Observing Maturing Source Rocks on Seismic Reflection

Data

T. Matava1 , R.G. Keys2 , S.E. Ohm3 , S. Volterrani4

ABSTRACT
Hydrocarbon generation in a source rock is a complex, irreversible phase change that
occurs when a source rock is heated during burial to change phase to a fluid. The fluid
density is less than the kerogen density so in a closed or partially closed system the
volume of the pore space occupied by fluids increases. Burial also increases the e↵ective
stress which leads to compaction and a significant reduction in porosity. The challenge
of identifying source rocks on seismic data then becomes di↵erentiating the smaller
porosity increase due to hydrocarbon formation from the larger porosity decrease as-
sociated with burial. We use a calibrated rock physics model to show that Vshale and
porosity data can be used to predict the compressional and shear wave velocities and
the density in wells over large sedimentary sections, including a source rock of variable
maturity. These well data and models show that the di↵erence between an immature
and mature source rock is an increase porosity (lower density) relative to compacting,
non-source rock sediments. We use these results to identify a potential source interval
in the Orphan Basin in Eastern Canada on 2D regional seismic data. We show that the
full stack amplitude response of a maturing source rock is significant during the main
phase of generation (0.2<transformation ratio<0.8) relative to surrounding sediments.
Regional scale consistency of the amplitude response with the kerogen maturity model
from an integrated basin simulator reduces exploration risk because the independence
of the thermal model from the seismic amplitude response. Finally, combining the
seismic response with the source rock maturity model provides insight into the likely
kerogen kinetics. Most applications require regional data sets to capture the matu-
rity window, however, applications are also possible around allochthonous salt where
geometries can lead to local changes in heat flow.

INTRODUCTION

Hydrocarbon generation from a kerogen in a source rock is expressed as the source rock
potential and is determined through the laboratory analysis of samples from wells (either
cuttings or core) or rock outcrops to measure the total organic carbon, TOC, and Hydro-
gm Carbon
carbon Index, HI (Law, 1999). The units of TOC are gm dry rock while HI has units of
mg Hydrocarbon
gm Carbon . The source rock potential scaled to useful units has the form
✓ ◆
3 ⇢dry
Volumehc (Barrels) = 6.28 x 10 (TOC)(HI) zA (1)
⇢hc

where ⇢ kg m 3 is the density, and z(m) is the source rock thickness and A(m2 ) is the
areal extent. Subscripts dry and hc refer to the dry bulk properties of the sediment and
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

the hydrocarbon fluid, respectively. Equation 1 is the basis for all hydrocarbon volume
calculations and is suitable for qualitatively examining the source potential in a basin, but
it is lacking for quantitative analysis because the transformation of a load bearing kerogen
to a fluid is not included in the analysis.
A change in phase of kerogen to a fluid is a complex set of parallel reactions in which the
kerogen and fluid properties vary through time. Two factors important to consider in this
phase change are the density di↵erence between the solid and fluid, and the temperature at
which the phase change occurs. The density is important because kerogen density, which
increases with increasing maturity (at constant stress), is significantly greater than the fluid
density. The density contrasts, when they occur in a closed system at constant stress, lead
to a significant increase in fluid volume and total volume. These density contrasts a↵ect
the mechanical properties of the volume including the bulk modulus and shear modulus
and, therefore, the seismic velocity of the volume. Temperature is an important factor
because the generation rate of hydrocarbons depends only on temperature and heating rate
(Appendix A). The temperature range over which hydrocarbon generation occurs is large.
Vernik and Nur (1990, 1992) previously identified through laboratory experiments that
the bulk density of the Bakken Formation, an important source rock in the Bakken Play,
correlates with TOC and that velocities are highly anisotropic. Their experiments made
clear the large di↵erences in fluid and kerogen density; however, their experiments were
performed at constant maturity and suggested that these volume changes may crack or
fracture the rocks to some degree. Carcione and Helle (1998) and Carcione (2000) develop
physical models for source rock velocity and attenuation that depends strongly on the
bedding angle. These models address the contrasts between the kerogen elastic moduli,
and shale and sediment moduli to show that source rocks could be observed in sediments
with reflection data. These models were largely isothermal and when experiments were
conducted at a non-constant temperature there was no control for hydrocarbon generation.
A number of laboratory methods have been developed which use detailed models of
kerogen distribution in a source to model the source rock elastic moduli to calculate bulk or
e↵ective elastic moduli. These models vary the moduli of the kerogen at di↵erent stages of
maturity to provide a measure of the bulk moduli as a function of temperature and maturity
(Carcione and Avseth, 2015; Zhao et al., 2016; Ibrahim and Mukerji, 2017; Suwannasri et al.,
2018a,b). Most of these methods generally use rock physics models to calculate the moduli
which means that models have the advantage of being bound by limits in porosity and in
certain cases velocities can satisfy Voight-Reuss or Hashin-Shtrikman bounds. For example,
Carcione and Avseth (2015) develop rock physics templates to calculate the elastic response
to maturing kerogens in North Sea wells in order to predict seismic attributes of source rocks
in a closed system. Kinetics with temperature dependent rate laws (Appendix A) are used
to model maturity changes of the source rocks, and smectite to illite in order to predict the
change in water, oil and gas saturation through time to a steady temperature gradient and
constant sedimentation rate in a closed system. The smectite to illite phase transformation is
a source of fluid which increases the pore pressure with increasing temperature and changes
in source rock maturity changed the oil and gas saturation. The smectite to illite phase
transformation also increases the sti↵ness of the sediment. Carcione and Avseth (2015)
do not explicitly include a constitutive in their calculations in order to relate changes in
porosity to changes in e↵ective stress through a compaction law. The material property
changes (e.g., rock and fluid moduli) to pressure and temperature changes are constant
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

through time by assuming a constant burial rate. The remaining models have similar limits
of application. Each of these models include a priori assumptions on the kerogen kinetics
and utilize correlations to estimate values of the the sti↵ness tensor. Finally, the rock
physics templates that result from these models are applied to well data only and synthetic
seismograms are not constructed to tie these models to seismic data at wells.
House and Edman (2019) suggest the porosity should change during kerogen conversion
and that overpressure should develop in the pore space. Seismic response is used with
neural networks to estimate geochemical properties of the source rock. Their interpretation
assumes a constant stress on the source rock.
Løseth et al. (2011, 2016) present seismic and well data which show the amplitude
response of several source rocks in the Norwegian North Sea varied with TOC. For example,
Løseth et al. (2011) first show that source rocks with thicknesses > 20 m could be identified
using only acoustic impedance and then show that the impedance value correlates with
TOC. The data presented by Løseth et al. (2011) correlates the amplitude response to TOC
but the depth range of the data is from 2-4.5 km, which suggests present day temperatures
range from 70-160 o C (Brigaud et al., 1992) or a vitrinite equivalent maturity range of
0.5-1.0 Ro(%). This maturity range is nearly the entire oil window. Løseth et al. (2011)
did observe a strong amplitude response with o↵set, but it is not known how maturity may
a↵ect this angle response.
In summary, a number of laboratory methods have been developed to model elastic
moduli and density of source rocks during the phase transformation from a solid kerogen to
a fluid. These methods assume either an isothermal system with variable e↵ective stress or a
constant stress system with variable temperature: none of these methods use a system with
independent changes in both temperature and e↵ective stress. Additionally, with exception
of Løseth et al. (2011) and House and Edman (2019), none of these methods are applied to
seismic data sets.
In this paper we first develop a mass balance in a representative elemental volume (REV)
of a source rock to show fundamental controls on volume due to hydrocarbon generation.
This section establishes a connection between the geochemical description of source rock
quality (TOC and HI) and the reaction kinetics which is a mathematical description of
source rock maturity (Appendix A). In the second section we use well data from the Spekk
formation in the Norwegian North Sea to show the combined e↵ects of compaction and
hydrocarbon generation on the rock properties. Rock physics models developed for these
wells show consistency between the velocity and density in a sediment package containing a
source rock a↵ected by heating during burial. Finally, an example from the Orphan Basin
in eastern Canada shows how these methods can be used to identify a source rock on seismic
data where porosity is a↵ected by both compaction and hydrocarbon generation.

MASS BALANCE OF HYDROCARBON GENERATION IN SOURCE

ROCKS

The geochemical approach of hydrocarbon potential (equation 1) is a bulk volume expression

which does not address the reaction rate, the kinetics of the reaction, nor does it lead to an
understanding of how generation of hydrocarbons can be observed in reflection seismic data.
Mass balances on a representative elemental volume (REV) are required to understand the
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

change in density associated with hydrocarbon generation.

The total mass, MT otal , in an REV is expressed as a sum of the mass of each component
in each phase in the REV

MT otal = Mrock + Mkt + Mhc

t
+ Mw , (2)

where superscript t refers to time (or o for initial values) and the subscript of components
in the solid phase are identified with subscripts k (kerogen), and rock (rock), and liquid
components are identified as w (water), and hc (hydrocarbon). Superscripts for time on
the total, rock and water masses are not included because they are assumed to be constant
in the closed system. Only the mass of the kerogen and hydrocarbons change in the closed
system and the hydrocarbon phases could be gas and or liquid but the particular phase
is not important for this analysis. Because the total mass is constant, equation 2 implies
that Mkt + Mhct = M o + M o . More detailed calculations could address miscibility between
k hc
components in phases with an Equation of State and in some cases this is important. For
example, miscibility of methane in water, is important for dry gas cases such as biogenic
gas (Duan et al., 1992).
Hydrocarbon generation is a one way, irreversible, reaction that depends only on time
and temperature (⌧ ). In the pre-generative state we assume no generated hydrocarbon
o = 0, until burial leads to
fluids are in place and no migrated hydrocarbons are present, Mhc
heating and the generation of hydrocarbons. Once kerogen begins to form a hydrocarbon
fluid, the mass of hydrocarbons is expressed as
t
Mhc = Mko Mkt
✓ ◆
Mkt
= Mko 1 (3)
Mko
= Mko R⌧t ,

where R⌧t is the transformation ratio at time t. The transformation ratio is a measure of
the conversion of kerogen to hydrocarbon fluids and is bounded by 0 and 1. When R⌧t = 0,
no hydrocarbons have been generated and when R⌧t = 1 then generation of hydrocarbons
is complete. When 0 < R⌧t < 1, continuous heating ( d⌧ dt > 0) is required for the reaction
t
to proceed and dR dt
⌧
0, Appendix A, which means that both Mkt and R⌧t are monotonic
functions of temperature. Finally, this definition for R⌧t is the same as in the general case
in which multiple fluid components are generated from a kerogen each with di↵erent kinetic
parameters. The general case is described in detail in Appendix A along with an example
of R⌧t (⌧ (t)) for constant heating rate d⌧
dt = C .

Conservation of mass on a closed REV requires that the total mass remain constant
during the phase change of kerogen to a fluid, or

MT otal = Mrock + Mkt + Mw + Mhc

t

MT otal = Mrock + Mko (1 RTt ) + Mw + Mko RTt . (4)

equation 4 can be written in terms of the time varying bulk density ⇢tb

1 ⇥ ⇤
⇢tb = ⇢rock Vrock + ⇢ok Vko (1 RTt ) + ⇢w Vwt + ⇢thc Vhc
t
, (5)
VTt otal
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

where V refers to the volume and

VTt otal = Vrock + Vkt + Vw + Vhc

t
. (6)

Introducing ratios of component volume fractions, , to simplify equation 5

t Vrock
rock =
VTt otal
t Vkt
k =
VTt otal
t Vw
w = (7)
VTt otal
Vhc t
t
hc =
VTt otal
t t t t
1 = rock + k + w + hc .

Each volume fraction has a superscript of time even though the mass may remain constant.
This is required because the total volume is no longer constant through time. The volume
fraction occupied by the fluid components, the porosity, is t = tw + thc . The volume
fraction of solid components is 1 t = t t
rock + k . Finally, substituting equation 7 into 5
leads to
⇢tb = ⇢rock 1 t
k
t
+ ⇢tk tk + ⇢w tw + ⇢thc thc . (8)

The mass of the closed system remains constant during the phase change; however, the
volume is not constant because kerogen changes phase to a fluid. The change in volume is
only associated with the kerogen and hydrocarbon densities, therefore:

V VTt otal VTootal

=
VTootal VTootal
Vkt + Vhct Vko
=
VTootal
 t t
1 Mk Mhc Mko
= +
VTootal ⇢tk ⇢thc ⇢ok
 t
1 Mk Mko RTt Mko
= +
VTootal ⇢tk ⇢thc ⇢ok

1 Mko ⇢ok Mkt ⇢o
= o o t o + tk RTt 1
VT otal ⇢k ⇢k Mk ⇢hc
 o
Vko ⇢k t ⇢ok t
= 1 R ⌧ + R 1
VTootal ⇢tk ⇢thc ⌧
✓ o ◆ ✓ ◆
o ⇢k ⇢ok t ⇢ok
= k R⌧ 1 . (9)
⇢thc ⇢tk ⇢tk

When R⌧t ⇡ 0, which corresponds to early maturity, the kerogen density changes are small
⇢o
and ⇢kt ⇡ 1. Therefore, V ⇡ 0. In the case that R⌧t ⇡ 1, which corresponds to late
k
⇢o ⇢o ⇢o
maturity, then ⇢kt > 1, but ⇢tk > ⇢kt , which means that V > 0 and is consistent with the
k hc k
total volume increasing on generation of hydrocarbons.
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

The change in pore volume, the volume occupied by hydrocarbon and pore fluids, can
be calculated to determine the e↵ect of relatively dense kerogen changing phase to a fluid.
Following a similar development as equation 9, the fractional change in volume occupied by
the pore fluids is expressed as
t + t o
hc w w
o
= o
w
✓ t ◆
1 Vhc t o
= + w w
o
w Vt
✓ T otal ◆
1 Vko ⇢ok t t o
= R + w
o
w VTt otal ⇢thc ⌧ w
✓ o ◆
1 VT otal Vko ⇢ok t
= R + tw o
o
w VTt otal VTootal ⇢thc ⌧ w

VTootal ok ⇢ok t t
w
= R ⌧ + 1
VTt otal ow ⇢thc o
w
VTootal ok ⇢ok t Vwt VTootal
= t t R ⌧ + 1
VT otal ow ⇢hc Vwo VTt otal
✓ o o ◆
VTootal k ⇢k t
= R⌧ + 1 1. (10)
VTt otal o ⇢t
w hc

Vwt
The ratio Vwo = 1 because the system is closed and the density of water is constant so the
Vo
volume of water is constant. The ratio VTt otal is calculated from equation 9 and equation 10
T otal
is similar to the total volume change during generation because the change in pore volume
in a closed system depends on the initial volume of the pore space occupied by pore fluid and
kerogen, the density di↵erence between kerogen and hydrocarbon and the transformation
ratio. Density di↵erences between kerogen and hydrocarbons depend on the particular
equation of states for these two components. In the case of the hydrocarbon fluids, they
depend on the kerogen kinetics (Appendix A) which controls the component feeds for the
fluids through time.
An example using equations 9 and 10 illustrates the interplay between kerogen volume,
fluid volume and total volume during the phase transformation of a kerogen to a fluid in a
closed system. Assume a saturated REV containing rock, kerogen and water and that rock
and water mass remains constant during heating such as in equations 2–8. The volume
fraction of the REV occupied by water and kerogen is 0.15 and 0.05. Additionally, assume
that the REV is heated at constant e↵ective stress and that kerogen transfers in bulk to a
single component hydrocarbon fluid. Finally, assume that the density of the kerogen and
hydrocarbons change linearly as the transformation ratio increases. For the kerogen density
changes, ranges are used that are consistent with Okiongbo et al. (2005) and Burnham
(2017).
Figure 1 shows the kerogen and fluid density changes with maturity as well at the
fractional change in the total volume and the change in volume occupied by fluids. The
total volume change is limited to ⇡ 4% initial volume because the kerogen only occupied
5% of the volume. The fraction of the pore volume occupied by water and hydrocarbon
fluids increases during generation from an initial value of 0.15 to 0.23 of the volume which
is an approximately a 50% increase in pore space in the volume. The increase volume of
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

the pore space occupied by fluids is due to the kerogen-fluid phase transformation and the
density di↵erence between the solid and fluid phases. In a real system, the REV is partially
open and hydrocarbons migrate to a trap; consequently, the change in pore volume is not
likely to be as large as in this example.
In summary, for a closed system at constant e↵ective stress, the mass balance on a
sediment, kerogen, hydrocarbon and water system, in which the solid kerogen is changing
phase to a hydrocarbon fluid, shows that generation increases the total volume of the REV
and the portion of the REV occupied by fluids, the porosity, also increases. In a partially
open physical system in which burial increases both the e↵ective stress and temperature, the
actual porosity change is a complicated relationship between e↵ective stress, temperature
and fluid composition which varies through time and depends on the kinetics and reaction
stoichiometry of the kerogen.

Wireline Logs Through a Source Interval

Burial leads to an increase in e↵ective stress and plastic, irreversible deformation of the
sediments but as we have shown the sediments undergoing burial are also a↵ected by changes
in temperature that leads to hydrocarbon generation in a source interval. Three wireline
logs penetrating the Spekk source rock in the Norwegian North Sea are used to identify
changes in the velocity and density of the sediments due to changes in e↵ective stress and
temperature. We model these changes with a single rock physics model that uses porosity
and Vshale to predict density, and compressional and shear velocity. These models allow us
to di↵erentiate compaction trends which decrease porosity from porosity increases resulting
from hydrocarbon generation. In our analysis we use the Xu-White rock physics model (Xu
and White, 1996) and the Keys and Xu (2002) dry rock approximation to the Xu-White
model. We refer to this rock physics model as the XW-DRA model. An assumption of the
XW-DRA model is that the pore spaces are connected and that fluids move in response to
gradients in e↵ective stress.
Logs from the Norwegian North Sea are available from the DISKOS Norwegian National
Data Repository and Geochemical Information (NPD, 2020). Three wells are used in this
study: 6407/9-1, 6406/2-1 and 6406/2-2 because they penetrate the Spekk source rock at
two di↵erent maturities. Well 6406/2-1 has a full set of logs to test the XW-DRA model
but does not have a log suite through the Spekk formation. Well 6406/2-2 is separated from
the 2-1 well by less than 5 km and also penetrates mature Spekk source rock (measured
vitrinite reflectance of 0.9-1.1%) but does not have a shear log. Well 6407/9-1 also does not
have a shear log but penetrates immature Spekk source rock (measured vitrinite reflectance
of 0.3%). Detailed maturity measurements for all three wells are shown in Appendix B.
Rock physics models were constructed for all three wells, Appendix C, using methods
outlined in Keys and Xu (2002) over the entire length of the logged intervals. A main result
of Appendix C is that the entire log interval can be modeled with a single rock physics
model that relates velocities and densities to the shale volume and porosity in the well.
This includes immature and mature sections of the Spekk source rock interval. In this
section we limit our interest to an approximately 500 m interval subset of these logs around
the Spekk source rock in wells 6407/9-1, Figure C3, and 6406/2-2, Figure C2 to better
understand the velocity, density and impedance changes associated with the Spekk source
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

Figure 1: Fractional change in volume and density in a closed REV as kerogen undergoes a
phase change from solid to a fluid at constant e↵ective stress. The left hand side are changes
in the total volume (equation 9) and the fluid volume (equation 10) and the density is the
right hand axis. Kerogen and hydrocarbon density are taken as linear with R⌧t . The initial
volume fraction of kerogen was ok = 0.05 and the initial porosity was ow = 0.15. The final
porosity is 0.23 suggesting a 50% increase in space occupied by the fluid, however, the total
volume only increased by ⇡ 4%.
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

rock at di↵erent maturities.

Figure 2 shows the velocity, density and impedance through the Spekk source rock in-
terval in well 6407/9-1. The Spekk source rock is immature in this well (Appendix B) and
trendlines show that the source interval follows an overburden compaction trend. Porosities
calculated in two di↵erent ways (Appendix C) also show a similar trend with the overbur-
den; however, there is more variability in these models. The DRA porosity (blue curve in
Figure 2) is the porosity calculated from the sonic log using the XW-DRA rock physics
model and the density porosity (black curve in Figure 2) from the density log. These data
indicate that a strong negative amplitude response is associated with the Spekk formation at
this location but this response is due to the properties of the Lyr formation rather than the
TOC in the Spekk formation which has been suggested as a control on amplitude response
Løseth et al. (2011).
Figure 3 shows the acoustic velocity, density and impedance measurements through the
Spekk source interval in well 6406/2-2. In this well the source rock is mature (Appendix
B) and compared to the 6407/9-1 well these data show that the overburden has compacted
significantly and approximately matches the Lyr formation, but the density of the Spekk
formation is now less than the overburden trend. There is only a small decrease in Spekk
velocity from the overburden trend. The acoustic impedance shows a significant negative
amplitude response will be associated with the Spekk interval, however, in this case the
density contrast is made more negative by the departure of the Spekk density from the
overburden trendline. We interpret this large decrease in density and smaller decrease in
velocity as a result of hydrocarbon generation which has increased the porosity of the source
rock interval through the transformation of load bearing kerogen to a fluid. The XW-DRA
model, Appendix C, shows that these processes are captured with the rock physics model
that relates an increase in porosity, resulting from a temperature driven phase change, to a
decrease in velocity and density.
The conversion of kerogen to a fluid is a complex, irreversible organic reaction involving
multiple hydrocarbon components. However, the ability to use a rock physics model to
capture the combined e↵ects of burial and a phase change suggests that the mechanical
response of the rock is a straightforward physical process. An increase in temperature
leads to the conversion of load bearing kerogen to a fluid which, in a closed REV, leads
to an increase in both the total volume and the fraction of the REV occupied by fluids.
In a partially open system a change in maturity reduces the e↵ective stress relative to
adjacent non-source rock sediment which makes the already negative amplitude response
of the source rock more negative. Additionally, the time derivative of equation 9 shows the
rate of increase in volume scales with the changing rate of maturity so a source rock rapidly
generating hydrocarbons should have a more negative reflection coefficient than source rock
slowly generating hydrocarbons.

Application: Orphan Basin

The mass balance formulation for a closed system and the well logs through the Spekk
formation at two distinct maturities show that an increase in porosity, a decrease in density,
leads to an amplitude response of a source interval during hydrocarbon generation. A
broad set of examples is possible to outline the utility of this approach to locating maturing
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

Figure 2: Data from the 6407/9-1 well in the Norwegian North Sea. Formations are labeled
with horizontal dashed lines. The red lines denote log data and the XW-DRA models are
removed from the velocity, density and impedance plates for clarity but they are shown
in Figure C3. The DRA porosity (blue curve on the porosity plate) is from the XW-DRA
model applied to the acoustic velocity. The density porosity (red curve on the porosity plate)
is the porosity calculated from the bulk density log. The caliper log is shifted uniformly by
145 units to fit into the Gamma Ray plot. Trend lines show that the velocity, density and
impedance of the Spekk formation is similar to the overburden. The Spekk source rock in
this well is immature.
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

Figure 3: Data for the 6406/2-2 well in the Norwegian North Sea. Formations are labeled
with horizontal dashed lines. The XW-DRA models are removed from the velocity, density
and impedance plates for clarity but they are shown in Figure C2. The DRA porosity
(blue curve on the porosity plate) is from the XW-DRA model applied to the acoustic
velocity. The density porosity (red curve on the porosity plate) is the porosity calculated
from the bulk density log. Trend lines show that both the velocity and density through the
Spekk formation are lower than the overburden trend. The Spekk source rock in this well
is mature.
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

source rocks but we focus on identifying an active source interval not currently considered
significant to show the existence of an additional petroleum system in the basin.
We have focused on the physical model to describe the processes occurring in a source
rock during generation; however, applying these methods to a petroleum system requires
a second physical model to describe the generation of hydrocarbons in a basin. Because
hydrocarbon generation is only a function of temperature, a thermal model is required
through geologic time to model the transformation of kerogen to a fluid. These models are
readily available as commercial software. The most rigorous of these models solve mass,
momentum and energy balances to forward model the temperature history in a basin using
a heat flow history at the base, a surface temperature history and Neumann conditions on
the side boundaries (Hantschel and Kauerauf, 2009). These physical models for tempera-
ture can be used to account for transient thermal e↵ects, such as cooling crust and rapid
sedimentation, which can lead to significantly di↵erent thermal histories than histories cal-
culated with steady state models. Temperature histories generated from a thermal model
are combined with equations A2–A3 to determine the transformation ratio, R⌧t , through
time.
Figure 4 shows the well data and the rock physics model (XW-DRA) for the Lona well
in the Orphan Basin. An approximately 100 m thick interval above the Cretaceous un-
conformity exhibits slow velocities, low density and high Poisson ratio which are attributes
commonly associated with source rocks. The rock physics model shows that the velocities
and density of a large section of the well, including a potential source rock section, can be
modeled using a single rock physics model using porosity and Vshale although in this loca-
tion the source rock is immature. Figure 4 shows that the below the unconformity the rock
physics model does not capture large shifts in the porosity and density. These are thought
to be zones containing calcite within the Jurassic section. Production in the Orphan Basin
is associated with a Jurassic source interval; consequently, little attention has been paid to
the Tertiary section in this area.
Approximately 450 km from the Lona well is the Baie Verte well that was drilled to test
a high below the Cretaceous Unconformity. The Baie Verte well report includes Rock-Eval
data, Figure 5, which shows that a fair to good quality source rock may be present above
the Tertiary unconformity. However, no additional TOC or HI data is available to confirm
the presence or quality of the source interval. The Baie Verte well was drilled in 1985 and
has only limited log data for rock physics modeling.
Figure 6 is a montage showing present day maturity of a Type II-III source rock from
a basin model constructed from a 2D seismic line, a plot of R⌧t extracted from the basin
model and amplitude on two reflectors in the 2D seismic data set, and maturity in plan
view for several 2D lines. The heavy black line on the 2D maturity model is the location of
the suspected source rock interval and is also where R⌧t model results are extracted from the
basin model. The heavy black line is also the location of the reflector amplitude from the
seismic data. Note that the source rock reflector amplitude closely tracks the transformation
ratio from the basin model. This is not the case for a slightly shallower amplitude. The
maturity in plan view is for three lines in the GrandSPAN seismic data set. The source
interval matures in the lows and transformation ratios closely follow the amplitude response.
The plan view of maturity, Figure 6, makes clear two additional points. First, the
area where the source interval is mature is proximal to a three dimensional seismic survey
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

Figure 4: Wireline logs and the Xu-White (dry rock approximation) rock physics model
fit to the data in the Lona well. Log data is in red and the XW-DRA fit to these data
is in blue. The log porosity data was computed from the density log and the XW-DRA
model was also used to compute the density and porosity. The Caliper log has been o↵set
to fit on the Gamma Ray plate and the two vertical lines are the limits used to compute
the Vshale . The base Tertiary unconformity is labeled with a horizontal dashed line. The
potential source rock just above 4500 m is identified only through its log character: slow
velocity, high porosity and Poisson ratio. Cuttings or other samples are not available for
analysis so identification of the source rock must be made inferred from seismic data.
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

Figure 5: Results of analysis from cutting samples from the Baie Verte well. The source rock
potential (S2 ) shows a gradual increase potential from 1000-3000 m depth, and produced
petroleum (S1 ) is very low. In the interval from 3000-3800 m source rock potential increases
to fair and good and at roughly 3300 m an increase in produced petroleum is observed
suggesting that the source rock has entered the generation window. Produced petroleum
is present down to the unconformity. The increase in S1 followed by concurrent decrease
in S2 follows the classical maturity profile for source rocks from immature to mature and
overmature, hence, the plot represents a strong circumstantial evidence for presence of a
source rock above the unconformity.
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

acquired to evaluate an approximately 650 km2 Class III AVO anomaly <2 km shallower
than the source interval. Current production in the Orphan Basin is from reservoirs sourced
from a Jurassic source rock and it is not possible to connect the Tertiary amplitude anomaly
to the Jurassic source rock with buoyancy driven migration. Figure 6 does not prove that
the amplitude response of the source interval and the AVO anomaly are connected, but the
proximity and current maturation suggests they are related. The second point to be made
is that it is difficult and cost prohibitive to acquire a 3D seismic survey that includes both
the AVO anomaly and mature zones of the suspected source rock. Regional two dimensional
data are required for this type of source rock analysis, but could easily be made part of a
regional petroleum system evaluation of a basin.

DISCUSSION

The well data shows that burial depth has a large impact on the sediment porosity. This
result is consistent with Hantschel and Kauerauf (2009) and Swarbrick (2012) in which
changes in vertical stress lead to compaction of the sediments. Figures 2 and 3 show
that ⇡3 km of sediment leads to an approximately 600 kg m 3 increase in density of the
sediment, or a loss of ⇡ 30 porosity units. Figures 2 and 3 shows that the porosity di↵erence
between the overburden and the Lange and Lyr formations decreases with increasing burial.
Trendlines to the data through the overburden and the Spekk formation show the e↵ects of
hydrocarbon generation on the velocity and density of the source rock in an immature case
and a mature case.
Our analysis shows that a load bearing kerogen undergoing a phase transformation
to form a fluid generates significant volume in the pore space. We observe this phase
change as a decrease in the reflector amplitude at the top of the source layer which can be
measured and modeled mainly as a decrease in density of the source interval. Figure 7 is
an illustration of the competing processes of porosity loss due to compaction and porosity
enhancement from hydrocarbon generation. The seismic reflection data indicate enhanced
porosity in the main phase of hydrocarbon generation and Figure 7 shows that hydrocarbon
generation leads to higher porosity than expected from compaction. This model for plastic
deformation of a source rock during burial and hydrocarbon generation contrasts with the
view that hydrocarbon generation leads to brittle failure of source rocks (Vernik and Nur,
1990). Additionally, because generation a↵ects porosity and the Poisson ratio of this interval
is elevated, the seismic response is limited to the full amplitude stack (Foster et al., 2010).
The Orphan Basin application combines the amplitude extraction of a known or sus-
pected source interval with a basin model and is essentially a combination of two physical
models. The phase transformation is the equivalent of a change in porosity that can be
used with a rock physics model to calculate the velocity and density response. The energy
balance and source rock kinetics incorporated in the basin model are used to calculate the
transformation ratio. The energy balance is constrained with well temperature data, heat
flow data and constraints on thermal conductivity. In the Orphan Basin, the two models
are combined to provide multiple lines of evidence for the existence of a source rock either
not previously observed, or not considered significant based on analysis of well data.
An interpretation from this work is that spatial changes in amplitude response of the
source rock suggests the kinetics may be more closely associated with a Type II-III source
rock than either a Type II or Type III source rock. If the source rock kinetics are assumed
 Matava et al (2021) in Press

Figure 6: A montage showing a present day maturity section for GrandSPAN Line 5800 along with a graph showing the amplitude
response of the source rock horizon and a shallower non-source rock horizon. This figure also shows, in plan view, the maturity models
of the source interval for other lines modeled as part of this study. The outline is of a Class III AVO anomaly <2 km shallower than the
source rock interval.
Observing Source Rocks on Seismic Data
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

to be Type II, then they are fast and the source rock maturity amplitude correlation is no
longer evident because the source rock is over mature. A similar conclusion is reached if
the source rock is assumed to have slow kinetics associated with a Type III source rock:
there should be no amplitude response because the source rock has not generated fluids
from the kerogen. Finally, the formulation of hydrocarbon generation (Appendix A) shows
that source rock quality is independent of source rock kinetics so the amplitude response
cannot be a measure source rock quality. However, a thin source interval with low TOC
and HI should be more difficult to observe than a thick source interval with high TOC and
HI.
Source rocks in frontier settings identified from an outcrop or well data can be upscaled
on the order of a billion times when applied to an entire basin. The ability to observe a
maturing source rock on seismic data can be used to demonstrate the existence of the source
rock away from well or outcrop control. Additionally, the transformation ratio depends
strongly on the kinetics of the source rock and these can be verified by comparing the
transformation ratio with the amplitude response. Finally, the rock physics models strongly
imply that the amplitude response is due to a change in porosity associated with maturity
that is independent of the basin model. The consistency between the amplitude response
suggested by the rock physics models and the maturity suggested by the basin models should
be interpreted as a reduction, but not removal, of exploration risk.
Finally, a significant risk element in play and prospect analysis is timing, which includes
the risk that a trap and seal are present prior to hydrocarbon generation and that the source
and trap are connected by a migration pathway. Migration pathways are not observed on
seismic data so the ability to identify a maturing source proximal to a trap with a favorable
geometry so that fluids can buoyantly migrate to a trap reduces exploration risk. However,
these methods should not be seen as removing the timing risk element from a prospect or
a play.

CONCLUSION

We combine physical models for porosity enhancement during source rock generation and
physical models for the propagation of seismic energy in a porous medium to show that
source rocks can be identified on seismic data. We show that amplitudes events are asso-
ciated with an increase in porosity relative to the surrounding sediments. Because these
tools are based on physical models calibrated to wells, they are robust and can be used to
constrain additional properties of the source rocks such as kerogen kinetics. Applying these
methods as part of an exploration workflow can better define risk elements associated with
source rock presence and timing. In special cases, these methods can be used to identify
new exploration plays.

ACKNOWLEDGEMENTS

Ion Geophysical Corporation provided the seismic data library, seismic interpretation and
basin modeling software which was the basis for this work. We appreciate that Ion agreed
to allow Figure 6 to be published with their data. D. Derbechar and M. Friday provided
the regional seismic interpretation framework on which the basin model and detailed inter-
pretation is based.
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

Three anonymous reviewers provided comments that significantly improved the material
presented in this paper. Their comments allowed us to refine the presentation of this
material and the conclusions drawn from our work. We thank them for their e↵orts on our
behalf.

APPENDIX A: SOURCE ROCK KINETICS

A change in phase of kerogen to a fluid is a complex set of parallel reactions in which

the kerogen properties and the generated fluid properties vary through time. In a typical
hydrocarbon mixture there are a large number of hydrocarbon components with many of
these components existing as both aliphatic and aromatic components. Additionally, early
components generated from a kerogen are more aliphatic than later components which are
more aromatic.
The dominant component of a hydrocarbon mixture on a mole fraction basis is methane
and it exhibits the greatest control on the viscosity and density of a bulk hydrocarbon
fluid. Because the physical properties of the mixture are of main interest, the complexity
of hydrocarbon generation is simplified to assume there is one solid kerogen component
and a smaller number of fluid components called pseudo-components because they lump
components of the mixture. In terms of the fluid pseudo-components, the generation of a
fluid from the kerogen is modeled as a series of parallel reactions with the form
!1
Mk ! ↵ 1 MC 1
!2
Mk ! ↵ 2 MC 2 5
!3
Mk ! ↵ 3 MC 6 10 (A1)
!4
Mk ! ↵4 MC10+ ,

where M is the mass, ! is the reaction rate, and ↵ represents the stoichiometric coefficient.
Numerical subscripts refer to the reaction number and subscripts k refer to kerogen and
Cj represents a pseudo component of the fluid with a range in carbon numbers (e.g., C1
corresponds to methane, CH4 ). This reaction network shown equation A1 has been used
historically but lacks the additional constraint of an equilibrium condition in the form of
equality in the chemical potential between products and reactants (Denbigh, 1961; Lewis
and Randall, 1961; Helgeson et al., 2009).
The mathematical approach used to describe reaction rate, !, in equation A1 takes the
form of a first-order reaction dependent only on temperature and time and has the form
dMk dMk d⌧
= = ! i Mk (A2)
dt d⌧ dt
where ⌧ is temperature, t is time, and d⌧
dt is the heating rate. The mass of the kerogen in
equations A1 and A2 ties with the source rock potential in equation 1 to conserve mass;
however, other formulations exist which constrain the generation to the amount of hydro-
gen in the kerogen (Schmoker, 1994), but these methods are not used in most volumetric
calculations. The rate law in equation A2 has the form of a decaying exponential in ! and
time. Because temperature is the only variable a↵ecting the reaction, solutions to equa-
tion A2 take the form of an Arrhenius reaction with activation energies and pre-exponential
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

constants (Hantschel and Kauerauf, 2009). This model of kerogen transformation has been
in practice for more than 30 years (Sweeney and Burnham, 1990) despite no equilibrium
condition and the potential for significant bias in its application (Matava et al., 2019).
Equation A2 requires an additional suite of laboratory analyses on source rock samples
to develop the activation and pre-exponential energies associated with a particular source
rock, usually referred to as the source rock kinetics. The analyses can be performed on a
full network of reactions (Equation A1) or a smaller subset of pseudo-components or even
a set of bulk component reactions. Behar et al. (1992) is an example of an experimental
program to develop kinetics of pseudo-components for source rocks in di↵erent basins and
these are required for a more physical estimate of fluid properties (e.g., density, and gas
oil ratio) generated through time from a source rock and how these a↵ect the fluids in a
prospective trap.
The extent of the kerogen reaction to form a fluid is measured as the transformation
ratio or
Mkt
R⌧t = 1 , (A3)
Mko
where R⌧t is the transformation ratio, Mk is the mass kerogen from equation A2 and su-
perscripts t and o refer to time and the initial masses of kerogen. The transformation ratio
is bounded by 0 and 1 with R⌧t = 0 corresponding to the initial case prior to hydrocarbon
generation and R⌧t = 1 corresponding to the end of generation.
Two additional points are required to fully understand the reaction network and trans-
formation ratio, equations A1-A3 . First, the reaction network is a one way reaction that
requires d⌧
dt > 0 to proceed. Second, Dow (1977) shows that the phase change reaction
after a period of cooling does not re-start until the temperature attained during subsequent
burial exceeds the maximum temperature prior to the start of cooling. The implication of
Dow (1977) is that the mass of products and reactants in the reaction network and R⌧t are
single valued, or monotonic functions for all time.
Figure A1 shows the transformation ratio through time for a typical source rock us-
ing the kinetic approach described in equations A2 and A3 assuming a constant heating
rate. Hydrocarbon generation occurs over a large temperature range but the main phase
of generation occurs when 0.2 < R⌧t < 0.8. Also shown in this figure is generation rate.
The largest generation rate occurs in the main phase of generation. Di↵erent source rock
kinetics may shift the temperature range in which R⌧t occurs and they may widen or narrow
the temperature range of the main phase of generation, but the overall shape of the curve
will be the same.

APPENDIX B: MATURITY OF MEASUREMENTS IN WELLS

6406/2-1, 6406/2-2, AND 6407/9-1

Maturity of Norwegian North Sea wells was determined from direct vitrinite reflectance
measurements. These data show that the 6407/9-1 well, Figure B1, is early mature while
the 6406/2-1, 6406/2-2 wells, Figure B2, are near the end of the maturity window. Note that
vitrinite reflectance measurements of maturity are global measures and the actual maturity
of the Spekk source rock depends on particular kinetics of this source interval which may
be greater or less than suggested by vitrinite reflectance.
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

Figure 7: Schematic e↵ects of burial on source rock generation and compaction for a three
layer system. The generation rate is highest during the main phase of generation, 0.2 <
R⌧t < 0.8, and this leads to the greatest delay in compaction. In this model, when a source
rock becomes over mature and the generation rate decreases, the porosity in the source
interval returns to the compaction trend and the amplitude response is no longer visible.

Figure A1: Transformation ratio, R⌧t , of a typical source rock due to heating at a constant
rate. The main phase of generation occurs when 0.2 < R⌧t < 0.8 and is where the generation
rate is largest. This curve is similar for all source rock types regardless of initial TOC and
HI. However, it can be shifted to the left or right due and the temperature range of the
main phase of generation can be larger or smaller than shown here.
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

Figure B1: Measured reflectance values in the 6409/9-1 in the Norwegian North Sea. Error
bars are one standard deviation of the individual measurements. These data were down-
loaded from the Norwegian Petroleum Directorate website (NPD, 2020). The measured
reflectance of the Spekk Formation in this well is approximately 0.3% which indicates the
source rock is immature. This well is essentially a vertical hole.
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

Figure B2: Measured reflectance values in the 6406/2-1 and 6406/2-2 wells in the Norwegian
North Sea. Error bars are one standard deviation of the individual measurements. These
data were downloaded from the Norwegian Petroleum Directorate website (NPD, 2020).
The two wells are separated by less than 5 km and the Spekk Formation was at the same
depth in each well to within meters. The measured reflectance of the Spekk Formation in
this well is ranges from 0.9-1.1% indicating the Spekk formation is mature. Both wells are
essentially vertical holes.
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

APPENDIX C: ROCK PHYSICS MODELING

Rock physics models are created for three logs to show that the compressional- and shear-
wave velocities and densities from well data, including a source interval at di↵erent matu-
rities, can be modeled with a relatively simple set of constant rock moduli, variable Vshale
and porosity. Once the relationship between velocity and porosity is established at di↵erent
maturities in well data, then impedance changes with maturity can be predicted and the
source rocks can be identified on seismic data based on their amplitude response at di↵erent
maturities. In this appendix we focus on well data from the northern North Sea which
includes the Spekk source interval. We use the entire well so that the Spekk source interval,
the main area of interest, can be seen as a subset of the entire well. The Xu-White model
rock physics model (Xu and White, 1996) applied to these wells closely follows methods
developed by Keys and Xu (2002) which they call the dry rock approximation or XW-DRA.
Keys and Xu (2002) express the compressional and shear wave velocities as a function
of the dry rock properties and the porosity. For a pore space saturated with a single phase
fluid (water) these relations are as follows:

⇢b = ⇢rock (1 w) + ⇢w w
q
Gd = Go (1 w)
p
Kd = Ko (1 w)
⇣ ⌘2
Kd
1 Ko
K = Kd + w 1 w Kd
Kw + Krock + 2
Krock
s
K + 43 Gd
Vp =
⇢b
s
Gd
Vs = ,
⇢b

where ⇢ is the density, is the porosity, G is the shear modulus, K is the bulk modulus,
and V is the velocity. Subscripts b, rock, w, p, s refer to the bulk mixture (water and solid),
solid rock with no porosity, water, P wave and S wave, respectively. The subscript d refers
to dry properties which, for the shear properties, is the same for the dry solid and water
saturated solid. Superscripts p and q are polarization coefficients.
Xu and White (1996) and the Keys and Xu (2002) dry rock approximation assume the
rock moduli are constant with both stress and temperature. The velocity and density depend
only on the porosity and a mixing parameter, typically taken as the shale content or Vshale .
Well data calibrated to the XW-DRA is guaranteed to produce velocities that satisfy Voigt-
Reuss bounds and velocities and densities that tend to solid rock velocities as the porosity
goes to zero. These rather simple models are in contrast to rock physics templates that
require numerous assumptions to describe in detail the components of sediment mixtures
some of which may be undergoing a phase transformation during hydrocarbon generation.
In this work, the rock physics model was first calibrated to Well 4 from Keys and Foster
(2012) using a dry rock Poisson ratio of 0.1 and then applied to the three North Sea wells.
Pore aspect ratios and polarization coefficients of this calibration are shown in Table C1.
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

Table C2 shows the endpoint rock properties used in the models for the three wells. These
values are consistent with Mavko et al. (2009) and suggest that a portion of the sti↵ rocks
are composed of carbonates which is consistent with data presented by Morgans-Bell et al.
(2001) in the Kimmeridge source rock which is the Spekk equivalent but in the United
Kingdom. Unfortunately, these data sets are not useful for assessing relative amounts of
carbonates and clastics in the well.
The well data in Figure C2 and Figure C3 contain plates showing wireline logs and results
of the rock physics modeling. The impedance track is from the density and compressional
velocity. The porosity plate contains two tracks with the DRA porosity calculated from the
XW-DRA model (Keys and Xu, 2002) while the porosity from density is calculated from
the density log using
⇢b = ⇢rock (1 ) + ⇢f , (C1)
where ⇢ is the density and is the volume fraction of the pore space to the total volume.
Subscripts b, rock, and f refer to the bulk, solid rock and fluid densities, respectively. In
the case of the solid rock, a linear mixing rule with Vshale is used to combine the two solid
rock components according to ⇢rock = Vshale ⇢c + (1 Vshale )⇢s where the compliant (c)
and sti↵ (s) rock component densities are shown in Table C2. The solid elastic moduli are
calculated from endpoint moduli with the same Vshale , using Voigt-Reuss averages.
The wells in Figures C2-C3 were separated into three layers for the rock physics mod-
eling. The first layer is to the top of the Spekk formation, the second layer is the Spekk
formation and the third layer is below the Spekk formation. The gamma ray log shows the
endpoint values used for calculating Vshale and Table C2 shows the endpoint moduli and
densities for each layer. The well was separated into layers to match trends in the velocity,
density and gamma ray curves.
The workflow for calibrating each layer within each log is as follows:

• Calibrate the rock physics model to the Mobil 4 well by using the dry rock Poisson
ratio, ↵ and the endpoint rock properties to determine, p and q values for the XW-
DRA rock physics model.
• Calculate Vshale for each layer in the three wells used in this study. The gamma ray
log is used to with the limits shown for each layer for this calculation.
• Invert the sonic log for porosity using the XW-DRA rock physics model.
• Forward model Vs and ⇢b using the XW-DRA.

Well 6406/2-1 has both shear data and acoustic data available and shows that the rock
physics model can be used to model the velocity and density data in the well over ⇡1500
m. Figure C2 shows that the rock physics model tracks the two velocities but does not
capture the variability of low density material in the underburden layer. Well 6406/2-2,
separated from the 2-1 well by <5 km, shows that the same rock physics model improves on
the density calculation but is without a shear velocity to fully test the model. Both of these
wells cut mature Spekk formation (Figure B2) while the Spekk formation in the 6407/9-1
well was much shallower and is immature (Figure B1). The XW-DRA rock physics model
for Well 6407/9-1 also matches the log data well over ⇡1500 m and together these models
show that a single rock physics model can be used to model moduli, velocities and densities
from porosity to capture broad changes in rock moduli due to compaction.
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

Table C1: Pore aspect ratios (↵) and polarization coefficients (p and q) for compliant and
sti↵ pores used in the the rock physics model calibrated to Well 4 (Keys and Foster, 2012)
in the North Sea data set. Endpoint rock moduli and densities, dry rock Poisson ratio and
↵ were used calculate the p and q for the XW-DRA model created for this well.

Compliant Sti↵
↵ 0.04 0.13
p 4.3815 13.3221
q 4.2891 11.0235

Table C2: Rock properties for each of the three layers in the rock physics models. Over-
burden is the layer above the Spekk source rock. The source layer is the Spekk formation.
Underburden is the layer below the base of the source interval. Vshale is used as the mixing
rule for sand and shales. These same rock properties were used in each well. Densities of
the fluids and solids are denoted by ⇢, the bulk modulus is K and the shear modulus G.
Densities and elastic moduli are assumed as constant with temperature. Values in this table
are consistent with Mavko et al. (2009), and suggest the overburden is a mixture of shale
and quartz, the source interval is a mixture of shale and carbonate and the underburden is
a mixture of shale and quartz and carbonate.

Fluid Compliant Sti↵

Layer ⇢
⇣ f⌘ Kf ⇣ ⇢c ⌘ Kc Gc ⇣ ⇢s ⌘ Ks Gs
kg kg kg
m3
(GPa) m3
(GPa) (GPa) m3
(GPa) (GPa)
Overburden 1100 2.2 2650 25 9 2800 39 33
Source Interval 1100 2.2 2550 25 9 2725 70 30
Underburden 1100 2.2 2650 25 12 2750 60 50
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

Figure C1: Log data and rock physics models for the 6406/2-1. The horizontal dashed lines
are the top and base of the Spekk formation. This log contains a full suite of logs to test
the calibration of the rock physics model developed for Well 4 from Keys and Foster (2012)
and in cases where the well log cannot be observed it is because the model overlies the log.
The Caliper track has been o↵set to fit into the Gamma Ray plate and below 5100 m the
Caliper log is not available. Only one layer is used on this model, the underburden layer,
and the two vertical bands on the gamma ray curve are the limits used for calculating the
Vshale . The main purpose of this log is to show that the rock physics model calibrated to a
well in the southern part of the North Sea can be used with these wells to predict velocity
and density from the porosity.
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

Figure C2: Log data and rock physics models for the 6406/2-2. This well is separated from
the 6406/2-1 well less than 5 km and is logged through the Spekk interval but does not
have a shear log. The horizontal dashed line is the top and base of the Spekk formation.
The Caliper log is placed in the same plate as the Gamma Ray log and the vertical bars
show the endpoints for the Vshale calculation. Three layers were used in this model and the
two vertical bands for each layer are shown on the gamma ray curve. The main purpose of
this log is to show that the rock physics model provides a good prediction of the velocity
and density from the porosity. Figure B2 shows that the Spekk formation has a vitrinite
reflectance of approximately 1% and is mature.
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

Figure C3: Log data and rock physics models for the 6407/9-1. This well also does not
contain a shear log but cuts the Spekk formation at a shallow depth and the formation is
immature, Figure B1. The Caliper log is placed in the same plate as the Gamma Ray log
and the vertical bars show the endpoints for the Vshale calculation. Three layers were used
in this model and the two vertical bands for each layer are shown. Figure B1 shows that
the maturity of the Spekk formation in this well is 0.3% and is immature.
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

REFERENCES
Behar, F., S. Kressmann, J. Rudkiewicz, and M. Vandenbroucke, 1992, Experimental sim-
ulation in a confined system and kinetic modelling of kerogen and oil cracking: Organic
Geochemistry, 19, no. 1-3, 173–189.
Brigaud, F., G. Vasseur, and G. Caillet, 1992, Thermal state in the north Viking Graben
(North Sea) determined from oil exploration well data: Geophysics, 57, no. 1, 69–88.
Burnham, A. K., 2017, Porosity and permeability of Green River oil shale and their changes
during retorting: Fuel, 203, 208–213.
Carcione, J., and H. B. Helle, 1998, Seismic attributes of petroleum source rocks, in SEG
Technical Program Expanded Abstracts 1998: Society of Exploration Geophysicists,
1004–1007.
Carcione, J. M., 2000, A model for seismic velocity and attenuation in petroleum source
rocks: An acoustic model for petroleum source rocks: Geophysics, 65, 1080–1092.
Carcione, J. M., and P. Avseth, 2015, Rock-physics templates for clay-rich source rocks:
Geophysics, 80, no. 5, D481–D500.
Denbigh, K., 1961, The principles of chemical equilibrium: with applications in chemistry
and chemical engineering: Cambridge University Press.
Dow, W. G., 1977, Kerogen studies and geological interpretations: Journal of Geochemical
Exploration, 7, 79–99.
Duan, Z., N. Møller, J. Greenberg, and J. H. Weare, 1992, The prediction of methane
solubility in natural waters to high ionic strength from 0 to 250 C and from 0 to 1600
bar: Geochimica et Cosmochimica Acta, 56, 1451–1460.
Foster, D. J., R. G. Keys, and F. D. Lane, 2010, Interpretation of AVO anomalies: Geo-
physics, 75, no. 5, 75A3–75A13.
Hantschel, T., and A. I. Kauerauf, 2009, Fundamentals of basin and petroleum systems
modeling: Springer Science & Business Media.
Helgeson, H. C., L. Richard, W. F. McKenzie, D. L. Norton, and A. Schmitt, 2009, A chemi-
cal and thermodynamic model of oil generation in hydrocarbon source rocks: Geochimica
et Cosmochimica Acta, 73, 594–695.
House, N., and J. Edman, 2019, Developments relating total organic carbon conversion in
unconventional reservoirs to 3D seismic attributes: Unconventional Resources Technology
Conference, Denver, Colorado, 22-24 July 2019, Society of Exploration Geophysics, 413–
431.
Ibrahim, M. A., and T. Mukerji, 2017, in Thermal maturation e↵ects on the elastic prop-
erties of organic-rich mudrocks: 3955–3961.
Keys, R. G., and D. J. Foster, 2012, in 1. A Data Set for Evaluating and Comparing Seismic
Inversion Methods: Society of Exploration Geophysicists, 1–12.
Keys, R. G., and S. Xu, 2002, An approximation for the Xu-White velocity model: Geo-
physics, 67, 1406–1414.
Law, C. A., 1999, Treatise of petroleum geology/handbook of petroleum geology: Explor-
ing for oil and gas traps. chapter 6: Evaluating source rocks: American Association of
Petroleum Geologists Special Volumes.
Lewis, G. N., and M. Randall, 1961, Thermodynamics: Rev. by Kenneth S. Pitzer and Leo
Brewer. 2d ed: McGraw-Hill.
Løseth, H., L. Wensaas, M. Gading, K. Du↵aut, and H. M. Springer, 2016, Method of
assessing hydrocarbon source rock candidate. (US Patent 9,244,182).
Løseth, H., L. Wensaas, M. Gading, K. Du↵aut, and M. Springer, 2011, Can hydrocarbon
Matava et al (2021) in Press Observing Source Rocks on Seismic Data

source rocks be identified on seismic data?: Geology, 39, no. 12, 1167–1170.
Matava, T., V. Matt, and J. Flannery, 2019, New insights on measured and calculated
vitrinite reflectance: Basin Research, 31, no. 2, 213–227.
Mavko, G., T. Mukerji, and J. Dvorkin, 2009, The rock physics handbook, second ed.:
Cambridge University Press.
Morgans-Bell, H. S., A. L. Coe, S. P. Hesselbo, H. C. Jenkyns, G. P. Weedon, J. E. Mar-
shall, R. V. Tyson, and C. J. Williams, 2001, Integrated stratigraphy of the Kimmeridge
Clay formation (Upper Jurassic) based on exposures and boreholes in south Dorset, UK:
Geological Magazine, 138, 511–539.
NPD, 2020, Norwegian petroleum directorate-diskos: www.npd.no.
Okiongbo, K. S., A. C. Aplin, and S. R. Larter, 2005, Changes in type II kerogen density as
a function of maturity: Evidence from the Kimmeridge Clay formation: Energy & fuels,
19, 2495–2499.
Schmoker, J. W., 1994, Volumetric calculation of hydrocarbons generated: Chapter 19:
Part IV. identification and characterization, in The petroleum system: From source to
trap, AAPG Memoir 60: AAPG, 323–3–26.
Suwannasri, K., T. Vanorio, and A. Clark, 2018a, in Data-driven elastic modeling of organic-
rich marl during maturation: Society of Exploration Geophysicists, 3473–3477.
Suwannasri, K., T. Vanorio, and A. C. Clark, 2018b, in Monitoring the changes in elastic
and transport properties of Eagle Ford marl upon maturation: Society of Exploration
Geophysicists, 3593–3597.
Swarbrick, R., 2012, Review of pore-pressure prediction challenges in high-temperature
areas: The Leading Edge, 31, 1288–1294.
Sweeney, J. J., and A. K. Burnham, 1990, Evaluation of a simple model of vitrinite re-
flectance based on chemical kinetics (1): American Association of Petroleum Geologists
Bulletin, 74, 1559–1570.
Vernik, L., and A. Nur, 1990, Ultrasonic velocity and anisotropy of petroleum source rocks:
The bakken formation, in SEG Technical Program Expanded Abstracts 1990: Society of
Exploration Geophysicists, 845–848.
——–, 1992, Ultrasonic velocity and anisotropy of hydrocarbon source rocks: Geophysics,
57, 727–735.
Xu, S., and R. E. White, 1996, A physical model for shear-wave velocity prediction1: Geo-
physical prospecting, 44, 687–717.
Zhao, L., X. Qin, D.-H. Han, J. Geng, Z. Yang, and H. Cao, 2016, Rock-physics modeling
for the elastic properties of organic shale at di↵erent maturity stages: Modeling elastic
properties of organic shale: Geophysics, 81, no. 5, D527–D541.