Surface controls on the characteristics of natural CO2 seeps: implications for engineered CO2 stores

CO₂ injected into rock formations for deep geological storage must not leak to surface, since this would be economically and environmentally unfavourable, and could present a human health hazard. In Italy natural CO₂ degassing to the surface via seeps is widespread, providing an insight into the various styles of subsurface ‘plumbing’ as well as surface expression of CO₂ fluids. Here we investigate surface controls on the distribution of CO₂ seep characteristics (type, flux and temperature) using a large geographical and historical data set. When the locations of documented seeps are compared to a synthetic statistically random data set, we find that the nature of the CO₂ seeps is most strongly governed by the flow properties of the outcropping rocks, and local topography. Where low‐permeability rocks outcrop, numerous dry seeps occur and have a range of fluxes. Aqueous fluid flow will be limited in these low‐permeability rocks, and so relative permeability effects may enable preferential CO₂ flow. CO₂ vents typically occur along faults in rocks that are located above the water table or are low permeability. Diffuse seeps develop where CO₂ (laterally supplied by these faults) emerges from the vadose zone and where CO₂ degassing from groundwater follows a different flow path due to flow differences for water and CO₂ gas. Bubbling water seeps (characterized by water bubbling with CO₂) arise where CO₂ supply enters the phreatic zone or an aquifer. CO₂‐rich springs often emerge where valleys erode into CO₂ aquifers, and these are typically high flux seeps. Seep type is known to influence human health risk at CO₂ seeps in Italy, as well as the topography surrounding the seep which affects the rate of gas dispersion by wind. Identifying the physical controls on potential seep locations and seep type above engineered CO₂ storage operations is therefore crucial to targeted site monitoring strategy and risk assessment. The surface geology and topography above a CO₂ store must therefore be characterized in order to design the most effective monitoring strategy.     

Polyphasic hydrothermal and meteoric fluid regimes during the growth of a segmented fault involving crystalline and carbonate rocks (Barcelona Plain,NE Spain)

A polyphasic tectonic‐fluid system of a fault that involves crystalline and carbonate rocks (Hospital fault, Barcelona Plain) has been inferred from regional to thin section scale observations combined with geochemical analyses. Cathodoluminescence, microprobe analyses and stable isotopy in fracture‐related cements record the circulation of successive alternations of hydrothermal and low‐temperature meteoric fluids linked with three main regional tectonic events. The first event corresponds to the Mesozoic extension, which had two rifting stages, and it is characterized by the independent tectonic activity of two fault segments, namely southern and northern Hospital fault segments. During the Late Permian‐Middle Jurassic rifting, these segments controlled the thickness and distribution of the Triassic sediments. Also, dolomitization was produced in an early stage by Triassic seawater at shallow conditions. During increasing burial, formation of fractures and their dolomite‐related cements took place. Fault activity during the Middle Jurassic–Late Cretaceous rifting was localized in the southern segment, and it was characterized by hydrothermal brines, with temperatures over 180°C, which ascended through this fault segment precipitating quartz, chlorite, and calcite. The second event corresponds to the Paleogene compression (Chattian), which produced exhumation, folding and erosion, favouring the percolation of low‐temperature meteoric fluids which produced the calcitization of the dolostones and of the dolomite cements. The third event is linked with the Neogene extension, where three stages have been identified. During the syn‐rift stage, the southern segment of the Hospital fault grew by tip propagation. In the relay zone, hydrothermal brines with temperature around 140°C upflowed. During the late postrift, the Hospital fault acted as a unique segment and deformation occurred at shallow conditions and under a low‐temperature meteoric regime. Finally, and possibly during the Messinian compression, NW‐SE strike‐slip faults offset the Hospital fault to its current configuration.     

Carbon dioxide controlled earthquake distribution pattern in the NW Bohemian swarm earthquake region,western Eger Rift,Czech Republic – gas migration in the crystalline basement

The ascent of magmatic carbon dioxide in the western Eger (Oh?e) Rift is interlinked with the fault systems of the Variscian basement. In the Cheb Basin, the minimum CO₂ flux is about 160 m³ h^?1, with a diminishing trend towards the north and ceasing in the main epicentral area of the Northwest Bohemian swarm earthquakes. The ascending CO₂ forms Ca‐Mg‐HCO₃ type waters by leaching of cations from the fault planes and creates clay minerals, such as kaolinite, as alteration products on affected fault planes. These mineral reactions result in fault weakness and in hydraulically interconnected fault network. This leads to a decrease in the friction coefficient of the Coulomb failure stress (CFS) and to fault creep as stress build‐up cannot occur in the weak segments. At the transition zone in the north of the Cheb Basin, between areas of weak, fluid conductive faults and areas of locked faults with frictional strength, fluid pressure can increase resulting in stress build‐up. This can trigger strike‐slip swarm earthquakes. Fault creep or movements in weak segments may support a stress build‐up in the transition area by transmitting fluid pressure pulses. Additionally to fluid‐driven triggering models, it is important to consider that fluids ascending along faults are CO₂‐supersaturated thus intensifying the effect of fluid flow. The enforced flow of CO₂‐supersaturated fluids in the transitional zone from high to low permeability segments through narrowings triggers gas exsolution and may generate pressure fluctuations. Phase separation starts according to the phase behaviour of CO₂‐H₂O systems in the seismically active depths of NW Bohemia and may explain the vertical distribution of the seismicity. Changes in the size of the fluid transport channels in the fault systems caused, or superimposed, by fault movements, can produce fluid pressure increases or pulses, which are the precondition for triggering fluid‐induced swarm earthquakes.     

Changes in chemical composition caused by water–rock interactions across a strike‐slip fault zone: case study of the Atera Fault,Central Japan

In some cases, water–rock interactions in fault zones can affect radionuclide migration. Here, we analyzed the chemical compositions of well‐exposed fault rocks from the strike‐slip Atera Fault, Central Japan, in order to understand the variability and behavior of major and selected trace elements. The fault zone has a 1.2‐m‐wide, smectite‐rich fault core and paired damage zones that developed within welded tuff on one side of the core and within granite on the other side. The 30‐cm‐wide, kaolinite‐rich fault gouge is developed in granite cataclasite, and it shows indications of the latest fault activity, while the 1.2‐m‐wide fault core appears to be older. Hydrogen and oxygen isotope ratios in the clay‐rich fault gouges, and carbon and oxygen isotope ratios in carbonates indicate that the two major clay‐rich zones formed in bedrock near the surface, consistent with observed deformation structures. Based on chemical analyses, we identified (1) a slight depletion in SiO₂, Na₂O, K₂O, and light rare earth elements at the edges of the 1.2‐m‐wide fault core, (2) a clear depletion in SiO₂, Na₂O, K₂O, and all rare earth elements except Eu in the 30‐cm‐wide fault gouge, and (3) an increase in CaO, MnO, and heavy rare earth elements across the entire 1.2‐m‐wide fault core. Findings (1) and (2) reflect water–rock interactions in the 1.2‐m‐wide fault core and in the 30‐cm‐wide fault gouge that resulted in the formation of smectite and kaolinite. Finding (3) reflects carbonate precipitation caused by the addition of basalt fragments from a nearby site to the 1.2‐m‐wide fault core during faulting, and subsequent sorption reactions of heavy rare earth elements via processes such as complexation with the carbonates.     

Faults and fault properties in hydrocarbon flow models

The petroleum industry uses subsurface flow models for two principal purposes: to model the flow of hydrocarbons into traps over geological time, and to simulate the production of hydrocarbon from reservoirs over periods of decades or less. Faults, which are three-dimensional volumes, are approximated in both modelling applications as planar membranes onto which predictions of the most important fault-related flow properties are mapped. Faults in porous clastic reservoirs are generally baffles or barriers to flow and the relevant flow properties are therefore very different to those which are important in conductive fracture flow systems. A critical review and discussion is offered on the work-flows used to predict and model capillary threshold pressure for exploration fault seal analysis and fault transmissibility multipliers for production simulation, and of the data from which the predictions derive. New flow simulation models confirm that failure of intra-reservoir sealing faults can occur during a reservoir depressurization via a water-drive mechanism, but contrary to anecdotal reports, published examples of production-induced seal failure are elusive. Ignoring the three-dimensional structure of fault zones can sometimes have a significant influence on production-related flow, and a series of models illustrating flow associated with relay zones are discussed.     

Hydraulic observations from a 1 year fluid production test in the 4000 m deep KTB pilot borehole

A long‐term pump test was conducted in the KTB pilot borehole (KTB‐VB), located in the Oberpfalz area, Germany. It produced 22 300 m³ of formation fluid. Initially, fluid production rate was 29 l min^?1 for 4 months, but was then raised to an average of 57 l min^?1 for eight more months. The aim of this study was to examine the fluid parameters and hydraulic properties of fractured, crystalline crusts as part of the new KTB programme ‘Energy and Fluid Transport in Continental Fault Systems’. KTB‐VB has an open‐hole section from 3850 to 4000 m depth that is in hydraulic contact with a prominent continental fault system in the area, called SE2. Salinity and temperature of the fluid inside the borehole, and consequently hydrostatic pressure, changed significantly throughout the test. Influence of these quantities on variations in fluid density had to be taken into account for interpretation of the pump test. Modelling of the pressure response related to the pumping was achieved assuming the validity of linear Darcy flow and permeability to be independent of the flow rate. Following the principle ‘minimum in model dimension’, we first examined whether the pressure response can be explained by an equivalent model where rock properties around the borehole are axially symmetric. Calculations show that the observed pressure data in KTB‐VB can in fact be reproduced through such a configuration. For the period of high pumping rate (57 l min^?1) and the following recovery phase, the resulting parameters are 2.4 × 10^?13 m³ in hydraulic transmissivity and 3.7 × 10^?9 m Pa^?1 in storativity for radial distances up to 187 m, and 4.7 × 10^?14 m³ and 6.0 × 10^?9 m Pa^?1, respectively, for radial distances between 187 and 1200 m. The former pair of values mainly reflect the hydraulic properties of the fault zone SE2. For a more realistic hydraulic study on a greater scale, program FEFLOW was used. Parameter values were obtained by matching the calculated induced pressure signal to fluid‐level variations observed in the KTB main hole (KTB‐HB) located at 200 m radial distance from KTB‐VB. KTB‐HB is uncased from 9031 to 9100 m and shows indications of leakage in the casing at depths 5200–5600 m. Analysis of the pressure record and hydraulic modelling suggest the existence of a weak hydraulic communication between the two boreholes, probably at depths around the leakage. Hydraulic modelling of a major slug‐test in KTB‐HB that was run during the pumping in KTB‐VB reveals the effective transmissivity of the connected formation to be 1 to 2 orders of magnitude lower than the one determined for the SE2 fault zone.     

Hydrodynamics and membrane seal capacity

The impact of hydrodynamic groundwater movement on the capacity of seals is currently in debate. There is an extensive record of publication on seals analysis and a similar history on petroleum hydrodynamics yet little work addresses the links between the two. Understanding and quantifying the effects of hydrodynamic flow has important implications for calibrating commonly used seal capacity estimation techniques. These are often based on measurements such as shale gouge, clay smear or mercury porosimitry where membrane sealing is thought to occur. For standard membrane seal analysis, seal capacity is estimated by quantifying capillary pressure‐related measurements and calibrating them with a large observational database of hydrocarbon column heights and measured buoyancy pressures. The seal capacity estimation process has historically been adjusted to account for a number of different generic trapping geometries. We define the characteristics of these geometries from a hydrodynamics viewpoint in order to fine‐tune the seal capacity calibration process. From theoretical analyses of several simplified trapping geometries, it can be concluded that generally, the high pressure side of the seal should be used as the water pressure gradient with which to calculate buoyancy pressure. Secondly, trap geometries where hydrocarbon is reservoired on both sides of a fault are not useful for estimating across fault seal capacity.     

Transient fluid flow in and around a fault

We present the results of simple numerical experiments in which we study the evolution with time of fluid flow around and within a permeable fault embedded in a less permeable porous medium. Fluid movement is driven by an imposed vertical pressure gradient. The results show that fluid flow is controlled by two timescales: τ_f = Sl²/κ_F and τ_F = Sl²/κ_M, where S is the specific storage of the porous material, l the length of the fault, and κ_M and κ_F are the hydraulic conductivities of the porous material and the fault, respectively. Fluid flow and the associated fluid pressure field evolve through three temporal stages: an early phase [t < τ_f] during which the initial fluid pressure gradient within the fault is relaxed; a second transient stage [τ_f < t < τ_F] when fluid is rapidly expelled at one end of the fault and extracted from the surrounding rocks at the other end leading to a reduction in the pressure gradient in the intact rock; a third phase [t < τ_F] characterized by a steady‐state flow. From the numerical experiments we derive an expression for the steady‐state maximum fluid velocity in the fault and the values of the two timescales, τ_f and τ_F. A comparison indicates excellent agreement of our results with existing asymptotic solutions. For km‐scale faults, the model results suggest that steady‐state is unlikely to be reached over geological timescales. Thus, the current use of parameters such as the focusing ratio defined under the assumption of steady‐state conditions should be reconsidered.     

The influence of poorly interconnected fault zone flow paths on spring geochemistry

Thermal springs commonly occur along faults because of the enhanced vertical permeability afforded by fracture zones. Field and laboratory studies of fault zone materials document substantial heterogeneities in fracture permeabilities. Modeling and field studies of springs suggest that spatial variations in permeability strongly influence spring locations, discharge rates and temperatures. The impact of heterogeneous permeability on spring geochemistry, however, is poorly documented. We present stable isotope and water chemistry data from a series of closely spaced thermal springs associated with the Hayward Fault, California. We suggest that substantial spatial variations observed in δ¹⁸O and chloride values reflect subsurface fluid transport through a poorly connected fracture network in which mixing of subsurface waters remains limited. Our measurements provide insight into the effect of fracture zone heterogeneities on spring geochemistry, offer an additional tool to intuit the nature of tectonically induced changes in fault zone plumbing, and highlight the need to consider local variations when characterizing fracture zone fluid geochemistry from spring systems with multiple discharge sites.     

From deep to shallow fluid reservoirs: evolution of fluid sources during exhumation of the Sierra Almagrera,Betic Cordillera,Spain

The combination of structural, geochemical and palaeotopographic data proves to be an efficient tool to understand fluid transfers in the crust. This study discriminates shallow and deep fluid reservoirs on both sides of the brittle–ductile transition under an extensional regime and points out the role of major transcurrent fault activity in this palaeohydrogeological setting. Palaeofluids trapped in quartz and siderite–barite veins record the transfer of fluids and metal solute species during the Neogene exhumation of the Sierra Almagrera metamorphic belt. Ductile then brittle–ductile extensional quartz veins formed from a deep fluid reservoir, trapping metamorphic secondary brines containing low‐density volatile phases derived from the dissolution of Triassic evaporites. During exhumation, low‐salinity fluids percolated within the brittle domain, as shown by transgranular fluid inclusion planes affecting previous veins. These observations indicate the opening of the system during Serravalian to early Tortonian times and provide evidence for the penetration of surficial fluids of meteoric or basinal origin into the upper part of the brittle–ductile transition. During exhumation, synsedimentary transcurrent tectonic processes occurred from late Tortonian times onwards, while marine conditions prevailed at the Earth's surface. At depth in the brittle domain, quartz veins associated with haematite record a return to high‐salinity fluid circulation suggesting an upward transfer fed from a lower reservoir. During the Messinian, ongoing activity of the trans‐Alboran tectono‐volcanic trend led to the formation of ore deposits. Reducing fluids caused the formation of siderite and pyrite ores. The subsequent formation of galena and barite may be related to an increase of temperature. The high salinity and Cl/Br ratio of the fluids suggest another source of secondary brine derived from dissolved Messinian evaporites, as corroborated by the δ³⁴S signature of barite. These evaporites preceded the main sea‐level drop related to the peak of the salinity crisis (5.60–5.46 Ma).