Thermal convection in faulted extensional sedimentary basins: theoretical results from finite-element modeling

Thermal convection has the potential to be a significant and widespread mechanism of fluid flow, mass transport, and heat transport in rift and other extensional basins. Based on numerical simulation results, large‐scale convection can occur on the scale of the basin thickness, depending on the Rayleigh number for the basin. Our analysis indicates that for syn‐rift and early post‐rift settings with a basin thickness of 5 km, thermal convection can occur for basal heat flows ranging from 80 to 150 mW m^?2, when the vertical hydraulic conductivity is on the order of 1.5 m year^?1 and lower. The convection cells have characteristic wavelengths and flow patterns depending on the thermal and hydraulic boundary conditions. Steeply dipping extensional faults can provide pathways for vertical fluid flow across large thicknesses of basin sediments and can modify the dynamics of thermal convection. The presence of faults perturbs the thermal convective flow pattern and can constrain the size and locations of convection cells. Depending on the spacing of the faults and the hydraulic properties of the faults and basin sediments, the convection cells can be spatially organized to align with adjacent faults. A fault‐bounded cell occurs when one convection cell is constrained to occupy a fault block so that the up‐flow zone converges into one fault zone and the down‐flow zone is centred on the adjacent fault. A fault‐bounded cell pair occurs when two convection cells occupy a fault block with the up‐flow zone located between the faults and the down‐flow zones centred on the adjacent faults or with the reverse pattern of flow. Fault‐bounded cells and cell pairs can be referred to collectively as fault‐bounded convective flow. The flow paths in fault‐bounded convective flow can be lengthened significantly with respect to those of convection cells unperturbed by the presence of faults. The cell pattern and sense of circulation depend on the fault spacing, sediment and fault permeabilities, lithologic heterogeneity, and the basal heat flow. The presence of fault zones also extends the range of conditions for which thermal convection can occur to basin settings with Rayleigh numbers below the critical value for large‐scale convection to occur in a basin without faults. The widespread potential for the occurrence of thermal convection suggests that it may play a role in controlling geological processes in rift basins including the acquisition and deposition of metals by basin fluids, the distribution of diagenetic processes, the temperature field and heat flow, petroleum generation and migration, and the geochemical evolution of basin fluids. Fault‐bounded cells and cell pairs can focus mass and heat transport from longer flow paths into fault zones, and their discharge zones are a particularly favourable setting for the formation of sediment‐hosted ore deposits near the sea floor.     

3D numerical modeling of hydrothermal processes during the lifetime of a deep geothermal reservoir

Understanding hydrothermal processes during production is critical to optimal geothermal reservoir management and sustainable utilization. This study addresses the hydrothermal (HT) processes in a geothermal research doublet consisting of the injection well E GrSk3/90 and production well Gt GrSk4/05 at the deep geothermal reservoir of Groß Schönebeck (north of Berlin, Germany) during geothermal power production. The reservoir is located between ?4050 to ?4250 m depth in the Lower Permian of the Northeast German Basin. Operational activities such as hydraulic stimulation, production (T = 150°C; Q = ?75 m³ h^?1; C = 265 g l^?1) and injection (T = 70°C; Q = 75 m³ h^?1; C = 265 g l^?1) change the HT conditions of the geothermal reservoir. The most significant changes affect temperature, mass concentration and pore pressure. These changes influence fluid density and viscosity as well as rock properties such as porosity, permeability, thermal conductivity and heat capacity. In addition, the geometry and hydraulic properties of hydraulically induced fractures vary during the lifetime of the reservoir. A three‐dimensional reservoir model was developed based on a structural geological model to simulate and understand the complex interaction of such processes. This model includes a full HT coupling of various petrophysical parameters. Specifically, temperature‐dependent thermal conductivity and heat capacity as well as the pressure‐, temperature‐ and mass concentration‐dependent fluid density and viscosity are considered. These parameters were determined by laboratory and field experiments. The effective pressure dependence of matrix permeability is less than 2.3% at our reservoir conditions and therefore can be neglected. The results of a three‐dimensional thermohaline finite‐element simulation of the life cycle performance of this geothermal well doublet indicate the beginning of thermal breakthrough after 3.6 years of utilization. This result is crucial for optimizing reservoir management. Geofluids (2010) 10 , 406–421     

Deep geothermal groundwater flow in the Seferihisar–Balçova area,Turkey: results from transient numerical simulations of coupled fluid flow and heat transport processes

The Seferihisar–Balçova Geothermal system (SBG) is characterized by complex temperature and hydrochemical anomalies. Previous geophysical and hydrochemical investigations suggest that hydrothermal convection in the faulted areas of the SBG and recharge flow from the Horst may be responsible for the observed patterns. A numerical model of coupled fluid flow and heat transport processes has been built in order to study the possible fluid dynamics of deep geothermal groundwater flow in the SBG. The results support the hypothesis derived from interpreted data. The simulated scenarios provide a better understanding of the geophysical conditions under which the different fluid dynamics develop. When recharge processes are weak, the convective patterns in the faults can expand to surrounding reservoir units or below the seafloor. These fault‐induced drag forces can cause natural seawater intrusion. In the Melange of the Seferihisar Horst, the regional flow is modified by buoyant‐driven flow focused in the series of vertical faults. As a result, the main groundwater divide can shift. Sealing caprocks prevent fault‐induced cells from being overwhelmed by vigorous regional flow. In this case, over‐pressured, blind geothermal reservoirs form below the caprocks. Transient results showed that the front of rising hot waters in faults is unstable: the tip of the hydrothermal plumes can split and lead to periodical temperature oscillations. This phenomenon known as Taylor–Saffman fingering has been described in mid‐ocean ridge hydrothermal systems. Our findings suggest that this type of thermal pulsing can also develop in active, faulted geothermal systems. To some extent, the role of an impervious fault core on the flow patterns has been investigated. Although it is not possible to reproduce basin‐scale transport processes, this first attempt to model deep groundwater geothermal flow in the SBG qualitatively supported the interpreted data and described the different fluid dynamics of the basin. Geofluids (2010) 10 , 388–405     

Analytical and numerical models of hydrothermal fluid flow at fault intersections

Fault intersections are the locus of hot spring activity and Carlin‐type gold mineralization within the Basin and Range, USA. Analytical and numerical solutions to Stokes equation suggest that peak fluid velocities at fault intersections increase between 20% and 47% when fracture apertures have identical widths but increase by only about 1% and 8% when aperture widths vary by a factor of 2. This suggests that fault zone intersections must have enlarged apertures. Three‐dimensional finite element models that consider intersecting 10‐ to 20‐m wide fault planes resulted in hot spring activity being preferentially located at fault zone intersections when fault zones were assigned identical permeabilities. We found that the onset of convection at the intersections of the fault zones occurred in our hydrothermal model over a narrow permeability range between 5 × 10^?13 and 7 × 10^?13 m². Relatively high vertical fluid velocities (0.3–3 m year^?1) extended away from the fault intersections for about 0.5–1.5 km. For the boundary conditions and fault plane dimensions used, peak discharge temperatures of 112°C at the water table occurred with an intermediate fault zone permeability of 5 × 10^?13 m². When fault plane permeability differed by a factor of 2 or more, the locus of hot spring activity shifted away from the intersections. However, increasing the permeability at the core of the fault plane intersection by 40% shifted the discharge back to the intersections. When aquifer units were assigned a permeability value equal to those of the fault planes, convective rolls developed that extend about 3 km laterally along the fault plane and into the adjacent aquifer.     

Heterogeneity effects on possible salinity-driven free convection in low-permeability strata

Although studies of free convection commonly focus on highly permeable strata, but numerical analyses indicate that density-driven free convection may also occur in heterogeneous low-permeability strata. Traditional Rayleigh number criteria are overly conservative in predicting thermohaline convection in these systems; so, numerical models are used to make inferences on the process. Simulations with stochastic realizations of permeability fields show that dense plumes can take preferential pathways to sink through generally low-permeability strata; patch analysis using percolation theory shows that the threshold permeability for the onset of free convection can be as low as 10⁻¹⁶ m² even with a mean permeability of 10⁻¹⁸ m². Threshold permeability for the percolation pathways decreases with increasing concentration gradient, vertical correlation length and the mean and variance of the permeability. The connectedness of relatively high-permeability zones is important in initiating and controlling plume fingers of free convection in both single-layer and sand-shale sequence models. Permeable units above and below are conducive to free convection through intervening low-permeability strata if buoyancy gradients exist. This heterogeneity is on scales that are difficult to sample by drilling and too localized to be simulated in regional models but may be significant in solute transport in these systems.     

Massive dolomitization of a Messinian reef in the Great Bahama Bank: a numerical modelling evaluation of Kohout geothermal convection

The hypothesis that Kohout thermal convection may have induced the massive dolomitization of the 60 m thick lowest more reefal unit in well Unda [top of Great Bahama Bank (GBB)] is evaluated through numerical modelling. A two‐dimensional (2‐D) section, including lithological and petrophysical data, together with datings for the sediments of the GBB, was used in the basin model TEMISPACK to reconstruct the history of the whole platform, with a focus on the reef unit. Simulations showed that during high sea‐level periods, Kohout convection is a valid mechanism in the settings of the GBB, although the convection cell remains flat in most cases because of high permeability anisotropy. This mechanism induces rapid fluid flow in the superficial as well as in the deeper parts of the platform, with velocities of at least two orders of magnitude higher than with compaction alone. Lithology appears as a strong control of fluid circulations at the margin scale through the permeability anisotropy, for which a critical value lies between values of 10 and 100. The reefal unit in Unda is part of a larger area determined by the lithologic distribution, in which flow velocities are significantly higher than in the rest of the platform. These velocities are high enough to bring the magnesium necessary to precipitate the observed amounts of dolomite, within durations in agreement with the available time of post‐reef deposition high sea level(s). However, neither fluid flow pattern nor flow velocities are able to explain the preferential massive dolomitization of the lower reef unit and the complete absence of dolomite in the upper one.     

Hydrothermal convection in and around mineralized fault zones: insights from two- and three-dimensional numerical modeling applied to the Ashanti belt, Ghana

Among hydrogeological processes, free convection in faults has been cited as a possible cause of gold mineralization along major fault zones. Here, we investigate the effects of free convection to determine whether it can cause giant orogenic gold deposits and their regular spatial distribution along major fault/shear zones. The approach comprises: (i) coupled two- and three-dimensional numerical heat- and fluid-flow simulations of simplified geological models; and (ii) calculation of the rock alteration index (RAI) to delineate regions where precipitation/dissolution can occur. Then, comparing the deduced alteration patterns with temperature distribution, potential areas of gold mineralization, defined by T  > 200°C and RAI < 0, are predicted. The models are based on the orogenic Paleoproterozoic ore deposits of the Ashanti belt in western Africa. These deposits occur in the most permeable parts of the fault zone, where the lateral permeability contrast is the highest. For a simple geometry, with a fault zone adjacent to a sedimentary basin half as permeable, we note a transition from three-dimensional circulation within the fault to a two-dimensional convective pattern in the basin far from the fault. Moreover, whereas two-dimensional undulated isotherms dominate in the basin, three-dimensional corrugated isotherms result from the preferred convective pattern within the fault, thus enhancing a periodic distribution of thermal highs and lows. In our most elaborate three-dimensional model with an imposed lateral permeability gradient, the RAI distribution indicates that fluid circulation in fault zones gives rise to a spatial periodicity of alteration patterns consistent with field data.     

The interplay of permeability and fluid properties as a first order control of heat transport, venting temperatures and venting salinities at mid-ocean ridge hydrothermal systems

While the fundamental influence of fluid properties on venting temperatures in mid-ocean ridge (MOR) hydrothermal systems is now well established, the potential interplay of fluid properties with permeability in controlling heat transfer, venting temperatures, and venting salinities has so far received little attention. A series of numerical simulations of fully transient fluid flow in a generic, across-axis model of a MOR with a heat input equivalent to magmatic supply at a spreading rate of 10 cm year⁻¹ shows a strong dependence of venting temperature and salinity on the system's permeability. At high permeability, venting temperatures are low because fluid fluxes are so high that the basal conductive heating cannot warm the large fluid masses rapidly enough. The highest venting temperature around 400°C as well as sub-seafloor fluid phase separation occur when the permeability is just high enough that the fluid flux can still accommodate all heat input for advection, or for lower permeabilities where advection is no longer capable to transfer all incoming magmatic heat. In the latter case, additional mechanisms such as eruptions of basaltic magma may become relevant in balancing heat flow in MOR settings. The results can quantitatively be explained by the 'fluxibility' hypothesis of Jupp & Schultz (Nature, 403 , 2000, 880), which is used to derive diagrams for the relations between basal heat input, permeability and venting temperatures. Its predictive capabilities are tested against additional simulations. Potential implications of this work are that permeability in high-temperature MOR hydrothermal systems may be lower than previously thought and that low-temperature systems at high permeability may be an efficient way of removing heat at MOR.     

Linked thermal convection of the basement and basinal fluids in formation of the unconformity‐related uranium deposits in the Athabasca Basin,Saskatchewan, Canada

World‐class unconformity‐related U deposits in the Athabasca Basin (Saskatchewan, Canada) are generally located within or near fault zones that intersect the unconformity between the Athabasca Group sedimentary basin rocks and underlying metamorphic basement rocks. Two distinct subtypes of unconformity‐related uranium deposits have been identified: those hosted primarily in the Athabasca Group sandstones (sediment‐hosted) and those hosted primarily in the underlying basement rocks (basement‐hosted). Although significant research on these deposits has been carried out, certain aspects of their formation are still under discussion, one of the main issues being the fluid flow mechanisms responsible for uranium mineralization. The intriguing feature of this problem is that sediment‐hosted and basement‐hosted deposits are characterized by oppositely directed vectors of fluid flow via associated fault zones. Sediment‐hosted deposits formed via upward flow of basement fluids, basement‐hosted deposits via downward flow of basinal fluids. We have hypothesized that such flow patterns are indicative of the fluid flow self‐organization in fault‐bounded thermal convection (Transport in Porous Media, 110, 2015, 25). To explore this hypothesis, we constructed a simplified hydrogeologic model with fault‐bounded thermal convection of fluids in the faulted basement linked with fluid circulation in the overlying fault‐free sandstone horizon. Based on this model, a series of numerical experiments was carried out to simulate the hypothesized fluid flow patterns. The results obtained are in reasonable agreement with the concept of fault‐bounded convection cells as an explanation of focused upflow and downflow across the basement/sandstone unconformity. We then discuss application of the model to another debated problem, the uranium source for the ore‐forming basinal brines.     

Computing geochemical mass transfer and water/rock ratios in submarine hydrothermal systems: implications for estimating the vigour of convection

A method of calculating chemical water/rock ratios is presented that enables the estimation of fluid velocities in open, flow‐through hydrologic systems. The approach is based on relating the gain/loss of a chemical species per kilogram of solid phase to the loss/gain of that species in the fluid phase, integrated across a specified length of the flowpath. After examining the underlying approximations of the approach using a one‐dimensional model of seawater moving through a basalt under nonisothermal conditions, the method is applied to representative zones within a two‐dimensional hydrothermal convective system. The method requires that regions within the flow system can be identified in which the direction of flow is steady for an extended period of time. Estimates of fluid velocity are spatial and temporal averages for the length of the flowpath used in the calculation. The location within the flow system and the nature of the alteration reactions determine which species can provide reliable values of the chemical water/rock ratio and useful estimates of fluid velocities. Over the length of the flowpath considered, the calculation of water/rock ratios works best when a species is controlled by a single reaction. Accurate estimates are favoured if the concentration profile of a species along the flowpath increases or decreases monotonically. If the length of the flowpath extends over more than one reaction zone, then erroneous estimates of the water/rock ratio and fluid velocity are more likely. Model calculations suggest that the quartz/silica system should provide reliable estimates for fluid velocity under a wide range of temperature and flow conditions, in particular in those regions of a system at or near quartz equilibrium, so that the aqueous silica concentration is buffered by quartz and correlated with the temperature distribution.