AUTHORS Gareth D. Jones ExxonMobil Development Company, 800 Bell St, Houston, Texas 77002; [email protected]Gareth Jones is a carbonate reservoir characteriza- tion specialist currently working for ExxonMobil Development Company in a multidisciplinary team investigating reservoir connectivity in a giant Middle East field. This article was completed while he was leader of the Carbonate Pore Systems Predic- tion team at ExxonMobil Upstream Research Com- pany. He has a B.Sc. degree in geology (University of London), an M.Sc. degree in hydrogeology (University of Birmingham) and received his Ph.D. from the University of Bristol, where he studied numerical modeling of fluid flow and diagenesis in carbonate platforms. Yitian Xiao ExxonMobil Upstream Research Company, 3120 Buffalo Speedway, Houston, Texas 77098; [email protected]Yitian Xiao is a senior research geoscientist in the Carbonate Architecture and Pore Systems Prediction Project at ExxonMobil Upstream Research Company. His current research focuses on the application of reactive transport models to predict carbonate and siliciclastic reservoir quality. He also maintains a strong research interest in applying computational chemistry to model source rock maturation and oil and gas generation. He has a B.S. degree in geo- chemistry and an M.S. degree in geophysics from China and received his M.Ph. degree and his Ph.D. in geochemistry from Yale University. He is an adjunct faculty in the Earth Science Department at Rice University. ACKNOWLEDGEMENTS We thank ExxonMobil and Tengiz field business partners (Kazakhoil, Tengizchevroil, and LukArco) for granting permission to publish the results of this study. AAPG reviewers David E. Eby, Paul M. (Mitch) Harris, and Rick P. Major are thanked for their suggestions and supportive comments. Joel Collins and Paul Hicks provided insightful reviews for ExxonMobil prior to article submission. Many col- leagues have improved our understanding of Tengiz and helped provide the information necessary to constrain this reactive transport study. In particular, we thank Tom Anderson, Steve Bachtel, Dan Carpenter, Joel Collins, Brent Francis, Sean Guidry, Paul Hicks, Peter Hillock, and Jim Weber. We appreciate the leadership of James Anderson, Tom Frantes, and Sherry Becker in supporting this study and their efforts in securing permission to publish. Laura Parnell assisted with drafting the figures. Geothermal convection in the Tengiz carbonate platform, Kazakhstan: Reactive transport models of diagenesis and reservoir quality Gareth D. Jones and Yitian Xiao ABSTRACT A fundamental challenge in carbonate reservoir characterization is predicting the spatial distribution of diagenesis. We used a reactive transport model to investigate the viability of geothermal convec- tion and associated patterns of diagenetic porosity modification in the Tengiz isolated carbonate platform reservoir. Before burial, forced convection generates significant calcite dis- solution (locally up to 45%) toward the platform center, minor cal- cite cementation (up to 0.4%) in the slope, and moderate calcite dissolution and cementation (up to 1.6%) in Serpukhovian bound- stone convective cells. The patterns and rates of diagenesis proved critically sensitive to specified vertical permeability. After burial with 200 m (660 ft) of salt, modeled subsurface temperature contrasts drive platform-scale free convection. Flow is hydraulically closed, but significant dissolution, up to 7.3% after 20 m.y., occurred in the Serpukhovian and Visean platform interior, and minor cementation up to 0.7% occurred toward the margin. A shale-filled salt-withdrawal basin, 500 m (1640 ft) deep, signifi- cantly modifies the subsurface temperature distribution and free convective flow. Ascending groundwaters beneath the withdrawal basin created a zone of calcite dissolution (up to 24.5% in 20 m.y.), with a mushroom geometry and minor cementation (up to 2.3% in 20 m.y.) in the distal platform interior and margin. Rates of dia- genesis are dramatically reduced with increasing overburden as com- paction retards convective flow. From a generic perspective, free convection persists if the salt overburden is substituted with shale, although flow is reversed, resulting in a different distribution of diagenesis. AAPG Bulletin, v. 90, no. 8 (August 2006), pp. 1251–1272 1251 Copyright #2006. The American Association of Petroleum Geologists. All rights reserved. Manuscript received December 19, 2005; provisional acceptance February 23, 2006; revised manuscript received March 15, 2006; final acceptance April 3, 2006. DOI:10.1306/04030605194
22
Embed
Geothermal convection in the Tengiz carbonate …...Tengiz carbonate platform, Kazakhstan: Reactive transport models of diagenesis and reservoir quality Gareth D. Jones and Yitian
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
AUTHORS
Gareth D. Jones � ExxonMobil DevelopmentCompany, 800 Bell St, Houston, Texas 77002;[email protected]
Gareth Jones is a carbonate reservoir characteriza-tion specialist currently working for ExxonMobilDevelopment Company in a multidisciplinary teaminvestigating reservoir connectivity in a giant MiddleEast field. This article was completed while hewas leader of the Carbonate Pore Systems Predic-tion team at ExxonMobil Upstream Research Com-pany. He has a B.Sc. degree in geology (Universityof London), an M.Sc. degree in hydrogeology(University of Birmingham) and received his Ph.D.from the University of Bristol, where he studiednumerical modeling of fluid flow and diagenesis incarbonate platforms.
Yitian Xiao is a senior research geoscientist in theCarbonate Architecture and Pore Systems PredictionProject at ExxonMobil Upstream Research Company.His current research focuses on the application ofreactive transport models to predict carbonate andsiliciclastic reservoir quality. He also maintainsa strong research interest in applying computationalchemistry to model source rock maturation andoil and gas generation. He has a B.S. degree in geo-chemistry and an M.S. degree in geophysics fromChina and received his M.Ph. degree and his Ph.D.in geochemistry from Yale University. He is anadjunct faculty in the Earth Science Department atRice University.
ACKNOWLEDGEMENTS
We thank ExxonMobil and Tengiz field businesspartners (Kazakhoil, Tengizchevroil, and LukArco) forgranting permission to publish the results of thisstudy. AAPG reviewers David E. Eby, Paul M. (Mitch)Harris, and Rick P. Major are thanked for theirsuggestions and supportive comments. Joel Collinsand Paul Hicks provided insightful reviews forExxonMobil prior to article submission. Many col-leagues have improved our understanding of Tengizand helped provide the information necessary toconstrain this reactive transport study. In particular,we thank TomAnderson, Steve Bachtel, Dan Carpenter,Joel Collins, Brent Francis, Sean Guidry, Paul Hicks,Peter Hillock, and Jim Weber. We appreciate theleadership of James Anderson, Tom Frantes, andSherry Becker in supporting this study and theirefforts in securing permission to publish. LauraParnell assisted with drafting the figures.
Geothermal convection in theTengiz carbonate platform,Kazakhstan: Reactive transportmodels of diagenesis andreservoir qualityGareth D. Jones and Yitian Xiao
ABSTRACT
A fundamental challenge in carbonate reservoir characterization is
predicting the spatial distribution of diagenesis. We used a reactive
transport model to investigate the viability of geothermal convec-
tion and associated patterns of diagenetic porosity modification in
the Tengiz isolated carbonate platform reservoir.
Before burial, forced convection generates significant calcite dis-
solution (locally up to 45%) toward the platform center, minor cal-
cite cementation (up to 0.4%) in the slope, and moderate calcite
dissolution and cementation (up to 1.6%) in Serpukhovian bound-
stone convective cells. The patterns and rates of diagenesis proved
critically sensitive to specified vertical permeability.
After burial with 200 m (660 ft) of salt, modeled subsurface
temperature contrasts drive platform-scale free convection. Flow
is hydraulically closed, but significant dissolution, up to 7.3% after
20m.y., occurred in the Serpukhovian andVisean platform interior,
and minor cementation up to 0.7% occurred toward the margin. A
shale-filled salt-withdrawal basin, 500 m (1640 ft) deep, signifi-
cantly modifies the subsurface temperature distribution and free
convective flow. Ascending groundwaters beneath the withdrawal
basin created a zone of calcite dissolution (up to 24.5% in 20 m.y.),
with a mushroom geometry and minor cementation (up to 2.3%
in 20 m.y.) in the distal platform interior and margin. Rates of dia-
genesis are dramatically reduced with increasing overburden as com-
paction retards convective flow. From a generic perspective, free
convection persists if the salt overburden is substituted with shale,
although flow is reversed, resulting in a different distribution of
diagenesis.
AAPG Bulletin, v. 90, no. 8 (August 2006), pp. 1251– 1272 1251
Copyright #2006. The American Association of Petroleum Geologists. All rights reserved.
Manuscript received December 19, 2005; provisional acceptance February 23, 2006; revised manuscriptreceived March 15, 2006; final acceptance April 3, 2006.
DOI:10.1306/04030605194
Simulations of geothermal convection provide a
physically viable model for the integration of direct dia-
genetic observations to augment predictions of reservoir
Figure 2. Hydrostratigraphicunits and model boundaryconditions prior to platformburial based on Figure 1. Dashedlines show location of twostratigraphic transects used inthis study for illustrating dia-genetic patterns. Note that al-though decompaction was notperformed in this study, ge-ometries generally reflect pre-compaction profiles.
1254 Geothermal Convection in the Tengiz Carbonate Platform, Kazakhstan
Rock and Fluid Properties
We specified a hydrostratigraphy in our platform-scale
numerical model based on the stratigraphic framework
and general distribution of facies summarized byWeber
et al. (2003). The Tengiz platform was subdivided into
four hydrostratigraphic units: (1) basement, (2) Devo-
nian, (3) Tournaisian andVisean, and (4) Serpukhovian
and Bashkirian (Figure 2). This stratigraphic frame-
work and the bathymetry-constrained environments of
deposition described by Weber et al. (2003) were used
to populate the model with four carbonate sediment
tion in the Tengiz platform (Figure 8). Oceanwaters are
drawn into the platform through the slope and discharge
through the submerged platform top (Figure 8A, B).
The hydrostratigraphy is a critical control on the style
andmagnitude of groundwater flow (Figure 8A, B). Flow
is focused in the Visean- to Bashkirian-age grainstone-
and packstone-dominated platform interior sediments,
where maximum horizontal velocities reach 2.6 m/yr
(8.5 ft/yr) (Figure 8B). In the muddier, more com-
pacted, and less permeableDevonian platform interior
sediments, the maximum flow velocity is 0.4 m/yr
(1.3 ft/yr), which reduceswith depth and distance from
the active platform margin to 0.01 m/yr (0.03 ft/yr)
(Figure 8B). In the platform slope and basinmudstones,
with relatively low permeability, the flow velocity is on
the order of 0.1 m/yr (0.3 ft/yr).
Three convection cells in the Serpukhovian bound-
stones modify the general flow pattern (Figure 8C).
The boundstone convection cells, which exhibit aspects
of both forced and free convection, can be described as
mixed (sensu Raffensperger and Vlassopoulos, 1999).
These margin convection cells extend vertically from
the ocean-sediment surface down to the top of the
Tournaisian andVisean stratigraphic unit (158–370m;
518–1213 ft) and have widths on the order of 500 m
(1640 ft) (Figure 8C). The cells are thus elliptical in
nature, with their long axes oriented in the horizontal
direction approximately parallel to bedding. These con-
vective cells are hydraulically open, with recharge and
discharge from and to the ocean allowing the replen-
ishment of reactants, which is important for mass
transport and diagenesis. Sensitivity simulations proved
the flow cell dimensions to be independent of numer-
ical grid-size refinements. The relatively high vertical
permeability of the microbial boundstones, specified to
characterize early syndepositional vertical fractures, is
Figure 7. Specified thermal conductivity-versus-porosity relationships for halite,limestone, and shale.
1258 Geothermal Convection in the Tengiz Carbonate Platform, Kazakhstan
the principal control on convection cell development in
the platform margin (Figure 8C). Sensitivity analyses
demonstrated that increasing the permeability anisot-
ropy of the microbial boundstones from 0.5 to 10,
reducing the vertical permeability, prevented convec-
tion cells from developing in the margin.
Convection dominates heat transport with geo-
thermally driven seawater, cooling the margin and
progressively warming along the flow path to elevate
temperatures in the platform interior (Figure 8A). Per-
meability, which is a critical control on flow rate and
the balance between conductive and convective heat
transport, can significantly modify the temperature dis-
tribution (Sanford et al., 1998). This is evident frommi-
nor temperature perturbations (<2jC) associated with
the convection cells developed in the Serpukhovian
Figure 8. Geothermal convection in the Tengiz platform prior to platform burial. (A) Temperature (contour interval of 2.5jC) and fluidvelocity vectors, which are not shown for every grid cell; (B) vertical Z fluid velocity vectors and temperature contours; (C) enlargementof free convection cell system in Serpukhovian boundstones. Fluid velocity arrows show only the direction, and the temperaturecontour interval is 0.5jC. Dashed yellow lines show the location of convection cells in the margin. (D) Calcite cement and dissolutionat 2 m.y. Note that to enhance visualization of diagenetic patterns, the color bar scale does not reflect the full range of the data, whichis discussed in the text. (E) Calcite cementation and dissolution along a Serpukhovian transect (see Figure 2) and (F) along a Viseantransect (see Figure 2).
vection diagenesis was critically dependent on the
specified vertical permeability in a similar manner to
Figure 9. Effect of reducing carbonate vertical permeability(compare to Figure 8). (A) Calcite cement and dissolutionat 2 m.y., temperature (contour interval of 2.5jC), and fluidvelocity vectors. (B) Calcite cementation and dissolution alongthe Serpukhovian transect (see text) and (C) along the Viseantransect. Note the reduction in dissolution in the distal platforminterior.
Jones and Xiao 1261
the unburied case (Figure 12). Reducing the vertical
permeability of grainstones, packstones, andmudstones
by one order of magnitude resulted in slower flow ve-
locities, less convection, and rates of calcite diagenesis
in the platform interior that were an order of magni-
tude or more lower than the higher vertical permeabil-
ity case (compare Figures 10, 12).
We tested the sensitivity of our results to the spec-
ified coordinate system by simulating geothermal con-
vection in a northeast-southwest transect of Tengiz
using Cartesian coordinates (Figures 1, 13). The salt
overburden was identical to that in Figure 10, and the
hydrostratigraphy is depicted in Figure 13A. Results
demonstrate that two large free convection cells that
were set up in the platform interior dominate flow.
Convective heat transport results in elevated tempera-
tures at the platform center, showing that this temper-
ature pattern, although accentuated with radial coordi-
nates, is not an artifact of the platform center no-flow
boundary condition (Figure 13B). Variations in the
magnitude and pattern of free convective flow in the
Serpukhovian boundstones is caused by differences in
their geometry, being better developed on the eastern
margin (Figure 13).
Figure 10. Geothermal convection after platform burial by shale and salt. (A) Artinskian shale; (B) Kungurian halite; (C) vertical Zfluid velocity and temperature (contour interval of 2.5jC); (D–F) calcite cementation and dissolution at 5, 10, and 20 m.y. Note that toenhance visualization of diagenetic patterns, the color bar scale does not reflect the full range of the data, which is discussed in the text.
1262 Geothermal Convection in the Tengiz Carbonate Platform, Kazakhstan
in fluid density drive two large-scale free convection cells
in a similar manner to the salt-only overburden case;
however, the direction of flow in each cell is reversed,
and the location and dimensions of the rising and falling
limbs have changed (compare Figures 10C, 14D). The
rising limbs of these cells are located beneath the with-
drawal basin, whereas the respective falling limbs are
located in the more distal platform interior and the
margin (Figure 14D). Flow velocities in the platform
interior and margin are relatively fast, ranging from
0.13 to 2.36m/yr (0.42 to 7.74 ft/yr) in the Tournaisian
and younger units, whereas circulation is considerably
weaker in theDevonian platform interior, where veloci-
ties range from 0.014 to 0.13 m/yr (0.045 to 0.42 ft/yr)
(Figure 14C).
Calcite diagenesis was simulated for a period of
20 m.y., in which no additional sedimentary layers
were added as described above. For simplicity, the com-
plex transient development of a withdrawal basin was
notmodeled. Calcite diagenesis is shown at 20m.y. and
along transects in the two stratigraphic layers used in
previous figures (Figures 14F, 15).
Figure 11. Calcite cementation and dissolution along (A) Ser-pukhovian transect and (B) Visean transect (see Figure 2).
Figure 12. Effect of reducing carbonate vertical permeability(compare to Figure 11). (A) Vertical Z fluid velocity and tem-perature (contour interval of 2.5jC). (B) Calcite cement anddissolution at 10 m.y. Note that to enhance visualization ofdiagenetic patterns, the color bar scale does not reflect the fullrange of the data, which is discussed in the text.
Jones and Xiao 1263
Reversal of flow in the convective cells and elevated
temperature beneath the withdrawal basin had a crit-
ical effect on the spatial distribution of diagenesis and
tion of up to 24.5% in 20 m.y. occurred in a cooling
zone that exhibited a mushroom geometry, with a
wide (1.7-km; 1.05-mi) vertical shaft extending from
the top of the Devonian to the Visean, before expand-
ing laterally in the Serpukhovian by more than 1 km
(0.6 mi) toward both the margin and platform interior
(Figures 14F, 15). In contrast to previous scenarios, ce-
mentation of up to 2.3% in 20 m.y. occurred inmost
of the Visean platform interior because of convective
warming. A second, weaker zone of cementation of up
to 0.6% in 20 m.y. occurred in the ascending limb of
the platform margin convective cell (Figures 14F, 15).
Increasing the depth of the overburden and shale-
filled withdrawal basin to 1800 m (5900 ft) resulted in
a similar pattern of convective flow. However, flow
velocities were significantly lower because of the de-
crease in sediment permeability in response to poros-
ity reduction by the additional overburden (Figure 16).
In the platform interior and margin, flow velocities
ranged from 0.03 to 0.16 m/yr in the Tournaisian and
younger units and from 10�4 to 0.03m/yr (3.3� 10�4
to 0.098 ft/yr) in the Devonian platform interior
(Figure 16C). Correspondingly, rates of dissolution
and cementation are severely reduced to a maximum
of 0.2% in 20 m.y., and significant changes in reser-
voir quality would require a long residence time under
these burial conditions (Figure 16D).
Free Geothermal Circulation: Shale Burial
Shale is also a common top seal for carbonate reser-
voirs (Greenlee and Lehmann, 1993). The contrast
between the thermal conductivity of carbonates and
shale is less than that of salt, but it may be sufficient to
drive geothermal convection in the burial environment
(Figure 7). We tested this hypothesis using the Tengiz
Figure 13. Geothermal convection for com-plete Tengiz cross section shown in Figure 1using Cartesian coordinates. (A) Hydrostratig-raphy; see Figures 3 and 10 for color legends.(B) Calcite cement and dissolution at 10 m.y.Note that to enhance visualization of diageneticpatterns, the color bar scale does not reflectthe full range of the data, which is discussed inthe text, temperature (contour interval of 2.5jC),and fluid velocity vectors.
1264 Geothermal Convection in the Tengiz Carbonate Platform, Kazakhstan
platform by specifying a shale basin fill instead of salt
for an otherwise identical simulation to that depicted
in Figure 10 (Figure 17A).
Results demonstrated that geothermally driven
platform-scale free convection would be feasible in the
Tengiz platformwith a shale burial scenario (Figure 17B).
A large platform-scale convection cell that extends lat-
erally from the platform interior to the Serpukhovian
boundstones, where it intersects a series of smaller con-
tive to the scenario with an all-salt basin fill, the direc-
tion of flow in this large convective cell is reversed,
being counterclockwise, with a 4-km (2.5-mi)-wide,
laterally extensive falling limb in the platform interior
and a narrower 2-km (1.2-mi) rising limb updip of the
Serpukhovian boundstones (Figure 17B). Flow is fo-
cused in the Visean- to Bashkirian-age platform inte-
rior, where velocities range from 0.03 to 2.12 m/yr
(0.098 to 6.95 ft/yr), with lower rates of flow that
range from 0.0014 to 0.57 m/yr (0.0045 to 1.87 ft/yr)
Figure 14. Geothermal circulation after platform burial and development of a 5.1-km (3.1-mi) shale-filled salt-withdrawal basin thatis 500 m (1640 ft) deep. (A) Halite, W = withdrawal basin; (B) shale; (C) horizontal X fluid velocity, velocity vectors, and temperature(contour interval of 2.5jC); (D) vertical Z fluid velocity, fluid velocity vectors, and temperature; (E) enlargement of vertical fluidvelocity distribution in platform; (F) calcite cementation and dissolution at 20 m.y. Note that to enhance visualization of diageneticpatterns, the color bar scale does not reflect the full range of the data, which is discussed in the text.
Jones and Xiao 1265
in the underlying less-permeable Devonian sediments
(Figure 17B).
Compared to the salt case, significant differences
exist in the temperature distribution, and the maxi-
mum temperature is 14jCwarmer (Figures 10C, 17B).
Temperatures are reduced in the platform interior and
elevated toward themargin because of convective cool-
ing and heating associated with the falling and rising
limbs of the dominant convection cell (Figure 17B).
Free convective cells described above generate local
minor temperature perturbations in the Serpukhovian
boundstones.
Although the pattern of diagenesis is distinctly dif-
ferent, the rates are comparable to the salt overburden
case, with the maximum amount of dissolution and ce-
mentation in 20m.y. being 11.5 and 4. 1%, respectively
(compare Figures 10E; 12; 17C, D).
DISCUSSION
The principal objective of this study was to investigate
geothermal convection and associated patterns of dia-
genetic porosity modification in the Tengiz platform.
In addition, our results provide a generic understand-
ing of geothermal convection in other isolated carbon-
ate platforms.
Figure 15. Calcite cement and dissolution along the two tran-sects in the Serpukhovian and Visean for a salt-withdrawal basincase (see Figure 14).
Figure 16. Geothermal circulation after development of a 1800-m (5900-ft)-deep salt-withdrawal basin. (A) Halite, W = withdrawalbasin; (B) shale; (C) Vertical Z fluid velocity and temperature; (D) Calcite cementation and dissolution at 20 m.y. (note the change incolor bar scales).
1266 Geothermal Convection in the Tengiz Carbonate Platform, Kazakhstan
Forced Geothermal Convection Prior to Burial
Simulations predict that, prior to burial, geothermally
driven forced convection of seawater occurred in the
Tengiz platform. This result is consistent with obser-
vations of active forced convection in many modern
carbonate platforms with shelf margins (Kohout et al.,
1977; Saller, 1984; Sanford et al., 1998). Based on the
rock record, previous studies have attributed cementa-
tion, dolomitization, and subcarbonate-compensation-
depth dissolution to forced convection in platform
margins (Saller, 1984). Results presented in this study
provide new insights on the spatial patterns and rates
of diagenesis by forced geothermal convection. Of par-
ticular interest is our prediction of dissolution in the
platform interior in response to ascending groundwater
along a flow path with an elevated geothermal gradient.
Dissolution is enhanced in isolated platforms with a
radial geometry, like Tengiz, because of flow conver-
gence toward the platform center (Sanford et al., 1998;
Jones et al., 2000). Results predict that in Tengiz, the
interior zone of dissolution can be extensive, forming
an asymmetric wedge that is several hundred meters
thick at the platform center and extends 4 km (2.5 mi)
toward the margin (Figure 8). Simulated dissolution
rates are locally relatively fast, up to 45% at 2m.y. in the
Serpukhovian, butmore typically less than 2% at 2m.y.
in the Visean and Tournaisian (Figure 8). Simulations
demonstrate the critical control of the hydrostratigra-
phy on the pattern and magnitude of flow, heat trans-
port, and associated diagenesis. Decreasing the ver-
tical permeability of platform interior sediments by
an order of magnitude dramatically reduced dissolu-
tion rates (Figure 9). The minimal diagenesis driven
by forced convection predicted in the Devonian sec-
tion at Tengiz further illustrated the control of per-
meability on the rates of transport-dominated diagen-
esis (Figure 8). Results demonstrated the importance of
initial permeability characterization in using reactive
transport models to predict reservoir quality.
In addition to the platform interior, dissolution in
the rising limbs and cementation in the associated fall-
ing limbs are predicted to occur in a mixed convective
cell system in the fractured Serpukhovian boundstones
(Figure 8). Rates of cementation and dissolution in
the boundstones are moderate, up to 1.6% in 2 m.y.,
but could be significant if diagenesis is focused in
fractures because aperture is a critical control on frac-
ture permeability (Anderson and Woessner, 1992).
Results also predict moderate cementation in the Tour-
naisian above the Devonian stratigraphic boundary that
Figure 17. Effect of a shale basin fill on geothermal convectionin Tengiz (compare to Figure 11). (A) Shale; (B) vertical Z fluidvelocity and temperature; (C) calcite cementation and dissolu-tion at 20 m.y.; and (D) calcite cementation and dissolutionalong transects in the Serpukhovian and Visean (at 10 m.y.).Note that to enhance visualization of diagenetic patterns, thecolor bar scale does not reflect the full range of the data, whichis discussed in the text.
Jones and Xiao 1267
extends from the margin to the distal platform inte-
rior (Figure 8C).
Calcite cementation is predicted in the margin and
slope region, which is consistent with observations above
the carbonate compensation depths inmodern platforms.
However, predicted rates of cementation are relatively
low, up to 0.4% in 2 m.y. Therefore, extensive slope
cementation would require a long residence time in a
forced convective system.Other processes, for example,
tidal pumping, submixing zone circulation, and/ormicro-
bial mediation, provide an alternative explanation for the
often-large volumes of calcite cement observed in car-
bonate margin and upper-slope deposits (Moore, 2001).
We did not investigate dolomitization, which is
uncommonly present in Carboniferous rocks at Tengiz,
becauseBasin2 does not include reaction kinetics (Weber
et al., 2003;Collins et al., 2006).Other studies of forced
convection, with reactive transport models, predict
dolomitization to occur coincident with the 60–70jCisotherm (Wilson et al., 2001). Extrapolation of this
result to Tengiz may explain the general absence of
dolomitization by geothermal convection because, at
the time of platform demise, we predict temperatures
in the Carboniferous section to be cooler than 60jC(Figure 8A).
Free Geothermal Convection after Burial
The occurrence of free convection after the burial of
the Tengiz platform is difficult to predict based on con-
flicting conclusions of free convection occurrence in na-
ture (Wood and Hewett, 1982; Hanor, 1987; Bjorlykke
et al., 1988). Variations of the Rayleigh criteria, a di-
mensionless number calculated to determine the onset
of free convection, were used in most of the previ-
ous studies (Lapwood, 1948; Nield, 1968; Wood and
Hewett, 1982). Groundwater-flow simulations pro-
vide a more informative prediction of free convection
because they are subject to less limiting assumptions,
particularly how the permeability heterogeneity and
geometry of the porousmedia are accounted for (Sharp
et al., 2001; Simmons et al., 2001). Furthermore, free
convection (non-Rayleigh) will be induced in sloping
layers where the isotherms are not horizontal, a con-
dition that is satisfied in ramps and rimmed shelf-
carbonate platform margins (Bories and Combarnous,
1973; Mullis, 1995; Heydari, 2000).
This study demonstrated that geothermally driven
free convection in the Tengiz platform is feasible after
burial (Figure 10). Free convective flow is hydrauli-
cally closed, but the redistribution of carbonate mass
can generate zones of enhanced reservoir quality, the so-
called ‘‘sweet spots,’’ which represent preferential dril-
ling targets. The general patterns of flow and diagenesis
are similar to forced convection prior to burial; how-
ever, the flow velocity and corresponding diagenetic
rates are lower, but, in 20 m.y., can still account for up
to 7.3% of calcite dissolution in the Serpukhovian plat-
form interior. The impact of diagenetic porosity modi-
fication by free convection in the burial realm will be
dependent on the timing of hydrocarbon charge and
the nature and rate of overburden loading.
Results presented demonstrated that the develop-
ment of shale-filled salt-withdrawal basins in the over-
burden significantly modifies free convection in the
to 7.3% after 20 m.y.) was predicted in the Serpu-
khovian and Visean platform interior.� Shale-filled salt-withdrawal basins in the overbur-
den significantly changed the modeled subsurface
temperature distribution and reversed the direction
of free convective flow. Ascending platform ground-
waters beneath the withdrawal basin created a zone
of calcite dissolution (up to 24.5% at 20m.y.) with a
mushroom geometry and minor calcite cementation
(up to 2.3% at 20m.y.) in the distal platform interior
and margin.� As the depth of burial increases, permeability re-
duction caused by compaction dramatically retards
convective flow and associated rates of diagenesis.� Free convection persisted if a shale overburden was
substituted for salt, although the general direction
of convective flow and diagenetic patterns were
reversed.
APPENDIX 1
Basin2TM
is a numerical model designed to trace through geo-logic time the evolution of groundwater-flow regimes in porousmedia, developed at the University of Illinois by Craig M. Bethkeet al. The Basin2 user’s guide (Bethke et al., 1999) is available for freedownload from the University of Illinois, Hydrogeology ProgramWeb site (www.geology.uiuc.edu/~bethke/index.htm). The Basin2user’s guide provides a complete description of the numerical code,model parameterization, example applications, and key references.
APPENDIX 2: EFFECTIVE PERMEABILITYANISOTROPY IN CARBONATE ROCKS
Permeability anisotropy (the ratio of horizontal to vertical per-meability) is derived from factors such as the alignment of grains,the interlayering of laminae of different permeability, and the ori-entation of fractures. Therefore, permeability anisotropy tends tovary with the scale of observation, becoming more pronounced assystem dimensions increase (Haldorsen, 1986). Many simulationsof groundwater flow, including this study, have demonstrated thatthe specified permeability anisotropy exerts a strong influence onthe pattern and magnitude of fluid flow, heat, and solute transport(Whitaker et al., 2004).
Measurement of core-scale permeability anisotropy in carbon-ate rocks reveals that horizontal permeability can average 10 toseveral hundred times the vertical permeability (McNamara and
Wardlaw, 1991; Thawer et al., 2000). However, the principal causeof permeability anisotropy in sequences of carbonate rocks is causedby the layering of different rock types, for example, interbeddedgrainstones and mudstones or evaporites, which cannot be measuredat the core-plug scale. An exception is the presence of open frac-tures, which enhance vertical permeability. Jones (2000) quantifiedthe effective permeability anisotropy of low-permeability mudstonebeds above and below a grainstone interval with a series of numericalflow simulations.
Flow was simulated through the domain ABCD (Figure 18).First, flow was simulated parallel to BC by specifying a pressuregradient from AB to CD and treating BC and AD as no-flowboundaries. Flow was subsequently simulated perpendicular to ABby specifying a pressure gradient fromAD to BC and treatingAB andCD as no-flow boundaries. The effective permeability anisotropyof the domain can then be calculated fromDarcy’s law as the ratio ofoutflow in directions BC and AB. The effect of mudstone continuityon permeability anisotropy was investigated by systematicallymodifying the hydrostratigraphy of the flow domain. The thick-nesses of the mudstone layers were held constant and represented10% of the total flow domain thickness; mudstone permeability(10� 5 d) was specified as isotropic, and a connectivity index (m:l)was defined as the ratio of the flow domain length to the mudstonebed length (Figure 18). The grainstone horizontal permeability (1 d)was held constant, whereas the grainstone permeability anisotropy
Figure 18. (A) Flow domain and hydrostratigraphy used tocalculate effective permeability anisotropy; (B) effect of mud-stone connectivity (m:l) and grainstone permeability anisotropy(kh/k v) on flow-derived effective permeability anisotropy (seetext for discussion) (Jones, 2000).
1270 Geothermal Convection in the Tengiz Carbonate Platform, Kazakhstan
(kh/k v) was systematically varied (by keeping the horizontal per-meability constant and reducing the vertical permeability) to encom-pass the range of values measured in grainstones and packstones atthe core scale (1, 10, 100, 1000).
Calculated values of effective permeability anisotropy are de-pendent on both the mudstone connectivity and the specified grain-stone permeability anisotropy (Figure 18). When the mudstone andgrainstones are specified as isotropic, a reduction in the mudstoneconnectivity (m:l) from 1 to 0.95 changed the effective permeabilityanisotropy by three orders of magnitude (Figure 18). Increasing thepermeability anisotropy of the grainstone reduces the overall effectof mudstone connectivity on the effective permeability anisotropy(Figure 18). However, when the mudstone connectivity is high(>0.9), the effective permeability anisotropy still increased by ordersof magnitude (Figure 18). Based on these results, we estimate themost likely range of effective permeability anisotropy in sequencesof carbonate rocks, at the scale of model grid cells used in this study,at 100–1000. The fractured Serpukhovian boundstones with ver-tically oriented enhanced permeability are an exception and are spec-ified to have permeability anisotropy of 0.5 based on rock propertiesused in reservoir simulations.
REFERENCES CITED
Anderson, M. P., and W. W. Woessner, 1992, Applied groundwatermodeling: Simulation of flow and advective transport: SanDiego, Academic Press, 381 p.
Bethke, C. M., 1985, A numerical model of compaction-drivengroundwater flow and heat transfer and its application to thepaleohydrology of intracratonic sedimentary basins: Journal ofGeophysical Research, v. 90, p. 6817–6828.
Bethke, C. M., 1989, Modelling subsurface flow in sedimentarybasins: Geologische Rundschau, v. 78, p. 129–154.
Bethke, C. M., 1998, The geochemist’s workbench: HydrogeologyProgram: Urbana–Champaign, University of Illinois, 224 p.
Bethke, C. M., M.-K. Lee, and J. Park, 1999, Basin modeling withBasin2 (release 4): Hydrogeology Program, Urbana–Champaign,University of Illinois, 205 p.
Bjorlykke, K., A. Mo, and E. Palm, 1988, Modelling of thermal con-vection in sedimentary basins and its relevance to diageneticreactions: Marine and Petroleum Geology, v. 5, p. 338–351.
Bories, S. A., and M. A. Combarnous, 1973, Natural convection in asloping layer: Journal of Fluid Mechanics, v. 57, p. 63–79.
Bredehoeft, J. D., 2003, From models to performance assessment:The conceptualization problem: GroundWater, v. 41, p. 571–577.
Carpenter, D. G., S. A. Guidry, J. D. Degraff, and J. Collins, 2006,Evolution of Tengiz rim/flank reservoir quality: New insightsfrom systematic, integrated core fracture and diagenesisinvestigations (abs.), in Giant hydrocarbon reservoirs of theworld: From rocks to reservoir characterization and modeling:AAPG Annual Meeting Abstract Volume, v. 15, p. 18.
Collins, J. F., A. M. Kenter, P. M. Harris, G. Kuanysheva, D. J.Fischer, and K. L. Steffen, 2006, Facies and reservoir qualityvariations in the late Visean to Bashkirian outer platform, rimand flank of the Tengiz buildup, Precaspian Basin, Kazakhstan(abs.), inGiant hydrocarbon reservoirs of the world: From rocksto reservoir characterization and modeling: AAPG AnnualMeeting Abstract Volume, v. 15, p. 21.
Esteban, M., and S. Q. Sun, 2002, Carbonate reservoirs: How im-portant is the late diagenesis? (abs.): AAPG Annual MeetingProgram, v. 11, p. A51.
Esteban, M., and C. Taberner, 2003, Secondary porosity develop-
ment during late burial in carbonate reservoirs as a result ofmixing and/or cooling of brines: Journal of Geochemical Ex-ploration, v. 78–79, p. 355–359.
Ewing, R. E., 1997, Aspects of upscaling in simulation of flow inporousmedia: Advances inWater Resources, v. 20, p. 349–358.
Goldhammer, R. K., 1997, Compaction and decompaction algo-rithms for sedimentary carbonates: Journal of SedimentaryResearch, v. 67, p. 26–35.
Greenlee, S. M., and P. J. Lehmann, 1993, Stratigraphic frameworkof productive carbonate buildups, in R. G. Loucks and J. F.Sarg, eds., Carbonate sequence stratigraphy recent develop-ments and applications: AAPG Memoir 57, p. 43–62.
Haldorsen, H. H., 1986, Simulator parameter assignment and theproblem of scale in reservoir engineering, in L. W. Lake andH. B. Carroll Jr., eds., Reservoir characterization: New York,Academic Press, p. 293–340.
Hanor, J. S., 1987, Kilometre-scale thermohaline overturn of porewaters in the Louisiana Gulf Coast: Nature, v. 327, p. 501–503.
Harrison, W. J., and L. L. Summa, 1991, Paleohydrology of the Gulfof Mexico Basin: American Journal of Science, v. 291, p. 109–176.
Heydari, E., 1997, Hydrotectonic models of burial diagenesis inplatform carbonates based on formation water geochemistryin North America sedimentary basins, in I. P. Montanez,J. M. Gregg, and K. L. Shelton, eds., Basin-wide diageneticpatterns: Integrated petrologic, geochemical and hydrologicconsiderations: SEPM Special Publication 57, p. 53–79.
Heydari, E., 2000, Porosity loss, fluid flow and mass transfer inlimestone reservoirs: Application to the Upper Jurassic Smack-over Formation, Mississippi: AAPG Bulletin, v. 84, p. 100–118.
Heydari, E., 2003, Meteoric versus burial control on porosity evol-ution in the Smackover Formation: AAPG Bulletin, v. 87,p. 1779–1797.
Hill, C. A., 1995, H2S-related porosity and sulfuric acid oil fieldkarst, in D. A. Budd, A. H. Saller, and P. M. Harris, Uncon-formities and porosity in carbonate strata: AAPG Memoir 86,p. 301–306.
Jones, G. D., 2000, Numerical modelling of saline groundwatercirculation in carbonate platforms: Ph.D. thesis, University ofBristol, United Kingdom, 330 p.
Jones, G. D., and Y. Xiao, 2005, Dolomitization, anhydrite ce-mentation and porosity evolution in a reflux system: Insightsfrom reactive transport models: AAPG Bulletin, v. 89, p. 577–601.
Jones, G. D., F. F. Whitaker, P. L. Smart, and W. E. Sanford,1998, The critical role of porosity/permeability transforma-tions for modelling regional groundwater circulation in car-bonate rocks (abs.): SEPM Research Conference, Fluid Flowin Carbonates: Interdisciplinary Approaches, Program withAbstracts.
Jones, G. D., F. F. Whitaker, P. L. Smart, and W. E. Sanford, 2000,Numerical modeling of geothermal and reflux circulation inEnewetak Atoll: Implications for dolomitization: Journal ofGeochemical Exploration, v. 69–70, p. 71–75.
Jones, G. D., F. F. Whitaker, P. L. Smart, and W. E. Sanford, 2004,Numerical analysis of seawater circulation in carbonate plat-forms: II. The dynamic interaction between geothermal andbrine reflux circulation: American Journal of Science, v. 304,p. 250–284.
Kaufman, J. K., 1994, Numerical models of fluid flow in carbonateplatforms: Implications for dolomitization: Journal of Sedi-mentary Research, v. A64, p. 128–139.
Kohout, F. A., H. R. Henry, and J. E. Banks, 1977, Hydrology relatedto geothermal conditions of the Floridan Plateau, inK. L. Smith
Jones and Xiao 1271
and G. M. Griffin, eds., The geothermal nature of the FloridanPlateau: Florida Department of Natural Resources Bureau ofGeology Special Publication 21, p. 1–34.
Lapwood, E. R., 1948, Convection of fluids in a porous medium:Proceedings of the Cambridge Philosophical Society, v. 44,p. 508–521.
Lee, M.-K., 1997, Predicting diagenetic effects of groundwater flowin sedimentary basins: A modeling approach with examples, inI. P. Montanez, J. M. Gregg, and K. L. Shelton, eds., Basin-wide diagenetic patterns: Integrated petrologic, geochemicaland hydrologic considerations: SEPM Special Publication 57,p. 3–14.
Lee, M.-K., and C. M. Bethke, 1994, Groundwater flow, late ce-mentation and petroleum accumulation in the Permian Lyonssandstone: AAPG Bulletin, v. 78, p. 217–237.
Lisovsky, N. N., G. N. Gogonenkov, and Y. A. Petzoukha, 1992,The Tengiz oil field in the Pre-Caspian Basin of Kazakhstan(former USSR)— Supergiant of the 1980s, in M. T. Halbouty,ed., Giant oil and gas fields of the decade 1978–1988: AAPGMemoir 54, p. 101–122.
Lucia, F. J., 1995, Rock fabric/petrophysical classification of car-bonate pore space for reservoir characterization: AAPG Bul-letin, v. 79, p. 1275–1300.
Machel, H. G., 1999, Effects of groundwater flow on mineral dia-genesis, with emphasis on carbonate aquifers: HydrogeologyJournal, v. 7, p. 94–107.
Machel, H. G., and J. H. Anderson, 1989, Pervasive subsurfacedolomitization of the Nisku Formation in central Alberta: Jour-nal of Sedimentary Petrology, v. 59, p. 891–911.
McNamara, L. B., and N. C. Wardlaw, 1991, Geological andstatistical description of the Westerose reservoir, Alberta: Bul-letin of Canadian Petroleum Geology, v. 39, p. 332–351.
Moore, C. H., 2001, Carbonate reservoirs: Porosity evolution anddiagenesis in a sequence stratigraphic framework: Netherlands,Elsevier, Developments in Sedimentology 55, 444 p.
Morrow, D. W., 1998, Regional subsurface dolomitization: Modelsand constraints: Geoscience Canada, v. 25, p. 57–70.
Morse, J. W., and R. S. Arvidson, 2002, The dissolution kinetics ofmajor sedimentary carbonate minerals: Earth-Science Reviews,v. 58, p. 51–84.
Morse, J. W., J. S. Hanor, and S. He, 1997, The role of mixing andmigration of basinal waters in carbonatemineral mass transport,in I. P. Montanez, J. M. Gregg, and K. L. Shelton, eds., Basin-wide diagenetic patterns: Integrated petrologic, geochemicaland hydrologic considerations: SEPM Special Publication 57,p. 41–50.
Mullis, A. M., 1995, Natural convection in porous permeable me-dia: Sheets, wedges and lenses: Marine and Petroleum Geol-ogy, v. 12, p. 17–25.
Nield, D., 1968, Onset of thermohaline convection in a porousmedium: Water Resources Research, v. 4, p. 553–560.
Phillips, O. M., 1991, Flow and reactions in permeable rocks: Cam-bridge, Cambridge University Press, 285 p.
Raffensperger, J. P., and D. Vlassopoulos, 1999, The potential forfree and mixed convection in sedimentary basins: Journal ofHydrogeology, v. 7, p. 505–520.
Saller, A. H., 1984, Petrologic and geochemical constraints on theorigin of subsurface dolomite, Enewetak Atoll: An example ofdolomitization by normal seawater:Geology, v. 12, p. 217–220.
Sanford, W. E., and L. F. Konikow, 1989, Simulation of calcite dis-solution and porosity changes in salt water mixing zones incoastal aquifers: Water Resources Research, v. 25, p. 655–667.
Sanford, W. E., F. F. Whitaker, P. L. Smart, and G. D. Jones, 1998,Numerical analysis of seawater circulation in carbonate plat-forms: I. Geothermal circulation: American Journal of Science,v. 298, p. 801–828.
Sarkar, J., J. A. Nunn, and J. S. Hanor, 1995, Free thermohalineconvection beneath allochthonous salt sheets: An agent for saltdissolution and fluid flow in Gulf Coast sediments: Journal ofGeophysical Research, v. 100, no. B9, p. 18,085–18,092.
Schmoker, J. W., and R. B. Halley, 1982, Carbonate porosity versusdepth: A predictable relation for South Florida: AAPGBulletin, v. 66, p. 2561–2570.
Sharp, J. M., T. R. Fenstemaker, C. T. Simmons, T. E. McKenna,and J. K. Dickinson, 2001, Potential salinity driven freeconvection in a shale-rich sedimentary basin: Example fromthe Gulf of Mexico Basin in south Texas: AAPG Bulletin, v. 85,p. 2089–2110.
Simmons, C. T., T. R. Fenstemaker, and J. M. Sharp Jr., 2001,Variable-density groundwater flow and solute transport inheterogeneous porous media: Approaches, resolutions andfuture challenges: Journal of Contaminant Hydrology, v. 52,p. 245–275.
Steefel, C. I., and A. C. Lasaga, 1994, A coupled model for transportof multiple chemical species and kinetic precipitation/dissolu-tion reactions with applications to reactive flow in single phasehydrothermal system: American Journal of Science, v. 294,p. 529–592.
Steefel, C. I., D. J. DePaolo, and P. C. Lichtner, 2005, Reactivetransport modeling: An essential tool and a new approach forthe earth sciences: Earth and Planetary Science Letters, v. 240,p. 539–558.
Thawer, R., A. Alhendi, Y. Al Mazroui, D. Boyd, T. Masuzawa,Y. Sugawara, C. Hollis, and B. Lowden, 2000, Controls on ver-tical and horizontal flow in a carbonate reservoir that impactgasflooding and waterflooding: Society of Petroleum EngineersPaper 87237, 10 p.
Vacher, H. L., 1988, Dupuit-Ghyben-Herzberg analysis of strip-island lenses: Geological Society of America Bulletin, v. 100,p. 580–591.
Weber, J. L., B. P. Francis, P. M. Harris, and M. Clark, 2003, Strat-igraphy, lithofacies and reservoir distribution, Tengiz field,Kazakhstan, inW.M.Ahr, P.M.Harris,W.A.Morgan, and I.D.Sommerville, eds., Permo-Carboniferous carbonate platformand reefs: SEPM Special Publication 78 and AAPGMemoir 83,p. 351–394.
Wendte, J., H. Quing, J. F. Dravis, S. L. O. Moore, L. D. Stasiuk,and G. Ward, 1998, High temperature saline (thermoflux)dolomitization of Swan Hills platform and bank carbonates,Wild River area, west-central Alberta: Bulletin of CanadianPetroleum Geology, v. 46, p. 210–266.
Whitaker, F. F., P. L. Smart, and G. D. Jones, 2004, Dolomitization:From conceptual to numerical models, in C. Braithwaite, G.Rizzi, and G. Darke, eds., The geometry and petrogenesis ofdolomite hydrocarbon reservoirs: Geological Society (London)Special Publication 235, p. 99–139.
Wilson, A. M., W. E. Sanford, F. F. Whitaker, and P. L. Smart, 2001,Spatial patterns of diagenesis during geothermal circulation incarbonate platforms: American Journal of Science, v. 301,p. 727–752.
Wood, J. R., and T. A. Hewett, 1982, Fluid convection and masstransfer in porous sandstones— A theoretical model: Geochi-mica et Cosmochimica Acta, v. 47, p. 1707–1713.
Xu, T., and K. Pruess, 2001, Modeling multiphase fluid flow andreactive geochemical transport in variably saturated fracturedrocks: 1. Methodology: American Journal of Science, v. 301,p. 16–33.
Yao, Q., and R. V. Demicco, 1997, Dolomitization of the Cambriancarbonate platform, southern Canadian Rocky Mountains: Do-lomite front geometry, fluid inclusion geochemistry, isotopicsignature and hydrogeological modeling studies: AmericanJournal of Science, v. 297, p. 892–938.
1272 Geothermal Convection in the Tengiz Carbonate Platform, Kazakhstan