Search

Characterization of karst systems and forecast of their state variables are essential for groundwater management and engineering in karst regions. These objectives can be met by the use of process-based discrete-continuum models (DCMs). However, results of DCMs may suffer from inversion nonuniqueness. It has been demonstrated that the joint inversion of observations regulated by different natural processes can tackle the nonuniqueness issue in groundwater modeling. However, this has not been tested for DCMs thus far. This research proposes a methodology for the joint inversion of hydro-thermo-chemo-graphs, applying to two small-scale sink-to-spring experiments at Freiheit Spring, Minnesota, USA. In order to address conceptual uncertainty, a multimodel approach was implemented, featuring seven mutually exclusive variants. Spring hydro-thermo-chemo-graphs, for all the variants simulated by MODFLOW-CFPv2, were jointly inverted using a weighted least squares algorithm. Subsequently, models were compared in terms of inversion and forecast performances, as well as parameter uncertainties. Results reveal the suitability of the DCM approach for simultaneous inversion and forecast of hydro-physico-chemical behavior of karst systems, even at a scale of meters and seconds. The estimated volume of the tracer conduit passage ranges from approximately 46–51 m3, which is comparable to the estimate from the flood-pulse method. Moreover, it was demonstrated that the thermograph and hydrograph contain more information about aquifer characteristics than the chemograph. However, this finding can be site-specific and should depend on the analysis scale, the considered conceptual models, and the hydrological state, which are potentially affected by minor unaccountable processes and features.

Serpentinization plays a fundamental role in the biogeochemical and tectonic evolution of the Earth and perhaps many other rocky planetary bodies. Yet, geochemical models still fail to produce accurate predictions of the various modes of serpentinization, which limits our ability to predict a variety of related geological phenomena over many spatial and temporal scales. Here, we use kinetic and reactive transport experiments to parameterize the brucite silicification reaction and provide fundamental constraints on SiO2 transport during serpentinization. We show that, at temperatures characteristic of the sub-seafloor at the serpentinite-hosted Lost City Hydrothermal Field (150 °C), the assembly of Si tetrahedra onto MgOH2 (i.e., brucite) surfaces is a rate-limiting elementary reaction in the production of serpentine and/or talc from olivine. Moreover, this reaction is exponentially dependent on the activity of aqueous silica ( a SiO 2 ( aq ) ), such that it can be calculated according to the rate law: Rate = 2.3 × 10 − 4 a SiO 2 ( aq ) 1.5 ( mol / m 2 / s ) . Calculations performed with this rate law demonstrate that both brucite and Si are surprisingly persistent in serpentinizing environments, leading to elevated Si concentrations in fluids that can be transported over comparatively large distances without equilibrating with brucite. Moreover, applying this rate law to an open-system reactive transport experiment indicates that advection, preferential flow pathways, and reactive surface area armoring can diminish the net rate of Si uptake resulting from this reaction even further. Because brucite silicification is a fundamentally rate-limiting elementary reaction for the production of both serpentine and talc from forsterite, our new constraints are applicable across the many environments where serpentinization occurs. The unexpected but highly consequential behavior of this simple reaction emphasizes the need for considering serpentinization and many other hydrothermal processes in a reactive transport framework whereby fluid, solute, and heat transport are intimately coupled to kinetically-controlled reactions.

Two flow-through experiments were conducted to assess serpentinization of intact dunite cores. Permeability and fluid chemistry indicate significantly more reaction during the second experiment at 200 °C than the first experiment at 150 °C. Permeability decreased by a factor of 2.4 and 25 during the experiments at 150 and 200 °C, respectively. Furthermore, hydrogen and methane concentrations exceeded 600 μmol/kg and 300 μmol/kg during the 200 °C experiment, and were one and two orders of magnitude higher, respectively, than the 150 °C experiment. Fe K-edge X-ray absorption near edge structure analyses of alteration minerals demonstrated Fe oxidation that occurred during the 200 °C experiment. Vibrating sample magnetometer measurements on post-experimental cores indicated little to no magnetite production, suggesting that the hydrogen was largely generated by the oxidation of iron as olivine was converted to ferric iron (Fe(III)) serpentine and/or saponite. Scanning electron microscopy images suggested secondary mineralization on the post-experimental core from the 200 °C experiment, portraying the formation of a secondary phase with a honeycomb-like texture as well as calcite and wollastonite. Scanning electron microscopy images also illustrated dissolution along linear bands through the interiors of olivine crystals, possibly along pathways with abundant fluid inclusions. Energy dispersive X-ray spectroscopy identified Cl uptake in serpentine, while Fourier transform-infrared spectroscopy suggested the formation of serpentine, saponite, and talc. However, no change was observed when comparing pre- and post-experimental X-ray computed tomography scans of the cores. Furthermore, (ultra) small angle neutron scattering datasets were collected to assess changes in porosity, surface area, and fractal characteristics of the samples over the ≈ 1 nm- to 10 μm-scale range. The results from the 200 °C post-experimental core generally fell within the range of values for the two pristine samples and the 150 °C post-experimental core that underwent negligible reaction, indicating that any change from reaction was smaller than the natural variability of the dunite. Even though there was little physical evidence of alteration, the initial stage of serpentinization at 200 °C was sufficiently significant to have a dramatic effect on flow fields in the core. Furthermore, this experiment generated significant dissolved hydrogen concentrations while simulating open system dynamics. Even though open systems prevent elevated hydrogen concentrations due to continual loss of hydrogen, we speculate that this process is responsible for stabilizing ferric Fe-rich serpentine in nature while also oxidizing more ferrous iron (Fe(II)) and cumulatively generating more hydrogen than would be possible in a closed system.

Equilibrium Fe isotope fractionation between chalcopyrite and dissolved Fe was determined in acidic chloride-bearing fluid at 350 °C and 500 bars. The study utilized deformable gold-cell technology, which allowed time-series sampling of solution during chalcopyrite recrystallization and isotope exchange. A key element of the experimental design involved the addition of anomalous dissolved 57 Fe to an on-going experiment as a means of determining the degree and rate of isotope exchange. Taking explicit account of imposed chemical and isotopic mass balance constraints of Fe in fluid and mineral (chalcopyrite) reservoirs, these data indicate that no more than 1000 h is required for the isotopically anomalous dissolved Fe reservoir to exchange completely with the coexisting chalcopyrite. The experimental calibration of the rate of Fe isotope exchange for the δ 57 Fe-spiked experiment provides critical insight for the time necessary to achieve Fe isotope exchange in two non-spiked, but otherwise identical experiments. The Fe isotope data indicate that the equilibrium fractionation between chalcopyrite and dissolved Fe, Δ 56 Fe Cpy-Fe (aq) , at 350 °C is small, 0.09 ± 0.17‰ (2 σ ), and is in good agreement with recent theoretical equilibrium predictions. Owing to the apparent rate of Fe isotope exchange at 350 °C, it is likely that chalcopyrite formed at high temperature deep-sea vents (black smoker systems) achieves isotopic equilibrium, and effectively records the Fe isotopic composition of the coexisting end-member hydrothermal fluid. Comparison of the experimental mineral– fluid equilibrium fractionation factors with conjugate chalcopyrite and dissolved Fe pairs sampled from high temperature hydrothermal vent systems at Axial Caldera and Main Endeavour Field (Juan de Fuca Ridge) are in agreement with this inference. The experimental data were further used to determine the mineral– mineral equilibrium Fe isotope fractionation between pyrite-chalcopyrite, Δ 56 Fe Pyr-Cpy , at 350 °C by combining previously determined pyrite-Fe 2+ (aq) equilibrium fractionation data with chalcopyrite-Fe 2+ (aq) from this study. The empirically determined Δ 56 Fe Pyr-Cpy value, 0.90 ± 0.34‰ (2 σ ), is consistent with theoretical predictions, and when coupled with mineral– fluid Fe isotope fractionation systematics and experimentally determined exchange rates, helps to delineate processes of sulfide mineralization in hydrothermal systems.

Greensands formations are globally abundant sedimentary rocks rich in Fe clays (typically glauconite) that commonly contain natural hydrocarbon accumulations and may be important reservoirs for geologic storage of anthropogenic CO2. Diagenesis in greensands is commonly accompanied by the conversion of primary glauconite to siderite (FeCO3), a process that could be exploited for the permanent trapping of CO2. Importantly, siderite formation after glauconite requires that the mostly oxidized Fe in the primary Fe clay minerals is reduced during diagenetic interactions. Here, we explore the effect of solution redox state on the stability of glauconite in sandstones with implications for the diagenetic and/or engineered formation of siderite. We performed two flow-through experiments on intact, glauconite-rich sandstone cores at 150 °C and 150 bar. Both experiments employed a 1 mol NaCl/kg, 0.1 mol NaHCO3/kg solution charged with ~0.58 mol CO2/kg solution, but the redox state of the injected fluid was manipulated between experiments in order to compare glauconite reactivity and siderite saturation state at oxidizing and reducing end-member conditions. After reaction with the oxidizing (O2 (aq) ≈ 6 μmol/kg) fluid, chemical and Mӧssbauer spectroscopic analyses indicate the production of Fe(III)-oxy/hydroxide minerals from glauconite, whereas, in the reducing (H2(aq) ≈ 5–40 mmol/kg) experiment, thermodynamic calculations and coupled chemical, mineralogical, and Mӧssbauer analyses suggest glauconite dissolution and precipitation of an Fe(II) mineral, likely siderite, and minor magnetite formation. These experimental results, along with thermodynamic calculations, confirm that solution redox state is the master variable dictating siderite formation in greensands.

Water temperature is a non-conservative tracer in the environment. Variations in recharge temperature are damped and retarded as water moves through an aquifer due to heat exchange between water and rock. However, within karst aquifers, seasonal and short-term fluctuations in recharge temperature are often transmitted over long distances before they are fully damped. Using analytical solutions and numerical simulations, we develop relationships that describe the effect of flow path properties, flow-through time, recharge characteristics, and water and rock physical properties on the damping and retardation of thermal peaks/troughs in karst conduits. Using these relationships, one can estimate the thermal retardation and damping that would occur under given conditions with a given conduit geometry. Ultimately, these relationships can be used with thermal damping and retardation field data to estimate parameters such as conduit diameter. We also examine sets of numerical simulations where we relax some of the assumptions used to develop these relationships, testing the effects of variable diameter, variable velocity, open channels, and recharge shape on thermal damping and retardation to provide some constraints on uncertainty. Finally, we discuss a multitracer experiment that provides some field confirmation of our relationships. High temporal resolution water temperature data are required to obtain sufficient constraints on the magnitude and timing of thermal peaks and troughs in order to take full advantage of water temperature as a tracer.

To investigate CO 2 Capture, Utilization, and Storage (CCUS) in sandstones, we performed three 150 °C flow-through experiments on K-feldspar-rich cores from the Eau Claire formation. By characterizing fluid and solid samples from these experiments using a suite of analytical techniques, we explored the coupled evolution of fluid chemistry, mineral reaction rates, and hydrogeochemical properties during CO 2 sequestration in feldspar-rich sandstone. Overall, our results confirm predictions that the heightened acidity resulting from supercritical CO 2 injection into feldspar-rich sandstone will dissolve primary feldspars and precipitate secondary aluminum minerals. A core through which CO 2 -rich deionized water was recycled for 52 days decreased in bulk permeability, exhibited generally low porosity associated with high surface area in post-experiment core sub-samples, and produced an Al hydroxide secondary mineral, such as boehmite. However, two samples subjected to ∼3 day single-pass experiments run with CO 2 -rich, 0.94 mol/kg NaCl brines decreased in bulk permeability, showed generally elevated porosity associated with elevated surface area in post-experiment core sub-samples, and produced a phase with kaolinite-like stoichiometry. CO 2 -induced metal mobilization during the experiments was relatively minor and likely related to Ca mineral dissolution. Based on the relatively rapid approach to equilibrium, the relatively slow near-equilibrium reaction rates, and the minor magnitudes of permeability changes in these experiments, we conclude that CCUS systems with projected lifetimes of several decades are geochemically feasible in the feldspar-rich sandstone end-member examined here. Additionally, the observation that K-feldspar dissolution rates calculated from our whole-rock experiments are in good agreement with literature parameterizations suggests that the latter can be utilized to model CCUS in K-feldspar-rich sandstone. Finally, by performing a number of reactive transport modeling experiments to explore processes occurring during the flow-through experiments, we have found that the overall progress of feldspar hydrolysis is negligibly affected by quartz dissolution, but significantly impacted by the rates of secondary mineral precipitation and their effect on feldspar saturation state. The observations produced here are critical to the development of models of CCUS operations, yet more work, particularly in the quantification of coupled dissolution and precipitation processes, will be required in order to produce models that can accurately predict the behavior of these systems.

© 2017 Elsevier Ltd The isotopic composition of methane is of longstanding geochemical interest, with important implications for understanding petroleum systems, atmospheric greenhouse gas concentrations, the global carbon cycle, and life in extreme environments. Recent analytical developments focusing on multiply substituted isotopologues (‘clumped isotopes’) are opening a valuable new window into methane geochemistry. When methane forms in internal isotopic equilibrium, clumped isotopes can provide a direct record of formation temperature, making this property particularly valuable for identifying different methane origins. However, it has also become clear that in certain settings methane clumped isotope measurements record kinetic rather than equilibrium isotope effects. Here we present a substantially expanded dataset of methane clumped isotope analyses, and provide a synthesis of the current interpretive framework for this parameter. In general, clumped isotope measurements indicate plausible formation temperatures for abiotic, thermogenic, and microbial methane in many geological environments, which is encouraging for the further development of this measurement as a geothermometer, and as a tracer for the source of natural gas reservoirs and emissions. We also highlight, however, instances where clumped isotope derived temperatures are higher than expected, and discuss possible factors that could distort equilibrium formation temperature signals. In microbial methane from freshwater ecosystems, in particular, clumped isotope values appear to be controlled by kinetic effects, and may ultimately be useful to study methanogen metabolism.

Four reactive flow-through laboratory experiments (two each at 0.1 mL/min and 0.01 mL/min flow rates) at 150°C and 150 bar (15 MPa) are conducted on intact basalt cores to assess changes in porosity, permeability, and surface area caused by CO2-rich fluid-rock interaction. Permeability decreases slightly during the lower flow rate experiments and increases during the higher flow rate experiments. At the higher flow rate, core permeability increases by more than one order of magnitude in one experiment and less than a factor of two in the other due to differences in preexisting flow path structure. X-ray computed tomography (XRCT) scans of pre- and post-experiment cores identify both mineral dissolution and secondary mineralization, with a net decrease in XRCT porosity of ∼0.7%–0.8% for the larger pores in all four cores. (Ultra) small-angle neutron scattering ((U)SANS) data sets indicate an increase in both (U)SANS porosity and specific surface area (SSA) over the ∼1 nm to 10 µm scale range in post-experiment basalt samples, with differences due to flow rate and reaction time. Net porosity increases from summing porosity changes from XRCT and (U)SANS analyses are consistent with core mass decreases. (U)SANS data suggest an overall preservation of the pore structure with no change in mineral surface roughness from reaction, and the pore structure is unique in comparison to previously published basalt analyses. Together, these data sets illustrate changes in physical parameters that arise due to fluid-basalt interaction in relatively low pH environments with elevated CO2 concentration, with significant implications for flow, transport, and reaction through geologic formations.

Four flow-through experiments at 150 °C were conducted on intact cores of basalt to assess alteration and mass transfer during reaction with CO2-rich fluid. Two experiments used a flow rate of 0.1 ml/min, and two used a flow rate of 0.01 ml/min. Permeability increased for both experiments at the higher flow rate, but decreased for the lower flow rate experiments. The experimental fluid (initial pH of 3.3) became enriched in Si, Mg, and Fe upon passing through the cores, primarily from olivine and titanomagnetite dissolution and possibly pyroxene dissolution. Secondary minerals enriched in Al and Si were present on post-experimental cores, and an Fe2O3-rich phase was identified on the downstream ends of the cores from the experiments at the lower flow rate. While we could not specifically identify if siderite (FeCO3) was present in the post-experimental basalt cores, siderite was generally saturated or supersaturated in outlet fluid samples, suggesting a thermodynamic drive for Fe carbonation from basalt-H2O-CO2 reaction. Reaction path models that employ dissolution kinetics of olivine, labradorite, and enstatite also suggest siderite formation at low pH. Furthermore, fluid-rock interaction caused a relatively high mobility of the alkali metals; up to 29% and 99% of the K and Cs present in the core, respectively, were preferentially dissolved from the cores, likely due to fractional crystallization effects that made alkali metals highly accessible. Together, these datasets illustrate changes in chemical parameters that arise due to fluid-basalt interaction in relatively low pH environments with elevated CO2.

Hydrothermal flow experiments of single-pass injection of CO2-charged brine were conducted on nine dolomite cores to examine fluid–rock reactions in dolomite reservoirs under geologic carbon sequestration conditions. Post-experimental X-ray computed tomography (XRCT) analysis illustrates a range of dissolution patterns, and significant increases in core bulk permeability were measured as the dolomite dissolved. Outflow fluids were below dolomite saturation, and cation concentrations decreased with time due to reductions in reactive surface area with reaction progress. To determine changes in reactive surface area, we employ a power-law relationship between reactive surface area and porosity (Luquot and Gouze, 2009). The exponent in this relationship is interpreted to be a geometrical parameter that controls the degree of surface area change per change in core porosity. Combined with XRCT reconstructions of dissolution patterns, we demonstrate that this exponent is inversely related to both the flow path diameter and tortuosity of the dissolution channel. Even though XRCT reconstructions illustrate dissolution at selected regions within each core, relatively high Ba and Mn recoveries in fluid samples suggest that dissolution occurred along the core's entire length and width. Analysis of porosity–permeability data indicates an increase in the rate of permeability enhancement per increase in porosity with reaction progress as dissolution channels lengthen along the core. Finally, we incorporate the surface area–porosity model of Luquot and Gouze (2009) with our experimentally fit parameters into TOUGHREACT to simulate experimental observations.

A controlled recharge event with multiple tracers was conducted on August 30, 2010. A pool adjacent to a sinkhole was filled with approximately 13,000 L of water. The water was heated, and salt, deuterium oxide, and uranine were added. The pool was then emptied into the sinkhole, and data were collected at Freiheit Spring approximately 95 m north of the sinkhole to monitor changes in discharge, temperature, conductivity/chloride, δD, uranine, and suspended sediment. This combined trace demonstrated the feasibility and utility of conducting superimposed physical, chemical, and isotopic traces. Flow peaked first at the spring and was followed by a suspended sediment peak; then essentially identical uranine, chloride, and δD peaks; and finally a temperature peak. The initial increase in flow at the spring recorded the time at which the water reached a submerged conduit, sending a pressure pulse to the spring at the speed of sound. The initial increase in uranine, chloride, and δD at the spring recorded the arrival of the recharge water. The initial change in temperature and its peak occurred later than the same parameters in the uranine, chloride, and δD breakthrough curves. As water flowed along this flow path, water temperature interacted with the aquifer, producing a delayed, damped thermal peak at the spring. The combination of conservative and nonconservative tracers illustrates unique pressure, advective, and interactive processes.

Summary A controlled recharge event was conducted with multiple tracers in a karst aquifer in southeastern Minnesota. A pool adjacent to a sinkhole was filled with approximately 13,000 L of water. After tracers were added and thoroughly mixed, the pool was emptied into the sinkhole. Data were collected at Freiheit Spring approximately 95 m north of the sinkhole to monitor spring responses. Flow peaked first at the spring, and suspended sediment peaked next. Then nearly identical uranine, chloride, and δD peaks occurred. Temperature was the last of the tracers to peak. The initial increase in flow at the spring recorded the time at which the water reached a submerged conduit, sending a pressure pulse to the spring at approximately the speed of sound in open water. The initial increase in uranine, chloride, and δD at the spring recorded the arrival of the recharge water. The initial change in temperature and its peak occurred later than the same features in the uranine, chloride, and δD breakthrough curves. As water flowed along this flow path, water exchanged heat with the aquifer, producing a lagged, damped thermal peak at the spring. The combination of hydraulic response and conservative and nonconservative tracers illustrates unique pressure, advective, and nonconservative processes. Geometrical properties of the flow system may be estimated using these tracers. By summing discharge between the time of the initial increase in stage produced by a pressure pulse in a fully phreatic flow path and the time of the chloride peak, the conduit volume is estimated as 47 ± 10% m3. Heat transport simulations were used to reproduce the modified thermal signal, and simulations with planar flow paths and hydraulic diameters of 7 and 8 cm produced the best fits to the observed temperature breakthrough curve. These volume and hydraulic diameter estimates together predict a bedding plane flow path that is 3.5 cm high by 9 m wide or 4 cm high by 8 m wide. The different tracers provide complementary information, and the combination of parameters provides useful constraints on flow path geometry.

Thermal patterns of karst springs and cave streams provide potentially useful information concerning aquifer geometry and recharge. Temperature monitoring at 25 springs and cave streams in southeastern Minnesota has shown four distinct thermal patterns. These patterns can be divided into two types: those produced by flow paths with ineffective heat exchange, such as conduits, and those produced by flow paths with effective heat exchange, such as small fractures and pore space. Thermally ineffective patterns result when water flows through the aquifer before it can equilibrate to the rock temperature. Thermally ineffective patterns can be either event-scale, as produced by rainfall or snowmelt events, or seasonal scale, as produced by input from a perennial surface stream. Thermally effective patterns result when water equilibrates to rock temperature, and the patterns displayed depend on whether the aquifer temperature is changing over time. Shallow aquifers with seasonally varying temperatures display a phase-shifted seasonal signal, whereas deeper aquifers with constant temperatures display a stable temperature pattern. An individual aquifer may display more than one of these patterns. Since karst aquifers typically contain both thermally effective and ineffective routes, we argue that the thermal response is strongly influenced by recharge mode.

Carbon sequestration experiments were conducted on uncemented sediment and lithified rock from the Eau Claire Formation, which consisted primarily of K-feldspar and quartz. Cores were heated to accentuate reactivity between fluid and mineral grains and to force CO(2) exsolution. Measured permeability of one sediment core ultimately reduced by 4 orders of magnitude as it was incrementally heated from 21 to 150 °C. Water-rock interaction produced some alteration, yielding sub-μm clay precipitation on K-feldspar grains in the core's upstream end. Experimental results also revealed abundant newly formed pore space in regions of the core, and in some cases pores that were several times larger than the average grain size of the sediment. These large pores likely formed from elevated localized pressure caused by rapid CO(2) exsolution within the core and/or an accumulating CO(2) phase capable of pushing out surrounding sediment. CO(2) filled the pores and blocked flow pathways. Comparison with a similar experiment using a solid arkose core indicates that CO(2) accumulation and grain reorganization mainly contributed to permeability reduction during the heated sediment core experiment. This suggests that CO(2) injection into sediments may store more CO(2) and cause additional permeability reduction than is possible in lithified rock due to grain reorganization.

[1] Simple mathematical models often allow an intuitive grasp of the function of physical systems. We develop a mathematical framework to investigate reactive or dissipative transport processes within karst conduits. Specifically, we note that for processes that occur within a characteristic timescale, advection along the conduit produces a characteristic process length scale. We calculate characteristic length scales for the propagation of thermal and electrical conductivity signals along karst conduits. These process lengths provide a quantitative connection between karst conduit geometry and the signals observed at a karst spring. We show that water input from the porous/fractured matrix is also characterized by a length scale and derive an approximation that accounts for the influence of matrix flow on the transmission of signals through the aquifer. The single conduit model is then extended to account for conduits with changing geometries and conduit flow networks, demonstrating how these concepts can be applied in more realistic conduit geometries. We introduce a recharge density function, ϕR, which determines the capability of an aquifer to damp a given signal, and cast previous explanations of spring variability within this framework. Process lengths are a general feature of karst conduits and surface streams, and we conclude with a discussion of other potential applications of this conceptual and mathematical framework.

[1] Previous studies, motivated by understanding water quality, have explored the mechanisms for heat transport and heat exchange in surface streams. In karst aquifers, temperature signals play an additional important role since they carry information about internal aquifer structures. Models for heat transport in karst conduits have previously been developed; however, these models make different, sometimes contradictory, assumptions. Additionally, previous models of heat transport in karst conduits have not been validated using field data from conduits with known geometries. Here we use analytical solutions of heat transfer to examine the relative importance of heat exchange mechanisms and the validity of the assumptions made by previous models. The relative importance of convection, conduction, and radiation is a function of time. Using a characteristic timescale, we show that models neglecting rock conduction produce spurious results in realistic cases. In contrast to the behavior of surface streams, where conduction is often negligible, conduction through the rock surrounding a conduit determines heat flux at timescales of weeks and longer. In open channel conduits, radiative heat flux can be significant. In contrast, convective heat exchange through the conduit air is often negligible. Using the rules derived from our analytical analysis, we develop a numerical model for heat transport in a karst conduit. Our model compares favorably to thermal responses observed in two different karst settings: a cave stream fed via autogenic recharge during a snowmelt event, and an allogenically recharged cave stream that experiences continuous temperature fluctuations on many timescales.

A single trace near the Fillmore-Winona County Border near the City of Rushford, Fillmore County, Minnesota. The trace began in Winona County and did not cross the county boundary. Primary goal of the study was to delineate the springshed feeding a cluster of trout streams. A collaborative effort between the Minnesota Department of Natural Resources and the University of Minnesota.

A summary of springshed delineation techniques using the integration of historic dye tracing information, borehole geophysics and chemistry measurements. Fountain, Mahoney, Waterhole, Wykoff, Starless River and Cold Spring springshed delineations included. Interpretations are subject to change as new traces and information becomes available. A collaborative effort between the University of Minnesota and the Department of Natural Resources.

A. J. Luhmann1, M. D. Covington2, J. M. Myre2, M. Perne2,3, S. W. Jones4, E. C. Alexander Jr.1, and M. O. Saar1,5 1University of Minnesota, Department of Earth Sciences, 310 Pillsbury Dr. SE, Minneapolis, Minnesota 55455, USA 2University of Arkansas, Department of Geosciences, 216 Ozark Hall, Fayetteville, Arkansas 72701, USA 3Jožef Stefan Institute, Department of Systems and Control, Jamova Cesta 39, Ljubljana, Slovenia 4527 Karrow St., Maryville, Tennessee 37803, USA 5ETH-Zurich, Geothermal Energy and Geofluids Group, Department of Earth Sciences, Zurich, Switzerland

Injection of cool CO2 into geothermally warm carbonate reservoirs for storage or geothermal energy production may lower near-well temperature and lead to mass transfer along flow paths leading away from the well. To investigate this process, a dolomite core was subjected to a 650 h, high pressure, CO2 saturated, flow-through experiment. Permeability increased from 10(-15.9) to 10(-15.2) m(2) over the initial 216 h at 21 °C, decreased to 10(-16.2) m(2) over 289 h at 50 °C, largely due to thermally driven CO2 exsolution, and reached a final value of 10(-16.4) m(2) after 145 h at 100 °C due to continued exsolution and the onset of dolomite precipitation. Theoretical calculations show that CO2 exsolution results in a maximum pore space CO2 saturation of 0.5, and steady state relative permeabilities of CO2 and water on the order of 0.0065 and 0.1, respectively. Post-experiment imagery reveals matrix dissolution at low temperatures, and subsequent filling-in of flow passages at elevated temperature. Geochemical calculations indicate that reservoir fluids subjected to a thermal gradient may exsolve and precipitate up to 200 cm(3) CO2 and 1.5 cm(3) dolomite per kg of water, respectively, resulting in substantial porosity and permeability redistribution.