Continuum-scale convective mixing in geological CO2 sequestration in anisotropic and heterogeneous saline aquifers

Deep saline aquifers are important geological formations for CO₂ sequestration. It has been known that dissolution of CO₂ increases brine density, which results in downward density-driven convection and consequently greatly enhances CO₂ sequestration. In this study, a continuum-scale lattice Boltzmann model is used to investigate convective mixing of CO₂ in saline aquifers. It is found that increasing permeability in either the vertical or horizontal direction accelerates the development of convective mixing. In a heterogeneous aquifer, increasing heterogeneity hampers the onset of convective mixing, because the heterogeneous permeability field results in a large portion of low-velocity region which reduces the instability of the system. The critical time for the onset of instability depends mainly on the coefficient of variation (COV) of the permeability field, and is insensitive to the correlation length. This implies that within the scale of critical time, mass transport is dominated by diffusion, and thus depends mainly on fine-scale heterogeneity controlled by COV. We derived an empirical formula for estimating the critical time, which leads to good estimates for all combinations of COV and correlation length. Fingering, channeling, and dispersion are the three mechanisms for mass transport. In dispersion, dissolved mass is approximately proportional to the square root of time, while in fingering and channeling it is approximately proportional to time. Mass transport by channeling depends significantly on permeability structure, while by fingering it is controlled by gravitational instability. It is also found that larger volumes of CO₂ can be stored in heterogeneous aquifers because of higher mass dissolution rates.     

Multiphysics modeling of CO2 sequestration in a faulted saline formation in Italy

The present work describes the results of a modeling study addressing the geological sequestration of carbon dioxide (CO₂) in an offshore multi-compartment reservoir located in Italy. The study is part of a large scale project aimed at implementing carbon capture and storage (CCS) technology in a power plant in Italy within the framework of the European Energy Programme for Recovery (EEPR). The processes modeled include multiphase flow and geomechanical effects occurring in the storage formation and the sealing layers, along with near wellbore effects, fault/thrust reactivation and land surface stability, for a CO₂ injection rate of 1 × 10⁶ ton/a. Based on an accurate reproduction of the three-dimensional geological setting of the selected structure, two scenarios are discussed depending on a different distribution of the petrophysical properties of the formation used for injection, namely porosity and permeability. The numerical results help clarify the importance of: (i) facies models at the reservoir scale, properly conditioned on wellbore logs, in assessing the CO₂ storage capacity; (ii) coupled wellbore-reservoir flow in allocating injection fluxes among permeable levels; and (iii) geomechanical processes, especially shear failure, in constraining the sustainable pressure buildup of a faulted reservoir.     

Elucidating geochemical response of shallow heterogeneous aquifers to CO2 leakage using high-performance computing: Implications for monitoring of CO2 sequestration

Predicting and quantifying impacts of potential carbon dioxide (CO₂) leakage into shallow aquifers that overlie geologic CO₂ storage formations is an important part of developing reliable carbon storage techniques. Leakage of CO₂ through fractures, faults or faulty wellbores can reduce groundwater pH, inducing geochemical reactions that release solutes into the groundwater and pose a risk of degrading groundwater quality. In order to help quantify this risk, predictions of metal concentrations are needed during geologic storage of CO₂. Here, we present regional-scale reactive transport simulations, at relatively fine-scale, of CO₂ leakage into shallow aquifers run on the PFLOTRAN platform using high-performance computing. Multiple realizations of heterogeneous permeability distributions were generated using standard geostatistical methods. Increased statistical anisotropy of the permeability field resulted in more lateral and vertical spreading of the plume of impacted water, leading to increased Pb²⁺ (lead) concentrations and lower pH at a well down gradient of the CO₂ leak. Pb²⁺ concentrations were higher in simulations where calcite was the source of Pb²⁺ compared to galena. The low solubility of galena effectively buffered the Pb²⁺ concentrations as galena reached saturation under reducing conditions along the flow path. In all cases, Pb²⁺ concentrations remained below the maximum contaminant level set by the EPA. Results from this study, compared to natural variability observed in aquifers, suggest that bicarbonate (HCO₃⁻) concentrations may be a better geochemical indicator of a CO₂ leak under the conditions simulated here.     

Effect of salt types on salt precipitation and water transport in saline sandy soil

Salt precipitation on the surface of porous media significantly affects water transport processes. Most studies on salt precipitation mainly focused on single salts, but in nature, salt precipitation usually occurs as mixtures. Consequently, information on the crystallization of salt mixtures and its effect on water transport remains scarce. This study investigated the precipitation of mixtures (the mass ratios of NaCl:Na₂SO₄ were 3:7, 5:5, and 7:3, respectively) of NaCl (typical efflorescence) and Na₂SO₄ (typical subflorescence) in the initially saturated sandy soil columns and its effect on evaporation and compared it with the cases of the two salts individually. The results showed that salt mixtures exhibited a mixed pattern of crystals including both efflorescence and subflorescence, and the efflorescence showed granular aggregation, unlike the mono-salts. The crystallization coverage of the salt mixtures was smaller than that of NaCl mono-salt; high (7:3) and low (5:5 and 3:7) proportions of NaCl led to larger and smaller crystallization coverage than that of Na₂SO₄ mono-salt, respectively. While the salt mixtures had less crystallization coverage than the mono-salts, they showed lower evaporation because the salt mixtures formed a denser crystallization structure of efflorescence-subflorescence-soil layer, this crystallization structure exhibited greater inhibition of water vapour diffusion, thus reducing evaporation. In addition, the crystallization of the salt mixtures with higher NaCl proportion afforded greater resistance of evaporation. The mixed crystallization pattern formed by the salt mixtures significantly enhances the crystallization resistance to evaporation.     

Tracking CO2 plume migration during geologic sequestration using a probabilistic history matching approach

Tracking the migration of the CO₂ plume is essential in order to better manage the operation of geologic sequestration of CO₂. However, the large cost of most monitoring technologies, such as time-lapse seismic, limits their application. We investigated the application of a probabilistic history matching methodology using routine measurements of injection data, which are affected by the presence of large-scale heterogeneities, as an inexpensive alternative to track the migration of CO₂ plume in an aquifer. The approach is demonstrated first through a synthetic example in which the ability to detect the presence of flow barriers was tested. In a second example, we applied our method to the In Salah field, one of the largest geological sequestration projects in the world, where the main direction of high permeability features was inferred. The accuracy and reproducibility of the results show that our approach for assisted history matching is an economic and viable option for plume monitoring during geologic CO₂ sequestration.     

Surface alteration of basalt due to cation-migration

Basaltic lava from Kilauea, Hawaii may have a red-brown surface, indicative of Fe-(hydr)oxides. This surface is not found where exposed to weathering, but at the interface between lava lobes, or in the interior of lava channels. We use several analytical techniques to determine how these Fe-(hydr)oxide surfaces may have developed. WDS-elemental distribution line profiles from the lava surface towards the lava′s interior detect an Fe-rich film of less than 5 μm thickness. Heat treatment of quenched, fresh lava samples of the same chemical composition between 600–1,090°C helps to replicate temperatures under which such an Fe-rich film might have formed. These experiments suggest that Fe-enrichment occurs above 1,020°C, whereas at lower temperatures Ca is enriched relative to Fe. One sample was treated below the glass transition temperature, at 600°C for 164 h. A depth profile with secondary neutral mass spectrometry shows an enrichment of Mg at the outer 50 nm of the glass surface. The formation of films requires cation migration, which is driven by an oxygen chemical potential between air and the reduced basalt (Fe²⁺/Fe³⁺ ratio of 13.3). The change of surface alteration from Mg to Ca film at lower temperatures, to predominantly Fe at high temperatures, is determined by a change of cation availability, largely controlled by crystallization that already occurs below 850°C, and volume crystallization that occurs above 925°C.     

The response of CO2 flux to rain pulses at a saline desert

As the substantial component of the ecosystem respiration, soil CO₂ flux is strongly influenced by infrequent and unpredictable precipitation in arid region. In the current study, we investigated the response of soil CO₂ flux to rain pulses at a saline desert in western China. Soil CO₂ flux was measured continuously during the whole growing season of 2009 at six sites. We found that there were remarkable changes in amplitude or diurnal patterns of soil CO₂ flux induced by rainfall events: from bimodal before rain to a single peak after that. Further analysis indicated that there is a significant linear relationship (P < 0.001) between soil CO₂ flux and soil temperature (T_soil). However, a hysteresis between the waveform of diurnal course of CO₂ flux and T_soil was observed: with soil CO₂ flux always peaked earlier than T_soil. Furthermore, a double exponential decay function was fitted to the soil CO₂ flux after rainfall, and total carbon (C) releases were estimated by numerical integration for rainfall events. The relative enhancement and total C release, in association with the rain pulses, was linearly related to the amount of precipitation. According to the size and frequency of rainfall events, the total amount of C release induced by rain pulses was computed as much as 7.88 g C·m^–2 in 2009, equivalent to 10.25% of gross primary production. These results indicated that rain pulses played a significant role in the carbon budget of this saline desert ecosystem, and the size of them was a good indicator of rain‐induced flux enhancement. Copyright © 2012 John Wiley & Sons, Ltd.     

High-resolution simulation and characterization of density-driven flow in CO2 storage in saline aquifers

Simulations are routinely used to study the process of carbon dioxide (CO₂) sequestration in saline aquifers. In this paper, we describe the modeling and simulation of the dissolution–diffusion–convection process based on a total velocity splitting formulation for a variable-density incompressible single-phase model. A second-order accurate sequential algorithm, implemented within a block-structured adaptive mesh refinement (AMR) framework, is used to perform high-resolution studies of the process. We study both the short-term and long-term behaviors of the process. It is found that the onset time of convection follows closely the prediction of linear stability analysis. In addition, the CO₂ flux at the top boundary, which gives the rate at which CO₂ gas dissolves into a negatively buoyant aqueous phase, will reach a stabilized state at the space and time scales we are interested in. This flux is found to be proportional to permeability, and independent of porosity and effective diffusivity, indicative of a convection-dominated flow. A 3D simulation further shows that the added degrees of freedom shorten the onset time and increase the magnitude of the stabilized CO₂ flux by about 25%. Finally, our results are found to be comparable to results obtained from TOUGH2-MP.     

降水过程对断层CO2气体异常排放响应的个例分析

为了从观测事实的角度揭示岩石圈的热异常对大气圈的影响,本文利用怀来后郝窑测点的断层气CO2排放观测数据和中国752站降水的逐日观测资料,分析了断层气CO2异常排放与降水事件的时空演变联系.个例分析结果表明,断层气CO2的异常排放会导致局地降水的增多.以1991年为例,伴随着断层气CO2排放的异常增加,在大范围降水负异常的背景下,CO2排放点周边区域出现显著的降水正异常（降水距平百分率大于零）区.同时,CO2异常幅度较大的时段,相应的降水正异常区的中心值也较大.此外,分析降水响应断层气CO2排放异常的时间还发现,断层气CO2异常排放对其后10天内的降水过程影响最为显著.     

Influence of small-scale heterogeneity on upward CO2 plume migration in storage aquifers

Recent advancements in experimental techniques allow sub-core-scale heterogeneities to be quantified in a high resolution. Based on the observations of heterogeneity distributions in natural core samples, we perform simulations to study the influence of small-scale heterogeneities on large-scale CO₂ migration during geological storage. We observe that even the heterogeneities at millimeter scale (the scale of a Representative Elementary Volume for sandstones) can affect large-scale buoyancy-driven upward CO₂ migration. For the representative examples we study, ignoring small-scale heterogeneities can lead to an overestimation of the migration speed by a factor of two.To analyze the cause of such overestimation, we introduce a dimensionless heterogeneity factor to characterize different levels of heterogeneity. The influence on CO₂ migration is quantified with respect to a variety of heterogeneity factors, correlation lengths, and fluid viscosity ratios for isotropic and anisotropic media. Our findings suggest that small-scale heterogeneities should not be ignored in core analysis, even for cores that appear relatively homogeneous and do not have distinguishable heterogeneous patterns. In addition, relative-permeability curves measured from core-flood experiments under high-flow-velocity conditions (a common practice to eliminate capillary end-effects) should not be directly used in modeling low-velocity CO₂ migration if small-scale heterogeneities are present.     

Cooling of the mesosphere and lower thermosphere due to doubling of CO2

A new parameterization of infrared radiative transfer in the 15-m CO₂ band has been incorporated into the Spectral mesosphere/lower thermosphere model (SMLTM). The parameterization is applicable to calculations of heating rates above approximately 15 km for arbitrary vertical profiles of the CO₂ concentration corresponding to the surface mixing ratio in the range 150–720 ppm. The sensitivity of the mesosphere and lower thermosphere (MLT) to doubling of CO₂ has been studied. The thermal response in the MLT is mostly negative (cooling) and much stronger than in the lower atmosphere. An average cooling at the stratopause is about 14 K. It gradually decreases to approximately 8 K in the upper mesosphere and again increases to about 40–50 K in the thermosphere. The cooling and associated thermal shrinking result in a substantial density reduction in the MLT that reaches 40–45% in the thermosphere. Various radiative, chemical, and dynamical feedbacks potentially important for the thermal response in the MLT are discussed. It is noted that the results of simulations are strikingly similar to observations of long-term trends in the MLT. This suggests that during the last 3–4 decades the thermal structure in the real upper atmosphere has undergone substantial changes driven by forcing comparable with that due to doubling of CO₂.     

Convective stability analysis of the long-term storage of carbon dioxide in deep saline aquifers

Deep saline aquifers are one of the most suitable geologic formations for carbon sequestration. The linear and global stability analysis of the time-dependent density-driven convection in deep saline aquifers is presented for long-term storage of carbon dioxide (CO₂). The convective mixing that can greatly accelerate the CO₂ dissolution into saline aquifers arises because the density of brine increases upon the dissolution of CO₂ and such a density difference may induce instability. The effects of anisotropic permeability on the stability criteria, such as the critical time for the appearance of convective phenomena and the critical wavelength of the most unstable perturbation, are investigated with linear and global stability analysis. The linear stability analysis provides a sufficient condition for instability while the global stability analysis yields a sufficient condition for stability. The results obtained from these two approaches are not exactly the same but show a consistent trend, both indicating that the anisotropic system becomes more unstable when either the vertical or horizontal permeability increases.     

A computational model for coupled multiphysics processes of CO2 sequestration in fractured porous media

In this paper, a computational model for the simulation of coupled hydromechanical and electrokinetic flow in fractured porous media is introduced. Particular emphasis is placed on modeling CO₂ flow in a deformed, fractured geological formation and the associated electrokinetic flow. The governing field equations are derived based on the averaging theory and the double porosity model. They are solved numerically with a mixed discretization scheme, formulated on the basis of the standard Galerkin finite element method, the extended finite element method, the level-set method and the Petrov–Galerkin method. The standard Galerkin method is utilized to discretize the equilibrium and the diffusive dominant field equations, and the extended finite element method, together with the level-set method and the Petrov–Galerkin method, are utilized to discretize the advective dominant field equations. The level-set method is employed to trace the CO₂ plume front, and the extended finite element method is employed to model the high gradient in the saturation field front. The proposed mixed discretization scheme leads to a convergent system, giving a stable and effectively mesh-independent model. The accuracy and computational efficiency of the proposed model is evaluated by verification and numerical examples. Effects of the fracture spacing on the CO₂ flow and the streaming potential are discussed.     

Migration of induced-infiltrated stream water into nearby aquifers due to seasonal ground water withdrawal

Analysis of stream-aquifer interaction due to ground water extraction has traditionally focused on the determination of the amount of water depleted in the stream. Less attention has been paid to the movement of infiltrated stream water inside aquifer, particularly for agricultural areas. This paper presents a method of using particle-tracking techniques to evaluate the transport of the leaked stream water in the nearby aquifers. Simple stream-aquifer conditions are used to demonstrate the usefulness of the analysis. Travel times, pathlines, and influence zones of stream water were determined between a stream and nearby pumping wells for seasonal ground water extraction areas. When water quantity is a concern, the analyses provide additional information about stream depletion; when water quality is an issue, they offer information for wellhead protection. Analyses were conducted for transient conditions, and both pumping and nonpumping periods were considered. According to the results from the simulation examples, migration of infiltrated stream water into the nearby aquifers is generally slow and most infiltrated stream water does not arrive at the pumping well at the end of a 90-day irrigation season. Infiltrated stream water may remain in the aquifer for several years before arriving at the pumping well. For aquifers with a regional hydraulic gradient toward streams, part of the infiltrated stream water may discharge back to streams during a recovery period.     

Effectiveness of subsurface pressure monitoring for brine leakage detection in an uncertain CO2 sequestration system

This work evaluates the detection sensitivity of deep subsurface pressure monitoring within an uncertain carbon dioxide sequestration system by linking the output of an analytical reduced-order model and first-order uncertainty analysis. A baseline (non-leaky) modeling run was compared against 10 different leakage scenarios, where the cap rock permeability was increased by factors of 2–100 (cap rock permeability from 10^?3 to 10^?1 millidarcy). The uncertainty variance outputs were used to develop percentile estimates and detection sensitivity for pressure throughout the deep subsurface as a function of space (lateral distance from the injection wells and vertical orientation within the reservoir) and time (years since injection), or P(x, z, t). Conditional probabilities were computed for combinations of x, z, and t, which were then used to generate power curves for detecting leakage scenarios. The results suggest that measurements of the absolute change in pressure within the target injection aquifer would not be able to distinguish small leakage rates (i.e., less than 50 × baseline) from baseline conditions, and that only large leakage rates (i.e., >100 × baseline) would be discriminated with sufficient statistical power (>99 %). Combining measurements, for example by taking the ratio of formation pressure in Aquifer 2/Aquifer 1, provides better statistical power for distinguishing smaller leakage rates at earlier times in the injection program. Detection sensitivity for pressure is a function of space and time. Therefore, design of an adequate monitoring network for subsurface pressure should account for this space–time variability to ensure that the monitoring system performs to the necessary design criteria, e.g., specific false-negative and false-positive rates.     

Vadose CO2 gas drives dissolution at water tables in eogenetic karst aquifers more than mixing dissolution

Most models of cave formation in limestone that remains near its depositional environment and has not been deeply buried (i.e. eogenetic limestone) invoke dissolution from mixing of waters that have different ionic strengths or have equilibrated with calcite at different _pCO₂ values. In eogenetic karst aquifers lacking saline water, mixing of vadose and phreatic waters is thought to form caves. We show here calcite dissolution in a cave in eogenetic limestone occurred due to increases in vadose CO₂ gas concentrations and subsequent dissolution of CO₂ into groundwater, not by mixing dissolution. We collected high‐resolution time series measurements (1 year) of specific conductivity (SpC), temperature, meteorological data, and synoptic water chemical composition from a water table cave in central Florida (Briar Cave). We found SpC, _pCO₂ and calcite undersaturation increased through late summer, when Briar Cave experienced little ventilation by outside air, and decreased through winter, when increased ventilation lowered cave CO_2(g) concentrations. We hypothesize dissolution occurred when water flowed from aquifer regions with low _pCO₂ into the cave, which had elevated _pCO₂. Elevated _pCO₂ would be promoted by fractures connecting the soil to the water table. Simple geochemical models demonstrate that changes in _pCO₂ of less than 1% along flow paths are an order of magnitude more efficient at dissolving limestone than mixing of vadose and phreatic water. We conclude that spatially or temporally variable vadose CO_2(g) concentrations are responsible for cave formation because mixing is too slow to generate observed cave sizes in the time available for formation. While this study emphasized dissolution, gas exchange between the atmosphere and karst aquifer vadose zones that is facilitated by conduits likely exerts important controls on other geochemical processes in limestone critical zones by transporting oxygen deep into vadose zones, creating redox boundaries that would not exist in the absence of caves. Copyright © 2014 John Wiley & Sons, Ltd.