ArticlePDF Available


Content may be subject to copyright.
K. U. Mayer
and K. T. B. MacQuarrie
University of British Columbia,
University of New Brunswick
Reactive Transport Modelling in
Sedimentary Rock: State-of-Science
NWMO TR-2007-12 December 2007
- ii -
Nuclear Waste Management Organization
22 St. Clair Avenue East, 6
Toronto, Ontario
M4T 2S3
Tel: 416-934-9814
- iii -
Reactive Transport Modelling in Sedimentary Rock: State-of-Science Review
NWMO TR-2007-12
December 2007
K. U. Mayer
and K. T. B. MacQuarrie
University of British Columbia,
University of New Brunswick
- iv -
This report does not necessarily reflect the views or position of the Nuclear Waste Management
Organization, its directors, officers, employees and agents (the “NWMO”) and unless otherwise
specifically stated, is made available to the public by the NWMO for information only. The contents of
this report reflect the views of the author(s) who are solely responsible for the text and its conclusions
as well as the accuracy of any data used in its creation. The NWMO does not make any warranty,
express or implied, or assume any legal liability or responsibility for the accuracy, completeness, or
usefulness of any information disclosed, or represent that the use of any information would not infringe
privately owned rights. Any reference to a specific commercial product, process or service by trade
name, trademark, manufacturer, or otherwise, does not constitute or imply its endorsement,
recommendation, or preference by NWMO.
- v -
Title: Reactive Transport Modelling in Sedimentary Rock: State-of-Science
Report No.: NWMO TR-2007-12
K. U. Mayer
and K. T. B. MacQuarrie
University of British Columbia,
University of New Brunswick
December 2007
To assess the suitability of sedimentary rock units for hosting deep geological repositories
(DGR) for used nuclear fuel, it is necessary to investigate the long term geochemical stability of
these formations. Multicomponent reactive transport modelling provides a viable method to
evaluate conceptual models and to assess parameter sensitivity in the context of rock-water
interaction, and can thus support and complement expensive field investigations. In this report,
hydrogeological and hydrogeochemical data are reviewed to identify the relevant transport and
reaction processes that control groundwater evolution and composition. In addition, previous
reactive transport modelling efforts in similar settings are reviewed to assess the current status
of reactive transport modelling in sedimentary rock formations. Based on this information,
model capabilities and formulation gaps are identified. In addition, recommendations are made
on how reactive transport modelling could be used most effectively to evaluate redox stability
and salinity evolution in sedimentary rock units in response to periods of glaciation and
deglaciation. It is found that reactive transport modelling has not been previously used to
assess such a scenario. However, modelling studies of seawater ingress and CO
sequestration in sedimentary rocks show promising results and suggest that modelling of the
geochemical evolution in a 2D-subsection of a sedimentary basin is a realistic goal. Although
there are advanced multicomponent reactive transport models that consider a wide range of the
required processes including aqueous speciation, ion exchange, mineral dissolution-
precipitation, microbially-mediated redox reactions, density coupling between flow and reactive
transport, as well as porosity and permeability changes, none of the currently available codes is
capable of capturing all processes of relevance. The MIN3P code belongs to this group of
codes and it is recommended that the capabilities of this model be expanded to include a) the
Pitzer ion interaction model, b) a modified formulation for microbially-mediated reactions that
accounts for inhibition as a function of salinity, c) a formulation for multicomponent and species-
dependent diffusion, and d) discretization methods that facilitate the generation of unstructured
grids that are better capable of dealing with irregular geometry and outcropping aquifer and
aquitard units. Furthermore, it is recommended that the enhanced code be used to investigate a
series of conceptual models for sedimentary units; within the context of these models, the effect
of parameter uncertainty can be evaluated with respect to recharge penetration depth, redox
stability and salinity evolution.
- vi -
- vii -
ABSTRACT .............................................................................................................................. v
INTRODUCTION ................................................................................................... 1
FORMATIONS ...................................................................................................... 2
2.1 HYDROGEOLOGICAL ASPECTS AND FLOW PROCESSES ............................ 2
ROCK SEQUENCES IN SOUTHWESTERN ONTARIO ....................................... 8
3.1.1 Horizontal Hydraulic Conductivities of Sandstone and Carbonate Rock Units.... 10
3.1.2 Vertical Hydraulic Conductivities in Confining Units (Shales) and Evidence for
Fracturing ............................................................................................................ 11
SETTINGS........................................................................................................... 13
ROCK SEQUENCES .......................................................................................... 14
5.1 GENERAL CONCEPTUAL MODEL ................................................................... 14
5.2 GOVERNING EQUATIONS ................................................................................ 15
6.2 SPECIES DEPENDENT MULTICOMPONENT DIFFUSION .............................. 18
6.3 OTHER PROCESSES AND DISCRETIZATION ISSUES ................................... 19
7. RECOMMENDATIONS AND FUTURE WORK .................................................. 20
7.1 MODEL DEVELOPMENT AND IMPROVEMENTS ............................................ 20
7.2 MODEL APPLICATIONS .................................................................................... 21
8. SUMMARY AND CONCLUSIONS ...................................................................... 22
REFERENCES ........................................................................................................................... 25
MODELLING IN SEDIMENTARY ROCK SEQUENCES .................................... 32
- viii -
Figure 1: Conceptual model of flow system evolution during glaciation/deglaciation conditions
involving a warm-based glacier in the northern section of the Michigan Basin. The left
diagram depicts the present day flow system. The ingress of glacial melt water into
outcropping aquifer units, displacement of basal brines, and discharge through overlying
shale units is displayed on the right for comparison (modified after McIntosh and Walter,
2005) ..................................................................................................................................... 2
Figure 2: Potential mechanisms of melt water infiltration a) ingress through subcrops or
outcrops of aquifers and flow through fractured confining units, b) ingress through subcrops
or outcrops of aquifers and displacement of resident groundwaters through aquifers (after
Person et al., 2007), c) vertical ingress through confining units (Siegel, 1991), d)
hypothetical 3D-effects during advancement of glacial lobe. Sedimentary basins as shown
in this figure are typically characterized by a spatial extent ranging from 100 - 500km. ....... 3
Figure 3: Cross-sectional view of region under consideration. Ordovician units include
sandstones and dolostones (Shadow Lake Fm), limestones (Trenton- Black River Gp (cf.
Simcoe Gp)), and shales (Blue Mountain Fm, Georgian Bay Fm, and Queenston Fm).
Silurian units include the Cabot Head Fm (shale), Manitoulin Fm (dolomite), Whirlpool Fm
(sandstone), a group of dolomites (Lockport Fm, Amabel Fm, Guelph, Fm), the Salina Fm
(evaporates with limestones), and Bass Island Fm (dolomite). The Bois Blanc Fm
(dolomite), Detroit River Gp (dolomite and limestones), Hamilton Gp (shales and
limestones) and Kettle Point Fm (shale) are Devonian in age. ............................................. 9
Figure 4: Comparison of hydraulic conductivity values for region of interest from a) Mazurek
(2004) and b) McIntosh et al., 2005. Hydraulic conductivities for the Devonian carbonate
rock units selected by McIntosh et al. (2005) compare well with the average values by
Mazurek (2004). Hydraulic conductivities for the Silurian-Ordovician carbonate rock units
used by McIntosh et al. (2005) are more than 3 orders of magnitude higher than average
values reported by Mazurek (2004). Hydraulic conductivities reported by McIntosh et al.
(2005) originate from a conference poster and cannot be traced back to original
references. .......................................................................................................................... 10
Figure 5: Schematic representation of modelling domains, shown in relation to a sedimentary
basin cross section, used in previous reactive transport studies. These studies include the
characterization of diffusive transport in rock formations considered for deep geologic
repositories, natural analogue studies for uranium migration and attenuation, several
modelling studies on geologic carbon sequestration, and the simulation of the interactions
between saline and fresh waters in coastal aquifer systems (see Appendix A). ................. 13
Figure 6: Conceptual model for groundwater evolution in response to melt water ingress. High
hydraulic gradients cause enhanced recharge of melt water containing high dissolved
oxygen (DO). Carbonate and evaporite mineral dissolution causes an increase in TDS and
fluid density, possibly associated with porosity and permeability increase. The presence of
reduced mineral phases and organic matter (OM) causes oxygen consumption, possibly
sulfate reduction and methanogenesis. Solid phase constituents are shown in grey boxes
(modified after McIntosh and Walter, 2005). ....................................................................... 15
Sedimentary rock units are currently being investigated for their potential to implement the deep
geological repository (DGR) concept for used nuclear fuel. For example, the sedimentary rocks
of southern Ontario possess many characteristics that are considered favourable for repository
siting and long-term waste isolation (Mazurek, 2004), including the potential presence of
reducing mineral phases and diffusion-controlled mass transport. Within the time frame of one
million years that is deemed necessary to demonstrate repository safety, significant climatic and
hydrologic changes may occur as a result of glaciation/deglaciation cycles. Evidence indicates
that past glaciations have previously modified groundwater flow systems (e.g. McIntosh and
Walter, 2005, 2006), and it is believed that future glacial cycles have the potential to cause
recharge of dilute glacial melt water containing relatively high concentrations of dissolved
oxygen (Guimerà et al., 1999, Sidborn, 2007), particularly in shallower flow systems within
sedimentary basins (McIntosh and Walter, 2005, Person et al., 2007). To demonstrate stability
within DGR host rocks, an understanding of far-field geochemical conditions and their evolution
over geologic time scales is required.
Reactive transport modelling is one approach that can be employed to assess the
geochemical and redox stability within sedimentary rock sequences. For example, reactive
transport modelling can clearly and convincingly demonstrate the degree to which dissolved
oxygen may be attenuated in the recharge region of a DGR host rock (e.g. Spiessl et al., 2007),
how rock-water interaction may cause groundwater salinity to vary along flow paths in aquifers,
or how diffusive transport of reactive solutes may evolve in the pore waters of low permeability
aquitard units (e.g. Pearson et al., 2002). These types of reactive transport investigations can
therefore complement field investigations and assist in supporting the safety case for a DGR.
However, before reactive transport modelling can be undertaken, the applicability of models to
sedimentary rock systems must be assessed and suitable conceptual models must be
The primary objectives of this report are to provide a review and assessment of the
applicability of reactive transport models for simulating long-term geochemical evolution in
sedimentary rock basins. The review deals primarily with the issue of past and potential future,
glacial melt water ingress into sedimentary rock units that could serve as a host for a DGR. As
noted above, reactive transport modelling can also be valuable for interpreting processes that
do not directly result from the ingress of glacial waters. To provide a sound basis for evaluating
model capabilities and limitations, both groundwater flow and geochemical reaction processes
are reviewed. Evidence is presented for past modification of groundwater flow systems during
glacial periods, and conceptual models for future flow system modifications and geochemical
reactions are proposed.
Previously published studies which have utilized reactive transport modelling to interpret
observations or make predictions of geochemical evolution in sedimentary rock sequences are
reviewed with the objective of identifying modelling approaches, data requirements, and
important feedbacks between reaction and flow processes. These studies therefore assist in
determining the capabilities that are required in reactive transport models.
This review for sedimentary rocks is similar in scope to a recent review of reactive
transport modelling in fractured crystalline rocks of the Canadian Shield (MacQuarrie and
Mayer, 2005). However, unlike the previous review, numerical formulations for reactive
- 2 -
transport models and a comparison of the capabilities of current reactive transport models are
not described in detail because these issues have been dealt with previously (e.g. MacQuarrie
and Mayer, 2005) and they are independent of the particular rock type. Also, because recent
reactive transport simulations in fractured crystalline rocks have been undertaken using the
MIN3P code (Mayer et al., 2002, Spiessl et al., 2007), when assessing model requirements,
outstanding modelling issues, and making recommendations for future modelling work in
sedimentary rocks, we do so by considering both the conceptual models proposed and the
current capabilities of the reactive transport code MIN3P.
To provide the necessary background for developing conceptual models for reactive transport in
sedimentary rock sequences, this section provides a summary of controlling hydrogeological
and geochemical processes that affect groundwater flow and reactive transport on the spatial
and time scales of interest.
Currently, the zone of active groundwater circulation in sedimentary basins is limited mostly to
glacial drift aquifers and shallow fractured rock (Figure 1, McIntosh and Walter, 2005, Person et
al., 2007), in part because the relatively flat landscape in many sedimentary basins does not
provide the topographic relief to drive groundwater flow to depths greater than 50-100 m
(McIntosh and Walter, 2005). In addition, the salinities of groundwaters at greater depth are
characteristic of brines (Wilson and Long, 1993a,b, McIntosh and Walter 2005, 2006) and the
resulting density gradient further inhibits groundwater flow into deeper units (Mazurek, 2004).
Figure 1: Conceptual model of flow system evolution during glaciation/deglaciation
conditions involving a warm-based glacier in the northern section of the Michigan Basin.
The left diagram depicts the present day flow system. The ingress of glacial melt water
into outcropping aquifer units, displacement of basal brines, and discharge through
overlying shale units is displayed on the right for comparison (modified after McIntosh
and Walter, 2005).
- 3 -
However, present day conditions are not necessarily reflective of past groundwater flow
dynamics. There is evidence that groundwater flow regimes have been reorganized on the
basin scale during periods of glaciation and deglaciation. During the Pleistocene, sedimentary
basins located in the northern United States and southern Canada, namely the Michigan,
Appalachian, Illinois, Forest City, Alberta and Williston basins, were covered periodically by ice
sheets of significant thickness (Person et al., 2007). For example, Peltier (2002) conducted a
series of simulations using the University of Toronto’s Glacial Systems Model that produced ice
sheet thicknesses of greater than 2km in southern Ontario, which is underlain by the Michigan
and Appalachian basins. In the case of a warm-based ice sheet, hydraulic heads at the ground
surface may reach up to 90% of the ice sheet thickness (Person et al., 2003, 2007), as was
observed under present day conditions for a basal drainage system of a West Antarctic ice
stream (Engelhard and Kamb, 1997). In this case, for an average ice sheet thickness of 1000 m,
observed water levels in boreholes that connected to the basal drainage system were typically
100 m below the ice surface. It can be hypothesized that these significant hydraulic heads
completely overwhelmed existing land topography and groundwater density stratification
(Boulton et al., 1993, Person et al., 2003, 2007). As a result, present-day groundwater divides
may not be relevant during periods of glaciation and deglaciation and the flow system may be
significantly altered (Figure 1). Conceptual two-dimensional simulations for a sedimentary basin
composed of four aquifers separated by aquitards suggest that past recharge rates under warm-
based glaciers may have exceeded present-date recharge by a factor of up to 10 (Person et al,
Figure 2: Potential mechanisms of melt water infiltration a) ingress through subcrops
or outcrops of aquifers and flow through fractured confining units, b) ingress through
subcrops or outcrops of aquifers and displacement of resident groundwaters through
aquifers (after Person et al., 2007), c) vertical ingress through confining units (Siegel,
1991), d) hypothetical 3D-effects during advancement of glacial lobe. Sedimentary basins
as shown in this figure are typically characterized by a spatial extent ranging from 100 -
- 4 -
Several conceptual models have been proposed to explain recharge into sedimentary
units during glacial periods. For example, McIntosh and Walter (2005, 2006) suggest that the
ingress of glacial melt water into Silurian-Devonian carbonate aquifers in the northern Michigan
basin took place through subcrops or outcrops at the basin margins. This concept implies that
recharge occurs directly into the permeable rock units, but not through the overlying confining
beds (Figure 1, Figure 2a).
It is clear that significant recharge can only occur if groundwater that is initially present
can be displaced by the recharging water. This implies that groundwater must either migrate
deeper into the rock units (Figure 2b), migrate laterally to a discharge point located in the same
unit (Figure 2d), or it must discharge to the ground surface after moving through overlying
aquitards (Figure 2a). Considering the low hydraulic conductivity of the confining beds in
sedimentary basins (mostly composed of shales), the potential for vertical discharge through
these units is limited. However, due to high hydraulic heads below ice sheets, significant vertical
gradients may develop, which may facilitate vertical discharge. In addition, mechanical loading
introduced by the weight of the advancing or retreating glacier may have caused fracturing of
the confining units thereby increasing their effective hydraulic conductivity and providing a
pathway for egress of groundwater (Person et al., 2007, Figure 2a).
For example, in the northern Michigan basin, there is evidence that meteoric water has
migrated into the Silurian and Devonian aquifers and subsequently into the overlying fractured
organic rich Devonian shale units (McIntosh and Walter, 2005, 2006). Glacial melt water is
estimated to comprise up to 50% of the shale formation water with an increasing contribution
towards the shale outcrop (Figure 1). Recharge into outcropping Silurian/Devonian aquifers is
also believed to have provided a pathway for ingress of melt water into the Illinois basin
(McIntosh and Walter, 2005) and the Williston basin (Grasby and Betcher, 2002). Ingress of
glacial melt water may also have affected deeper units (i.e. Cambrian/Ordovician rock units), as
evidenced by anomalous fluid pressures (McIntosh and Walter, 2005, 2006 and references
Person et al. (2003) used this concept to explain the relatively dilute water composition
in sedimentary units in the Atlantic continental shelf. In a modelling study, dilute water
compositions present today in the shelf deposits could only be reproduced by assuming
enhanced recharge during periods of glacial melt water production. McIntosh et al. (2005)
conducted density-dependent flow and transport simulations for the northern Michigan Basin
under glacial boundary conditions. Their results show that high hydraulic heads during basal
melting could result in penetration to depths of ~ 300 m in rocks at the basin margin, consistent
with field observations.
In a more recent modelling study, Person et al. (2007) suggested an alternative
conceptual model, which assumed basin wide flow and displacement of the groundwater
towards the opposite end of the basin margin (Figure 2b). This conceptual model was discussed
for the Williston basin and a sensitivity analysis was conducted that assumed that hydraulic
conductivity declined with depth; however, the decrease was limited to one order of magnitude
and porosity was assumed to be constant with depth (Person et al., 2007).
Siegel et al (1991) suggested yet another concept and proposed that recharge into the
Cambrian-Ordovician aquifer of Iowa (Illinois and Forest City basins) occurred vertically
downward through hundreds of metres of confining beds under the Des Moines glacial lobe.
This scenario was deemed more likely than infiltration through aquifer outcrops at distant
locations in southern Minnesota. Vertical recharge was assumed to occur predominately
- 5 -
through vertical sets of fractures (Figure 2c) and numerical modelling based on reasonable
aquitard parameters (bulk hydraulic conductivity: K = 10
, φ = 0.001) suggests that this
hypothesis is plausible. Simulations conducted by Boulton et al. (1994) for a two-dimensional
cross-section of a major aquifer system in the Netherlands also indicated that vertical ingress
through thick aquitards (300-1000m) is possible under a warm-based ice sheet. The simulated
recharge quantity was significant, considering that subglacial catchments for groundwater flow
extended from the ice divide to the glacier margin (on the order of 1000 km). The simulations
showed high horizontal gradients during glacial melt water production (one to two orders of
magnitude larger than under present day conditions), large vertical gradients in some regions,
and ingress of melt water to depths of approximately 1,500 m. However, it must be noted that
these simulations assumed a specified flux in a region of basal melting that extended over 100’s
of km and were based on a hydraulic conductivity for the main aquitard unit of approximately 5 x
, which is orders of magnitude higher than typical values for clay and shale aquitards (1
x 10
, Mazurek, 2004).
One mechanism for glacial melt water recharge and discharge which, to the best of our
knowledge, has not been discussed to date is the introduction of melt water caused by 3-D
effects during the presence of a glacial lobe over a sedimentary basin (Figure 2d). While high
hydraulic heads may exist underneath the ice sheet, hydraulic heads adjacent to the glacial lobe
are limited by ground surface elevation or the water level of surface water bodies. Recharge
may therefore occur into a permeable sedimentary unit and may be driven out laterally into
zones of lower hydraulic head within the same unit. Unlike the scenarios shown in Figures 2a
and 2b, this case does not require flow through low permeability confining units. In addition,
kilometre-deep displacement of formation waters (as shown in Figure 2b), which is questionable
due to depth-dependent decreases in formation permeability, fluid density contrasts between
fresh recharge water and resident brines, and the lack of an exfiltration zone, is not required if
lateral discharge can occur.
Following complete retreat of continental ice sheets, it can be envisioned that the flow
regime in the aquifer units continues to evolve. Zones of overpressurization or displaced water
of high salinity may provide the driving force for upward movement of groundwater. Evidence for
this type of groundwater regime is provided by the presence of saline springs along basin
margins (Grasby and Betcher, 2002 as referenced by McIntosh and Walter, 2005, Ferguson et
al., 2007). Person et al. (2007) also report reflux of cold climate recharge for the Williston basin;
however, the chemical composition of the discharging water has been altered to a brine due to
rock-water interactions. In addition, upward migration may also be driven by gas production at
depth, which may increase hydraulic head in the region of gas generation and may induce
basin-scale groundwater flow towards the basin margins (Mazurek, 2004).
Evidence for past glacial melt water recharge, as discussed above, is provided by anomalous
O and δD signatures of groundwater in comparison to precipitation formed under current
climate conditions. Groundwater to depths of up to 300 metres in Silurian-Devonian aquifers
and shales along the northern margin of the Michigan Basin have δ
O of -15 to -10‰ and δD
values of -100 to -75 ‰ (McIntosh and Walter, 2005, 2006), indicating Pleistocene recharge into
these regions. Siegel et al. (1991) also provided evidence for Pleistocene recharge into the
Cambrian-Ordovician aquifer of Iowa. Most recently, Person et al. (2007) conducted a review
- 6 -
that compiled isotopic data from various sedimentary basins and suggested that recharge of
Pleistocene waters is a common feature at the margins of sedimentary basins.
On the other hand, data presented by Dollar et al. (1991) for southwestern Ontario
supports recent (< 10,000 years) ingress of freshwater only into Devonian units (to a depth of <
100 m). Unlike the situation in the northern part of the Michigan basin (McIntosh and Walter,
2005, 2006), available oxygen isotopic data suggests that deeper units (Silurian, Ordovician and
Cambrian) have not been affected by recharge of meteoric water. These units show a signature
of very long residence times and in situ-rock-water interaction (Mazurek, 2004).
In summary, although ingress of meteoric water seems to be common, isotopic data
suggests maximum ingress depths of approximately 300 m.
The initially low salinity of recharging freshwater may be significantly increased due to rock-
water interaction (e.g. Wilson and Long, 1993a, McIntosh and Walter, 2005), a process which
may completely overwhelm the freshwater signature. The presence of high salinity groundwater
at depth therefore does not provide conclusive evidence for the absence of glacial melt water
recharge. This argument is supported by previous studies, which suggest that rock-water
interaction has played a dominant role in determining the composition of deep basal brines.
Evaporation alone was found to be insufficient for explaining observed water compositions in
many cases (e.g. Wilson and Long, 1993a,b, McIntosh and Walter, 2006). The main
geochemical processes affecting major ion composition could include carbonate mineral
dissolution, dolomitzation (Wilson and Long, 1993a,b, McIntosh and Walter, 2006),
dedolomitization (McIntosh and Walter, 2006), halite dissolution (Wilson and Long, 1993a,b),
anhydrite dissolution (McIntosh and Walter, 2006), ion exchange of Na with Ca (Wilson and
Long, 1993a, McIntosh and Walter, 2006), and to a lesser degree dissolution-precipitation of Al-
silicate phases (Wilson and Long, 1993a, McIntosh and Walter, 2006).
Dissolution and precipitation reactions for the carbonate minerals calcite and dolomite
are given by the following:
22)( HCOMgCaHCOCaMg (2)
In groundwaters that are relatively depleted in Ca, but enriched in Mg, it is possible that calcite
dissolves and dolomite precipitates, leading to the process of dolomitization (Wilson and Long,
1993a,b, Mazurek, 2004, McIntosh and Walter, 2006):
)(2 CaCOCaMgCaCOMg (3)
This reaction is important because it can enhance porosity and hydraulic conductivity of the
sediments (Weaver et al., 1995) and can provide a feedback on flow and transport processes.
- 7 -
The salinity of the groundwater is mainly controlled by the dissolution of evaporite
minerals including gypsum:
22 ++
anhydrite (McIntosh and Walter, 2006):
SOCaCaSO (5)
and halite (e.g.: Wilson and Long, 1993a,b):
+ ClNaNaCl (6)
Ion exchange reactions with clays may have an additional effect on groundwater
composition (McIntosh and Walter, 2006), although this class of reactions does not significantly
affect salinity:
++ NaXCaCaXNa 22
It can be expected that the reactions described by equations 1-7 will also take place in
response to the infiltration of glacial melt water. These reactions will significantly increase the
salinity of the groundwater and its density, while maintaining a circum-neutral pH value. As
previously mentioned, some groundwaters that have been identified as Pleistocene recharge,
based on their isotopic signature, have attained the geochemical composition of a brine (e.g.
McIntosh and Walter, 2005, 2006). Considering that changes in salinity are also associated with
changes in fluid density, a direct feedback between geochemical reactions and flow and
transport processes can be expected.
In terms of redox state, the significant organic carbon content typically present in
sedimentary formations, the abundance of hydrocarbon deposits, and the occurrence of
methanogenic conditions, suggest the development of highly reducing conditions at shallow
depths (McIntosh and Walter, 2005, 2006). In this type of environment, oxygen in recharge
water is rapidly consumed by reaction with organic carbon contained in the sediments:
++ (8)
Depending on the depositional history, it can also be expected that pyrite, other sulfide
minerals, and Fe(II)-containing mineral phases supply further redox reduction capacity (Dollar et
al., 1991, McIntosh and Walter, 2005). For example, the oxidation of pyrite also consumes
++ (9)
For the Michigan basin, McIntosh and Walter (2006) showed that pyrite oxidation is occurring in
the shallowest portion of the aquifers.
At greater depth, sulfate reduction and methanogenesis are the dominant redox
processes in sedimentary basins (Wilson and Long, 1993, Martini et al., 1998, McIntosh et al.,
- 8 -
2002, 2004, McIntosh and Walter, 2005, 2006). Sulfate reduction may occur if sulfate rich
groundwater is displaced into regions with high labile organic carbon content:
Sulfate may also originate from the oxidation of pyrite by oxygen (equation 9). If sulfate
becomes depleted, and organic carbon is still present, it is likely that further organic matter
degradation by fermenters and methanogenic bacteria takes place:
2 COCHOCH + (11)
However, microbially mediated reactions are inhibited under highly saline conditions,
which may explain the persistence of sulfate over long periods of time despite the presence of
organic matter. Recent work by Martini et al. (1998) and McIntosh and Walter (2005, 2006)
provide evidence that the ingress of glacial melt water into Silurian/Devonian aquifers of the
northern Michigan basin and its subsequent ingress into the overlying Antrim shale has
enhanced microbial activity and has caused significant microbially-mediated gas production
(predominantly methane). In this case, the groundwater signature remained relatively fresh due
to the lack of soluble mineral phases such as halite, which facilitated the enhancement of
microbial activity and led to the generation of economically viable gas deposits in the Michigan
basin (McIntosh and Walter, 2006).
In summary, the literature suggests that rock-water interaction will increase the salinity of
ingressing glacial melt water. The degree of salinity gain is controlled by sediment composition
and the availability of soluble minerals such as halite and other evaporate mineral phases. The
literature also suggests a strong redox buffer capacity likely capable of buffering glacial melt
water that may contain a significant degree of excess air with dissolved O
elevated by a factor of up to five (MacQuarrie and Mayer, 2005).
The general hydrogeological and geochemical concepts discussed above are now put into the
context of the Canadian portion of the Michigan Basin. This is done to provide a more definitive
example of how groundwater flow and geochemical reactions may be invoked to constrain the
hydrogeological and geochemical evolution at a specific site. This sedimentary basin has been
selected for discussion because of the relative abundance of data that has been assembled by
others for the region underlying southwestern Ontario (e.g. Mazurek, 2004) and for the northern
part of the Michigan basin (McIntosh and Walter, 2005, 2006). However, with sufficient data
similar site-specific conceptualizations could be developed for other sedimentary basins.
As pointed out by Dollar et al. (1991), the Michigan basin is a remarkably circular feature
consisting of symmetrical bands of outcropping aquifer and aquitard units, which supports the
translation of flow and reactive transport concepts from the region of the basin investigated by
McIntosh and Walter (2005, 2006) to southwestern Ontario. However, differences may exist in
aquifer hydraulic conductivity and fracturing of the aquitards in these two regions of the basin,
which may affect system evolution in the two regions in different ways. In addition, it is also
important to evaluate if the nature of glaciation has been similar in both regions, considering
that recharge will be very different under warm-based and polar-based glaciers (Person et al.,
- 9 -
2007). Recently, a comprehensive geosynthesis of the existing geoscientific literature for
Southern Ontario has been conducted (Mazurek, 2004). The hydrogeological and
hydrogeochemical information for southwestern Ontario and its interpretation with respect to
Quaternary glaciations (Mazurek, 2004) provides the framework for discussing the applicability
of the concepts put forward by McIntosh and Walter (2005, 2006) for the northern Michigan
Figure 3: Cross-sectional view of region under consideration. Ordovician units
include sandstones and dolostones (Shadow Lake Fm), limestones (Trenton- Black River
Gp (cf. Simcoe Gp)), and shales (Blue Mountain Fm, Georgian Bay Fm, and Queenston
Fm). Silurian units include the Cabot Head Fm (shale), Manitoulin Fm (dolomite),
Whirlpool Fm (sandstone), a group of dolomites (Lockport Fm, Amabel Fm, Guelph, Fm),
the Salina Fm (evaporates with limestones), and Bass Island Fm (dolomite). The Bois
Blanc Fm (dolomite), Detroit River Gp (dolomite and limestones), Hamilton Gp (shales
and limestones) and Kettle Point Fm (shale) are Devonian in age.
A cross-section of the sedimentary units of southwestern Ontario (Figure 3; after
Mazurek, 2004) shows that several major formations overlie the Precambrian crystalline
basement, and that these gradually dip to the southwest towards the center of the Michigan
basin. The stratigraphic sequence includes Middle and Upper Ordovician units (limestones and
shales), as well as Silurian and Devonian units. In the region of the Algonquin Arch, which
underlies southwestern Ontario, pre-existing Cambrian and Lower Ordovician rocks have been
eroded prior to deposition of the younger rock formations (Ziegler and Longstaffe, 2000). Middle
and Upper Ordovician rocks include the units that have recently been identified as most suitable
for the purpose of long-term radioactive waste management (Mazurek, 2004). Ordovician
limestones constitute most of the Trenton and Black River Groups (cf. Simcoe Gp.), while
Ordovician shale units include the Blue Mountain Formation, Georgian Bay Formation and
Queenston Formation (Mazurek, 2004).
- 10 -
To apply the concepts introduced in section 2 in a site specific context, it is useful to compare
the hydraulic conductivities of the sedimentary rock units in southwestern Ontario to the values
reported for other sedimentary basins, and most importantly for the northern Michigan basin
(McIntosh and Walter, 2005, 2006). Such an assessment will help to develop a framework for
evaluation of the ingress mechanisms depicted in Figure 2. For rock units that can potentially
act as aquifers (sandstones and carbonates), a comparison of horizontal hydraulic
conductivities is informative, while vertical hydraulic conductivities and, in particular, evidence
for the presence of vertical fractures are most relevant for the confining units (typically shales).
Figure 4: Comparison of hydraulic conductivity values for region of interest from a)
Mazurek (2004) and b) McIntosh et al., 2005. Hydraulic conductivities for the Devonian
carbonate rock units selected by McIntosh et al. (2005) compare well with the average
values by Mazurek (2004). Hydraulic conductivities for the Silurian-Ordovician carbonate
rock units used by McIntosh et al. (2005) are more than 3 orders of magnitude higher
than average values reported by Mazurek (2004). Hydraulic conductivities reported by
McIntosh et al. (2005) originate from a conference poster and cannot be traced back to
original references.
3.1.1 Horizontal Hydraulic Conductivities of Sandstone and Carbonate Rock Units
McIntosh et al. (2005) employed a density dependent flow and transport model to simulate
ingress of glacial melt water into the sedimentary units of the northern Michigan basin under
glacial loading. These simulations predicted fairly deep ingress of melt water (~ 300 m) into
Devonian, Silurian, Ordovician and Cambrian rock units. Figure 4 presents the hydraulic
conductivity values that were used by McIntosh et al. (2005) for the respective units. These
values can be compared to a compilation of hydraulic conductivity data for sedimentary rocks
underlying southwestern Ontario (Mazurek, 2004). The data presented by Mazurek (2004)
- 11 -
stems from in-situ packer tests in boreholes and are generally considered conservative
estimates of hydraulic conductivity (i.e. actual hydraulic conductivities may be lower).
While geometric means (shown as black bars in chart by Mazurek (2004), Figure 4) for
all shale formations and Devonian carbonate rock units are in good agreement, McIntosh et al.
(2005) have assigned significantly larger hydraulic conductivity values for the Silurian and
Ordovician rock formations (up to 3 orders of magnitude higher, these values originate from a
conference poster and cannot be traced back to original references). These are the units which
were simulated to have some of the deepest ingress of glacial melt water (McIntosh et al.,
2005). It must be noted that hydraulic conductivity values for limestones compiled by Mazurek
(2004) also encompass more shallow units, and the maximum values for Silurian-Ordovician
carbonate rock units were < 10
with a geometric mean on the order of 10
Figure 6-9 in Mazurek, 2004). This provides further support that the high hydraulic conductivity
values used by McIntosh may not be applicable for the Ordovician and Silurian rocks of
southwestern Ontario. However, the review by Mazurek (2004) also provides evidence that
Ordovician limestones may locally be more permeable. The enhanced hydraulic conductivity is
believed to be caused by local dolomitization along faults (Mazurek, 2004). Such a local
hydraulic conductivity enhancement is supported by isotopic data, which suggests mixing of
waters between sandstone units and Ordovician limestones (see section 3.2, Dollar et al., 1991)
McIntosh et al. (2005) also used a relatively high hydraulic conductivity of 2.0 x 10
m s
for the basal sandstone unit (3 orders of magnitude larger than the highest measured sandstone
value in southwestern Ontario). The basal sandstone units in southwestern Ontario are thin and
restricted to the Shadow Lake formation. In addition, there are facies changes in the Shadow
Lake formation across southern Ontario with gradation to a dolostone towards the region where
this unit outcrops. Although Cambrian sandstones are present in the northern part of the
Michigan basin, these were eroded in the region of the Algonquin Arch prior to deposition of the
overlying units (Ziegler and Longstaffe, 2000). As a result, Cambrian units do not outcrop at the
basin margin in the region north of Toronto. The hydraulic conductivity reported in the
compilation by Mazurek (2004) for sandstones is based on a single unit located in Silurian rocks
(Whirlpool formation), and may not be representative for sandstones contained in the Shadow
Lake Fm. In regions where the Cambrian sandstones were not eroded, large flows obtained
from wells support relatively high permeabilities in these rocks (Dollar et al., 1991).
3.1.2 Vertical Hydraulic Conductivities in Confining Units (Shales) and Evidence for
Considering the low hydraulic conductivity of the confining beds, the presence of fractures and
fault zones is a key factor for vertical migration through aquitards. The presence of fractures
could provide a major control for the overall evolution of the flow field during glacial melt water
ingress (Figures 2a and 2c). As summarized by Mazurek (2004), Sanford et al. (1985)
developed a conceptual fracture network model for southwestern Ontario. The presence of
three fracture systems was hypothesized; however, only a small density of faults was predicted
for the Bruce Megablock when compared to the Niagara Megablock. In addition only one of the
three fracture systems is thought to be prevalent in the Bruce Megablock (Sanford et al., 1985).
Reactivation of the fracture zones is believed to be restricted to orogenic events, the most
recent of which occurred approximately 250 million years ago during the late Paleozoic-early
Mesozoic (Sanford et al., 1985). Although some evidence exists for the presence of neotectonic
fractures in the Lockport Formation of southwestern Ontario, these features are associated with
the retreat of the Niagara escarpment (Gross and Engelder, 1991). Fluid overpressures have
- 12 -
been identified in Silurian and Devonian formations in southern Ontario (Raven et al., 1992,
Novakowski and Lapcevic, 1988). The fact that supernormal pressures exist suggest that
fracturing is limited at present time and that the bulk hydraulic conductivity is very low (Raven et
al., 1992). Overall, except for one local example in the Devonian units overlying the Salina
formations (Weaver et al., 1995), there is very little evidence for vertical fracturing of the
confining units and the hydraulic conductivity both in vertical and horizontal directions is likely
low. This suggests that the potential for vertical flow through aquitards, as has been depicted in
Figure 2a and c, is low.
Information on the isotopic composition of groundwater is also useful to determine if infiltration
of glacial melt water has occurred in southwestern Ontario. In this context, it must be
acknowledged that extensive data sets, such as those for the northern Michigan basin, do not
currently exist for the Ontario portion of the basin margin. In southwestern Ontario, groundwater
with a Pleistocene isotopic recharge signature is most abundant in very shallow Quaternary
aquifers (maximum depth of 50 m) that overlie the older sedimentary rocks (Desaulniers et al.,
1981, 1986, Husain et al., 2004). As noted in section 2.2., Dollar et al. (1991) suggest that
Pleistocene recharge may also have occurred into the Dundee Formation (Middle Devonian,
Figure 3). It is worthwhile noting that all groundwater with cold water isotopic signatures from
Devonian formations were sampled at depths of less than 100 m, in regions where these
formations subcrop. The ingress mechanism is not discussed; however, it is suggested that the
dissolution of evaporite deposits in the underlying Salina Formation may have caused the
formation of collapse breccias within the Devonian units. The presence of these collapse
structures may have made the Dundee Formation more accessible to ingress of glacial melt
water (Dollar et al., 1991, Mazurek, 2004).
For the remaining units (Silurian, Ordovician and Cambrian), the isotopic signature and
chemical composition of groundwater sampled from oil and/or gas producing reservoirs does
not provide evidence for past ingress of glacial melt water (Dollar et al., 1991). In addition, most
units show characteristic chemical and isotopic composition indicating very limited flow across
unit boundaries. An exception is a limited number of brine samples from hydrocarbon-producing
wells within the Cambrian and lower Ordovician rock formations (Dollar et al., 1991, Mazurek,
2004) which show overlap in their stable water isotopic signatures. This suggests local
exchange occurred between the units through a possible hydraulic or diffusive connection.
The long term geochemical evolution of groundwater is largely controlled by the composition of
the sedimentary rock sequences hosting the groundwaters (McIntosh and Walter, 2006). As a
result, groundwater in southwestern Ontario has been affected by processes similar to those
identified in other sedimentary basins including the northern region of the Michigan basin. In the
case of future melt water ingress, the salinity of recharge waters is expected to be controlled by
the presence of halite and other highly soluble mineral phases. In southwestern Ontario, these
mineral phases are only abundant in the Silurian Salina Formation (Dollar et al., 1991). On the
other hand, carbonate minerals are ubiquitous, and even the shale units contain a significant
carbonate mineral fraction (Mazurek, 2004). It can be expected that equilibration of groundwater
with these phases will maintain pH values at circum-neutral levels. Organic carbon and reducing
- 13 -
mineral phases including pyrite and Fe(II) bearing chlorite are abundant in the Ordovician
formations (Mazurek, 2004), suggesting that redox processes are similar to those in the
northern Michigan basin. Both sulfate reduction and methanogenesis cause the production of
fairly insoluble gases. For example, dissolved hydrogen sulfide has been identified in the
Queenston Formation (Mazurek, 2004).
As demonstrated by the review of recent literature presented in Appendix A, multicomponent
reactive transport modelling in sedimentary basins has been rather limited, and relatively few
studies have dealt with the long-term evolution of geochemical conditions in sedimentary rock
environments. One notable exception is the study by Pearson et al. (2002), who applied one-
dimensional reactive transport modelling to the Opalinus Clay formation in a specific attempt to
better interpret the dilution of sea water in the formation by fresh water present in the bounding
A number of studies have used reactive transport simulations to investigate CO
sequestration in geological settings that share some similarities with the sedimentary rock
sequences being considered for waste repositories. These modelling investigations are
typically localized to either consider processes in the shale cap rocks, or the immediate vicinity
of the sandstone-shale interface (e.g. Gaus et al., 2005, Xu et al., 2005, Figure 5). These
studies have, however, highlighted the importance of having detailed initial mineralogical
characterization in aquitard formations (see Appendix A for details).
Figure 5: Schematic representation of modelling domains, shown in relation to a
sedimentary basin cross section, used in previous reactive transport studies. These
studies include the characterization of diffusive transport in rock formations considered
for deep geologic repositories, natural analogue studies for uranium migration and
attenuation, several modelling studies on geologic carbon sequestration, and the
simulation of the interactions between saline and fresh waters in coastal aquifer systems
(see Appendix A).
Although no previous studies have directly considered the ingress of dilute glacial melt
waters into sedimentary basins, several large scale simulation studies (Figure 5) have focused
on the mixing of different water types in carbonate aquifers with spatial and time scales (e.g.
10,000 to 1 million years) that are relevant to nuclear waste isolation. In addition, the influence
- 14 -
of variable fluid densities (Rezaei et al., 2005), heat transport, and high ionic strength on
aqueous geochemistry have been considered in some studies (Jones and Xiao, 2005). As
noted by the review in Appendix A, one relatively common focus of previous studies has been
the simulation of porosity evolution in aquifers and aquitards, and in some cases the feedback
between porosity and permeability (Jones and Xiao, 2005).
In this section a general framework is developed that describes the interactions between fluid
flow and the geochemical evolution of groundwater in sedimentary basins subjected to glacial
cycles. This model integrates information for sedimentary basins and also includes information
from previous reactive transport modelling studies in sedimentary rocks (Appendix A). The key
governing equations that are required to simulate such a system are also summarized.
Based on the review conducted in section 2, it appears possible that glacial melt water with a
fresh water signature, and elevated O
concentration, may recharge into outcropping carbonate
and/or sandstone formations at the margins of sedimentary basins (Figure 6). To what depth
this recharge migrates is dependent on the driving force (i.e. hydraulic gradients) imposed by
the water table in an overlying warm-based ice sheet during periods of glaciation, and the
permeabilities of the various rock formations. If the driving force and hydraulic conductivities are
sufficiently high, dense brines which are present at depth and are stagnant under present day
conditions may be mobilized and displaced. Theoretically, displacement within the aquifer may
result in groundwater flow to greater depth or in lateral directions. Alternatively, displacement
may occur upwards through overlying confining units. It must be clarified that little evidence
exists for upward migration of brines in the geologic past.
To what degree geochemical reactions modify water quality during the recharge process
will depend on the composition of the rock units. Generally, the dissolution of carbonate
minerals is expected to keep the pH in a circum-neutral range. Overall, mineral dissolution,
including the dissolution of soluble phases, such as anhydrite, gypsum and halite will increase
the ionic strength of the solution, and as a result also the solution density. If solution density is
affected significantly, the fluid flow may be impacted leading to a direct coupling between
reaction processes, transport and flow. Groundwater composition may be further affected by ion
exchange reactions. An additional feedback mechanism between reactions, flow and transport
may be caused by porosity and permeability enhancement or reduction due to mineral
dissolution-precipitation reactions.
Furthermore, it is likely that organic matter and reduced mineral phases such as pyrite,
chlorite and other Fe(II)- and sulfide-bearing phases lead to the consumption of oxygen that is
dissolved in the recharge water (Figure 6). Other redox reactions such as sulfate reduction or
methane generation may also play a role. A feedback may exist between fluid density and the
rate of these reactions, considering that highly saline conditions tend to strongly inhibit microbial
activity (McIntosh and Walter, 2005).
- 15 -
Figure 6: Conceptual model for groundwater evolution in response to melt water
ingress. High hydraulic gradients cause enhanced recharge of melt water containing
high dissolved oxygen (DO). Carbonate and evaporite mineral dissolution causes an
increase in TDS and fluid density, possibly associated with porosity and permeability
increase. The presence of reduced mineral phases and organic matter (OM) causes
oxygen consumption, possibly sulfate reduction and methanogenesis. Solid phase
constituents are shown in grey boxes (modified after McIntosh and Walter, 2005).
To describe reactive transport in sedimentary basins, it is necessary to consider the coupling
between density-dependent flow and transport and geochemical reactions (e.g. Rezaei et al.,
2005, Jones and Xiao, 2005). The governing equation for density-dependent fluid flow in
saturated porous media can be written as (Bear, 1979):
() ( )
is the porosity [L
void L
porous medium],
is the fluid density [ML
], q
is the
specific discharge vector [LT
], and Q
represents fluid sources/sinks [T
]. The corresponding
reactive transport equations for N
aqueous components can be written in terms of the total
component concentration T
() ( ) ( )
where D
is the dispersion tensor, Q
[mol dm
porous medium] and Q
[mol dm
medium] are internal source and sink terms from kinetic intra-aqueous and heterogeneous
reactions, and Q
[mol dm
porous medium] is an external source and sink term (Mayer et
- 16 -
al., 2002). This equation also implicitly considers equilibrium reactions through the total
concentration terms T
(Mayer et al., 2002, MacQuarrie and Mayer, 2005).
The fact that fluid density is directly related to solution chemistry can be expressed as a
linear function of total dissolved solute concentrations c
(Voss and Souza, 1987, Kharaka et
al.,1988, Guo and Langevin, 2002):
is the freshwater density and ρ/c
is a unitless constant, approximately 0.688
(Kharaka et al.,1988). The total solute concentration c
is calculated based on the sum of the
total concentrations T
. Alternatively, more sophisticated relationships can be used which are
based on the Pitzer ion interaction model (Monnin, 1989). In any case, this relationship directly
accounts for the coupling between fluid flow and reactive solute transport.
The reduction in aquifer hydraulic conductivity due to the dissolution and precipitation of
minerals can be described by a number of different permeability-porosity relationships
(MacQuarrie and Mayer, 2005). A model commonly used is provided by the Carman-Kozeny
relationship (e.g.: Le Gallo et al. 1998):
where k
and k
represent the current and initial media permeabilities. Additional functional
relationships can be included, for example to describe reaction rates as a function of salinity.
This section assesses the general applicability and limitations of reactive transport codes for
modelling rock-water interactions, including the attenuation of oxygen and mineral precipitation-
dissolution reactions and related porosity changes for the more dilute, shallow groundwater in
sedimentary formations and for the highly saline groundwater found at depth.
Numerous models exist that are capable of simulating the transport processes and
biogeochemical reactions (dissolution-precipitation, ion exchange, and microbially mediated
oxidation reduction reactions) relevant in sedimentary basins (e.g.: review by MacQuarrie and
Mayer, 2005). However, considering the significant salinity gradients that may be present, a
coupled treatment of geochemical reactions and density dependent flow and transport is
needed and fewer codes meet these requirements (Freedman and Ibaraki, 2002, Lichtner et al.,
2004, Kim et al., 2004, Mao et al., 2006, Henderson and Mayer, 2007). Although these codes
are theoretically suitable, they have not been used to date to evaluate the long-term
geochemical evolution in sedimentary basins. The effect of dissolution-precipitation reactions on
porosity and permeability has also been implemented in reactive transport codes (e.g., Steefel
and Lasaga, 1994, Bethke, 2002, Lichtner et al. 2004, Henderson and Mayer, 2007). In
addition, the capability of these models to deal with strong salinity contrasts, locally diffusion-
controlled transport, microbially mediated gas production, and relatively complex geometries
needs to be evaluated.
- 17 -
An important aspect of reactive transport modelling is the appropriate evaluation of activity
coefficients both under dilute and highly saline conditions. Most reactive transport codes rely on
the Davis equation or some form of the extended Debye-Huckel equation to determine the
activity of dissolved ions as a function of ionic strength. However, these activity relationships are
only applicable in relatively dilute solutions up to an ionic strength similar to that of sea water.
Highly saline solutions and brines found in sedimentary basins require an alternative approach
for calculating the activities of dissolved ions. One of the most common approaches is the Pitzer
ion interaction model (Pitzer, 1973, Pitzer, 1991). The Pitzer ion interaction model is based on a
virial expansion, which reduces to a modified Debye-Huckel formulation at low concentrations
(Lichtner and Felmy, 2003). In addition to the effects of the formation of complexes and ion
pairing, activity coefficients calculated using the Pitzer formalism account for electrostatic
interactions and ion hydration effects (Lichtner and Felmy, 2003).
The Pitzer equations have been implemented in geochemical equilibrium models to
extend their applicability to high ionic strength solutions. For example, the Pitzer formalisms
were introduced into the PHREEQE code, resulting in PHRQPITZ (e.g. Plummer et al., 1988,
Plummer and Parkhurst, 1990). More recently, the Pitzer equations were also successfully
implemented in a number of reactive transport codes (Bethke, 2002, Lichtner and Felmy, 2003,
Bachler and Kohl, 2005, Zhang et al., 2005) and in version 2.12 of PHREEQC (Parkhurst and
Appelo, 1999), which is also available in PHAST (Parkhurst et al., 2004). Although the Pitzer
ion-interaction model is complex in nature, implementing the equations in reactive transport
models does not pose a major difficulty. The activity coefficients are calculated on a nodal basis
and can be lagged a time step behind the reactive transport calculations (Lichtner and Felmy,
2003), which simplifies the solution algorithm while not significantly affecting the accuracy of the
results. In addition, a numerical implementation of the Pitzer equations is available as public
domain software through the USGS (PHRQPITZ, e.g. Plummer et al., 1988, Plummer and
Parkhurst, 1990). This computer code and the associated database can be modified and
implemented in other reactive transport models. Applications of reactive transport models that
include activity corrections based on the Pitzer equations include the assessment of
contaminant transport at the Hanford tank farm (Lichtner et al., 2004, Zhang et al., 2005),
dolomitization in a regional carbonate aquifer (Jones and Xiao, 2005), and the simulation of a
geothermal energy system (Bachler and Kohl, 2005).
One of the main limitations of the Pitzer approach is that a large number of ion
interaction parameters are required. These parameters can only be obtained through extensive
laboratory experimentation over a range of pressures and temperatures (Lichtner and Felmy,
2003). As a result, parameters are only available for a select group of ions. However, progress
has been made during recent years and in addition to the available ion interaction parameters
for major ions, Pitzer coefficients have been obtained for select Si and Al species (Park and
Englezos, 1999, Felmy et al., 2001). Moller et al. (2007) report further progress with respect to
the inclusion of Al and Si species into the Pitzer equations, and data for select trace elements
are also becoming available (Lichtner and Felmy, 2003).
For the reactions that control salinity in sedimentary basins (equations 1-6), the present
set of Pitzer parameters already provides a suitable foundation for reactive transport
calculations. Combined with the fact that source code is available, there appear to be no major
obstacles for implementing the Pitzer ion interaction equations into reactive transport models.
- 18 -
For the investigation of the long-term evolution of groundwater geochemistry in sedimentary
basins, it is also not anticipated that the computational burden associated with using the Pitzer
activity correction will be limiting.
Due to the low permeability of aquitard units, mass transport in these rocks is often diffusion-
controlled (Mazurek, 2004). Although diffusive transport is commonly described by Fick’s law, it
is well known that its applicability is limited, in particular in electrolyte solutions and in fine
grained porous media rich in clay minerals (Oelkers, 1996, Appelo and Wersin, 2007).
Fick’s law assumes that diffusion is solely a function of a concentration gradient, and
that interactions between different solutes do not take place. This is not strictly true because
many ions present in solution carry a charge. As a result, the migration of ions by diffusion
causes a charge imbalance, resulting in an electrical field that drives electrochemical migration
to counteract the charge imbalance (Lichtner, 1996). The generation of a charge imbalance due
to diffusive migration is aggravated because each ion has a unique diffusion coefficient
(Lichtner, 1996), leading to different rates of migration. Failure to account for the effect of
electrical potential and interactions between the ions can lead to large errors in mass balance
calculations (Oelkers, 1996, Li et al., 2007). In extreme cases, “diffusion” of a charged ion can
occur in the direction against its concentration gradient. This is of particular importance in
regions of steep concentration gradients, such as boundaries between aquifer and aquitard
units and in solutions with high ionic strength. As a result, to adequately describe
multicomponent transport in diffusion-dominated systems, a formulation is required that includes
species-dependent diffusion coefficients and an electrochemical migration term (Lichtner, 1996,
Oelkers, 1996, Giambalvo et al., 2002).
Additional complications are introduced if diffusion occurs in fine grained deposits
containing reactive mineral surfaces with high surface areas. In these materials, charged ions
participate in surface complexation reactions. Most mineral surfaces are negatively charged,
which favors surface complexation of cations. Some cations do not form inner-sphere surface
complexes at the mineral surface, but remain predominantly in the diffusive double layer (DDL),
a layer of stagnant water surrounding the minerals. The ions remain mobile in this region and
are still able to diffuse along a concentration gradient. These processes enhance cation
migration, but limit the migration of negatively charged species by anion exclusion (Appelo and
Wersin, 2007). The thickness of the diffusive double layer is a function of salinity and decreases
with increasing ionic strength (Stumm and Morgan, 1996, Boudreau et al , 2004). However,
according to Appelo and Wersin (2007), even for solutions with an ionic strength similar to
seawater, a significant fraction of cation migration may occur in the DDL.
Despite the shortcomings of Fick’s law, multicomponent diffusion formulations have not
been used commonly in reactive transport modelling studies. For example, Pearson et al.
(2002) conducted a reactive transport study for the Opalinus Clay under diffusion-controlled
conditions based on Fick’s law. The authors of this work acknowledge that the mismatch of
observed and simulated data and uncertainties of the model predictions may be partially
attributable to the use of a single diffusion coefficient for all species. More recently, Appelo and
Wersin (2007) used a modified version of PHREEQC to simulate multicomponent diffusive
transport of tritium, iodide, and sodium in Opalinus clay and accounted for surface complexation
reactions. Laboratory data and simulations demonstrated that tortuosity was increased for
iodide (a negatively charged ion) due to anion exclusion, while it was lowered for sodium
- 19 -
(positively charged) due to enhanced migration in the diffusive double layer. Other applications
of multicomponent diffusion models have focused on migration of Cl
, Na
, K
and OH
concrete (Truc, 2000, Lorente et al, 2003).
Mathematical formulations for multicomponent diffusion controlled by electrochemical
gradients, including species-dependent diffusion coefficients, and considering surface
complexation are available (Oelkers, 1996, Truc et al., 2000, Appelo and Wersin, 2007). The
development effort associated with the implementation of multicomponent diffusion into reactive
transport models is considered moderate.
The occurrence of glaciation and deglaciation events contributes additional complexity to the
system under consideration. The advance of an ice sheet over sedimentary rocks affects the
stress regime in the subsurface, which may in turn affect porosity and hydraulic conductivity of
both aquifer and aquitard units. The weight of the ice sheet may also result in a change in land
surface elevation due to compression of the lithosphere with subsequent rebound following
deglaciation. Together, these processes cause transient changes in boundary conditions and
system properties which may affect the flow regime and reactive transport. However, to date
there have been no attempts to implement or couple these processes within a reactive transport
modelling context. Considering the significant complexity associated with simulating glacial
compression and rebound, as well as multicomponent reactive transport, it does not appear to
be warranted to develop a combined model at this stage. The development effort is expected to
be significant, and computational requirements would likely be prohibitive and not allow
beneficial use of such a model.
As pointed out by Martini et al. (1998) and McIntosh et al. (2002, 2004), the generation
of biogenic gas (predominantly methane) may occur in sedimentary rock systems, in particular if
fresh water ingress occurs into units with abundant organic matter. Although gas solubility
increases with depth, exsolution may occur if gas generation is significant. Under these
conditions, it may be possible that resident groundwater is displaced, that gas is transported
towards the ground surface along bedding planes or fracture systems, and that the relative
permeability for water flow is reduced. An adequate description of these processes would
require a multi-phase flow approach coupled with a reactive transport model. Modelling
techniques and formulations for implementing multi-phase flow are available; however, the effort
of implementation into a reactive transport model is significant because gas migration and
generation need to be coupled directly. The TOUGH-REACT code (Xu and Pruess, 2001),
which has been applied at the Yucca Mountain site (Spycher et al., 2003), is probably the most
advanced reactive transport code that contains multi-phase capabilities. However, simulations
that couple multi-phase flow and reactive transport on the time and spatial scales of interest in
the current context are beyond today’s computing capabilities. To adequately describe gas
generation, it will also be necessary to account for the fact that microbially mediated reactions
are inhibited in the presence of highly saline solutions or brines. Including inhibition as a
function of salinity is not a common feature in reactive transport codes; however, provided that a
functional relationship is available that adequately describes this dependence, its numerical
implementation is expected to be straightforward.
Flow, transport and reaction processes can also be affected by heat flow, which is not
often included in reactive transport models. Temperature is directly coupled to fluid density;
however, for the strong salinity gradients observed in sedimentary basins, it can be expected
- 20 -
that salinity-induced density gradients will overwhelm temperatureinduced density changes. The
effect of temperature on mineral solubilities and biogeochemical reaction rates may be more
significant. Although the simulation of heat flow may be required for the investigation of reactive
transport in the near field of a nuclear waste repository (Spycher et al., 2003), it may be
sufficient to base simulations on a depth-dependent temperature profile (e.g. Mayer et al., 2006)
if the temperature field is a function of depth, for example determined by a geothermal gradient.
Formulations for heat flow and conduction in the subsurface area available and the
implementation effort is considered moderate.
Lastly, most reactive transport models developed to date rely on an orthogonal grid
structure, which is typically chosen to facilitate local mass conservation – a requirement for
reactive transport codes. However, such a discretization scheme makes it awkward, although
not impossible, to discretize sedimentary basins including multiple aquitard and aquifer
formations. This is particularly true near basin margins, where these units outcrop. A more
suitable formulation would be provided by a control volume finite element discretization (e.g.
Durlofsky, 1994), which to the best of our knowledge has not been incorporated to date in
multicomponent reactive transport models. The effort for implementing an unstructured grid is
considered moderate to significant.
Although modelling studies have been conducted to evaluate the influence of glacial recharge
into sedimentary basins (Boulton et al., 1994, Person et al., 2003, McIntosh et al., 2005, Person
et al., 2007), none of these studies have considered the effect of geochemical reactions on
groundwater salinity, pH or redox conditions. The major recommendation of this review is that
work needs to be initiated in this area to, as mentioned in Section 1, provide a means to
demonstrate how groundwater composition could evolve in relation to hydrological perturbations
that may result from glaciation of sedimentary rock sequences.
Three major categories of recommendations are made which relate to a) model
developments and improvements, b) model applications to evaluate the conceptual models
discussed in this report, and c) model parameters and boundary conditions that are required to
constrain the simulations.
None of the computer codes currently available is capable of simulating reactive transport in
sedimentary basins over time scales of > 10,000 years, while including all processes introduced
above. The review conducted by MacQuarrie and Mayer (2005) compared different reactive
transport modelling techniques and model implementations and demonstrated that the MIN3P
code (Mayer et al., 2002) has comparable capabilities to other state-of-the-science reactive
transport models. The MIN3P code has previously been used to assess redox stability in
crystalline rock of the Canadian Shield (Spiessl et al., 2007), and we recommend to further
improve this model to meet the needs for simulations in sedimentary systems. Specifically within
the context of the MIN3P code, there is a need to address several of the outstanding issues
discussed in Section 6. Namely, we are recommending that the code be improved to handle the
following situations:
- 21 -
High ionic strength solutions – as identified throughout this report, the presence of brines
in sedimentary formations, or the evolution of fresh water towards a brine, means that
there is a need to have specific formulations for activity coefficients and ionic strength
effects. Implementation of the Pitzer formulations currently appears to be the most
appropriate way to deal with this need.
Species dependent diffusion – the inclusion of species dependent diffusion has been
identified as important to adequately describe diffusion dominated transport, particularly
when dealing with low hydraulic conductivity units. Diffusion coefficients among species
vary by a factor of up to 10 (Lichtner, 1996), diffusive transport may be significantly
affected by electrochemical forces, under extreme conditions causing diffusion against
the concentration gradient (Oelkers, 1996), and interactions with charged mineral
surfaces in fine grained deposits causes significant differences in migration
characteristics between anions and cations (Appelo and Wersin, 2007). These
considerations suggest that using a single diffusion coefficient for all species, and
ignoring electrochemical migration and the effect of surface charge can lead to an
erroneous assessment of transport behaviour under diffusion controlled conditions. It is
recommended that species dependent diffusion be implemented in MIN3P because this
will allow the code to more comprehensively deal with diffusion-controlled reactive
Inhibition of microbial activity as a function of salinity – an interesting and potentially
important finding of the review is that high salinity solutions may act to retard microbial
activity. Such an effect may have significant implications for simulating reactive
transport in sedimentary rock sequences and it is recommended that formulations which
simulate this effect should be developed and implemented in MIN3P. In this context, it
can also be evaluated to what degree rates of kinetically controlled inorganic reactions
are affected by changes in fluid density.
Other outstanding issues which have been identified in Section 6 will also require future
model developments and improvements. Perhaps most importantly is the issue of discretization
of domains to adequately deal with sedimentary rock sequences. While the current orthogonal
discretization capabilities of MIN3P would allow for simulation of local scale problems (e.g.
diffusive-reactive transport within a single horizontal aquitard unit), the simulation of larger scale
aquifer-aquitard systems, which dip inward from the basin margin, will benefit from an improved
discretization method. It is therefore recommended that work be initiated to implement a method
of discretization that will allow for simulation of gently-dipping sedimentary units.
Application of reactive transport models to evaluate the potential conceptual models presented
in this report is recommended; however, it must be acknowledged this cannot be done in
isolation from other studies that are currently aimed at understanding ice sheet dynamics and
large-scale regional groundwater flow in sedimentary basins. For example, to conduct reactive
transport modelling studies with boundary conditions that adequately reflect the current
understanding of future permafrost conditions and glaciation/deglaciation cycles in southern
Ontario, it will be necessary to investigate more thoroughly the findings that have been obtained
using the University of Toronto’s Glacial Systems Model. To evaluate potential 3D effects
created by ice sheet lobes that could develop over a portion of the sedimentary basin (e.g.
Figure 2d), it may be necessary to conduct more detailed coupled ice sheet-regional
groundwater flow modelling in addition to reactive transport modelling. However, the simulation
of reactive transport in 3D, on a basin-wide scale, is currently not recommended due to the
- 22 -
significant computational demands of such a model. Reactive transport modelling should focus
on assessing geochemical evolution and potential feedbacks between flow, transport and
reaction processes in 1D and 2D scenarios. Information gained from these simulations will be
useful to assess overall system evolution on a basin-wide scale.
Reactive transport modelling studies in 1D (e.g. for simulating solute profiles in single
aquitards) or 2D (as depicted in Section 5) are considered both computationally feasible and
informative from the point of view of supporting DGR stability. It is therefore recommended that
reactive transport modelling initially be applied to 2D cross sections that focus on the units of
interest at the basin margin (as shown in Figures 2a and c), similar to the approaches taken by
Person et al. (2003) and McIntosh et al. (2005).
Due to the large number of processes involved, reactive transport simulations require a
significant number of input parameters. It is necessary to define the geometry, material
parameters, and initial conditions in each of the sedimentary units. In addition, it is required to
define boundary conditions for flow and transport and the parameters that control the progress
of biogeochemical reactions. The following provides a more detailed summary of the required
input parameters:
Geometry and thicknesses of various sedimentary formations.
Time frame of simulation.
Hydraulic conductivities and porosities of various sedimentary units, possibly including
spatial or depth-dependent variation within these units.
Boundary conditions for flow (hydraulic heads or recharge fluxes) and transport
(recharge water composition) for present day and future conditions. The boundary
conditions must account for the effect of ice sheet advance and retreat including the
potential development of permafrost conditions.
Initial conditions including groundwater composition and sediment composition
(mineralogy, organic carbon content, and cation exchange capacity) parameters need to
be specified for the various sedimentary units.
Equilibrium constants and rates of biogeochemical reactions.
The various uncertainties that have been raised with respect to the potential conceptual
models, for example the vertical hydraulic conductivities of shale units and local permeability
enhancements in limestone units, could then be investigated in a quantitative manner. Likewise,
issues related to density distribution and fluid composition, sediment composition (e.g.
mineralogy and TOC - information which is currently sparse), and porosity and permeability
distributions could be feasibly explored in 2D reactive transport simulations.
This review was conducted to determine if multicomponent reactive transport modelling can
contribute to the assessment of the long-term geochemical evolution and stability in
sedimentary rock systems. A scenario of specific interest is the infiltration of glacial melt water
into rock units of sedimentary basins during glaciation and deglaciation cycles. A literature
review has shown that there is evidence for the past occurrence of melt water infiltration to
- 23 -
several 100’s m depth in a number of sedimentary basins. However, the lack of evidence for
deep melt water ingress in southwestern Ontario suggests that processes controlling melt water
recharge in that region may differ from other parts of the Michigan basin. The literature shows
that geochemical reactions are indeed of importance and control the compositional evolution of
the infiltrating melt water, both in terms of salinity and redox state. In addition, diffusion-
controlled transport in low permeability units may also be affected by rock-water interaction due
to anion exclusion and preferential migration of cations in the diffusive double layer. In both
cases, multicomponent reactive transport modelling is required to explain changes in
groundwater composition on a mechanistic level and to adequately describe solute transport.
The suitability of existing model formulations for assessing the systems of interest was
also evaluated. In terms of geochemical reactions that are of key relevance in sedimentary
basins, adequate model formulations are currently available. The most important reactions
include aqueous complexation, ion exchange, mineral dissolution-precipitation and microbially
mediated oxidation-reduction reactions. Due to the strong salinity contrasts, it is deemed
necessary to employ a model that considers the coupling between geochemical reactions and
density-dependent flow and transport. To adequately describe diffusive transport in low
permeability units, it appears that a formulation that is based on a single average diffusion
coefficient for all species is insufficient. Diffusion coefficients among species vary by an order of
magnitude, diffusive transport may be affected by electrochemical gradients, and interactions
with the diffusive double layer in fine grained deposits cause differences in migration
characteristics between cations, anions, and uncharged species. To rigorously describe
diffusion dominated transport a multicomponent description with species-dependent diffusion
coefficients, electrochemical migration, and surface complexation is required.
To our knowledge, multicomponent reactive transport modelling has not yet been used
to investigate the evolution of sedimentary rock sequences in response to glaciation events.
However, simulations of CO
-sequestration and CO
-escape from aquifers used for
sequestration, as well as seawater infiltration into freshwater aquifers show similar
characteristics in terms of geochemical complexity, although these studies have been
conducted on a smaller scale. The success of these simulations nevertheless suggests that it is
feasible to also conduct reactive transport modelling for the scenario of glacial melt water
infiltration. For example, reactive transport modelling can be used to clearly and convincingly
demonstrate the degree to which dissolved oxygen may be attenuated in the recharge region of
a DGR host rock and how rock-water interaction transforms freshwater recharge into
groundwater with characteristics of a brine. In terms of diffusion modelling in low permeability
media, multicomponent and species-dependent formulations have recently been applied with
significant differences in diffusion rates as a function of solution composition and the role of
surface complexation reactions (Appelo and Wersin, 2007).
Although multiple reactive transport codes with advanced capabilities exist, none of the
currently available models meets all requirements for simulating the geochemical evolution of
sedimentary basins in response over time periods of > 10,000 years. The MIN3P code (Mayer
et al., 2002) belongs to a select group of state-of the science reactive transport codes
(MacQuarrie and Mayer, 2005) that meet many of the requirements and already contains a
geochemistry-density coupling and a relationship to account for permeability and porosity
changes. In addition, this code has previously been used to study redox stability in crystalline
rocks of the Canadian Shield (Spiessl et al., 2007). To ensure continuity and to improve the
capabilities of the MIN3P code for sedimentary rock systems, it is recommended to implement
a) the Pitzer ion interaction model, b) a modified formulation for microbially mediated reactions
that accounts for inhibition as a function of salinity, c) a formulation for multicomponent and
- 24 -
species-dependent diffusion, and d) discretization methods that facilitate the generation of
unstructured grids that are better capable of dealing with irregular geometry and outcropping
aquifer and aquitard units.
Multicomponent reactive transport modelling provides an approach to investigate a
series of possible conceptual models and assess parameter sensitivity for groundwater
evolution in sedimentary rocks. Similar modelling studies would also be useful for
reconstructing past melt water infiltration events, and visualizing how groundwater salinity and
redox state may have evolved. These studies could be supported by available
paleohydrogeologic data.
- 25 -
Appelo C.A.J., and P. Wersin, 2007. Multicomponent diffusion modeling in clay systems with
application to the diffusion of tritium, iodide, and sodium in opalinus clay, Env. Sci.
Technol., 41:5002-5007.
Bachler, D., T. Kohl, 2005. Coupled thermal-hydraulic-chemical modelling of enhanced
geothermal systems, Geophys. J. Internat. 161:533-548.
Bear, J., 1979. Hydraulics of Groundwater, McGraw-Hill, New York, 569pp.
Bethke, C. M., 2002. The Xt2 model of transport in reacting geochemical systems:
Hydrogeology program: Urbana, Illinois, University of Illinois, 89 p.
Boudreau B.P., F.J.R. Meysman, and J.J. Middelburg, 2004. Multicomponent ionic diffusion in
porewaters: Coulombic effects revisited, Earth Plan. Sci. Let., 222:653-666.
Boulton, G.S., T. Slot, K. Blessing, P. Glasbergen, T. Leijnse, and K. van Gijssel, 1993. Deep
circulation of groundwater in overpressured subglacial aquifers and its geological
consequences. Quaternary Science Reviews, 12:739-745.
Desaulniers, D.E., J.A. Cherry, and P. Fritz, 1981. Origin, age and movement of pore water in
argillaceous quaternary deposits at four sites in southwestern Ontario. J. Hydrology, 50:
Desaulniers, D.E., R. S. Kaufmann, J.A. Cherry, and H.W. Bentley, 1986. 37Cl-35Cl variations
in a diffusion-controlled groundwater system. Geochim. Cosmochim. Acta, 50: 1757-
Dollar, P.S., S.K. Frape, and R.H. McNutt, 1991. Geochemistry of formation waters,
southwestern Ontario, Canada and Southern Michigan U.S.A.: Implications for origin and
evolution, Ontario Geoscience Research Grant Program, Grant No. 249, Ontario
Geological Survey, Open file report 5743, 72p.
Durlofsky, L.J., 1994. Accuracy of mixed and control-volume finite element approximations to
Darcy velocity and related quantities, Water Resour. Res., 30:965-973.
Engelhardt, H., and B. Kamb, 1997. Basal hydraulic system of a West Antarctic ice stream:
Constraints from borehole observations, J. Glaciol., 43:207-230
Felmy, A.R., H. M., Cho, J. R. Rustad, M.J. Mason, 2001. An aqueous thermodynamic model
for polymerized silica species to high ionic strength, J. Solution Chem., 30:509-525.
Ferguson, G.A.G., R. N. Betcher, S. E. Grasby, 2007. Hydrogeology of the Winnipeg formation
in Manitoba, Canada, Hydrogeol. J., 15:573-587.
Freedman, V., and M. Ibaraki, 2002. Effects of chemical reactions on density-dependent fluid
flow: on the numerical formulation and the development of instabilities, Adv. Water
Resour. 25:439-453.
- 26 -
Gaus, I., M. Azaroual M, and I. Czernichowski-Lauriol, 2005. Reactive transport modelling of the
impact of CO
injection on the clayey cap rock at Sleipner (North Sea). Chemical
Geology, 217 (3-4): 319-337.
Giambalvo, E.R., C.I. Steefel, A.T. Fisher, N.D. Rosenberg, and C.G. Wheat, 2002. Effect of
fluid-sediment reaction on hydrothermal fluxes of major elements, eastern flank of the
Juan de Fuca Ridge, Geochim. Cosmochim. Acta., 66:1739-1757.
Grasby, S., and R. Betcher, 2002. Regional hydrogeochemistry of the carbonate rock aquifer,
southern Manitoba, Can. J. Earth Sci., 39:1053-1063.
Gross, M.R., and T. Engelder, 1991. A case for neotectonic joints along the Niagara
escarpment, Tectonics, 10:631-641.
Guimerà J., L. Duro, S. Jordana and J. Bruno. 1999. Effects of ice melting and redox front
migration in fractured rocks of low permeability. Svensk Kärnbränslehantering AB.
Report no. TR-99-19.
Guo, W., and C. D. Langevin, 2002. User’s Guide to SEAWAT: A Computer Program for
Simulation of Three-Dimensional Variable density Ground-Water Flow, U.S. Geological
Survey Techniques of Water-Resources Investigations Book 6, Chapter A7, 77 p.
Henderson, T.H., and K. U. Mayer, 2007. Multicomponent reactive transport modeling of
density-driven chemical oxidation of chlorinated solvents, SIAM Geosciences
Conference, Santa Fe, March 19-23.
Husain, M.M., J.A. Cherry, and S.K. Frape, 2004. The persistence of a large stagnation zone in
a developed regional aquifer, southwestern Ontario. Can. Geotech. J., 41: 943-958.
Jones, G.D., and Y.T. Xiao, 2005. Dolomitization, anhydrite cementation, and porosity evolution
in a reflux system: Insights from reactive transport models. AAPG Bulletin, 89 (5): 577-
Kharaka, Y., W Gunter , P. Aggarwal, E. Perkins, and J. Debraal, 1988. SOLMINEQ.88: A
computer program for geochemical modeling of water-rock interactions, U.S. Geol.
Survey, Open File Report 88-4227.
Kim, J., F.W. Schwartz, T.F. Xu, H. Choi, and I.S. Kim, 2004. Coupled processes of fluid flow,
solute transport, and geochemical reactions in reactive barriers, Vadose Zone J. 3:867-
Le Gallo, Y., O. Bildstein, and E. Brosse. 1998. Coupled reaction-flow modeling of diagenetic
chnages in reservor permeability, porosity and mineral compositions. J. Hydrol. 209:366-
Li, L., C.A. Peters, and M.A. Celia, 2007. Reply to “Comment on upscaling geochemical reaction
rates using pore-scale network modeling” by Peter C. Lichtner and Qinjun Kang.
Advances in Water Resources, 30: 691-695.
Lichtner, P.C., 1996. Continuum formulation of multicomponent-multiphase reactive transport,
Rev. Mineralogy, 34:1-81.
- 27 -
Lichtner P.C., and A.R. Felmy, 2003. Estimation of Hanford SX tank waste compositions from
historically derived inventories, Comp. Geosci., 29:371-383.
Lichtner P.C., S. Yabusaki, K. Pruess, and C. I. Steefel. 2004. Role of competitive cation
exchange on chromatographic displacement of cesium in the vadose zone beneath the
Hanford S/SX tank farm. Vadose Zone J., 3:203-219.
Lorente S., M. Carcasses M, J. P. Ollivier JP, 2003. Penetration of ionic species into saturated
porous media: the case of concrete, Int. J. Energy Res., 27:907-917.
MacQuarrie, K.T.B., and K.U. Mayer, 2005, Reactive transport modeling in fractured rock: A
state-of-the-science review, Earth Science Reviews, 72(3-4), 189-227.
Mao, X., H. Prommer, D.A. Barry, C. D. Langevin, B. Panteleit, and L. Li, 2006. Three-
dimensional model for multi-component reactive transport with variable density
groundwater flow, Env. Model. Software, 21: 615-628.
Martini, A.M., L. M. Walter, J. M. Budai, T. C. W. Ku, C. J. Kaiser, and M. Schoell, 1998. Genetic
and temporal relations between formation waters and biogenic methane: Upper
Devonian Antrim Shale, Michigan Basin, USA, Geochim. Cosmochim. Acta, 62:1699-
Mayer, K.U., E.O. Frind, and D.W. Blowes, 2002. Multicomponent reactive transport modeling in
variably saturated porous media using a generalized formulation for kinetically controlled
reactions, Water Resour. Res., 38, 1174, doi: 10:1029/2001WR000862.
Mayer, K.U., S.G. Benner, and D.W. Blowes, 2006. Process-based reactive transport modeling
of a permeable reactive barrier for the treatment of mine drainage, J. Contam. Hydrol.,
Mazurek, P., 2004. Long-term used fuel waste management – Geoscientific review of the
sedimentary sequence in southern Ontario. Nuclear Waste Management Organization
Background Paper 6-12, 99pp.
McIntosh, J.C., L. M. Walter, and A. M. Martini, 2002. Pleistocene recharge to midcontinent
basins: Effects on salinity structure and microbial gas generation, Geochim. Cosmochim.
Acta., 66:1681-1700.
McIntosh, J.C., L. M. Walter, and A. M. Martini, 2004. Extensive microbial modification of
formation water geochemistry: Case study from a Midcontinent sedimentary basin,
United States, Geol. Soc. Am. Bul., 116:743-759.
McIntosh, J.C., and L.M. Walter, 2005. Volumetrically significant recharge of Pleistocene glacial
meltwaters into epicratonic basins: Constraints imposed by solute mass balances.
Chem. Geol., 222: 292-309.
McIntosh, J.C., G. Garven, and J. S. Hanor, 2005. Modeling variable-density fluid flow and
solute transport in glaciated sedimentary basins, Poster presented at AGU Fall Meeting,
- 28 -
Abstract H11D-1310. Poster available at:
McIntosh, J.C., and L.M. Walter, 2006. Paleowaters in Silurian-Devonian carbonate aquifers:
Geochemical evolution of groundwater in the Great Lakes region since the late
Pleistocene. Geochim. Cosmochim. Acta, 70: 2454-2479.
Moller, N., C. Christov, and J. Weare, 2007. Thermodynamic model for predicting interactions of
geothermal brines with hydrothermal aluminum silicate minerals. Proc.: 32
on Geothermal Reservoir Engineering, Stanford University, California, Jan. 22-24.
Monnin, C., 1989. An ion interaction model for the volumetric properties of natural waters:
density of the solution and partial molal volumes of electrolytes to high concentrations at
25 C, Geochim. Cosmochim. Acta, 53:1177–1188.
Novakowski, K.S., and P.A. Lapcevic, 1988. Regional hydrogeology of the Silurian and
Ordovician sedimentary rock underlying Niagara Falls, Ontario, Canada. J. Hydrol., 104:
Oelkers, E.H., 1996. Physical and chemical properties of rocks and fluids for chemical mass
transport calculations, Rev. Mineralogy, 34:131-191.
Park, H., and P. Englezos, 1999. Thermodynamic modeling of sodium aluminosilicate formation
in aqueous alkaline solutions, Ind. Eng. Chem. Res., 38:4959-4965.
Parkhurst, D.L., and C.A.J. Appelo, 1999, User's guide to PHREEQC (Version 2) - A computer
program for speciation, batch-reaction, one-dimensional transport, and inverse
geochemical calculations: U.S. Geological Survey Water-Resources Investigations
Report 99-4259, 312 p.
Parkhurst, D.L., K.L. Kipp, P. Engesgaard, and S.R. Charlton, 2004, PHAST - A program for
simulating ground-water flow, solute transport, and multicomponent geochemical
reactions: U.S. Geological Survey Techniques and Methods 6–A8, 154 p.
Pearson, F.J., D. Arcos, A. Bath, J.Y. Boisson, A.M. Fernández, H.-E. Gäbler, E. Gaucher, A.
Gautschi, L. Griffault, P. Hernán, and H.N. Waber, 2002, Geochemistry of Water in the
Opalinus Clay Formation at the Mont Terri Laboratory, Technical Report 2003-03,
Peltier, W. R., 2002. A design basis glacier scenario, Ontario Power Generation Report No:
Person, M., B. Dugan, J.B. Swenson, L. Urbano, C. Stott, J. Taylor, and M. Willett, 2003.
Pleistocene hydrogeology of the Atlantic continental shelf, New England, GSA Bulletin,
115: 1324-1343.
Person, M., J. McIntosh, V. Bense, and V.H. Remenda, 2007. Pleistocene hydrology of North
America: The role of ice sheets in reorganizing groundwater flow systems. Rev.
Geophys., submitted June 2006.
- 29 -
Pitzer, K.S., 1973. Thermodynamics of electrolytes. I. Theoretical basis and general equations,
J. Phys. Chem., 77:268-277.
Pitzer, K.S., 1991. Activity coefficients in electrolyte solutions, 2
ed. CRC Press, Boca raton,
FL, 542pp.
Plummer, L.N., D.L. Parkhurst, G.W. Fleming, and S.A. Dunkle, 1988. A computer program
incorporating Pitzer's equations for calculation of geochemical reactions in brines: U.S.
Geological Survey Water-Resources Investigations Report 88-4153, 310 p.
Plummer, L.N., and D.L. Parkhurst, 1990. Application of the Pitzer Equations to the PHREEQE
geochemical model, in Melchior, D.C., and Bassett, R.L., eds., Chemical modeling of
aqueous systems II: American Chemical Society Symposium Series 416, Washington,
D.C., American Chemical Society, p. 128-137.
Raven, K.G., K.S. Novakowski, R.M. Yager, and R.J. Heystee, 1992. Supernormal fluid
pressures in sedimentary rocks in southern Ontario – western New York State. Can.
Geotech J., 29: 80-93.
Rezaei, M., E. Sanz, E. Raeisi, C. Ayora, E. Vazquez-Sune, and J. Carrera, 2005. Reactive
transport modeling of calcite dissolution in the fresh-salt water mixing zone, J.
Hydrology, 311 (1-4): 282-298.
Sanford, B.V., F.J. Thompson, and G.H. McFall, 1985. Plate tectonics – A possible controlling
mechanism in the development of hydrocarbon traps in southwestern Ontario, Bulletin of
Canadian Petroleum Geology, 33:52-71.
Sidborn, M. 2007, Modelling long-term redox processes and oxygen scavenging in fractured
crystalline rocks, Doctoral thesis in Chemical Engineering, KTH, Stockholm, Sweden.
Siegel, D.I., 1991. Evidence for dilution of deep, confined ground water by vertical recharge of
isotopically heavy Pleistocene water, Geology, 19:433-436.
Spiessl, S.M., K.T.B. MacQuarrie, and K.U. Mayer, 2007, Identification of key parameters
controlling dissolved oxygen migration and attenuation in fractured crystalline rocks. J.
Contaminant Hydrology, in press, accepted manuscript, available on-line 15 September
Spycher, N.F., E.L. Sonnenthal, and J.A. Apps, 2003. Fluid flow and reactive transport around
potential nuclear waste emplacement tunnels at Yucca Mountain, Nevada, J. Contam.
Hydrol., 62:653-673.
Steefel, C. I., and A.C. Lasaga. 1994. A coupled model for transport of multiple chemical
species and kinetic precipitation/dissolution reactions with application to reactive flow in
single phase hydrothermal systems. Am. J. Sci. 294:529-592.
Stumm, W., and J.J. Morgan, 1996. Aquatic Chemistry, 3rd ed., Wiley-Interscience, Wiley, 1022
- 30 -
Truc, O., J.P.Ollivier, and L.O. Nilsson, 2000. Numerical simulation of multi-species transport
through saturated concrete during a migration test - MsDiff code, Cement and Concrete
Res., 30:1581-1592.
Voss, C.I., and W.R. Souza, 1987. Variable density flow and solute transport simulation of
regional aquifers containing a narrow freshwater-saltwater transition zone, Water
Resour. Res. 23:1851-1866.
Weaver, T.R., S.K. Frape, and J.A. Cherry, 1995. Recent cross-formational fluid flow and mixing
in the shallow Michigan Basin, GSA Bulletin, 107: 697-707.
Wilson, T.P., and D.T. Long, 1993. Geochemistry and isotope chemistry of Michigan Basin
brines: Devonian formations. Appl. Geochem., 8: 81-100.
Wilson, T.P., and D.T. Long, 1993. Geochemistry and isotope chemistry of Ca-Na-Cl brines in
Silurian strata, Michigan Basin, U.S.A. Appl. Geochem., 8: 507-524.
Xu, T.F., and K. Pruess, 2001. Modeling multiphase non-isothermal fluid flow and reactive
geochemical transport in variably saturated fractured rocks: 1. Methodology, Am. J. Sci.,
Xu, T.F., J.A. Apps, and K. Pruess, 2005. Mineral sequestration of carbon dioxide in a
sandstone-shale system. Chemical Geology, 217 (3-4): 295-318.
Zhang, G.X., Z. P. Zheng, and J. M. Wan, 2005. Modeling reactive geochemical transport of
concentrated aqueous solutions, Water Resour. Res., 41:W02018, doi: 10.1029/
Ziegler, K., and F. J. Longstaffe, 2000. Multiple episodes of clay alteration at the
Precambrian/Paleozoic unconformity, Appalachian Basin: Isotopic evidence for long-
distance and local fluid migrations, Clays and Clay Minerals, 48:474-493.
- 31 -
- 32 -
A.1 REVIEW OF REACTIVE TRANSPORT MODELLING STUDIES ................................... 34
A.2 REFERENCES ................................................................................................................ 38
- 33 -
- 34 -
Several recent studies are reviewed here to provide examples of how reactive transport
models have been applied in sedimentary rock systems, and to gain an appreciation of the
model capabilities that may be required. Because there are few examples that deal directly with
issues related to deep geological repositories (DGR), studies from other fields such as CO
sequestration have also been considered.
One of the few studies that have directly employed reactive transport modelling for
interpreting geochemical data from a potential DGR host rock is that of Pearson et al. (2002).
They applied the PHREEQC code to simulate 12 to 15 million years of 1D diffusive transport
coupled with geochemical reactions in the Opalinus Clay formation at the Mont Terri Rock
Laboratory in northern Switzerland. The objective of the modelling was to better understand the
transport and reaction processes that were responsible for dilution of sea water, which was
assumed to be initially present in the claystone pores, by fresh groundwaters that has
subsequently invaded adjacent formations. The reactions that were considered included
(Pearson et al., 2002): dissolution-precipitation of quartz, calcite, dolomite, pyrite and celestite,
plus ion exchange of Na
, K
, Ca
, Mg
and Sr
. The measured data and modelling results
were in reasonable agreement for the major cations, with the exception of Mg
, while there
was a systematic disagreement between the simulated results and measurements for carbonate
and pH. Although the reasons for these disagreements were not fully resolved, Pearson et al.
(2002) have noted that the model should be considered preliminary because of over-simplified
assumptions including invariant boundary conditions, no flow or solute transport through the
bottom of the sequence, and the use of a single value of diffusion coefficient for all species.
Gurban et al. (2003) have applied two different modelling approaches, hydrochemical
mixing and mass balance (M3 code) and reactive transport (HYTEC code), to two natural
analogue reactor zones, Bangombé and OK84 at Okélobondo. Although the two sites differ
with respect to the depth of the uranium reaction zone, the geological materials at both consist
of sandstones and pelites, which in some locations are rich in manganese. Both sites have
waters of relatively low ionic strength and issues related to activity corrections were not
discussed. The results of the modelling generally agree with the limited field data, in that
uranium is initially oxidized at Okélobondo, but then quickly attenuated down gradient by
inorganic reactions, while at Bangombé a redox buffer zone that is rich in organic matter
consumes dissolved oxygen and protects uraninite from dissolution. The authors conclude that
the two modelling approaches considered are complementary. The mixing and mass balance
model can be useful for an initial assessment of the field data and for determining the likely
biogeochemical reactions causing deviations from ideal mixing of end members, while the
reactive transport model gives a spatial and temporal description of the reactions and can be
used to test several hypotheses regarding uranium transport.
Reactive transport modelling has recently been employed in several studies related to
the injection (sequestration) of CO
into sedimentary rock sequences. Gaus et al. (2005)
investigated the major reactions occurring when CO
is injected into a saline aquifer which is
overlain by up to 250 m of shale cap rock. The reactive transport domain was limited to the
shale cap rock, and it was assumed that dissolved CO
would migrate vertically upward by
diffusion from a supercritical CO
gas bubble trapped at the sandstone-shale interface. Gaus et
al. (2005) assumed that the cap rock mineralogy was homogeneous, consisting primarily of
quartz (21.5% by mass), mica/illite (24.7% by mass), and kaolinite (18%), with an assumed
initial porosity of 5%. The modelling investigation was performed in stages, starting with batch
modelling (assuming equilibrium or kinetic or reactions) and advancing to 1D reactive transport
- 35 -
modelling. Rate constants were selected from the literature, while reactive mineral surface
areas were estimated from geometric equations for spheres and a “scaling factor” of 0.001 (ratio
of reactive surface area to geometric surface area). Gaus et al. (2005) used the same diffusion
coefficient for all species (based on the value for CO
) and argue that this approach does not
affect the accuracy of the results because the variation in diffusion coefficients among various
aqueous species is an order of magnitude smaller than the range they computed for the
effective diffusion coefficient of CO
. In addition, although Gaus et al. (2005) note that the
Pitzer approach for activity coefficients would be the optimal choice for this problem, this
approach was not applied because aluminum speciation was considered crucial and aluminum
speciation with the Pitzer approach was not currently possible. Development of aluminum
speciation within the context of the Pitzer interaction equations appears to be ongoing (e.g.
Moller et al., 2007).
For the equilibrium batch modelling of CO
injection, the results suggest large changes
in the clay mineral volumes (and porosity), with the dissolution of smectites and illite and
precipitation of chalcedony, kaolinite, K-feldspar and large amounts of calcite. When compared
to the results of the kinetic batch model at a time of 15,000 years, the equilibrium batch model
results are questionable, and the authors caution that the results of equilibrium batch models
may be erroneous for assessing the progress of reactions in the shale cap rock.
The PHREEQC code was employed by Gaus et al. (2005) to conduct 1D diffusion-
reaction simulations. The results show that if plagioclase is assumed to be present only as
albite, then CO
migrates at a rate that is comparable to non-reactive diffusion; however, if
plagioclase is assumed to be present as a 50:50 albite:anorthite, then CO
migration into the
cap rock is significantly retarded by reaction with anorthite. Reaction with anorthite limits the
migration of elevated CO
concentrations to within the lower metre of the cap rock after 3000
years, and the maximum porosity decrease is less than 3%. Not surprisingly, Gaus et al. (2005)
conclude that having access to detailed mineralogical data when assessing the impact of
geochemical reactions on the safety of CO
sequestration remains crucial. Although the
feedback of porosity changes on effective diffusion coefficients was not considered in the study,
the computed decrease in shale porosity can be expected to cause a reduction in effective
diffusion coefficients, which would serve to further limit CO
migration into the shale unit.
Xu et al. (2005) have applied the reactive fluid flow and geochemical transport code
TOUGHREACT to analyze mass transfer between sandstone and shale layers and CO
immobilization through carbonate precipitation. They considered a simplified horizontally
bedded sandstone-shale system over a time period of 100,000 years, with the sandstone
represented by one model grid block (i.e. perfect mixing) and 1D diffusion being the main
transport process in the shale bed. Specific attention was paid to critically evaluating and
updating the thermodynamic data for carbonates, chlorite, and kerogen. Even though the
salinity of the formation water they were simulating was about 1 mol NaCl/L (i.e. I ~ 1), Xu et al.
(2005) used the extended Debye–Hückel equation to compute activity coefficients (with the
exception of CO
Simulation results indicate that most CO
sequestration occurs in the sandstone, with
the major CO
trapping minerals being dawsonite and ankerite. Interesting reversals in
transport occur for certain species; for example, Fe
and Ca
initially diffuse from the shale
into the sandstone to supply reactants for precipitation of siderite and ankerite, but later (~
100,000 years) the diffusion of these species reverses direction. Xu et al. (2005) conclude that
the sequestration time depends on the rates of mineral dissolution and precipitation, which are
- 36 -
products of the kinetic rate constant and reactive surface area. Scaling kinetic rates for all
minerals by the same factor is equivalent to scaling the time coordinate.
Although the CO
sequestration simulations have been conducted in geological settings
that share some similarities with the sedimentary rock sequences being considered for waste
repositories, the modelling investigations are typically localized to either consider processes in
the shale cap rocks, or the immediate vicinity of the sandstone-shale interface. This is
appropriate considering the localized nature of CO
injection at significant depths. It does not
appear that any long-term changes in surface boundary conditions have been considered in
sequestration modelling investigations; however, Xu et al. (2005) suggest that regional
groundwater flow may be sufficiently slow so as to maintain chemical gradients within the
sandstone aquifers.
Reactive transport modelling at larger scales, and considering additional factors such as
density-dependent flow, has been conducted to investigate the mixing of saline and fresh water
in coastal carbonate aquifers or deep hydrothermal systems. Based on the findings from a 1D
reactive transport investigation, which showed that saturation indices cannot be used to predict
the extent of calcite dissolution when mixing controls the overall rate of dissolution, Rezaei et al.
(2005) simulate seawater-freshwater mixing in a 2D coastal carbonate (i.e. calcite) aquifer.
Their study closely parallels earlier work by Sanford and Konikow (1989) and considered fluid
density variability, reactive transport, and porosity enhancement by calcite dissolution. Results
from the RETRASO model indicate that the maximum dissolution of calcite occurs near the
saline side of the groundwater discharge zone because of a very active convection cell that
develops there, which enhances mixing and thus reaction rates (Figure A.1). A maximum
porosity increase rate of 24% every 10,000 years occurs in this area of the domain; however,
over the majority of the mixing zone the rate of porosity increase is between 1 and 5% per
10,000 years. The feedback between porosity and hydraulic conductivity, which was not
considered in the modelling, may tend to further focus the zone(s) of dissolution. In this
advection-dominated flow system the overall amount of calcite dissolution tends to increase with
larger values of dispersivity because larger values will enhance the mixing rate; however, the
changes in porosity enhancement as a result of variations in dispersivity were relatively minor
(Rezaei et al., 2005).
Figure A.1: Example of porosity evolution due to mixing of fresh water and sea water
in a coastal aquifer system (Rezaei et al., 2005). The solid contours indicate the porosity
increase in % during a simulation period of 10,000 years, while the dashed lines are
contours of % sea water.
- 37 -
Corbella et al. (2003) also used the RETRASO model to simulate a relatively high
temperature system (150
C) in which a dilute fluid and a brine were mixed. The main
conclusion of this work was that the mixing of hydrothermal solutions of different chemistry, and
saturated with respect to calcite, causes both precipitation and dissolution of this mineral in
separate zones of the domain.
Jones and Xiao (2005) have investigated dolomitization, anhydrite precipitation, and
porosity evolution in regional carbonate aquifers. They applied the isothermal Xt2 code
(Bethke, 2002) which solves the coupled governing equations of transport (fluid mass and
solute mass) and the geochemical system (mass balance, equilibrium reactions, and kinetic
reactions) in two dimensions. The modelling scenario considered a hypothetical two-
dimensional (500 m deep by 10 km wide) carbonate aquifer, initially containing sea water which
is displaced by brines of differing chemistry. The geochemical model included three solid
species (calcite, dolomite, and anhydrite) and 32 aqueous species. The Pitzer equations were
employed for calculating mineral equilibria in solutions more concentrated than sea water.
Because they noted a relatively rapid change in porosity, Jones and Xiao (2005) included a
feedback with the aquifer permeability.
Starting with a base case scenario, Jones and Xiao (2005) varied parameters
individually to investigate the sensitivity of the dolomite (precipitated), anhydrite (generally
precipitated), and porosity distributions for simulation times up to approximately 1 million years.
For the range of parameters investigated, they found that dolomitization was critically-
dependent on the reflux (i.e. intruding brine groundwater) velocity and the reactive surface area.
Temperature was also important because of its effect on the reaction rates. Dolomitization
resulted in a porosity increase of 8% (with an initial porosity of 35%) that is consistent with the
mole-for-mole replacement theory favored by Weyl (1960). The increase in porosity by
dolomitization was offset by the precipitation of anhydrite, which occluded porosity by 0–22%
ahead of the dolomitization reaction front (Jones and Xiao, 2005). The feedback relationship
between porosity and permeability was a relatively moderate control on the evolution of
dolomitization; however, it was noted that predicting crystal size from mineral reactions is
beyond the capabilities of current reactive transport models, which thus limits the ability to
predict permeability from porosity (Jones and Xiao, 2005).
In studies that have focused on porosity enhancement in carbonate aquifers, it appears
that it has not been common to consider interactions with adjacent aquitard units. Again,
because these units would be expected to have little influence on the processes occurring in the
more permeable carbonate aquifers, this may be a reasonable simplification. Such a
conceptual model may be applicable to the investigation of long-term changes in boundary
conditions in sedimentary sequences, and could constitute a first-order modelling approach if
aquitard units can be assumed to isolate individual aquifers. However, the interaction of aquifers
and aquitards will increase in importance if shale units have increased permeabilities caused,
for example, by fracturing.
This review of recent literature indicates that several studies have considered various
aspects of reactive transport modelling that are relevant to investigation of waste disposal in
sedimentary rock sequences of southern Ontario. The study of Pearson et al. (2002) has
considered the effects of diffusive transport and reactions driven by fresh water ingress into
aquifers bounding a potential host rock (Opalinus Clay), while modelling conducted within the
context of CO
sequestration has considered in detail the reactive transport processes occurring
in shale aquitard units and highlighted the importance of having detailed initial mineralogical
characterization. Several studies have also focused on the mixing of different water types in
carbonate aquifers with spatial and time scales that are of relevance to nuclear waste isolation
- 38 -
(e.g. 10,000 to 1 million years). In addition, the influence of variable fluid densities, heat
transport, and high ionic strength on aqueous geochemistry have been considered in some
studies. One common theme of previous studies is that of simulating porosity evolution in
aquifers and aquitards, and in some cases the feedback between porosity and permeability.
Species-dependent diffusion coefficients have not been commonly applied in reactive transport
modelling in sedimentary rock sequences; however, Li et al. (2007) note that large errors can be
introduced in saline solutions by using a fixed diffusion coefficient, although the effect of
ignoring electrochemical migration appears to be minor.
Bethke, C. M., 2002. The Xt2 model of transport in reacting geochemical systems:
Hydrogeology program: Urbana, Illinois, University of Illinois, 89 p.
Corbella M., C. Ayora, and E. Cardellach, 2003. Dissolution of deep carbonate rocks by fluid
mixing: a discussion based on reactive transport modeling. J. Geochemical Exploration,
78-9: 211-214.
Gaus I., M. Azaroual M, and I. Czernichowski-Lauriol, 2005. Reactive transport modelling of the
impact of CO
injection on the clayey cap rock at Sleipner (North Sea). Chemical
Geology, 217 (3-4): 319-337.
Gurban, I., M. Laaksoharju, B. Madé, and E. Ledoux, 2003. Uranium transport around the
reactor zone at Bangombe and Okelobondo (Oklo): examples of hydrogeological and
geochemical model integration and data evaluation. J. Contaminant Hydrology, 61(1-4):
Jones G.D., and Y.T. Xiao, 2005. Dolomitization, anhydrite cementation, and porosity evolution
in a reflux system: Insights from reactive transport models. AAPG Bulletin, 89 (5): 577-
Li, L., C.A. Peters, and M.A. Celia, 2007. Reply to “Comment on upscaling geochemical reaction
rates using pore-scale network modeling” by Peter C. Lichtner and Qinjun Kang.
Advances in Water Resources, 30: 691-695.
Moller, N., C. Christov, and J. Weare, 2007. Thermodynamic model for predicting interactions of
geothermal brines with hydrothermal aluminum silicate minerals. Proc.: 32
on Geothermal Reservoir Engineering, Stanford University, California, Jan. 22-24.
Pearson, F.J., D. Arcos, A. Bath, J.Y. Boisson, A.M. Fernández, H.-E. Gäbler, E. Gaucher, A.
Gautschi, L. Griffault, P. Hernán, and H.N. Waber, 2002, Geochemistry of Water in the
Opalinus Clay Formation at the Mont Terri Laboratory, Technical Report 2003-03,
Rezaei, M., E. Sanz, E. Raeisi, C. Ayora, E. Vazquez-Sune, and J. Carrera, 2005. Reactive
transport modeling of calcite dissolution in the fresh-salt water mixing zone, J.
Hydrology, 311 (1-4): 282-298.
- 39 -
Sanford, W.E., and L.F. Konikow, 1989. Simulation of calcite dissolution and porosity changes
in saltwater mixing zones in coastal aquifers. Wat. Resour. Res., 25: 655-667.
Weyl, P.K., 1960. Porosity through dolomitization: Conservation of mass requirements, J.
Sedimentary Petrology, 30: 85
Xu T.F., J.A. Apps, and K. Pruess, 2005. Mineral sequestration of carbon dioxide in a
sandstone-shale system. Chemical Geology, 217 (3-4): 295-318.
... Thus, modeling is a crucial tool to predict the response of systems under certain conditions [Hunter et al., 1998]. Corresponding models are an essential tool in investigating the coupled transport of fluids and reactive substances through porous media and the resulting chemical reactions in the pores [Steefel et al., 2005;MacQuarrie and Mayer, 2005;Xu et al., 2006]. ...
Geochemical processes in subsurface reservoirs affected by microbial activity change the material properties of porous media. This is a complex biogeochemical process in subsurface reservoirs that currently contains strong conceptual uncertainty. This means, several modeling approaches describing the biogeochemical process are plausible and modelers face the uncertainty of choosing the most appropriate one. Once observation data becomes available, a rigorous Bayesian model selection accompanied by a Bayesian model justifiability analysis could be employed to choose the most appropriate model, i.e. the one that describes the underlying physical processes best in the light of the available data. However, biogeochemical modeling is computationally very demanding because it conceptualizes different phases, biomass dynamics, geochemistry, precipitation and dissolution in porous media. Therefore, the Bayesian framework cannot be based directly on the full computational models as this would require too many expensive model evaluations. To circumvent this problem, we suggest performing both Bayesian model selection and justifiability analysis after constructing surrogates for the competing biogeochemical models. Here, we use the arbitrary polynomial chaos expansion. We account for the approximation error in the Bayesian analysis by introducing novel correction factors for the resulting model weights. Thereby, we extend the Bayesian justifiability analysis and assess model similarities for computationally expensive models. We demonstrate the method on a representative scenario for microbially induced calcite precipitation in a porous medium. Our extension of the justifiability analysis provides a suitable approach for the comparison of computationally demanding models and gives an insight on the necessary amount of data for a reliable model performance.
... Therefore, a very small amount of CO 2 can enter the caprock, which in turn limits the reaction rate of mineralisation in the reservoir, and may potentially alter the porosity and permeability due to induced degradation. On the other hand, for the permeable host rock, the advection of flow is more dominant (at the presence of pressure gradient), meaning larger amounts of CO 2 can pass through, and consequently the impact of long-term reaction and mineral trapping is significant [189,190]. Wang and Peng [191] developed a numerical model to simulate the CO 2-brine interaction in the fracture network, and evaluated the caprock sealing efficiency based on deformation, gas diffusion, advection and sorption of CO 2. It was revealed that the diffusion process results in initial swelling and later shrinking of the shale matrix through sorption of CO 2 and alters the porosity/permeability of the fracture network. However, in their model geochemical reaction kinetics were not implemented, and should be considered to further improve the accuracy of the simulations. ...
Carbon capture and storage (CCS) has been identified as an urgent, strategic and essential approach to reduce anthropogenic CO2 emissions, and mitigate the severe consequences of climate change. CO2 storage is the last step in the CCS chain and can be implemented mainly through oceanic and underground geological sequestration, and mineral carbonation. This review paper aims to provide state-of-the-art developments in CO2 storage. The review initially discussed the potential options for CO2 storage by highlighting the present status, current challenges and uncertainties associated with further deployment of established approaches (such as storage in saline aquifers and depleted oil and gas reservoirs) and feasibility demonstration of relatively newer storage concepts (such as hydrate storage and CO2-based enhanced geothermal systems). The second part of the review outlined the critical criteria that are necessary for storage site selection, including geological, geothermal, geohazards, hydrodynamic, basin maturity, and economic, societal and environmental factors. In the third section, the focus was on identification of CO2 behaviour within the reservoir during and after injection, namely injection-induced seismicity, potential leakage pathways, and long-term containment complexities associated with CO2-brine-rock interaction. In addition, a detailed review on storage capacity estimation methods based on different geological media and trapping mechanisms was provided. Finally, an overview of major CO2 storage projects, including their overall outcomes, were outlined. This review indicates that although CO2 storage is a technically proven strategy, the discussed challenges need to be addressed in order to accelerate the deployment of the technology. In addition, beside the necessity of techno-economic aspects, public acceptance of CO2 storage plays a central role in technology deployment, and the current ethical mechanisms need to be further improved.
... During the last three decades, the study of reactive transport in fractured rocks has attracted a large interest in the environmental and physical sciences, both in interdisciplinary theoretical and applied fields [Bryant et al., 2001]. In particular, research focused on transport in low-porosity rocks hosting nuclear waste repositories [Tang et al., 1981;Sudicky and Frind, 1982;Lever and Bradbury, 1985;Banwart et al., 1999;Smellie and Karls- son, 1999;Park et al., 2001;MacQuarrie and Mayer, 2005;H€ oltt€ a et al., 2008;Dideriksen et al., 2010;MacQuarrie et al., 2010;Hartley et al., 2015;Tsang et al., 2015]. Furthermore, redox front characteristics help to understand pollutant transport in porous [Cribbin et al., 2014] or fractured rocks [Akagawa et al., 2006]. ...
Full-text available
See for a read-only version on Wiley's site: .......................................................... A pavement outcrop with excellent exposure of spatial relationships among joints, veins, small offset normal faults and associated alteration halos (redox fronts) provided an opportunity to compare predictions of analytical models for reaction front propagation in a fracture–matrix system with a real field situation. The results have important implications for fluid flow and pollutant transport through a fractured medium. The alteration halos observed suggest that all joints of different sets and most small faults are conductive to meteoric water at shallow depth. On the other hand, veins are local barriers to mass transport by diffusion. By using petrologic and petrophysical data, analytical modeling, and the width of the alteration halos, it was possible to estimate when the fracture network was open to fluid flow. The inferred time for fluid flow and diffusion through the fracture network is sensitive to the porosity n of the rock matrix used in the analytical solutions: 2200 ± 500 years with n = 0.08, 4600 ± 900 years with n = 0.05, and 16000 ± 4000 with n = 0.02. The second and third age determinations are consistent with the landscape evolution of the area since the end of the last Wűrmian ice age and with the timing required to fill the fractures observed in outcrop. We suggest that analytical modeling is an important tool for the determination of transport and reaction time scales in fractured formations where it is constrained by a robust petrophysical and chemical properties dataset.
... [3] Rigidity may possibly be valid for fractured rocks system [e.g., Peters and Klavetter, 1988;Therrien and Sudicky, 1996;MacQuarrie and Mayer, 2005]; however, in most natural soils the pore system is dynamically changing due to swelling and shrinkage depending on the water contents and due to soil biota activities among other factors [e.g., Horn, 2004;Richards and Peth, 2009]. Such changes in pore volume and pore size distribution including deformations in pore geometry result in the occurrence of separate porosity compartments. ...
We present follow-up work to previous work extending the classical rigid (RGD) approach formerly proposed by Gerke and van Genuchten, to account for shrinking effects (SHR) in modeling water flow and solute transport in dual-permeability porous media. In this study we considered three SHR scenarios, assuming that aggregate shrinkage may change either: (i) the hydraulic properties of the two pore domains, (ii) their relative fractions, or (iii) both hydraulic properties and fractions of the two domains. The objective was to compare simulation results obtained under the RGD and the SHR assumptions to illustrate the impact of matrix volume changes on water storage, water fluxes, and solute concentrations during an infiltration process bringing an initially dry soil to saturation and a drainage process starting from an initially saturated soil. For an infiltration process, the simulated wetting front and the solute concentration propagation velocity, as well as the water fluxes and water and solute exchange rates, for the three SHR scenarios significantly deviated from the RGD. By contrast, relatively similar water content profiles evolved under all scenarios during drying. Overall, compared to the RGD approach, the effect of changing the hydraulic properties and the weight of the two domains according to the shrinkage behavior of the soil aggregates induced a much more rapid response in terms of water fluxes and solute travel times, as well as a larger and deeper water and solute transfer from the fractures to the matrix during wetting processes.
... Model parallelization is also required to migrate from a simplified two-dimensional cross-sectional model to a full three-dimensional approach, considering that the effect of dimensionality has been identified as an important issue for evaluating the impact of glaciations on sedimentary basins (Person et al. 2012). For example, the presence of structural arches where aquifers crop out at the surface over long lateral distances is believed to be critical to the infiltration of recharge (Mayer & MacQuarrie 2007). ...
Deep sedimentary basins are complex systems that over long time scales may be affected by numerous interacting processes including groundwater flow, heat and mass transport, water-rock interactions, and mechanical loads induced by ice sheets. Understanding the interactions among these processes is important for the evaluation of the hydrodynamic and geochemical stability of geological CO2 disposal sites, and is equally relevant to the safety evaluation of deep geologic repositories for nuclear waste. We present a reactive transport formulation coupled to thermo-hydrodynamic and simplified mechanical processes. The formulation determines solution density and ion activities for ionic strengths ranging from fresh water to dense brines based on solution composition, and simultaneously accounts for the hydro-mechanical effects caused by long-term surface loading during a glaciation cycle. The formulation was implemented into the existing MIN3P reactive transport code (MIN3P-THCm), and was used to illustrate the processes occurring in a two-dimensional cross section of a sedimentary basin subjected to a simplified glaciation scenario consisting of a single cycle of ice sheet advance and retreat over a time period of 32,500 years. Although the sedimentary basin simulation is illustrative in nature, it captures the key geological features of deep Paleozoic sedimentary basins in North America, including interbedded sandstones, shales, evaporites, and carbonates in the presence of dense brines. Simulated fluid pressures are shown to increase in low hydraulic conductivity units during ice sheet advance due to hydro-mechanical coupling. During the period of deglaciation, Darcy velocities increase in the shallow aquifers and to a lesser extent in deeper high-hydraulic conductivity units (e.g., sandstones) as a result of the infiltration of glacial meltwater below the warm-based ice sheet. Dedolomitization is predicted to be the most widespread geochemical process, focused near the freshwater/brine interface. For the illustrative sedimentary basin, the results suggest a high degree of hydrodynamic and geochemical stability.This article is protected by copyright. All rights reserved.
... These wormholes facilitate the flow of hydrocarbons from the reservoir to the wellbore, which can help enhance production (Schechter 1992;Williams et al. 1979;Economides et al. 1993). Numerous experimental and theoretical studies have demonstrated that the shape and structure of these wormholes depend on the combined effect of acid-injection rate, heterogeneities in permeability/porosity of rock, surface reaction rate, and dispersion caused by molecular diffusion and pore-scale phenomena (Daccord 1987;Fredd and Fogler 1998;Gong and El-Rabaa 1999;Bazin 2001;Mayer et al. 2002;Xu et al. 2003;Golfier et al. 2002;Kang et al. 2006;Panga et al. 2005;Steefel et al. 2005;MacQuarrie and Mayer 2005;Kalia and Balakotaiah 2007;Tardy et al. 2007;Lungwitz et al. 2007;Farshbaf Zinati et al. 2007;Chang et al. 2008;Cohen et al. 2008;Szymczak and Ladd 2009;Meakin and Tartakovsky 2009;Noiriel et al. 2009;McDuff et al. 2010;Detwiler 2010;De Oliveira et al. 2012;Furui et al. 2012;Ratnakar et al. 2013;Maheshwari et al. 2013;Maheshwari and Balakotaiah 2013a;Elkhoury et al. 2013;Gharbi et al. 2013;Hao et al. 2013;Ovaysi and Piri 2014;Menke et al. 2015;Maheshwari et al. 2015). For instance, at a very low injection rate (such that the characteristic time scale for reaction is very low compared with acid transport), acid continues dissolving the entire face of the rock, causing facial dissolution. ...
Full-text available
Polymer-based (gelled or in-situ gelled) and emulsified acids have been used for matrix acidization of carbonate reservoirs for several years. Gelled and emulsified acids are typically used for acidization of high-temperature carbonate reservoirs because of their lower reaction rate as compared with nongelled/emulsified acids, resulting in deeper penetration of acid, whereas in-situ gelled acid is used for acid diversion. Literature review indicates that several laboratory-scale experimental studies have been performed to analyze the effect of acid gelation and emulsion on carbonate acidization as compared with nongelled/emulsified acids. However, there are very few modeling or quantitative theoretical studies regarding carbonate acidization with gelled and emulsified acids that can be tested at laboratory or field scale. More specifically, a theoretical analysis of the effect of transport and rheological properties (i.e., shear-thinning behavior) of gelled and emulsified acids on the acidization process is not available in the literature. Therefore, the primary objective of this study is to analyze the effect of transport and rheological properties of gelled and emulsified acids on carbonate acidization in three dimensions, which can help in terms of design of gelled- and emulsified-acid properties to achieve lower leakoff rate and deeper penetration of wormholes. The authors present 3D numerical simulations of carbonate acidization with hydrochloric acid (HCl), gelled acid, and emulsified acid by use of a two-scale-continuum model. By use of this model, the effect of transport and rheological properties of these non-Newtonian acids on the acidization curve and dissolution pattern is analyzed and compared with the available laboratory-scale experimental data. It has been observed from the numerical simulations that a lower amount of acid is necessary to breakthrough, and thinner wormholes are formed for both gelled and emulsified acids compared with neat HCl. Additionally, acidization remains in the optimum dissolution regime for a large variation in terms of acid-injection rate for both gelled and emulsified acids compared with neat HCl. Finally, the authors develop a wormholing criterion for acids, the rheological behavior of which can be described by the power law. This criterion can be used to estimate the optimum injection rate for vuggy and nonvuggy carbonates.
... In 2007, a state-of-science review of reactive transport modelling in sedimentary rocks was completed (Mayer and MacQuarrie, 2007). The review revealed that reactive transport modelling had previously not been used to assess glaciation processes in sedimentary rocks. ...
Many countries worldwide are investigating the use of advanced fuels and fuel cycles for purposes such as increasing the sustainability of the nuclear fuel cycle, or decreasing the radiological impact of used fuel. One common metric used to assess the radiological impact to humans of fuels placed in a repository is the total radiotoxicity of the fuel, but this approach does not take into account how engineered and natural (i.e., rock) barriers can remove many radiotoxic nuclides from ground water before they reach the surface. In this study, we evaluate the potential radiological dose consequences of advanced fuels in the context of a full system model simulation for release and transport from a repository, transport through the surrounding geosphere, release to the biosphere and dose consequences for the target critical group. Heavy water moderated reactors, such as the CANDU® reactor, are well-suited to the use of advanced fuels, and the post-closure performance of a deep geological repository for spent natural uranium fuel from them has already been studied. For this study, two advanced fuels of current interest were chosen: a TRUMOX fuel designed to recycle plutonium and minor actinides and thereby reduce the amount of these materials going into disposal, and a plutonium thorium-based fuel whose main goal is to increase sustainability by reducing uranium consumption. The impact of filling a deep geological repository, of identical design to that for natural uranium, with used fuel from these fuel cycles was analyzed. It was found that the two advanced fuels analyzed had dose rates, to a hypothetical critical group of humans living above the repository, which remained a factor of 170 to 340 lower than the current acceptance limit for releases, while being 5.3 (for TRUMOX) and 2.6 (for thorium-plutonium) times higher than those of natural uranium. When the dose rates are normalized to total energy produced, the repository emissions are comparable. In this case, the maximum dose rates were found to be 6% lower for the TRUMOX fuel, and 16% higher for the plutonium thorium fuel, than for the used natural uranium fuel.
The governing equations and mathematical models describing CO2 spreading and trapping in saline aquifers and the related hydro-mechanical and chemical processes were described in Chapt. 3. In this chapter, the focus is on methods for solving the relevant equations. The chapter gives an overview of the different approaches, from high-fidelity full-physics numerical models to more simplified analytical and semi-analytical solutions . Specific issues such as modeling coupled thermo-hydro-mechanical-chemical processes and modeling of small-scale processes , such as convective mixing and viscous fingering , are also addressed. Finally, illustrative examples of modeling real systems, with different types of modeling approaches, are presented.
The objective of the presented study is to assess the evolving mine water quality of uranium mines abandoned between 1958 and 1992 in the Czech Re-public. The sampling proved that actual uranium concentrations in mine waters did not in most cases exceed 0.45 mg/L. Uranium concentrations in the discharges from the adits abandoned more than 40 years ago were below MCL of US EPA for drinking water. Special attention is given to the most recently abandoned Mine Olší. Upon the geochemical modelling results using Geochemist’s Workbench, three time phases of mine water evolution can be defined at this deposit.
Full-text available
The recent profusion of microscopic characterization methods applicable to Earth Science materials, many of which are described in this volume (Anovitz and Cole 2015, this volume; Noiriel 2015, this volume), suggests that we now have an unprecedented new ability to consider geochemical processes at the pore scale. These new capabilities offer the potential for a paradigm shift in the Earth Sciences that will allow us to understand and ultimately quantify such enigmas as the apparent discrepancy between laboratory and field rates (White and Brantley 2003) and the impact of geochemical reactions on the transport properties of subsurface materials (Steefel and Lasaga 1990, 1994; Steefel and Lichtner 1994; Xie et al. 2015). It has only gradually become apparent that many geochemical investigations of Earth materials have suffered (perhaps inadvertently) from the assumption of bulk or continuum behavior, leading to volume averaging of properties and processes that really need to be considered at the individual grain or pore scale. For example, a relationship between reaction-induced porosity and permeability change can perhaps be developed based on bulk samples, but ultimately a mechanistic understanding and robust predictive capability of the associated geochemical and physical processes will require a pore-scale view. The question still arises: Do we need pore-scale characterization and models in geochemistry and mineralogy? The laboratory–field rate discrepancy (or enigma) is a good example of where a pore-scale understanding may provide insights not easily achievable with bulk characterization and models. If the reasons for this apparent discrepancy between laboratory and field rates cannot be explained, then it appears unlikely that scientifically defensible and quantitative models for a number of important Earth Science applications ranging from chemical weathering and its effects on atmospheric CO2, to subsurface carbon sequestration, to nuclear waste storage, to contaminant remediation and transport, …
Full-text available
It is generally assumed that meltwater from the base of ice sheets is discharged in a relatively thin subglacial zone. Whereas this may be true for ice sheets resting on impermeable beds, many ice sheets, such as the glacial period ice sheets of North America and Europe, flowed over extensive aquifers. A theory is developed which suggests that high rates of meltwater discharge into these aquifers would have completely reorganised their flow fields, producing integrated patterns of glacially pressurised flow controlled by the continental-scale form of the ice sheet surface rather than the small-scale topographic basins which determine modern aquifer extent. The theory is applied to the aquifers which underlay the Saalian ice sheet in The Netherlands, where it is shown that potential gradients and groundwater flow velocities would have developed which were two orders of magnitude greater than modern values and that the dominant flow vectors would have been normal to those of the modern flow. Thus, glacial/interglacial cycling in areas which have suffered periodic glaciation during the late Cenozoic may have experienced alternating phases of greater and lesser flow energy. It represents another example of climatically-di'iven cyclical change in the earth. Under highly energised glacial conditions, potential gradients mucl~ larger than modern values may have produced many common features of sediment disruption, such as diapirs, liquifaction structures and pipes, and forms such as glacial 'doughnuts' and pock marls, which have hitherto been explained by other processes. Deep and penetrative flushing of aquifers by glacial meltwater may have left a distinctive geochemical signal in them which may be used to test the theory. Gases such as methane, generated at shallow depth, may have been trapped beneath the glacial 'cap-rock', and may also have played an important role in these processes.
Full-text available
Pressure and tracer measurements in boreholes drilled to the bottom of Ice Stream B, West Antarctica, are used to obtain information about the basal water conduit system in which high water pressures are developed.These high pressures presumably make possible the rapid movement of the ice stream. Pressure in the system is indicated by the borehole water level once connection to the conduit system is made. On initial connection, here also called “breakthrough” to the basal water system, the water level drops in a few minutes to an initial depth in the range 96–117 m below the surface. These water levels are near but mostly somewhat deeper than the floation level of about 100 m depth (water level at which basal water pressure and ice overburden pressure are equal), which is calculated from depth-density profiles and is measured in one borehole. The conduit system can be modelled as a continuous or somewhat discontinuous gap between ice and bed; the thickness of the gap δ has to be about 2 mm to account for the water-level drop on breakthrough, and about 4 mm to fit the results of a salt-tracer experiment indicating downstream transport at a speed of 7.5 mm s−1. The above gap-conduit model is, however, ruled out by the way a pressure pulse injected into the basal water system at breakthrough propagates outward from the injection hole, and also by the large hole-to-hole variation in measured basal pressure, which if present in a gap-conduit system with δ = 2 or 4 mm would result in unacceptably large local water fluxes. An alternative model that avoids these objections, called the “gap opening” model, involves opening a gap as injection proceeds: starting with a thin film, the injection of water under pressure lifts the ice mass around the borehole, creating a gap 3 or 4mm wide at the ice/bed interface. Evaluated quantitatively, the gap-opening model accounts for the volume of water that the basal water system accepts on breakthrough, which obviates the gap-conduit model. In order to transport basal meltwater from upstream it is then necessary for the complete hydraulic model to contain also a network of relatively large conduits, of which the most promising type is the “canal” conduit proposed theoretically by Walder and Fowler (1994): flat, low conduits incised into the till, ∼0.1 m deep and perhaps ∼1 m wide, with a flat ice roof. The basal water-pressure data suggest that the canals are spaced ∼50–300 m apart, much closer than R-tunnels would be. The deepest observed water level, 117 m, is the most likely to reflect the actual water pressure in the canals, corresponding to a basal effective pressure of 1.6 bar. In this interpretation, the shallower water levels are affected by loss of hydraulic head in the narrow passageway (s) that connect along the bed from borehole to canal(s). Once a borehole has frozen up and any passageways connecting with canals have become closed, a pressure sensor in contact with the unfrozen till that underlies the ice will measure the pore pressure in the till, given enough time for pressure equilibration. This pressure varies considerably with time, over the equivalent water-level range from 100 to 113 m. Basal pressure sensors 500 m apart report uncorrelated variations, whereas sensors in boreholes 25 m араrt report mostly (but not entirely) well-correlated variations, of unknown origin. In part of the record, remarkable anticorrelated variations are interspersed with positively correlated ones, and there are rare, abrupt excursions to extreme water levels as deep as 125 m and as shallow as 74 m. A diurnal pressure fluctuation, intermittently observed, may possibly be caused by the ocean tide in the Ross Sea. The lack of any observed variation in ice-stream motion, when large percentagewise variations in basal effective pressure were occurring according to our data, suggests that the observed pressure variations are sufficiently local, and so randomly variable from place to place, that they are averaged out in the process by which the basal motion of the ice stream is determined by an integration over a large area of the bed.
Chlorite and illite are commonly associated with ubiquitous secondary K-rich feldspar in the rocks located immediately above and below the Precambrian-Paleozoic unconformity in southwestern Ontario, and elsewhere in the mid-continent of North America. This alteration assemblage is attributed to long-distance migration of hot brines driven westward by orogenic processes originating along the eastern seaboard of North America. The δD and δ18O values of chlorite and illite, plus K-Ar dates for secondary K-rich feldspar and illite, were used to determine the nature, origin, and timing of the fluids that altered Precambrian granites and their overlying rocks in southwestern Ontario. The δ18O values of the chlorite-forming fluids are best explained by initial hot brines (≥ 150°C) evolved mostly from seawater. Secondary K-rich feldspar formation followed shortly thereafter, as the fluids cooled and perhaps mixed with meteoric water. Regional migration of the brines was induced by Taconic orogenic events to the east. The hydrogen and oxygen isotopic compositions for the secondary illite of the early to mid-Carboniferous indicate its crystallization from local meteoric water at low temperatures (40-55°C). Infiltration of local meteoric water into the Paleozoic and uppermost altered Precambrian rocks occurred during uplift, erosion, and subaerial exposure of local arches in southern Ontario. The local basement reactivation and associated secondary illite formation in this portion of the North American hinterlands was likely a distal expression of east-coast Acadian and Alleghanian orogenic activity.
Recent geological studies of SW Ontario and adjacent E segments of the St Lawrence Platform strongly suggest that the succession of Paleozoic sedimentary and tectonic events recorded therein may have been triggered and controlled by plate motions and associated orogenic events centred at or beyond the margins of the craton. Reconstructions of Paleozoic depositional and tectonic processes along the E rim of the Michigan Basin in Ontario, and immediately adjacent regions of the Canadian Shield, supported by studies of satellite imagery, indicate that basement uplift during and subsequent to that time was transmitted through the crust by vertical rotation (tilting) of fault-bounded megablocks. The fault-block readjustment that took place on and marginal to the Algonquin-Findlay arch trend in SW Ontario provided the structural control for the origin and development of Cambrian, Ordovician, Silurian and Devonian oil and gas traps.-from Authors