Susan D. Hovorka

Gas-water-rock interactions in Frio Formation following CO2 injection: Implications for the storage of greenhouse gases in sedimentary basins

To investigate the potential for the geologic storage of CO2 in saline sedimentary aquifers, 1600 t of CO2 were injected at 1500 m depth into a 24-m-thick sandstone section of the Frio Formation, a regional brine and oil reservoir in the U.S. Gulf Coast. Fluid samples obtained from the injection and observation wells before CO2 injection showed a Na-CaCl‐type brine with 93,000 mg/L total dissolved solids (TDS) at near saturation with CH4 at reservoir conditions. Following CO2 breakthrough, samples showed sharp drops in pH (6.5‐5.7), pronounced increases in alkalinity (100‐3000 mg/L as HCO3) and Fe (30‐1100 mg/L), and significant shifts in the isotopic compositions of H2O, dissolved inorganic carbon (DIC), and CH4. Geochemical modeling indicates that brine pH would have dropped lower but for the buffering by dissolution of carbonate and iron oxyhydroxides. This rapid dissolution of carbonate and other minerals could ultimately create pathways in the rock seals or well cements for CO2 and brine leakage. Dissolution of minerals, especially iron oxyhydroxides, could mobilize toxic trace metals and, where residual oil or suitable organics are present, the injected CO2 could also mobilize toxic organic compounds. Environmental impacts could be major if large brine volumes with mobilized toxic metals and organics migrated into potable groundwater. The d 18 O values for brine and CO2 samples indicate that supercritical CO2 comprises ;50% of pore-fluid volume ;6 mo after the end of injection. Postinjection sampling, coupled with geochemical modeling, indicates that the brine gradually will return to its preinjection composition.

The U-tube: A novel system for acquiring borehole fluid samples from a deep geologic CO2 sequestration experiment

A novel system has been deployed to obtain geochemical samples of water and gas, at in situ pressure, during a geologic CO2 sequestration experiment conducted in the Frio brine aquifer in Liberty County, Texas. Project goals required high-frequency recovery of representative and uncontaminated aliquots of a rapidly changing two-phase (supercritical CO2-brine) fluid from 1.5 km depth. The data sets collected, using both the liquid and gas portions of the downhole samples, provide insights into the coupled hydro-geochemical issues affecting CO2 sequestration in brine-filled formations. While the basic premise underlying the U-Tube sampler is not new, the system is unique because careful consideration was given to the processing of the recovered two-phase fluids. In particular, strain gauges mounted beneath the high-pressure surface sample cylinders measured the ratio of recovered brine to supercritical CO2. A quadrupole mass spectrometer provided real-time gas analysis for perfluorocarbon and noble gas tracers that were injected along with the CO2. The U-Tube successfully acquired frequent samples, facilitating accurate delineation of the arrival of the CO2 plume, and on-site analysis revealed rapid changes in geochemical conditions.

Process‐based approach to CO2 leakage detection by vadose zone gas monitoring at geologic CO2 storage sites

A critical issue for geologic carbon sequestration is the ability to detect CO2 in the vadose zone. Here we present a new process-based approach to identify CO2 that has leaked from deep geologic storage reservoirs into the shallow subsurface. Whereas current CO2concentration-based methods require years of background measurements to quantify variability of natural vadose zone CO2, this new approach examines chemical relationships between vadose zone N2, O2, CO2, and CH4 to promptly distinguish a leakage signal from natural vadose zone CO2. The method uses sequential inspection of the following gas concentration relationships: 1) O2 versus CO2to distinguish in-situ vadose zone background processes (biologic respiration, methane oxidation, and CO2 dissolution) from exogenous deep leakage input, 2) CO2 versus N2 to further distinguish dissolution of CO2 from exogenous deep leakage input, and 3) CO2 versus N2/O2 to assess the degree of respiration, CH4 oxidation and atmospheric mixing/dilution occurring in the system. The approach was developed at a natural CO2-rich control site and successfully applied at an engineered site where deep gases migrated into the vadose zone. The ability to identify gas leakage into the vadose zone without the need for background measurements could decrease uncertainty in leakage detection and expedite implementation of future geologic CO2 storage projects.

The impact of geological heterogeneity on CO2 storage in brine formations: a case study from the Texas Gulf Coast

Abstract Geological complexities such as variable permeability and structure (folds and faults) exist to a greater or lesser extent in all subsurface environments. In order to identify safe and effective sites in which to inject CO2 for sequestration, it is necessary to predict the effect of these heterogeneities on the short- and long-term distribution of CO2. Sequestration capacity, the volume fraction of the subsurface available for CO2 storage, can be increased by geological heterogeneity. Numerical models demonstrate that in a homogeneous rock volume, CO2 flowpaths are dominated by buoyancy, bypassing much of the rock volume. Flow through a more heterogeneous rock volume disperses the flow paths, contacting a larger percentage of the rock volume, and thereby increasing sequestration capacity. Sequestration effectiveness, how much CO2 will be sequestered for how long in how much space, can also be enhanced by heterogeneity. A given volume of CO2 distributed over a larger rock volume may decrease leakage risk by shortening the continuous column of buoyant gas acting on a capillary seal and inhibiting seal failure. However, where structural heterogeneity predominates over stratigraphic heterogeneity, large columns of CO2 may accumulate below a sealing layer, increasing the risk of seal failure and leakage.

Inverse Modeling of Water-Rock-CO2 Batch Experiments: Potential Impacts on Groundwater Resources at Carbon Sequestration Sites

This study developed a multicomponent geochemical model to interpret responses of water chemistry to introduction of CO2 into six water-rock batches with sedimentary samples collected from representative potable aquifers in the Gulf Coast area. The model simulated CO2 dissolution in groundwater, aqueous complexation, mineral reactions (dissolution/precipitation), and surface complexation on clay mineral surfaces. An inverse method was used to estimate mineral surface area, the key parameter for describing kinetic mineral reactions. Modeling results suggested that reductions in groundwater pH were more significant in the carbonate-poor aquifers than in the carbonate-rich aquifers, resulting in potential groundwater acidification. Modeled concentrations of major ions showed overall increasing trends, depending on mineralogy of the sediments, especially carbonate content. The geochemical model confirmed that mobilization of trace metals was caused likely by mineral dissolution and surface complexation on clay mineral surfaces. Although dissolved inorganic carbon and pH may be used as indicative parameters in potable aquifers, selection of geochemical parameters for CO2 leakage detection is site-specific and a stepwise procedure may be followed. A combined study of the geochemical models with the laboratory batch experiments improves our understanding of the mechanisms that dominate responses of water chemistry to CO2 leakage and also provides a frame of reference for designing monitoring strategy in potable aquifers.

Lithostratigraphy and geochronology of fills in small playa basins on the Southern High Plains, United States

Playa basins are small depressions (typically ≤1.5 km 2 ) on the Southern High Plains of northwestern Texas and eastern New Mexico. There are about 25 000 playas in the region; they lie on the Blackwater Draw Formation (Pleistocene), a widespread eolian deposit, and locally on the Ogallala Formation (Miocene‐ Pliocene). Understanding the lithostratigraphy and chronostratigraphy of the fill in the basins is important because it should (1) provide clues to the origin and evolution of playas, which have been long debated; (2) yield a paleoenvironmental record for the region; and (3) aid in understanding the history and future of the regional aquifer because playas are the principal source of recharge. Data from 19 playa basins, combined with published data from 4 other basins, show that the basin fill is composed of six distinctive facies: (1) lacustrine mud; (2) lacustrine carbonate; (3) lacustrine delta deposits; (4) eolian sand and silt; (5) eolian loam; and (6) accretionary eolian deposits (Blackwater Draw Formation). Mud deposited under ponded conditions is the most common facies and is the surficial deposit on the floors of most playas, often producing Vertisols. The carbonate was precipitated under lacustrine conditions and is another common facies and surface deposit. Delta deposits are common near the basin margins. Well-sorted layers of eolian sand and silt and poorly sorted eolian loam occur locally above, within, or below the lacustrine deposits. The modern basins in all study areas are locally or completely inset into the Blackwater Draw Formation, supporting the interpretation that the basins are at least in part erosional features. In larger basins with thicker fills, generally coincident with thicker Blackwater Draw Formation, the formation interfingers with the lacustrine fill. Dating is based on radiocarbon ages from the fill in 12 basins and from lunettes adjacent to 5 basins. All dated basins were present at the end of the Pleistocene and some were present in some form throughout the Pleistocene. Lacustrine mud and other clastic deposits accumulated in the late Quaternary and locally much earlier, showing that at least some basins contained water throughout the time of human occupation of the region. Dating of eolian sediments supports other data indicative of aridity and wind deflation in the early and late Holocene. The lacustrine carbonate is late Pleistocene or older and its paleoenvironmental significance is unknown. These lithostratigraphic and chronostratigraphic relationships show that some basins have a prolonged history as depressions, persisting in more or less the same location as the High Plains surface aggraded by eolian addition (Blackwater Draw Formation) throughout the Pleistocene. Sizes of the basins varied through time as they were encroached upon by the Blackwater Draw Formation, enlarged by fluvial, lake margin, and eolian erosion, were filled and reexposed, or were buried. Some basins are newly formed on the High Plains surface and have no apparent predecessors. The only evidence for subsidence beneath the basins is gently warped fill in the basins on the northern margin of the region, known to be affected by salt dissolution in Paleozoic bedrock. Pedogenic carbonates typically are absent from or beneath the basin fill, due to focused recharge through the basins. Playa basins probably have been a ubiquitous component of the High Plains landscape through much of the Quaternary.

Origin of Satin Spar Veins in Evaporite Basins

ABSTRACT Satin spar (fibrous gypsum) veins occur in rocks overlying evaporites in the Amadeus Basin, Australia; Appalachian Basin, USA; Cheshire Basin, England; Elk Point Basin, Canada; Palo Duro Basin, USA; Paradox Basin, USA, and Zechstein Basin, England. These antitaxial veins, which are characterized by central partings and near-vertical fibers, fill horizontal and inclined fractures. Most satin spar veins occur in highly fractured or brecciated clastic strata that overlie or are associated with bedded halite, anhydrite, and gypsum. Saline springs are commonly present, indicating that halite dissolution is active. The similarities in morphology and occurrences of these veins suggest a common genesis. Satin spar veins result from several simultaneously active processes. Recharge of low-salinity surface water results in dissolution of shallow ( 200 m to 750 m deep) halite, development of cavernous porosity, and formation of extensional fractures in the rock column overlying salt-dissolution zones. Greater solubility of anhydrite than gypsum at low temperatures and at salinities below halite saturation causes hydration of anhydrite to gypsum, which takes place without significant volume change. Once groundwater is saturated with respect to gypsum, any further anhydrite hydration and solution must result in supersaturation, and the excess CaSO4 carried by ground water flowin from dissolution zones precipitates as gypsum in open, high-permeability fractures.

Halite pseudomorphs after gypsum in bedded anhydrite; clue to gypsum-anhydrite relationships

ABSTRACT Halite pseudomorphs after gypsum crystals preserve depositional textures and record the early diagenetic history of anhydrite beds in cyclic Permian evaporites of the Texas Panhandle. The well-documented sedimentologic setting and good preservation of fabrics permit identification of a sequence of diagenetic processes in gypsum. Gypsum (rather than anhydrite) was the dominant sulfate precipitated, even at salinities near halite saturation. Large gypsum crystals grew vertically on floors of brine pools, surrounded by a matrix of autochthonous transported gypsum sand and silt. All gypsum has been diagenetically altered to anhydrite ± halite, but the diagenetic fabrics vary depending on position within regressive evaporite cycles. Because the composition and texture of the primary sediment were similar throughout the anhydrite bed, these different diagenetic fabrics are interpreted as the result of changes in the processes of replacement of gypsum by anhydrite and halite as the shallow burial environment became increasingly saline. In the lower parts of anhydrite beds, diagenesis in low-salinity brines favored obliteration of primary sedimentary structures by formation of nodules. In the middle parts of beds, introduction of high-salinity brine into the shallow burial environment favored alteration of gypsum to anhydrite and preservation of pseudomorphs after large gypsum crystals. At the top of anhydrite beds, introduction of halite-saturated brines into gypsum sediments on the floor of the brine pool resulted in pervasive replacement of gypsum by halite. Petrogr phic relationships indicate that most or all of the primary gypsum was altered to anhydrite in the very shallow (< 2 m) burial environment under the influence of saline diagenetic brines.

Fluid Inclusions in Bedded Permian Halite, Palo Duro Basin, Texas: Evidence for Modification of Seawater in Evaporite Brine-Pools and Subsequent Early Diagenesis

ABSTRACT The chemical composition of fluid inclusions trapped within the marine-dominated Permian evaporite sequence of the Palo Duro Basin, Texas, departs significantly from that expected in the evolution of seawater through evaporative concentration and mineral precipitation. The origin of these fluids is evaluated in the context of the sedimentary and diagenetic environments in which the halites evolved. The petrography and fabric show that many of the halites escaped advanced recrystallization, suggesting that the fluids entrapped in such samples are original brines. The seawater-derived brine-pool and diagenetic brines from which the halites precipitated were modified as a result of fluctuations in water level on an evaporite flat and of a complex hydrological system through which seawate was fed into the basin and refluxed. The ionic ratios between Ca, Mg and SO4 are influenced by dolomitization and associated reactions of interstitial brines with carbonates adjacent to the halite beds. Formation and subsequent dissolution of efflorescent evaporite crusts that are not in equilibrium with subaqueously precipitated evaporites shift the ratios of Mg, Na, K, and Cl to Br away from the expected path of evaporative seawater concentration. Syndepositional alteration of clay and feldspar minerals probably also occurred, but these effects cannot be evaluated from the present data. The variability and complexity of brine evolution within an uncomplicated evaporite setting, such as that of the Palo Duro Basin, has important implications for interpretation of the origin and history of evaporite-influenced brines in other basins. Laboratory evaporation experiments and analyses of salt pan brines only serve as a basic reference for the evaporation path of seawater; they ignore the early diagenetic processes, and hence taken alone are inadequate analogs for ancient marine evaporite environments.

Testing efficiency of storage in the subsurface: frio brine pilot experiment

Can we demonstrate that subsurface storage is an effective method of reducing emissions of CO2 to the atmosphere? The Frio Brine Pilot Experiment is designed to test storage performance of a typical subsurface environment in an area where large-volume sources and sinks are abundant, near Houston, Texas, USA. We employed extensive pre-experiment characterization and modeling to identify significant factors that increase or decrease risk of leakage from the injection zone. We then designed the experiment to focus on those factors, as well as to test for presence or absence of events that are not expected. A fully developed reservoir model of heterogeneous reworked fluvial sandstones of the Frio Formation documents three-dimensional compartmentalization of the injection horizon by faulting associated with salt-dome intrusion and growth. Modeling using the TOUGH2 simulator showed that a significant source of uncertainty for subsurface performance of injected CO2 is residual CO2 saturation during storage. If initial displacement of water during injection is efficient and capillary effects create the expected residual saturation of 30 percent CO2, the volume occupied by the plume will be limited, and long-term storage can be expected even in an open system. If, however, during injection, CO2 moves out from the injection well along high-permeability pathways, it may not contact most pores, and residual saturation will have a smaller effect on storage. Our experiment is therefore designed to monitor plume geometry and CO2 saturation near the injection well and closely spaced observation well. Leakage out of the injection zone as a result of well engineering or other flaws in the seal is also monitored in the sandstone immediately overlying the injection zone and at the surface using multiple techniques. Permitting strategies include cooperation among two State agencies, as well as Federal NEPA assessment, because of the innovative aspects of the experiment.