Hg)

OF MICROCRACKS

Some sedimentary rocks such as quartzites and

most igneous rocks have almost no porosity. For
these rocks the velocity is determined by the
mineral composition, provided an extensive system of microcracks is not present. In the absence
of microcracks, the elastic parameters can be
accurately estimated by use of the theories of
Voigt (1928) and Reuss (1929) and the known
elastic constants of the crystals. It has been
demonstrated by Hill (1952) that the Voigt and
Reuss methods give upper and lower bounds to
velocities for aggregates of crystals that are randomly oriented. A relevant example is provided
by quartz crystals, and the results are given in
Table 3.
The effect of microcracks can be illustrated by
the behavior of gabbros upon heating, as demonstrated by Ide (1937). Also, our experimental
data for a gabbro as depicted in Figure 2 shows
that the untreated, dried rock has a P-wave

oil

Methane,
(76

ft/sec

25 OC

Brine,

Air,

0.998

2.239

(distilled)

P-wavevelocity,

This is the relationship applicable for mixtures of

any two fluids. For clays and shales having very
high water content, k and p approach zero, and
the velocity of the formation approaches that of
the fluid. It is noteworthy that for clays and
shales the water is bound to the fine grained
microstructure so that water content is a more
pertinent term than porosity for use in describing
the rock structure. Again, the properties of the
rock frame are difficult to characterize. Permeabilities of clays and shales, of course, are very
much lower than those of reservoir sands.

20,00(
DRY
SAMPLE
BEFORE
TREATING
HEAT

z
0

(FEW

MICROCRACKS)

,
*
:
0

GABBRO

c
Y

FROM

BRAZIL

POROSITY

1.7 %

DENSITY

3.5

r 15,004
.

,*,/x-+:

/V

t
u
0

DRY

>:

-HEAT
(SEVERE

SAMPLE
TREATING

AFTER
Al
750C

MICR~CRACKING

INFERRED)

5
z
:
10,ooc

I
IO00

I
2000
AXIAL

I
3000
PRESSURE,

I
4000

I
5000

I
6000

PSI

FIG. 2. Effect of microcracks on velocity of gabbro.

774

Gardner,

Gardner,

and

Gregory

porosity and mineral composition aione determine

velocity. Under this condition, a time-average
relationship has been found empirically to interrelate velocity and rock parameters for a fairly
wide range of porosities:

The parameters 1~ and V,qrgive the dependence

on fluid velocity and mineral composition; 4 is
:
13000
the fractional porosity. Wide experience with
s
both in-situ determinations and laboratory exc
perimentation supports the general applicability
t;
a
of this relationship for most sedimentary rocks,
;
particularly when the fluid content is brine. This
9000 relationship is compatible with Gassmanns
SAND PACK
_
theory in recognizing that the elastic moduli of
9=.20
7000
the frame increase as porosity decreases.
The parameter L.%fis equal to the value of V
as C#Iapproaches zero. We have noted earlier that
1000 t-----i
0
!lOOO
4000
6000
1000
10000
the theories of Voigt and Reuss can be used to
PRESSURE, PSI
estimate the velocity for this extreme case. For
FIG. 3. P-wave velocity versus skeleton pressure
example, as given in Table 3, the velocity 1rAtffor
for brine-saturated carbonates, sandstone, and sand
aggregates of quartz crystals should lie between
pack.
19,300 and 20,300 ft/sec. Such a value has been
found
to be very satisfactory for deep clean sandvelocity of 18,700 ft/scc which increases to
stones. A value of 1,~f between 22,000 and 23,000
20,800 ft/sec with axial stress. In contrast, the
ft/scc similarly is found to be applicable for deep
treated rock (heated to 750C and cooled) has a
P-wave velocity of 11,000 ft/scc which increases limestones.
For formations at relatively shallow depths
with pressure but even at 6000 psi does not exceed
the influence of microcracks, and therefore pres16,000 ft/scc. Heating is presumed to introduce a
sure, cannot be ignored without introducing sigsystem of microcracks caused by the unequal
nificant errors. The time-average equation can be
expansion of the minerals. Pressure tends to
retained if the parameter l,+, is regarded as an
close the cracks, but a very large pressure would
empirical constant with a value less than the
be required to reestablish the original framework.
Voigt-Reuss values. In other words, we can asThe typical observation that nonporous rock
sume that the traveltime (which is the quantity
velocity increases with pressure thus is attributusually measured) is a linear function of porosity
able in substantial measure to the presence at low
at any depth with the coefficients of the linear
pressures of microcracks which are diminished
at higher pressures. A similar relationship also form to be chosenby consideration of suitable data.
4s a special illustration of the use of the timeholds true for the more porous rocks. Figure 3
average equation, in Figure 4 we have plotted
shows examples of velocity versus pressure for
some laboratory data for cows from a depth of
rocks spanning a range of porosity and mineral
about 5000 it. The cores were confined at a
composition. In the case of the packing of quartz
skeleton pressure of 3000 psi with brine in the
grains, the microcracks arc, presumably, the
pores to simulate the original environment. The
contacts between grains.
two principal minerals in the rock (calcite and
TIME-AVERAGE
RELATIONSHIP
quartz in the form of tripolitc) are intimately
At
pressures corresponding to those of decpcr
mixed in relative proportions varying from apsediments in situ, the influence of variation in
proximateI), SO percent calcite-50 percent quartz
pressure on velocity becomes small, and then
to 80 percent calcite-20 percent quartz. A petroI-

Formation

velocity

graphic analysis of the cores indicated that the

lower porosity samples (also from the upper part
of the formation) appeared to have a continuous
calcite matrix, whereas the higher porosity samples appeared to have a continuous quartz
(tripolite) matrix. It can bc seen that the lower
porosity data points can be approximated by a
time-average line with I,M= 22,500 ft/sec, which
is a velocity suitable for a calcite matrix; the
higher porosity data points can be approximated
by a time-average line with L,+f= 19,200 ft/sec,
which is a velocity suitable for a quartz matrix.
It is interesting that the data appear to separate
on these two lines according to the mineral that
is predominantly the continuous phase. In this
investigation we found no correlation between
velocity and the concentration by volume of the
minerals.
It is also of interest to note that when the
traveltimes at high pressure (10,000 psi) were
plotted against porosity, no separation of the
data along two lines could be detected. This may
indicate that velocity measurements on cores at
pressures appropriate to the depth of the formation contain more useful information than
measurements at an arbitrarily high pressure.
OVERBURDEN

AND

FLUID

PRESSURE

The overburden pressure is usually defined as

and

Density

775

the vertical stress caused by all the material, both

solid and fluid, above the formation. An average
value is 1.0 psi for each foot of depth, although
small departures from this average have been
noted. The fluid pressure is usually defined as the
pressure exerted by a column of free solution that
would be in equilibrium with the formation. The
reference to a free solution is signiiicant when
dealing with clays or shales with which other
pressures such as osmotic, swelling, etc., can be
associated. The normal fluid pressure gradient is
frequently assumed to be ,165 psi for each foot of
depth, although large departures from this value
occur in high-pressure shales.
The skeleton or frame pressure of a rock is the
total external pressure less the fluid pressure. The
elastic parameters of the skeleton increase as the
skeleton pressure increases, and a corresponding
increase in velocity is observed. The increase in
elastic parameters is attributable to the reactions
at the intergranular contacts and the closure of
microcracks as the skeleton pressure increases.
Hence, when both overburden pressure and formation fluid pressure are varied, only the difference between the two has a significant influence
on velocity.
A set of our velocity data that confirms this
assertion is given in Figure 5. When the skeleton
pressure P or the difference between overburden
and fluid pressure is increased, the velocity increases; when the difference between overburden
and fluid pressure remains constant, the velocity
remains constant.
RECENT

\o,

FIG. 4. Velocity versus porosity for samples

of quartz-calcite rocks under 3000 psi confining
pressure.

BASINS

In this section we consider a sedimentary basin

which illustrates how Y-wave velocity is affected
by many of the factors discussed above. Wells in
a number of young basins typically will penetrate
successiveLayersof sand and shale that may range
in age from recent to lower Eocene. This provides us an opportunity to study the effects of
age, pressure, depth, porosity, and fluid content
for a fairly constant matrix material.
The uppermost sedimentary layers are unconsolidated, and the porosity varies mainly with the
grain size distribution and clay content. The
velocity is only slightly greater than that of sea
water. With increasing depth the velocity increases partly because the pressure increases and
partly because cementation occurs at the grain-

776

Gardner,
I

WATER

1000

5.

Gregory
I

LESS

4000

SANDSTONE

SATURATED

EXTERNAL

FIG.

and
I

= SKELETON

PRESSURE

Gardner,

PRESSURE
INTERNAL

6000
PRESSURE,

= EXTERNAL
FLUID

TOTAL

PRESSURE

8000

10000

PSI

Velocity through a water-saturated sandstone core as a function

of skeleton pressure.

to-grainconfkxts. CemenfAtion is the more important factor. It has been shown by Maxwell
(1960) that the rate of cementation depends also
on the rate of flow and the composition of fluids
flowing through the pores as well as on temperzture.
The rapid increase of velocity with depth
normally continues until the time-average velocity is approached. Below this depth, the layers
behave like other well-consolidated rock and the
velocity depends mainly on porosity.
It is for the shallower layers that the fluid content, i.e., water, oil, or gas has an appreciable
effect.
The solid curve in Figure 6 shows a representative curve of velocity versus depth for brinesaturated, in-situ sands based on some sonic log
and electric log data. The dotted curve is from
laboratory data for fresh, unconsolidated, water
saturated packings of quartz sand grains at pressures corresponding to the depth. Thus, the dotted
curve indicates what would happen to sands if
they were buried without consolidation or cementation, and the divergence of these curves is
attributable to these effects. The dashed curve
shows the time-average velocity calculated using
the average porosity read from well logs. At the
shallower depths the actual velocity is less than
the Lime-average, but below about XUOOft the
agreement is close.

Figure 7 illustrates the results~fcr sandsin more

detail. For any depth the traveltime can be
approximated by a linear function of porosity;
below 8000 it this linear function coincides with
the time-average equation.
HIGH-PRESSURE

SHALES

In the wells referred to above, the fluid pressure

Moo0
0

1 I1
I I I I I I,
I
I
*ooo
6ooo
lclow
1.000
ILOCITI, FfII PEPSECOND

FIG. 6. Velocity as a function of depth showing

consolidation effect for in-situ tertiary sands. For
ccrmparison,~ the ve!ocities
of experimentd
smd
packs at pressurescorresponding to these depths are
also shown.

Formation

Velocity

increased at a rate of about 0.5 psi per foot of

depth. However, it sometimes happens that high
fluid pressure zones are encountered, i.e., zones
in which the fluid pressure is well above that
given by the normal gradient with depth. Such
wells provide an opportunity to study the relation
between velocity, the difference between overburden pressure and fluid pressure, and consolidation with depth.
Hottmann
and Johnson (1965) presented
pertinent data for velocity measured in shales
and the corresponding fluid pressure and depth.
When there was no excess fluid pressure, they
found that the interval traveltime AT, in microseconds per foot, decreased with depth Z, in feet,
according to the formula
Z = A -

B log, AT,

(3)

where A = 82,776, and B = 15,695. They also give

data from wells that penetrated zones with abnormally high fluid pressure. We have found that
all these data can be correlated by the equation

(+;F)13Zz/3=

A-Blog,*T,

(4)

where PO= overburden pressure; YF= fluid pres-

FIG.

7.

P-wave
transit times from well logs
sands in young sedimentary
basins.

for

and

Density

777

sure; a = normal overburden pressure gradient;

and p=normal fluid pressure gradient. Figure 8
illustrates this correlation.
For a normally prcssurcd section equation (4)
reduces to equation (3). One interesting feature
of equation (4) is that both the pressure difference
PO-PF
and the depth Z are present. The factor
Z2j3 may be interpreted as the effect of increased
consolidation with depth. For sands the effect of
consolidation outweighs the effect of pressure.

UNCONSOLIDATED

GAS

SANDS

In a shale-sand sequence some sands may contain oil or gas, but the overlying shale may not
contain either, except for the amount dissolved
in water. The reflection coefficient at such an
interface may be influenced appreciably by the
fluids in the formation, if the depth is not too
great. Gassmanns equation can be used to estimate the magnitude of the possible effect of the
presence of oil or gas in a sand upon its velocity
when the velocity of the brine-saturated sand is
known.
By using appropriate values in equation (la)
for parameters other than the skeleton moduli of
a given brine-saturated sand, we can determine
the relative skeleton modulus k/a. The values of
the parameters used are deduced in part from
known wave velocities equations, (lb) and (lc),
and a specification that LV= Sk with S having a
representative value of 2.0. Suitable values for
,&or L/k can be used for different fluid filling of
the pores, and the corresponding values of M
and the P-wave velocity calculated using equations (1).
The results of some such calculations are given
in Figure 9, where velocity is plotted versus depth
for sands saturated with either brine, oil, or gas.
The dashed curve illustrates typical values of
velocity versus depth for shales saturated with
brine. It is clear that in the first few thousand
feet the reflection coefficient at the boundary
between a shale and a sand will be significantly
greater if the sand contains gas than if the sand
contains brine. This observation might possibly
be of practical significance when there is a lateral
change from a brine-filled sand to a gas-filled
sand. However, at considerable depths the reflection coefficient becomes almost independent of
the nature of the fluid content.

778

Gardner,

Gardner,

and

Gregory

5
;4 100
2

2
5
90
111

CALCULATED
EQUATION

BY
3

CALCULATED
EQUATION

BY
4

4
i
80

90

100

110

CALCULATED

8. Transit

FIG.

120

TRANSIT

130

time

140

IS0

~-SEC./FOOT

time relations for high-pressure shales [data from Hottmann

and Johnson (1965)].

vr

4,000

8.000

12,Doo

16,000

18,000

COMPUTED
rl
Y
Y

DEPARTURES
FROM

CURVE

I
:
;

2.ooo

4,000

6,000
VELOCITY,

FIG.

9. Velocity

FEET

8.ooo
PER

10,ooo

12,000

14,000

SECOND

versus depth for shale and for

different fluids.

in-situ

sands containing

!%rnletion
REFLECTION

f&city

COEFFICIENTS

Acoustic impedance contrasts govern seismic

reflection coefficients at a plane interface between
two media. Thus,
R =

PlVl

P2V2
(5)

p1v1+

PYVZ

the ampl&de of t!le reflected tvave when

the incident wave has unit amplitude and is perpendicular to the interface.
Peterson et al (1955) showed that for practical
purposes this formula can be approximated by
gives

R = $ Zn (p1V~/p~Vr).

(6)

The empirical relationship between density

and velocity depicted in Figure 1,

p = .23V.25,

(7)

is a fair average for a large number of laboratory

and field observations of different brine-saturated
rock types (excluding evaporites). Combining
this with Petersons relation, WChave

R = a(1.25 In V,/Vz).

(8)

In a general way then, density commonly

varies with velocity so that its effects upon reflection coefficients is fairly satisfactorily taken into
account by multiplying the reflection coefficient
due to velocity contrast by a factor 1.25. Departures from this rule exist as evidenced by scatter
of observation points and may in some cases be
significant. The multiplying factor 1.25 also
increases the relative amplitudes of multiple
reflections over that estimated using velocity
contrasts alone.
CONCLUSIONS

have attempted to show that the P-wave

velocity in the upper layers of the earth (depths
less than 25,000 ft) varies systematically with different factors, although an absolute prediction of
velocity is seldom possible. The following summary of the effects appears to be generally valid.
We

(1) Gassmanns theory is typically valid for

sedimentary rocks in interrelating elastic constants, densities, and P-wave velocities for
different rock components and for the consolidated whole rock. An important component,
however, is the frame or skeleton, which may

have a wide range of c!astic parameters. These

parameters ordinarily can be characterized only
through Gassmanns theory, and this fact
limits the usefulness of the theory for practical
applications.
(2) Microcracks can be present in a rock
(within the rock skeleton) and materially reduce the P-wave velocity of the whole rock.
Pressure can close them and cause the velocity
to increase. The elastic parameters of rocks
without microcracks can be estimated by using
the theories of Voigt and Reuss and the known
elastic constants of the crystals.
(3) The well known time-average relationship empirically relates velocity and porosity
for a moderately wide range of rock types and
formation fluids when the rock is under a substantial pressure.
(4) The effective pressure governing the
elastic properties of the skeleton of porous
sedimentary rocks is the difference between the
total external pressure (or overburden pressure)
and the internal fluid pressure. Increase in the
skeleton pressure increases the elastic reactions
at intergranular interfaces and the velocity of
the whole rock.
(5) It is shown that in a recent basin, the
increase in velocity of sands with depth is substantially greater than the increase in velocity
of a sample of comparable sand subjected to
pressure in the laboratory. The difference is
mainly attributed to in-situ cementation of
sand grains with geologic time
(6) For high-pressure shales, the skeleton
pressure is markedly subnormal and may approach zero. The associated velocities are also
greatly reduced but have values consistent
with the existing skeleton pressure.
(7) For unconsolidated shallow gas sands,
some computations using Gassmanns theory
indicate a substantial difference in P-wave
velocity from that of the same sand filled with
brine. Some possibility of recognizing this
effect in seismic reflections exists.
(8) A simple systematic relationship exists
between the velocity and density of many sedimentary rocks in situ. For these rocks the
empirical relationship permits estimation of
reflection coefficients from velocity information
alone.

780

Gardner,

Gardner,

REFERENCES

Biot, M. A., 1956, Theory of propagation of elastic

waves in fluid-saturated porous solid: Parts I and II:
J. Acoust. Sot. .4m. v. 28, p. 1688191.
Clark, Sydney P., Editor, 1966, Handbook of physical
constants: GSA Memoir 97.
Faust, L. Y., 1953, A velocity function including lithologic variation: Geophysics, v. 18, p. 271-288.
Gassmann, F:, 1951, Ueber die Elastizitat porijser
Medien: Vrerteljahrsschrift der Naturforschenden
Ges., Zurich, v. 96, p. l-23.
Hill, R. W., 1952, The elastic behavior of a crystalline
gtrgate:
Proc. Phys. Sot. (London), A65, p. 349Hottmann, C. E., and Johnson, R. K., 1965, Estimation
of formation pressures from log-derived shale properties: J. Petr. Tech., v. 17, p. 717-722.

and

Gregory

Ide, John M., 1937?The velocity of sound in rocks and

glassesas a functron of temperature: J. Geology, v. 45,
p. 689-716.
Maxwell, John C., 1960, Experiments on compaction
and cementation of sand: GSA Memoir 79, p. 105132.
Peterson, R. A., Fillippone, W. R., and Coker, F. B.,
1955, The synthesis of seismograms from well log
data: Geophysics, v. 20, p. 516538.
Reuss, A., 1929, Berechnung der Fleissgrenze von
Mischkristallen auf Grund der Plastizitatsbedingung
$i; Einkristalle: 2. Angew. Math. Mech., v. 9, p. 49,
Voigt, W., 1928, Lehrbuch der Kristallphysik: Leipzig,
B. G. Teuhner.
White, J. E., 1965, Seismic waves: Radiation, transmission and attenuation: New York, McGraw-Hill
Book Co., Inc., p, 132.