Geophysical Prospecting, 2005, 53, 481–496

Ultrasonic velocities of North Sea chalk samples: influence of porosity,

fluid content and texture
Birte Røgen,1 Ida L. Fabricius,1∗ Peter Japsen,2 Christian Høier,2 † Gary Mavko3
and Jacob M. Pedersen4
1 Environment & Resources DTU, Bygningstorvet Building 115, Technical University of Denmark, 2800 Kongens Lyngby, Denmark,
2 GeologicalSurvey of Denmark and Greenland (GEUS), Øster Voldgade 10, 1350 Copenhagen K, Denmark, 3 Department of Geophysics,
School of Earth Sciences, Stanford University, Stanford, CA 94305-2215, USA, and 4 Ødegaard A/S, Titangade 15, 2200 Copenhagen N,
Denmark

Received November 2002, revision accepted August 2004

ABSTRACT
We have studied 56 unfractured chalk samples of the Upper Cretaceous Tor Forma-
tion of the Dan, South Arne and Gorm Fields, Danish North Sea. The samples have
porosities of between 14% and 45% and calcite content of over 95%. The ultrasonic
compressional- and shear-wave velocities (V P and V S ) for dry and water-saturated
samples were measured at up to 75 bar confining hydrostatic pressure corresponding
to effective stress in the reservoir. The porosity is the main control of the ultrasonic
velocities and therefore of the elastic moduli. The elastic moduli are slightly higher for
samples from the South Arne Field than from the Dan Field for identical porosities.
This difference may be due to textural differences between the chalk at the two loca-
tions because we observe that large grains (i.e. filled microfossils and fossil fragments)
that occur more frequently in samples from the Dan Field have a porosity-reducing
effect and that samples rich in large grains have a relatively low porosity for a given
P-wave modulus. The clay content in the samples is low and is mainly represented
by either kaolinite or smectite; samples with smectite have a lower P-wave modulus
than samples with kaolinite at equal porosity. We find that ultrasonic V P and V S of
dry chalk samples can be satisfactorily estimated with Gassmann’s relationships from
data for water-saturated samples. A pronounced difference between the V P /V S ra-
tios for dry and water-saturated chalk samples indicates promising results for seismic
amplitude-versus-offset analyses.

ified limestone data, Castagna, Batzle and Kan (1993) pro-

INTRODUCTION
Knowledge of the relationships between the petrophysical
properties of reservoir rocks and the parameters that can be
observed with geophysical methods has increased during the
last decades, but knowledge of the acoustic properties of chalk
is still limited, even though several studies have been pub-
lished. Some of the earlier studies are as follows. For unspec-
∗ E-mail: il@er.dtu.dk
†Present address: DONG/A/S, Agern Allé 24–26, 2970 Hørsholm,
Denmark.
vided a second-order polynomial fit to the V P /V S relationship.
Lind (1997) found that Wyllie’s time-average model (Wyllie,
Gregory and Gardner 1958) can be modified to give good
estimates of the porosity of chalk from velocity data by appli-
cation of the concept of critical porosity. Raiga-Clemenceau,
Martin and Nicoletis (1988) suggested an acoustic formation
factor as an alternative to Wyllie’s time-average equation and
provided empirical constants for calcite, among other miner-
als. Walls, Dvorkin and Smith (1998) fitted a modified up-
per Hashin–Shtrikman (MUHS, Nur et al. 1998) bound to
log data on chalk from the Ekofisk Field, resulting in an


C 2005 European Association of Geoscientists & Engineers 481
482 B. Røgen et al.
empirical porosity–velocity relationship. The effect of clay on
the porosity–velocity relationship in chalk has not been stud-
ied extensively, but an effective-medium model for chalk where
the clay is mainly in suspension was suggested by Fabricius
(2003). Empirical relationships between sonic-wave velocities,
porosity and clay content have been established for sandstone
(e.g. Han, Nur and Morgen 1986; Wilkens et al. 1986; Yan
2002). Han et al. (1986) found that as little as 1% of clay
significantly reduces the elastic moduli of sandstones.
In this study our aim is to establish a systematic data foun-
dation for the interpretation of porosity and fluid content of
chalk from acoustic data. Using laboratory measurements on
core material from three Danish hydrocarbon fields, we in-
vestigate how fluid content, texture and mineralogy influence
the relationship between porosity and ultrasonic velocity in
unfractured chalk.
To obtain samples with different burial and hydrocarbon-
filling histories while maintaining high purity and low strati-
graphic variation, we selected 56 chalk samples from the Tor
Formation of the Chalk Group from the Dan Field (25 oil-
zone samples), the South Arne Field (29 oil-zone samples) and
the Gorm Field (two water-zone samples) in the Danish North
Sea (Jørgensen 1992; Mackertich and Goulding 1999; Hurst
1983) (Fig. 1). The Tor Formation is of Maastrichtian age and
is the main oil-producing formation in these fields. According
to information from field operators, present effective stress on
reservoir rocks is nearly equal for the three fields, because the
Figure 1 Location map.

C 2005 European Association of Geoscientists & Engineers, Geophysical Prospecting, 53, 481–496
present difference in burial depths of about 900 m is counter-
acted by differences in overpressure.
Based on the work of, for example, Scholle (1977), we ex-
pect compaction to be the primary porosity-controlling factor
in chalk, and textural variations to be the secondary porosity-
controlling factor. An important textural factor is the grain-
size distribution which we have assessed by image analysis
of backscatter electron micrographs. Irrespective of miner-
alogy, particles of clay and fine-silt size are termed ‘mud’
or ‘matrix’ and particles of coarse-silt and sand size (larger
than 20 µm in diameter) are termed ‘large grains’. The calcite
‘mud’ consists mainly of nannofossils whereas the large grains
are mainly microfossils and fossil fragments. The term ‘large
porosity’ covers intrafossil and moldic porosity. The chalk
samples are classified according to depositional texture either
as mudstones (<10% large grains) or wackestones (>10%
large grains) (Dunham 1962) (Fig. 2). The proportion of min-
erals not soluble in hydrochloric acid is termed the ‘insolu-
ble residue’ and consists mainly of particles of clay size. Even
though the amount of insoluble residue is less than 5%, it can
have a significant influence on the petrophysical properties of
the chalk (Røgen and Fabricius 2002).
We investigate the effect of pore fluid on the elastic proper-
ties of chalk by comparing data from dry core samples and
from water-saturated core samples. Gassmann (1951) sug-
gested a low-frequency model for fluid substitution in a porous
medium relating the elastic moduli of rock saturated with dif-
ferent fluids. In line with Borre (1998), Wang (2000), Borre
and Fabricius (2001), Gommesen, Mavko and Mukerji (2002)
and Japsen et al. (2004), we apply Gassmann’s relationships
(e.g. Mavko, Mukerji and Dvorkin 1998) to test whether
dry moduli can be predicted from measurements on water-
saturated chalk samples.
Similarly to Walls et al. (1998), Japsen et al. (2000) used
a modified upper Hashin–Shtrikman bound to model the
velocity–porosity relationship for the Tor Formation, based
on log data from two wells in the Dan and South Arne
Fields. We apply a modified upper Hashin–Shtrikman bound
to ultrasonic laboratory data on chalk taken from the same
two fields and partly from the same wells. To avoid textu-
ral influence only samples classified as mudstone according to
Dunham (1962) are used for this model.
In this study, we show that the influence of fluid (water
versus dry) on the ultrasonic velocities of North Sea chalk
can be estimated with reasonable accuracy using Gassmann’s
relationships. We also show that the presence of large grains
(microfossils and fossil fragments) and smectite influences the
elastic properties for North Sea chalk.

DEPOSITIONAL TEXTURE RECOGNIZABLE
Original components not bound together during deposition Original components were
Contains mud Lacks mud bound together during
(particles of clay and fine-silt size) and is
Mud-supported Grain-supported grain-
Less than More than supported
10% grains 10% grains
Mudstone Wackestone Packstone
Figure 2 Classification of carbonate rocks according to depositional texture (Dunham 1962). Grains are defined as particles with a diameter
larger than 20 µm, and the term ‘mud’ covers smaller (mainly carbonate) particles of clay and fine-silt size. All samples in this study are classified
as either mudstone or wackestone.
METHODS
The sample-selection criteria were to cover both the main trend
and the scatter in a cross-plot of porosity and permeability
for the available chalk samples. In order to avoid complicat-
ing factors, samples with fractures or other visible inhomo-
geneities were avoided. We used 1.5-inch vertical plugs.
Plugs were Soxhlet-cleaned with methanol and toluene for
salt and hydrocarbons and then dried at 110◦ C. Porosity was
measured as helium-porosity with an uncertainty at 0.1%
point. Gas permeability was measured at a differential pres-
sure between 0 and 1 bar, with a confining sleeve pressure of
400 psi. Uncertainty of permeability data is 4%.
The calcite content was measured by adding a surplus of
0.5 m hydrochloric acid to the powdered sample, and subse-
quent titration by 0.5 m Na-hydroxide solution. The miner-
alogical composition of the insoluble residue was identified
by X-ray diffraction (Cu-Kα) of orientated, subsequently gly-
colated (3 days, 60◦ C) and heated (500◦ C, 2 hours) samples.
A Philips diffractometer with graphite monochromator and
automatic divergence slit was used. Thin sections were used
for a visual classification of depositional texture according to
Dunham (1962) (Fig. 2).
Quality control
As a test of sample homogeneity, the gas permeability of a sam-
ple can be compared with the estimated matrix permeability
predicted from the Kozeny equation. Due to the homogene-
ity and connectivity of the pore space, the Kozeny equation
is valid for chalk (Mortensen, Engstrøm and Lind 1998). We
can use the Kozeny equation to test whether the measured
permeability is true matrix permeability or is dominated by
fracture permeability. The Kozeny equation expresses a rela-
tionship between liquid permeability kl [m2 ], porosity ϕ [frac-
tion], specific surface area relative to bulk volume S [m2 /m3 ]

C 2005 European Association of Geoscientists & Engineers, Geophysical Prospecting, 53, 481–496
Ultrasonic velocities of North Sea chalk samples 483
DEPOSITIONAL
TEXTURE NOT
RECOGNIZABLE
deposition
(Subdivide according to
classifications designed to
bear on physical texture
or diagenesis.)
Crystalline
Grainstone Boundstone
Carbonate
and a porosity-dependent factor c. The liquid permeability can
be approximated by the Klinkenberg correction. An empirical
conversion from measured gas permeability kg to Klinkenberg
permeability kk of chalk was given by Mortensen et al. (1998).
Thus, the Kozeny equation is given by
ϕ3
kl = c(ϕ) ≈ kk = 0.52kg1.083 , (1)
S2
where c(ϕ) can be calculated from ϕ (Mortensen et al. 1998).
In the porosity interval, 20–50%, c(ϕ) varies from 0.21 to
0.25. The Kozeny equation implies that permeability is lower
for chalk rich in clay and silica compared with clean chalk,
due not only to sorting (poorer sorting reduces porosity) but
also to the higher surface area (i.e. higher resistance to flow)
(Røgen and Fabricius 2002).
For quality control using the Kozeny equation, the specific
surface area was obtained by the BET method using nitrogen
(99.999% pure) as adsorbate (Brunauer, Emmet and Teller
1938). Prior to the BET measurements, the crushed samples
were outgassed at 70◦ C using nitrogen as a carrier gas.
In four samples, a high measured gas permeability compared
with the estimated matrix permeability may indicate fractures
in the samples (A782, B216, N169 and P177, Fig. 3). The
four samples that appear fractured in the permeability mea-
surements do not appear to have lower P-wave moduli than
unfractured samples when measured at 75 bar hydrostatic
confining pressure (Fig. 14a).
The velocity of all samples was measured at 25 bar, 50 bar
and 75 bar hydrostatic confining pressures to test for pressure-
dependent behaviour. Borre and Fabricius (2001) found that
velocities stabilize as the effective stress increases to above
half the in situ vertical effective stress. Our data confirm this
by showing a velocity increase as low as 4% associated with
an increase in hydrostatic confining pressure from 25 to the
approximate in situ effective stress of 75 bar (Fig. 4). This
484 B. Røgen et al.

10 (a)
Estimated matrix permeability [mD]

Dry
6 Saturated
1

vP [km/s]
4
0.1 South Arne 23% porosity
Dan
Gorm 41% porosity
2
1:1
0.01
0.1 1 10 100
0
Measured gas permeability [mD] 0 25 50 75 100
Figure 3 Estimated matrix permeability versus measured gas per- Hydrostatic confining pressure [bar]
meability. The estimated matrix permeability was calculated using
Kozeny’s equation and an empirical relationship between gas and (b)
Klinkenberg permeability (Mortensen et al. 1998). In four samples
Dry
(A782, B216, N169 and P177), a high measured gas permeability 6 Saturated
compared with the estimated matrix permeability may indicate frac-
tures in the samples.

vS [km/s]
4

shows that the measurements at 75 bar hydrostatic confining

pressure are comparable with in situ conditions. 2 23% porosity
The fluid-substitution relationships of Gassmann apply for 41% porosity

slightly moist rock samples (Mavko et al. 1998). To test if the

0
dry ultrasonic measurements were made on samples that were 0 25 50 75 100
too dry, three samples were re-measured at room moisture Hydrostatic confining pressure [bar]
conditions. This resulted in a slight reduction in the velocities,
Figure 4 Example of pressure dependence of velocity measurements
but the reduction was smaller than the estimated standard
for samples with low and high porosities. (a) Compressional-wave
deviations. Therefore it was concluded that the dry samples velocity (V P ); (b) shear-wave velocity (V S ). The increase in velocity is
were not too dry and thus the measurements are valid input approximately 4% between 25 bar and 75 bar hydrostatic confining
to Gassmann’s relationships. pressure. Eighty-seven per cent of the samples have an increase in
velocity lower than 6%.

Ultrasonic measurements on chalk samples

Ultrasonic compressional- and shear-wave velocities were
measured on 56 samples. All samples were measured in a dry
state but, due to time constraints, only 34 samples were mea-
sured when saturated with calcite-equilibrated tap-water. The
ultrasonic velocity V of the samples was calculated from the
transit time t it takes an ultrasonic P- or S-wave to travel a
certain distance d (i.e. the sample length) through a sample,
i.e.
d porosity and the mass of the dry and the saturated sample
V = , (2)
t − tdelay
where tdelay is the system delay time. The system delay time was
determined by measuring the transit time without any plugs
and on three aluminium plugs of different lengths. Transit
times for P- and S-waves were measured on a digital oscil-
loscope. The P- and S-wave data were saved digitally for later
automated analysis, using the arrival picker software devel-
oped by Ødegaard A/S. Readings of traveltime for both wave
types were taken as first peak. The P- and S-wave transducers
have a centre frequency of 700 kHz. The confining pressure
was increased gradually to the next pressure step over a time
period of 30 minutes. When unloading the core holder after
testing, the confining pressure was decreased gradually from
75 bar to 0 bar over a time period of 90 minutes. For water-
saturated plugs, the saturation was calculated from helium
with an uncertainty of 1% point. The water saturation was
higher than 98%.
For fluid substitution, we applied Gassmann’s relationships
in the form given by Mavko et al. (1998):
Ksat Kdry Kfluid
= + , Gsat = Gdry , (3)
K0 − Ksat K0 − Kdry ϕ(K0 − Kfluid )


C 2005 European Association of Geoscientists & Engineers, Geophysical Prospecting, 53, 481–496

where K0 and Kfluid are the mineral and fluid bulk moduli,
Ksat and Kdry are the effective bulk moduli of saturated and
dry rock, respectively, ϕ is the porosity, and Gsat and Gdry
are the effective shear moduli of saturated and dry rock, re-
spectively. For isotropic elastic materials, the elastic moduli
are expressed by the compressional- and shear-wave veloci-
ties, V P and V S , respectively, and the bulk density ρ. The P-
wave modulus M is given by M = ρV 2P , the shear modulus
G is given by G = ρV 2S and the bulk modulus K is given by
K = M − 4/3G.
Standard deviations of sample length, transit time and den-
sity were estimated as 0.1 mm, 0.09 µs and 0.02 g/cm2 , re-
spectively. Standard deviations of ultrasonic velocities and of
elastic moduli were estimated by error propagation. The esti-
mated standard deviations s are as follows: s(V P ) < 0.08 km/s,
s(V S ) < 0.03 km/s, s(M) < 1.76 GPa, s(G) < 0.41 GPa and
s(K) < 1.84 GPa, where M is the P-wave modulus, G is the
shear modulus and K is the bulk modulus.
Microtextural analysis of chalk samples
by image analysis
Image analyses for this study were performed on backscat-
ter micrographs, following the method for determination of
grain-size distribution described by Røgen, Gommesen and
Fabricius (2001). In this method, grain-size distributions are
defined by cross-sectional area in a two-dimensional plane.
Vertically orientated samples with a height and width of ap-
proximately 1 cm were impregnated with epoxy and polished
before imaging in a scanning electron microscope. A standard
imaging procedure was followed for each sample. This in-
cludes an overview image of the sample to check for inho-
mogeneities in the sample or in preparation. Then four image
fields are chosen along the centre-line in the vertical direc-
tion with spacing of 2 mm. Each image field is represented by
two images. One image at large magnification (representing
60 µm × 80 µm) shows only the chalk matrix, the insoluble
residue and the matrix porosity. The second image is a con-
centric image at 10 times smaller magnification (representing
600 µm × 800 µm) showing microfossils, chalk clasts and
shell debris as well as large porosity (intraparticular and
moldic porosity), but not resolving the matrix. By combin-
ing the information from each of the magnifications and as-
suming a homogeneous matrix, we obtain information cor-
responding to 100 images at the large magnification for
each image field of the small magnification. Examples of
grain-size distributions from image analysis are shown in
Fig. 5.

C 2005 European Association of Geoscientists & Engineers, Geophysical Prospecting, 53, 481–496
Ultrasonic velocities of North Sea chalk samples 485
Modified upper Hashin–Shtrikman model
The empirical modified upper Hashin–Shtrikman model
(MUHS) model describes how the bulk and shear moduli
(KMUHS and GMUHS ) increase as porosity ϕ is reduced from the
high-porosity end member ϕ max to zero porosity. This model
is given by the equations,
1 1 − ϕmax ϕmax
= + , (4)
KRmax Ks Kf
KMUHS
ϕ/ϕmax
= Ks + ,
(KRmax − Ks )−1 + (1 − ϕ/ϕmax ) (Ks + 4/3Gs )−1
(5)
ϕ/ϕmax
GMUHS = Gs +
, (6)
−1 2(1−ϕ/ϕmax )(Ks +2Gs )
(−Gs ) + 5Gs (Ks +4/3Gs )
MMUHS = KMUHS + 4/3GMUHS . (7)
At the high-porosity end member, the bulk modulus is given
as the modulus at the Reuss suspension curve (KRmax ) and the
shear modulus is zero. Ks and Gs denote the estimated bulk
and shear moduli of the solid, respectively, and Kf denotes
the bulk modulus of the fluid. The modified upper Hashin–
Shtrikman model is fitted to bulk, shear and P-wave moduli
at the same time, resulting in a set of estimated parameters:
Ks , Gs and ϕ max .
R E S U LT S A N D D I S C U S S I O N
Overall porosity–depth trend
A general loss of porosity with depth may be observed in
chalk (Scholle 1977; Borre and Fabricius 1998; Japsen 1998).
This trend reflects compaction of the chalk due to an increase
in effective stress. However, textural variations of the chalk
overshadow this general trend in the depth intervals studied
(Fig. 6).
Influence of porosity on ultrasonic velocity
We found negative correlations between porosity and P- and
S-wave velocities for dry as well as saturated samples (Fig. 7;
Table 1). In the porosity range 14–45%, we found V P in
the range 4.3–1.9 km/s and V S in the range 2.6–1.3 km/s;
these results are similar to those of Borre (1998). We found
that the V P /V S ratios are distinctly different for dry and satu-
rated samples (Fig. 8). For dry samples, we observe a negative
correlation with porosity, whereas for saturated samples the
486 B. Røgen et al.

Mudstone
10

1
0.1

10
0.01

100

1000
0.001

10000

100000
Wackestone
10

10
0.1

100

1000
0.01

10000
0.001

100000
Wackestone
10

0
1

10
0.1
0.01

100
0.001

1000

10000

100000

Grain area [µm2]

200 µm

Figure 5 Backscatter micrographs representing 600 µm × 800 µm and final grain-size distributions of a mudstone (< 10% grains) and two
wackestones (> 10% grains) (micrographs at large magnification showing matrix particles are not shown). The isolated peak to the left in the
histograms represents insoluble residues that are not resolved in the images. The mudstone contains hollow microfossils. The two wackestones
contain filled microfossils.

correlation between V P /V S and porosity is positive. For dry compared at, for example, 30% porosity, the V P /V S ratios in
samples, the V P /V S ratio ranges from 1.70 to 1.46 in the this study are similar to those for saturated samples found in
porosity range 14–45%. For saturated samples, the V P /V S ra- other studies, but they are slightly smaller than those found
tio ranges from 1.80 to 2.10 in the same porosity range. When for dry samples in other studies (Table 2).


C 2005 European Association of Geoscientists & Engineers, Geophysical Prospecting, 53, 481–496
 Ultrasonic velocities of North Sea chalk samples 487

Porosity [%] Porosity [%] ships between V P and V S for dry and saturated samples (Figs 8
0 25 50 0 25 50 and 10). The bulk and P-wave moduli are higher for saturated
x00 y00 samples than for dry samples, whereas the shear moduli for
saturated samples are slightly lower than for the dry samples;
this is probably a result of fluid–solid interactions weakening
20
x20 y20
20 grain contacts (Fig. 11; compare Japsen et al. 2004). This ef-
fect was also observed by Assefa, McCann and Sothcott (2003)
for oolitic limestone. When we apply Gassmann’s relationship
Relative depth [m]

Relative depth [m]

x40
40 y40
40 (3) to data for saturated samples and predict dry moduli, we
obtain good agreement with the measured dry data (Fig. 12).
The Gassmann-estimated P- and S-wave velocities of dry sam-
60
x60 60
y60 ples compare well with the measured velocities of dry samples
(Fig. 13). In general, velocities are slightly underestimated by
Gassmann’s relationship, although V P values of high-porosity
80
x80 80
y80 samples are an exception with higher estimated than measured
A
B
values. This is in agreement with the local flow model of a
M
C N rock with medium-to-low aspect ratio pores (Mavko et al.
x100
100 y100
100 1998; Wang 2000). When defining the Gassmann deviation
Figure 6 Porosity versus vertical depth below uppermost sample for as ((V Gassmann − V measured )/V measured ) (Wang 2000), an average
the South Arne Field (wells A, B and C) and the Dan Field (wells M and underestimation of dry V P of 2% and of dry V S of 4% is
N). Normal compaction resulting in porosity reduction for increasing observed (Fig. 13). This observation is consistent with results
depth can be observed within wells A, C and N. The variations are
obtained by Borre (1998) and Borre and Fabricius (2001) for
interpreted as variations in texture (here focused on the proportion of
large grains, clay mineralogy and cementation). chalk samples from the Gorm and Tyra Fields.
In line with the results of Borre (1998), we conclude that
5
Gassmann’s relationships give good estimates of velocities of
4 dry samples when applied to ultrasonic measurements on
water-saturated chalk. In contrast, Wang (1997) found that
3
v [km/s]

V P of carbonates is underestimated by Gassmann’s relation-

2 ships by up to 25%. Carbonate rocks can be heterogeneous
vp(saturated)
vp(dry)
1
vs(dry)
vs(saturated)
0 study. The reason that Gassmann’s relationships give good re-
0 10 20
Porosity [%]
Figure 7 P- and S-wave velocities versus porosity of 33 saturated and
55 dry chalk samples.
Influence of fluid content on velocity
We observe that V P values for saturated samples are equal to
V P values for dry samples with the highest velocities, but are
slightly higher for samples with the lowest velocities (i.e. high
porosities) (Fig. 9; Table 1). This increase in saturated V P val-
ues can be seen for samples with porosities above 35% (V P
below 3 km/s) and is due to the increasing influence of fluid
properties on acoustic properties at high porosities. For shear-
wave velocities we observe lower velocities for saturated sam-
ples than for dry samples, in accordance with the difference in
densities (Fig. 9). Consequently, we observe distinct relation-
and have high permeability so the results of Wang (1997) may
not be comparable with the pelagic chalk we deal with in this
30 40 50
sults for chalk may be the homogeneity and low permeability
of the chalk, where individual particles and pores are so much
smaller than the applied wavelength that the ultrasonic wave
experiences the chalk as a homogeneous medium. Matrix par-
ticles are about 3 µm in diameter, the largest observed grain
is approximately 100 µm in diameter, and the applied fre-
quency results in wavelengths between 700 µm and 6000 µm.
Gassmann’s relationships represent a low-frequency theory, so
when we observe that it applies to ultrasonic data, it probably
also applies to the lower-frequency logging data and seismic
data.
Influence of microtexture on velocity
Most Dan Field samples have a lower P-wave modulus for
a given porosity than the corresponding South Arne Field


C 2005 European Association of Geoscientists & Engineers, Geophysical Prospecting, 53, 481–496
 488 B. Røgen et al.
Table 1 Data for 56 chalk samples of the Tor Formation in the North Sea

Carbonate Specific
Gas content surface Matrix <3.16 µm2 >316 µm2 X-ray data VP VS VP VS Water
Sample Porosity permeability (% of area porosity % (rel. to % (rel. to (dry) (dry) (sat) (sat) saturation
no. (%) (mD) solids) (m2 /g) Texture (%) solid) solid) Quartz Kaolinite Smectite Other (km/s) (km/s) (km/s) (km/s) (%)

A755 24.5 0.52 97.9 1.75 M 25 16 5 2 3 b, i 4.05 2.42 4.08 2.24 99

A759 27.2 1.42 97.2 1.75 M 28 25 8 2 2 f, i- s, p 3.59 2.15 – – –
A760 29.3 1.15 98.2 1.90 W 30 13 14 1 3 f, i 3.48 2.16 3.60 1.98 99

C

A761 30.4 1.48 98.6 1.70 W 32 14 12 2 2 1 b, f, i, p – – 3.47 1.86 99

2005 European Association of Geoscientists & Engineers, Geophysical Prospecting, 53, 481–496

A763 28.0 1.47 98.7 1.53 W 30 12 15 2 3 f, i 3.52 2.18 3.60 1.98 100
A767 30.9 2.15 98.9 1.57 M 33 15 11 2 i-s 3.47 2.12 3.53 1.96 100
A771 32.1 2.41 98.7 1.68 W 34 16 13 1 b, f, i, i-s 3.33 2.03 3.40 1.82 100
A774 28.0 1.34 99.0 1.62 M 30 20 8 3 f, i 3.66 2.23 – – –
A779 22.6 0.40 99.3 1.17 M 23 15 5 3 f, i-s, p 3.97 2.37 – – –
A780 30.9 1.99 99.3 1.49 M 32 16 6 3 1 f, i-s, p 3.38 2.08 – – –
A782 24.8 3.28 99.6 1.23 M 25 15 2 3 1 f, i-s, p 3.68 2.25 – – –
A783 20.1 0.96 98.7 1.29 M 21 16 3 3 f, i-s, p 3.98 2.42 – – –
B216 13.5 0.54 93.1 2.04 W 15 20 15 3 2 f, i-s, p 4.30 2.60 – – –
B220 30.1 1.25 92.1 2.13 M 30 26 2 2 2 f, i 3.33 2.07 3.33 1.82 100
B236 35.5 2.52 95.6 1.63 M 37 23 9 3 2 f, i-s, p 3.15 2.05 – – –
B240 37.6 4.07 96.7 1.54 M 39 23 9 3 2 f, i-s 2.57 1.68 – – –
B244 37.4 3.50 97.1 1.60 M 39 21 8 3 2 f, i-s, p 2.86 1.81 – – –
B260 41.1 4.10 96.0 1.92 M 41 25 5 2 2 f, i, p 2.45 1.62 2.63 1.37 100
B264 40.2 4.51 97.4 1.94 M 40 26 3 2 2 b, f, i 2.43 1.58 2.58 1.34 100
B274 38.0 3.62 97.5 1.40 M 38 20 4 3 2 i-s 2.68 1.76 – – –
C121 42.5 8.36 96.6 1.77 M 42 23 7 2 2 i, p 2.09 1.42 2.44 1.25 100
C128 42.7 7.17 96.3 1.73 M 43 21 7 2 2 i, p 2.06 1.37 2.37 1.13 100
C136 41.4 5.76 96.5 1.71 M 42 22 6 2 2 i, p 2.07 1.39 2.43 1.22 100
C151 43.7 7.40 95.9 1.42 M 43 34 1 2 2 i, p 2.20 1.44 2.48 1.26 99
C162 43.2 8.87 96.9 1.62 M 44 23 5 3 2 f, i, p 2.32 1.51 – – –
C176 39.5 4.78 96.1 1.65 M 40 25 4 2 2 f, i, p 2.70 1.73 – – –
C200 45.0 8.60 97.0 1.37 M 44 43 1 2 2 i, p 1.98 1.3 2.31 1.12 99
C204 41.3 5.93 97.4 1.78 M 43 17 9 2 2 f, i, p 1.96 1.34 2.36 1.19 100
C212 42.7 6.29 96.9 1.90 M 43 25 4 2 2 b, f, i, p 2.06 1.39 2.43 1.23 100
 
C
2005 European Association of Geoscientists & Engineers, Geophysical Prospecting, 53, 481–496

Table 1 (Continued)

Carbonate Specific
Gas content surface Matrix <3.16 µm2 >316 µm2 X-ray data VP VS VP VS Water
Sample Porosity permeability (% of area porosity % (rel. to % (rel. to (dry) (dry) (sat) (sat) saturation
no. (%) (mD) solids) (m2 /g) Texture (%) solid) solid) Quartz Kaolinite Smectite Other (km/s) (km/s) (km/s) (km/s) (%)

M001 24.0 0.96 95.0 1.19 W 26 21 15 2 2 f, i, p 3.28 2.11 3.36 1.82 99

M002 27.1 1.27 96.9 1.28 W 29 14 13 3 1 f, p 3.51 2.13 3.57 – 99
M003 23.9 0.52 95.6 1.41 M 25 16 10 2 2 f, i 3.39 2.12 3.52 1.89 100
M004 28.0 0.55 95.8 1.63 W 29 16 13 2 2 i, p 2.97 1.89 3.11 1.62 100
M005 18.7 0.44 93.9 1.87 M 19 30 8 2 2 f, i, p 3.67 2.32 3.71 2.01 95
M006 23.7 0.63 97.6 1.22 M 25 12 9 2 2 p 3.83 2.33 3.91 2.14 100
M008 29.9 3.79 96.6 1.19 M 32 19 12 3 1 f, i, p 3.02 1.91 3.14 1.70 100
M014 25.8 1.28 96.1 1.49 M 26 27 6 2 2 f, i, p 3.18 2.01 3.20 1.74 100
M015 29.9 2.69 97.8 1.08 M 32 12 12 2 2 p 3.11 1.97 3.25 1.75 100
M017 31.3 1.83 97.1 1.27 W 39 12 32 3 1 p 3.10 1.93 3.22 1.72 100
M020 29.4 1.60 98.7 1.16 W 32 11 16 2 2 i, p 3.09 1.97 3.26 1.75 100
M021 30.3 2.58 97.4 1.14 M 31 23 8 2 2 f, i, p 3.18 1.95 3.22 1.70 99
M022 29.0 1.37 96.4 1.45 W 32 13 15 2 2 i, p 2.93 1.89 3.09 1.67 99
M030 29.7 1.28 98.4 1.31 M 30 13 6 2 2 i, p 3.14 1.96 3.23 1.74 100
M031 31.5 3.28 98.9 1.05 M 34 9 12 3 1 i 2.84 1.84 2.97 1.60 99
M032 28.8 2.29 97.1 1.41 M 30 21 9 3 1 f, i, p 3.21 2.00 3.27 1.75 100

Ultrasonic velocities of North Sea chalk samples 489

M035 32.3 3.23 97.9 1.24 M 33 25 5 2 2 f, i, p 2.84 1.81 2.95 1.57 100
M037 31.7 2.32 98.0 1.17 M 32 12 3 3 1 i, p 2.96 1.85 3.11 1.66 100
N166 26.7 1.91 98.7 1.28 W 33 14 28 2 2 fl, i, p 2.99 1.91 – – –
N167 26.5 2.45 98.8 1.00 W 31 14 21 1 3 f, fl, i 3.32 2.07 – – –
N169 25.6 5.31 98.1 1.26 W 31 16 23 2 2 f, i 3.19 2.00 – – –
N170 26.9 2.24 98.3 1.18 W 30 16 19 2 2 fl, p 3.10 1.95 – – –
N171 25.8 1.66 97.4 1.34 M 27 24 7 2 2 i, p 3.26 2.06 – – –
N172 23.5 1.30 98.5 1.11 W 28 13 21 2 2 i, p 3.46 2.17 – – –
N175 22.2 0.94 97.9 1.33 W 26 17 17 2 3 i, p 3.67 2.23 – – –
P176 23.1 0.83 97.9 1.29 W 28 18 21 2 2 b, f, i 4.02 2.37 – – –
P177 20.8 34.03 98.5 1.06 W 27 14 30 2 1 3 i, p 4.18 2.46 – – –

The letter in sample numbers refers to the well. Wells A, B and C are in South Arne Field; wells M and N are in Dan Field; well P is in Gorm Field. Depositional texture (Dunham 1962): M denotes
mudstone, W denotes wackestone (see Fig. 2). Grain-size distribution from image analysis is given as proportion of solids with a cross-sectional area smaller than 3.16 µm2 and proportion of solids with
a cross-sectional area larger than 316 µm2 . Key for X-ray data: 1, present; 2, contributor; 3, dominating; 4, sole detected. For other minerals: b, barite; f, feldspar; fl, fluorite; i, illite; i-s, illite–smectite
interstratified; p, pyrite. Velocities are measured at a hydrostatic confining pressure of 75 bar. Water saturation of dry samples is 0–1%.
490 B. Røgen et al.

2.5
Saturated
6
Dry

vP (saturated) [km/s]
2
4
vP/vS [-]

1.5 2
Calcite (Mavko et al . 1998)
1:1
0
1 0 2 4 6
0 10 20 30 40 50 vP (dry) [km/s]
Porosity [%]

Figure 8 V P /V S ratio versus porosity for dry and saturated samples. 6

In the porosity range, 14–45%, dry samples have V P /V S ratios from

vS (saturated) [km/s]
1.70 to 1.46 and saturated samples have V P /V S ratios from 1.80 to
2.10. Compare with V P /V S of 1.94 for calcite (Mavko et al. 1998). 4

Table 2 V P /V S ratio at 30% porosity from different studies, with

origin of chalk samples 2

This study Borre (1998) Hvid (1998) 1:1

0
V P /V S @ Dan and South Gorm Copenhagen
0 2 4 6
30% porosity Arne Fields Field limestone
vS (dry) [km/s]

Dry 1.5–1.6 1.7–1.9 1.6–2.1 Figure 9 Velocity for saturated versus dry samples. Mineral endpoint
Saturated 1.8–1.9 1.8–2.1 – indicated with asterisk at highest velocity and fluid endpoint indicated
with asterisk at lowest velocity. See text for explanation.
4
Dry
samples (Fig. 14a; Table 1). This may be due to textural differ- Saturated
ences where we observe that the samples from the Dan Field 3
contain smectite and many are classified as wackestones. In
vS [km/s]

contrast, the samples from the South Arne Field are domi-
nated by mudstone texture and contain kaolinite. Due to the
limited amount of large porosity (less than 2%) present in the
studied samples, we cannot reach a conclusion on the effect of
large porosity on the elastic properties (Fig. 14d).
In backscatter micrographs we observe that the microfossils
appear hollow in some samples and filled with cement in other
samples (Fig. 15). The samples with filled microfossils do not
generally have pore-filling cement in the matrix. The presence
of large grains in the form of filled microfossils in a mud-
supported chalk reduces porosity and indicates softer rather
than stiffer chalk, compared with a pure mudstone with the
same porosity (Fig. 14b). The porosity-reducing effect of filled
microfossils is thus more significant than a possible sediment-
stiffening effect. This may also be due to the differences in clay
mineralogy associated with this texture.
The samples have high calcite content, mostly over 95%,
and all samples contain quartz. Despite the low clay content
we observe that the mineralogy of the clay is related to the
Castagna
2
1
0
0 2 4 6
vP [km/s]
Figure 10 Relationship between V P and V S for both dry and satu-
rated samples. Mineral and fluid endpoints are indicated with aster-
isks (Mavko et al. 1998). Note the distinct relationships between V P
and V S for dry and saturated samples. The measured V S values of
saturated samples are slightly higher than V S values estimated from
V P values by application of the empirical relationship for limestone
of Castagna et al. (1993). Observe that the empirical relationship of
Castagna et al. (1993) does not fit the endpoints.
stiffness of the samples. Smectite is present in the softest sam-
ples which also tend to have wackestone texture (from the
Dan Field), whereas kaolinite is present in the stiffest sam-
ples, which are dominated by mudstone texture (from the


C 2005 European Association of Geoscientists & Engineers, Geophysical Prospecting, 53, 481–496
 Ultrasonic velocities of North Sea chalk samples 491

25 25
1:1 1:1

K(Gassmann dry) [GPa]

20 20
K(saturated) [GPa]

15 15

10 10

5 5

0 0
0 5 10 15 20 25 0 5 10 15 20 25
K(dry) [GPa] K(dry) [GPa]
15 15
1:1 1:1

G(Gassmann dry) [GPa]

G(saturated) [GPa]

10 10

5 5

0 0
0 5 10 15 0 5 10 15
G(dry) [GPa] G(dry) [GPa]
40
40
1:1
1:1
M(Gassmann dry) [GPa]
M(saturated) [GPa]

30
30

20
20

10
10

0
0
0 10 20 30 40
0 10 20 30 40
M(dry) [GPa]
M(dry) [GPa]
Figure 11 Elastic moduli of saturated samples versus moduli of dry
Figure 12 Estimated moduli of dry samples calculated from saturated
samples. Bulk (K), shear (G) and P-wave (M) moduli are plotted in
samples using Gassmann’s relationships versus measured moduli of
the upper, middle and lower figures, respectively. The shear moduli
dry samples. Bulk (K), shear (G) and P-wave (M) moduli are plot-
(G) of the saturated samples are slightly lower than those of the dry
ted in the upper, middle and lower figures, respectively. In general,
samples, whereas the bulk (K) and P-wave moduli of water-saturated
Gassmann’s relationships give good estimates of the elastic moduli of
samples are higher than those of dry samples due to the influence of
dry chalk samples calculated from saturated samples. Note, however,
the moduli of the fluid present. The estimated standard deviations of
the increasing error for the shear modulus with increasing velocity
measurements of K, G and M are 1.84 GPa, 0.41 GPa and 1.76 GPa,
(see Fig. 13). The correlation coefficients between measured dry mod-
respectively.
uli and Gassmann-predicted dry moduli are 0.99, 0.96 and 0.98 for
K, G and M, respectively.

South Arne Field) (Fig. 14c). Variations in clay mineralogy

can be due either to geographical depositional differences or in the ϕ–M domain is 5 percentage points. The differences in
to temperature-related diagenesis. The volume of clay minerals moduli for samples with different dominant clay mineralogies
is less than 1% (Røgen and Fabricius 2002) whereas the dif- can have several explanations. Kaolinite has higher mineral
ference in porosity between the kaolinite and smectite trends moduli than smectite, but with the small volumes present, this


C 2005 European Association of Geoscientists & Engineers, Geophysical Prospecting, 53, 481–496
492 B. Røgen et al.

(a) (c)
5 5
Gassmann vP(dry) [km/s]

Gassmann vS(dry) [km/s]

4 4

3 3

2 2

1 1
1 2 3 4 5 1 2 3 4 5
Measured vP(dry) [km/s] Measured vS(dry) [km/s]

(b) (d)
6 6
Gassmann deviaton of vP [%]

Gassmann deviaton of vS [%]

4 4
2 2
0 0
-2 -2
-4 -4
-6 -6
-8 -8
-10 -10
10 20 30 40 50 10 20 30 40 50
Porosity [%] Porosity [%]

Figure 13 (a) Gassmann-estimated V P (dry) versus measured V P (dry). (b) Gassmann deviation of V P versus porosity. The Gassmann deviation
is defined as ((V Gassmann − V measured )/V measured ) (Wang 2000). Gassmann’s relationship underestimates V P on average by 2%. (c) Gassmann-
estimated V S (dry) versus measured V S (dry). (d) Gassmann deviation of V S versus porosity. Gassmann’s relationship consequently underestimates
V S (on average by 4%).

is hardly the main reason. Smectite may have a softening effect
on the calcite frame that kaolinite does not have and may thus
be an indicator of a softer calcite frame.
Empirical velocity–porosity relationship
In order to establish a texturally independent moduli–porosity
relationship for chalk, a modified upper Hashin–Shtrikman
bound was fitted to the elastic moduli (bulk, shear and P-wave)
and porosity data for 37 dry mudstone samples. Mudstones
are a better choice than wackestones because mudstones con-
stitute a more homogeneous group with respect to grain-size
distribution. The bulk and shear moduli of the fluid (air) were
set at 150 kPa and zero, respectively. The estimated parameters
defining the model are bulk and shear moduli at zero poros-
ity of 40 GPa and 28 GPa, respectively, and a high-porosity
end member at 52%. Correlation coefficients are 0.93, 0.96
and 0.95 for bulk, shear and P-wave moduli, respectively. The
empirical value for elastic properties at 0% porosity is lower
than the mineral value for calcite, and is not well determined
when low-porosity samples are not present in the data set. It
can be seen that the same correlation coefficients are obtained
with the modified upper Hashin–Shtrikman model fitted to
log data (Japsen et al. 2000) or even with a simple linear re-
gression (Fig. 16). The linear regression model is: Kdry = 27 −
0.6ϕ; Gdry = 19 − 0.4ϕ; Mdry = 53 − ϕ. We conclude that
in the present case, there is no justification for preferring the
modified upper Hashin–Shtrikman model to a linear model.
CONCLUSIONS
On the basis of the experimental results from 56 chalk samples
from the Tor Formation in the Dan, South Arne and Gorm


C 2005 European Association of Geoscientists & Engineers, Geophysical Prospecting, 53, 481–496
 Ultrasonic velocities of North Sea chalk samples 493

(a) (c)
50 50
South Arne Kaolinite
Dan Smectite
40 Gorm 40
Other
High permeability
M(dry) [MPa]

M(dry) [MPa]
30 30

20 20

10 10

0 0
10 20 30 40 50 10 20 30 40 50
Helium porosity [%] Helium porosity [%]

(b) (d)
50 50
Mudstones Up to 1% large porosity
Wackestones > 1% large porosity
40 40
M(dry) [MPa]

M(dry) [MPa]
30 30

20 20

10 10

0 0
10 20 30 40 50 10 20 30 40 50
Helium porosity [%] Helium porosity [%]

Figure 14 P-wave moduli for dry samples versus helium porosity. (a) Fields are marked; (b) depositional texture; (c) dominant clay mineral;
(d) amount of large porosity. The South Arne and Dan Field samples form almost separate trends that may be explained by differences in texture.
The increased content of filled microfossils in wackestones compared with mudstones reduces porosity and indicates softer rather than stiffer
chalk, even in the limited amounts present in the studied samples. Clay minerals are present in such small amounts that the porosity-reducing
effect can be discounted. The differences in moduli for samples with different dominant clay mineralogies can have two explanations: firstly,
that kaolinite has higher mineral moduli than smectite; secondly, either that smectite has a softening effect on the calcite frame that kaolinite
does not have, or the opposite, that kaolinite has a stiffening effect on the calcite frame that smectite does not have. The limited amount of large
porosity in hollow microfossils present in these samples prevents conclusions about a possible effect. Four samples that appear fractured in the
permeability measurements are marked in (a), and do not appear to have lower P-wave moduli than the unfractured samples.

Figure 15 Backscatter micrographs of samples with hollow microfossils to the left and with filled microfossils to the right. Each image represents
800 µm × 600 µm. Together with fossil fragments, the microfossils are termed large grains (visible in thin section and larger than 316 µm2 in
image grain-size analysis). Large grains are seen as white areas, large porosity as black areas and the matrix particles are not resolved. Matrix
particles are nannofossils, small fossil fragments and insoluble residue of silt and clay size (smaller than 3.16 µm2 in image grain-size analysis).


C 2005 European Association of Geoscientists & Engineers, Geophysical Prospecting, 53, 481–496
494 B. Røgen et al.
Figure 16 Empirical models and measured data of moduli versus he-
lium porosity for dry mudstone samples (less than 10% grains). In
black: plug data, modified upper Hashin–Shtrikman model (MUHS)
for plug data (full line) and linear regression (dotted line). In grey: Dan
Field log data and corresponding modified upper Hashin–Shtrikman
model (Japsen et al. 2000). Correlation coefficients between any of
the three models (MUHSplug , MUHSlog , linear regression) and mea-
sured values of elastic moduli of core plugs are 0.93, 0.96 and 0.95
for bulk (K), shear (G) and P-wave modulus (M), respectively. Error
bars indicate the estimated standard deviations of measurements of K,
G and M; these are 1.84 GPa, 0.41 GPa and 1.76 GPa, respectively.

C 2005 European Association of Geoscientists & Engineers, Geophysical Prospecting, 53, 481–496
hydrocarbon fields in the North Sea, we conclude that porosity
is the main factor controlling the ultrasonic velocity of North
Sea chalk, and that a decrease in ultrasonic velocity due to
large grains and smectite can be detected as a secondary factor.
We also conclude that Gassmann’s relationships are applicable
to ultrasonic measurements on North Sea chalk.
Gassmann’s relationships give good estimates of the elastic
moduli of dry samples when predicted from ultrasonic V P and
V S data from water-saturated samples. For 32 samples within
the porosity range 14–45%, dry V P and V S are underestimated
by 2% and 4% on average, respectively, with 10% maximal
error.
Within the porosity range 14–45%, samples from the South
Arne Field are found to have slightly higher elastic moduli
than samples from the Dan Field, 90 km away. This observa-
tion is surprising, considering the high carbonate content of
all samples, and we suggest that the difference is related to
textural differences between the chalk at the two locations, as
indicated by differences in the large-grain content (microfos-
sils and fossil fragments) and differences in clay mineralogy.
For a given velocity, the presence of filled microfossils indi-
cates lower porosity. For a given velocity, samples with smec-
tite tend to have lower porosity than samples with kaolinite.
Indurated water-zone samples from the Gorm Field with pore-
filling cement have relatively high P-wave velocities for a given
porosity.
Comparison of ultrasonic V P and V S data shows negative
correlation between porosity (18–45%) and ultrasonic veloc-
ity for dry as well as water-saturated chalk samples. A modi-
fied upper Hashin–Shtrikman model was fitted to the data for
37 dry samples classified as mudstone, but linear regression of
moduli and porosity data gave an equally good fit. Within the
porosity range 14–45%, the V P /V S ratio is found to show neg-
ative correlation with porosity for dry samples, and positive
correlation with porosity for water-saturated samples. This
difference between the V P /V S ratio of dry and saturated sam-
ples indicates that fluid differences may be detected through
amplitude-versus-offset analyses of seismic data. The results
of this study indicate that separation of porosity and fluid ef-
fects may be possible on seismic data obtained from chalk.
This could have important implications for exploration and
reservoir monitoring of hydrocarbon reservoirs in chalk.
ACKNOWLEDGEMENTS
This paper presents results from the ‘Rock Physics of Chalk’
research project. The project was funded by the Danish Energy
Research Programme, EFP-98. Core material, as well as log

data and seismic data from recent wells, were kindly provided
by the operators of the Dan and the South Arne Fields, Mærsk
Olie og Gas AS and Amerada Hess A/S, respectively. For tech-
nical assistance, we thank the staff at GEUS Core Laboratory,
at the Department of Physics, DTU, and at Environment &
Resources, DTU. We also thank Mark Sams, Kelvin Gylden-
holm and an anonymous reviewer for thorough reviewing and
suggestions for improvements.
REFERENCES
Assefa S., McCann C. and Sothcott J. 2003. Velocities of compres-
sional and shear waves in limestones. Geophysical Prospecting 51,
1–13.
Borre M. 1998. Ultrasonic velocity of North Sea Chalk –predicting sat-
urated data from dry. In: Nordic Petroleum Technology Series: IV,
Research in Petroleum Technology (ed. M.F. Middleton), pp. 71–
98. Nordisk Energiforskningsprogram, Ås, Norway.
Borre M. and Fabricius I.L. 1998. Chemical and mechanical processes
during burial diagenesis of chalk: an interpretation based on spe-
cific surface data of deep-sea sediments. Sedimentology 45, 755–
769.
Borre M. and Fabricius I.L. 2001. Ultrasonic velocities of water satu-
rated chalk from Gorm field, Danish North Sea: sensitivity to stress
and applicability of Gassmann’s equation. In: Nordic Petroleum
Technology Series: V, Research in Petroleum Technology (ed.
I.L. Fabricius), pp. 1–18. Nordisk Energiforskningsprogram, Ås,
Norway.
Brunauer A., Emmet P.H. and Teller E. 1938. Adsorption of gases in
multimolecular layers. Journal of the American Chemical Society
60, 309–319.
Castagna J.P., Batzle M.L. and Kan T.K. 1993. Rock physics –
the link between rock properties and AVO response. In: Offset-
Dependent Reflectivity – Theory and Practice of AVO Analysis (eds
J.P. Castagna and M.M. Backus), pp. 135–171. Investigations in
Geophysics 8. Society of Exploration Geophysicists.
Dunham R.J. 1962. Classification of carbonate rocks according to de-
positional texture. In: Classification of Carbonate Rocks (ed. W.E.
Ham), pp. 108–123. Memoir 1. American Association of Petroleum
Geologists.
Fabricius I.L. 2003. How burial diagenesis of chalk sediments
controls sonic velocity and porosity. AAPG Bulletin 87, 1755–
1778.
Gassmann F. 1951. Über die Elastizität poröser Medien. Viertel-
jahrsschrift der Naturforschenden Gesellschaft in Zürich 96, 1–
23.
Gommesen L., Mavko G. and Mukerji T. 2002. Fluid substitution
studies for North Sea chalk logging data. 72nd SEG meeting, Salt
Lake City, USA, Expanded Abstracts, 340–343.
Han D., Nur A. and Morgen D. 1986. Effects of porosity and clay
content on wave velocities in sandstones. Geophysics 51, 2093–
2107.
Hurst C. 1983. Petroleum geology of the Gorm field, Danish North
Sea. In: Petroleum Geology of the Southeastern North Sea and the

C 2005 European Association of Geoscientists & Engineers, Geophysical Prospecting, 53, 481–496
Ultrasonic velocities of North Sea chalk samples 495
Adjacent Onshore Areas (eds J.P. Kaasschieter and T.J.A. Reijers),
pp. 157–168. Geologie en Mijnbouw 62. Royal Geological and
Mining Society of the Netherlands, The Hague.
Hvid J.M. 1998. Influence of depositional texture and porosity on
ultrasonic wave propagation in Danian Limestone from Eastern
Denmark. In: Nordic Petroleum Technology Series: IV, Research in
Petroleum Technology (ed. M.F. Middleton), pp. 125–154. Nordisk
Energiforskningsprogram, Ås, Norway.
Japsen P. 1998. Regional velocity-depth anomalies, North Sea chalk.
A record of overpressure and Neogene uplift and erosion. AAPG
Bulletin 82, 2031–2074.
Japsen P., Bruun A., Fabricius I.L., Rasmussen R., Vejbcek O.V.
and Pedersen J.M. et al. 2004. Influence of porosity and pore
fluid on acoustic properties of chalk: AVO response from oil,
South Arne Field, North Sea. Petroleum Geoscience 10, 319–
330.
Japsen P., Wagner H., Gommesen L. and Mavko G. 2000. Rock
physics of chalk: modelling the sonic velocity of the Tor Formation,
Danish North Sea. 62nd EAGE conference, Glasgow, Scotland, Ex-
tended Abstracts, P0170.
Jørgensen L.N. 1992. Dan Field; Denmark, Central Graben, Danish
North Sea. In: AAPG Treatise of Petroleum Geology, Atlas of Oil
and Gas Fields (eds N.H. Foster and E.A. Beaumont), pp. 199–218.
American Association of Petroleum Geologists.
Lind I. 1997. A modified Wyllie equation for the relationship between
porosity and sonic velocity of mixed sediments and carbonates from
the Caribbean Sea. In: Nordic Petroleum Technology Series: III,
Research in Petroleum Technology (ed. M.F. Middleton), pp. 123–
137. Nordisk Energiforskningsprogram, Ås, Norway.
Mackertich D.S. and Goulding D.R.G. 1999. Exploration and ap-
praisal of the South Arne field, Danish North Sea. In: Petroleum
Geology of Northwest Europe: Proceedings of the 5th Conference
(eds A.J. Fleet and S.A.R. Boldy), pp. 959–974. Geological Society,
London.
Mavko G., Mukerji T. and Dvorkin J. 1998. The Rock Physics Hand-
book. Cambridge University Press.
Mortensen J., Engstrøm E. and Lind I. 1998. The relation among
porosity, permeability, and specific surface of chalk from the Gorm
Field, Danish North Sea. SPS Reservoir Evaluation and Engineer-
ing, 245–251.
Nur A., Mavko G., Dvorkin J. and Galmudi D. 1998. Critical poros-
ity: a key to relating physical properties to porosity in rocks. The
Leading Edge 17, 357–362.
Raiga-Clemenceau J., Martin J.P. and Nicoletis S. 1988. The con-
cept of acoustic formation factor for more accurate porosity deter-
mination from sonic transit time data. The Log Analyst 29, 54–
60.
Røgen B. and Fabricius I.L. 2002. Influence of clay and silica on per-
meability and capillary entry pressure of chalk reservoirs in the
North Sea. Petroleum Geoscience 8, 287–293.
Røgen B., Gommesen L. and Fabricius I.L. 2001. Grain size distribu-
tions of chalk from image analysis of electron micrographs. Com-
puter and Geosciences 27, 1071–1080.
Scholle P.A. 1977. Chalk diagenesis and its relation to petroleum ex-
ploration; oil from chalks, a modern miracle? AAPG Bulletin 61,
982–1009.
496 B. Røgen et al.
Walls J.D., Dvorkin J. and Smith B.A. 1998. Modeling seismic velocity
in Ekofisk Chalk. 68th SEG meeting, New Orleans, USA, Expanded
Abstracts, 1016–1019.
Wang Z. 1997. Seismic properties of carbonate rocks. In: Carbonate
Seismology (eds I. Palaz and K.J. Marfurt), pp. 29–52. Geophysical
Developments Series 6. Society of Exploration Geophysicists.
Wang Z. 2000. The Gassmann equation revisited: comparing labo-
ratory data with Gassmann’s predictions. In: Seismic and Acous-
tic Velocities in Reservoir Rocks: 3, Recent Developments (eds Z.
Wang and A. Nur), pp. 8–23. Geophysics Reprint Series 19. Society
of Exploration Geophysicists.

C 2005 European Association of Geoscientists & Engineers, Geophysical Prospecting, 53, 481–496
Wilkens R.H., Simmons G., Wissler T.M. and Caruso L. 1986. The
physical properties of a set of sandstones – Part III. The effects of
fine-grained pore-filling material on compressional wave velocity.
International Journal of Rock Mechanical Mining Science and Ge-
omechanics Abstracts 23, 313–325.
Wyllie M.R.J., Gregory A.R. and Gardner G.H.F. 1958. An experi-
mental investigation of factors affecting elastic wave velocities in
porous media. Geophysics 23, 459–493.
Yan J. 2002. Reservoir parameter estimation from well log and core
data: a case study from the North Sea. Petroleum Geoscience 8,
63–69.