TY - JOUR
AU - Rybach,, L
AB - Abstract Increased interest in geothermal energy has led to the need for more sophisticated analyses of available geothermal resources for heating and power production. The evaluation methods have traditionally used 1D models, but progress in computing power allows us now to perform joint interpretations that include large-scale effects of geological, topographical and hydrogeological structures. This paper describes the steps taken in such an integrated assessment. The evaluation of geothermal potential is actually finalized for one of the most populated areas of Switzerland. The methodology accounts for individual utilization scenarios and is based on various temperature data that have been systematically collected over many years. The state-of-the-art compilation involves comprehensive 3D regional geological and thermal models. An excellent data fit with >20 boreholes in northern Switzerland is achieved using the calculation scheme presented here. Zones of significant convective flow are identified and the flow velocity is quantified. In contrast to the results obtained from earlier geothermal resource assessments, the existing geological, hydrogeological and petrophysical data are included in the full 3D numerical evaluation. The results obtained for a regional scale of a well-documented subsurface area in northern Switzerland are displayed in terms of geothermal productivity and energy. The calculations identify the topmost crystalline basement as a most promising structure for geothermal exploitation with predicted maximum doublet productivities of >100 MWt. The annually extractable energy of 13 PJ km-3 easily covers a broad range of Swiss energy needs. The atlas can be further expanded to cover utilization schemes not treated here, such as ground-coupled heat pump systems. geothermal resources, numerical modelling, temperature data, geological visualization 1. Introduction Geothermal energy is becoming an attractive alternative for many conventional fuel-based energy utilizations. In a time when political discussion favours regenerative energies for future energy scenarios, geothermal energy may become a vital option since it offers the advantage of providing base load. In parallel, the development of technologies and exploitation of geothermal reservoirs became more significant in the last decade. Current review studies provide evidence for large economical competitiveness of both low-enthalpy utilization (‘direct use’) for heating and high-enthalpy systems for cogeneration (combined heat and power (CHP)) or pure power generation. The installed worldwide capacity on direct use has more than tripled in the period between 1995 and 2005, attaining >28 GWt (with >250 PJ a-1) in 2005 (Lund et al2005). High-enthalpy systems develop especially in countries next to the Pacific Rim. In 2005, the installed worldwide electric capacity was 8900 MWe producing 57 TWh a-1 (Bertani 2005). In recent years, efforts to tap into high-enthalpy resources in low-temperature areas have been significant using the enhanced geothermal systems (EGS) approach, which has developed on the basis of the hot dry rock (HDR) concept. There is currently no geothermal power generation in Switzerland, but this technology gets much public attention. New perspectives are important for local authorities to evaluate the potential for future energy scenarios. In Central Europe, several resource evaluations have been performed. For Switzerland, the possible extractable geothermal power was evaluated in the past to be >6 GWt from aquifers (namely the Upper Marine Molasse, Cretaceous Malm layer and Buntsandstein) and >10 GWt from crystalline basement in the area of the Swiss Molasse Basin (Rybach 1982). Jung et al (2002) concluded in a recent study on the possibilities of geothermal electricity production that 95% of the German resources are related to crystalline rock, 4% to fault zones and only 1% to aquifer systems. It was estimated that a total of 1200 EJ (∼300 000 TWh) could be made available, corresponding to 600 times the present annual electricity demand. The goal of the present geothermal resource analysis is to provide useful information for future project planning for public funding and industrial activities. The study includes key features of the future Swiss geothermal resource atlas with a detailed subsurface model of possible geothermal resources in northern Switzerland. For the investigation of potential utilization sites, a new procedure has been developed on the basis of available data from geological structures, temperature logging, petrophysical evaluations and hydraulic testing. Therewith, it expands earlier potential assessments that focused mostly on a purely thermal data interpretation. Our approach is mainly restricted to the evaluation of lithological settings and physical conditions and is therefore complementary to other technical assessments (Sanyal and Butler 2004). 2. Geothermal resource classification 2.1. Definitions In order to classify geothermal energy, the different definitions of resources will shortly be outlined. In general, the total available subsurface geothermal energy is estimated using a volumetric method (i.e., Muffler and Cataldi (1978)): with EHIP being the thermal energy stored in the subsurface, generally referred to as ‘heat in place’, ρcP the heat capacity of rock, V the volume, Tprod the temperature of the produced fluid and Treinj the temperature of the re-injected fluid. Since this evaluation is only related to the heat content of the subsurface, the energy can be enormous even for low temperature resources. For instance, cooling a 1 km3 volume of rock at 5 km depth by 100 °C releases a thermal energy of 230 PJ (or 64 000 GWh). This would correspond to 0.2% of the total Swiss energy needs (Swiss Federal Statistical Office 2003). However, EHIP has only little economic significance since the energy needs to be extracted by a transport medium (generally water) that is not universally present. Convection represents a highly efficient heat transfer mechanism in a reservoir. Therefore, permeable sedimentary or crystalline reservoir structures are the most favourable sites (e.g., Sanyal and Butler (2004)). Besides the fact that high permeability has a favourable impact on the operation conditions (low reinjection pressure and high flow rate), the higher thermal efficiency represents an additional advantage of convective systems. But even under optimum conditions, only a fraction of the stored energy can be produced. The recovery factor, R, expresses this difference by with Eg being the thermal energy produced or the utilizable heat. Muffler and Cataldi (1978) have highlighted the effect of porosity for liquid systems and additionally of temperature of steam systems on the recovery factor. Lavigne (1978) defined this factor on the basis of 1D and 2D models. With similar considerations, Jung et al (2002) obtained values for R between 5.6% and 7.3% for pure conductive situations. In a review of existing production data, Williams (2004) has estimated values between 8% and 21% for operating highly convective systems in the USA. In an earlier assessment of the Swiss potential, Eg was estimated to be 5 PJ annually for low-enthalpy systems and ∼500 PJ for high enthalpy (Rybach 1992a). If the produced thermal energy should be converted into electric energy, the thermodynamic conversion ideally represented by the Carnot cycle needs to be accounted for: with Ee being the geothermally generated net electric energy and η the temperature-dependent efficiency factor. The efficiency factor describes the net electric energy gain from hot liquid or steam produced, and does not account for the energy needed for operation. For low-temperature resources, organic Rankine cycles (ORC) are now standard technology, but new energy conversion schemes (Kalina, hybridization) are likely to be improved and widespread in future. Conversion is currently possible at T > 100 °C, with efficiencies increasing at higher temperatures. At the lower temperature limit, η is in the range between 3% (Hudson 1998) and 9% (Jung et al2002) reflecting local site conditions and technological developments. Conversion is expected to improve up to 13% at higher temperatures (T = 250 °C). The investigations of Seitz (2004) and Köhler (2005) demonstrate the effect of utilization temperature and careful selection of the ORC working fluid on system efficiencies. The generally low efficiency level indicates that the ‘waste heat’ from the geothermal conversion is ideal for CHP utilization and cascading use (Rybach and Kohl 2004). 2.2. Resource classes The breakdown of geothermal resources should comply with various criteria (McKelvey 1972). Current technology as well as possible future developments need to be considered. Here, a scheme is proposed that accounts for resources available today and in the future. In the simplest case, the sum of all resource classes would correspond to the total utilizable heat in an area. However, this notion suggests implicitly that a future resource could be exploited with a recovery factor that is predicted today. But in reality, the recovery factor will vary as a function of the assumed extraction scenario. Therefore, the scheme presented herein does not allow explicitly summing up all energies from the different resource classes. The selection criteria for the definition of a class are subsurface parameters such as temperature, hydrogeological parameters and technological considerations such as conversion possibilities. Under low-temperature conditions the individual geothermal resource classes should differ from those under high-temperature conditions (Sanyal 2005). Even if not all classes are treated in this paper, we propose the following classification: Class 1—near-surface resource. The present geothermal utilization in Switzerland—as published in the biennial geothermal statistical census (Signorelli et al2004b)—strongly reflects the population density and covers therefore only a small part of the available near-surface resources. Most near-surface utilization originates from ground-coupled heat pump (GCHP) systems that are presently limited to about 200 m depth. In 2004, the Swiss utilization for direct use amounted annually to 70 GJ m-2. Although this category is not covered by the present atlas, it may be included in future extensions. Class 2—low-temperature, hydrothermal resources. These are categorized here as deep aquifer systems that can be used for heat production applying today's standard technology. For the exploitation of deep aquifers, minor stimulation efforts can be assumed resulting in an increase of ambient transmissivity by a factor of 2. Similarly to class 1, the actual utilization is often improved by heat pump systems. Typical criteria are hydrogeologically favourable conditions, temperature (below the economical level of power generation, T < 100 °C) and depth (minimum 200 m to separate this class from GCHP or groundwater heat pump systems). Once the results of a finalized resource atlas for Switzerland are available, the well-documented present Swiss utilization (Signorelli et al2004b) can be compared to the available potential. Using surface GIS data (population density, district heating systems, etc), appropriate locations for the development could be proposed and more closely evaluated. Class 3—hydrothermal resources requiring major technology use. This class includes deep geothermal systems that require high-cost exploitation technologies. Under current economic considerations, electricity production by EGS may be most favourable together with continuing cascade use to reduce the geothermal waste heat. The extensive application of stimulation techniques (hydraulic fracturing or acidization) is certainly necessary to create economic reservoirs. We assume a transmissivity increase by a factor of 10 for crystalline rock and 2.5 for sedimentary rock. Minimum reservoir conditions are met by (isotropic) permeabilities of 10-15 m2 and temperatures of 100 °C that represent current limitations set by the ORC technology. The maximum depth is ∼5 km, set by economic consideration of today's drilling costs. Since we focus on the evaluation of this resource class, the necessary simplifications of this regional assessment are worth mentioning. As in all other classes, further local investigations are required to elaborate on the suitability of a specific resource. Class 4—resources for predictable technology development. Like class 3, this class includes high-enthalpy systems that are, however, not yet economically viable. The electricity production in these areas will become economical when major technological progress takes place. Future concepts should include development of cheaper drilling and exploration technology, hybridization of geothermal energy (i.e. with biomass), application of Kalina cycles, stimulation improvement etc. For this class, we expect transmissivity enhancements by stimulation techniques by factors of 10 and 100 for sediments and crystalline rock, respectively. This would allow hot fluids to be produced at poor hydrogeological conditions (permeability > 10-16 m2). Also lower temperature areas (T > 90 °C) and drilling depths down to 7 km should be included in this class. Class 5—long-term exploitable resource. These are areas that cannot be exploited with any predictable technological progress. This would include in particular deep-seated reservoirs, deeper than 7 km and low permeability zones (k < 10-16 m2). The focus of the present investigation is on classes 2 and 3. 3. Procedure 3.1. General The resources will be defined in terms of three key factors: hydrogeology, geothermal conditions and system layout (i.e., doublet systems). The subsurface structure is derived from a 3D geological model. GIS data from the current surface utilization (distribution of industrial and living complexes) can be additionally included. In this paper we describe the results of a survey in northern Switzerland; other areas of the Swiss Plateau will be investigated in the coming years. Of note is that our approach allows a 3D characterization of subsurface structures and evaluates the thermal field by full 3D numerical finite element (FE) methods, based on a calibrated temperature model. The general procedure of the resource assessment is as follows: Elaboration of a regional 3D geological model. Assessment of hydraulic data from potential aquifer systems. Spatial discretization of the geological model into a numerical FE scheme. Interpretation of thermal borehole data and development of a regional thermal calibration model. Assessment of the technically available resource potential, Eg, based on hydraulic, thermal and geometrical considerations. Combination of results for specific geological units with current surface utilization. Kohl et al (2003) have described the basic methodology of integrating geological data in detail. Here, the procedure is outlined briefly. The subsurface in northern Switzerland has been systematically explored for nuclear waste disposal studies. Many seismic sections, geological interpretations and borehole data are available and well documented (see figure 1). As a general trend, the crystalline basement dips to southeast from the Black Forest in the north (where it reaches the surface) towards the Alpine flexure zones. It is intersected by numerous fault systems of various strike angles and by Permo-Carboniferous troughs (PCTs). A dominant PCT, hidden below Mesozoic sediments, was discovered during the nuclear waste research exploration. In the study area, the crystalline rock and the sedimentary Mesozoic layers outcrop in a NW–SE sequence: the granitic basement can be found further to the NW and the younger Tertiary molasse to the SE. The individual structures are identified from a total of 37 seismic profiles and 63 boreholes. Moreover, among the 33 boreholes identifying the Upper Muschelkalk layers, only 13 reach the crystalline basement. Clearly, with increasing depth the information becomes scarce. We will concentrate more specifically on well-known aquifer systems and the crystalline basement, i.e. on resource classes 2 and 3. Surface geology and the locations of boreholes and seismic profiles can be seen in figure 1 Figure 1. Open in new tabDownload slide Investigation area of northern Switzerland with seismic profiles (lines) and boreholes. Boreholes with only geological information (dots) and with additional temperature data (squares) are shown. The large red rectangles identify the boundaries of the three 3D models. Figure 1. Open in new tabDownload slide Investigation area of northern Switzerland with seismic profiles (lines) and boreholes. Boreholes with only geological information (dots) and with additional temperature data (squares) are shown. The large red rectangles identify the boundaries of the three 3D models. Gocad software is used to build a 3D structural model of the geological units (Mallet 1992). The flexible modelling tools and the mathematical interpolation capabilities of the code especially were consistent with the pre-defined project goals (Andenmatten Berthoud and Kohl 2003). With this tool, a first spatial discretization of the geological model into a FE numerical scheme could be conducted. The present procedure uses tetrahedral elements that can exactly reproduce fault zones or geological interfaces. The discretization of the geological model is the key to the interpretation of subsurface temperature fields. Figure 1 also illustrates the boundaries of the three models that have been extracted from the regional geological model. They extend over an area of 40 × 60 km (model I) and 50 × 56 km (models II and III). The vertical extension is down to 10 000 m below sea level. As an example, the geological structure of model I and the tetrahedron discretization are shown in figure 2. Figure 2. Open in new tabDownload slide Geological structures of the Gocad model (left) and tetrahedral element mesh (right). Example for area of model I. Figure 2. Open in new tabDownload slide Geological structures of the Gocad model (left) and tetrahedral element mesh (right). Example for area of model I. In order to define the subsurface resources in terms of productivity, the porous media concept of Gringarten (1978) has been adopted, permitting the estimation of flow production for geothermal doublet systems. Although numerous geothermal systems in Switzerland are exploited by single wells, we use doublet systems for our analysis. They fulfil the present utilization concepts (maintaining water balance, less chemical precipitation), albeit at higher cost (drilling, pumping power for re-injection). By the application of the Gringarten solution, two key parameters, hydraulic transmissivity and geometrical system layout, are included in the present analysis. Using the specific formulations of Gringarten (1978), fluid flow rate can be calculated in an iterative process as with Tr being the aquifer's transmissivity, Δh the maximum variation of hydraulic head in the borehole, Δt the time period with no sensitive temperature decline (i.e., 30 yr throughout this paper), c a factor expressing the ratio of fluid and aquifer heat capacity, Δz the reservoir height, rw the borehole radius (i.e. 0.1 m throughout this paper) and Qi the flow rate at the ith iteration loop. Furthermore, the distance between the injection and production boreholes, and therewith the reservoir geometry, is also estimated. From the hydraulic productivity, a geothermal productivity can be calculated using the temperature difference, ΔT, between the produced (ambient) and the re-injected fluid temperature: The geothermal energy produced over a time span, Δt, is then determined as with (ρcp)f being the specific heat of the fluid and Q the extracted flow rate (calculated from equation (4)). From this, the recovery factor, R, is calculated following equation (2). Gringarten (1978) highlighted that this factor depends mostly on the geometrical distance to neighbouring doublets. In this paper, we assume a rectangular geometry for doublet systems defined by a factor 2 in length and a factor 1 in width of the distance between the injection and production. Since equation (4) provides a solution for productivity even at very low transmissivities, an assumption about the system geometry is introduced. It is assumed that a distance of 200 m between fluid production and reinjection has to be exceeded to allow geothermal production. Below this value, Eg (and therewith also R) is set automatically to zero. The resource assessment is based on hydraulic transmissivity and temperature. By applying equations (4)–(6), the results of the hydrogeologic and the thermal model are merged into a single productivity value at every location of the 3D area. However, the spatial representation of the model parameters is not necessarily analogous, i.e. the thermal conductivity distribution does not need to be identical to the transmissivity pattern. Thus, based on the same geological assessment, different approaches were chosen to accommodate relevant hydrogeological or thermal information. The individual assumptions are described below. A new resource evaluation code GeoProduct has been written that allows us to combine to the multiple tasks of this evaluation. These are, for example, defining input routines into the Gocad model, checking spatial discretization on consistency, writing the FE input scheme, performing productivity calculations and superimposing GIS data for the visualization of identical graphical programs. GeoProduct is a comprehensive collection of sub-programs that use a uniform database and ensure the consistency of the various input parameters for a given model. Its analytical capabilities could be further extended to project cost assessments for drilling and operational conditions. 3.2. Hydrogeological settings of northern Switzerland The major structures of geothermal interest and the available hydrogeological data have been described in detail by Signorelli et al (2004a) and will be only summarized herein. The following hydrogeological units have been treated specifically: the crystalline basement, altered upper crystalline zone (including Buntsandstein layers), Permo-carboniferous trough (PCT) and the well-known Upper Muschelkalk, Upper Malm and Upper Marine Molasse aquifers. In the geological model, these units are exactly positioned with an orientation and a height reconstructed from the top and the bottom parts of the adjacent structures. In addition, several well-known fault zones have been included in the geological model. They are major NW-SE striking Hercynian faults originating in the southern Black Forest and Tabular Jura: the Eggberg Fault, Vorwald Fault, Zurzach Fault, Neuhausen Fault and Randen Fault. The southern and northern border faults of the PCT are also included. They can be observed at the surface as the Rafz-Marthalen flexure zone and the Baden-Irchel-Herdern Lineament (Müller et al2001). This model does not extend to areas with faults associated with the Rhine Graben. The remaining areas are also treated but they do not play an important role for hydrothermal systems due to their low hydrogeological relevance. The thrusted molasse located in the SE of model III (figure 1) is not accounted for specifically. However, future findings of currently unknown structures are easily integrated into this scheme. Hydrogeological data from the boreholes used are scarce. For instance, the Upper Malm aquifer is described by only two transmissivity values. Transmissivity and hydrogeological characteristics for potential aquifers from Signorelli et al (2004a) are summarized below: The topmost 500 m of the granitic basement are generally strongly fractured and altered immediately beneath the sedimentary cover with maximum hydraulic conductivity values in the range of 10-6–10-7 m s-1. The Buntsandstein is a thin 5–25 m layer that has nearly identical hydrogeological parameters as the adjacent crystalline rock. The Upper Muschelkalk extends between 40 and 120 m depth with a few sections of extremely high hydraulic conductivity (10-4 m s-1). The Upper Malm is less suitable for geothermal production with some maximum values of 10-6 m s-1. Its thickness increases towards the east from 160 to 360 m. Note, this aquifer is not identical to the well-karstified Lower Malm that has favourable conditions in Bavaria (Germany). The Upper Marine Molasse (OMM) is used presently for geothermal heating projects. It has good hydrogeological characteristics up to K = 10-4 m s-1. The thickness of the OMM increases towards the Alpine front to 1000 m. 3.3. Thermal modelling and temperature uncertainty Numerical modelling represents the basis for the spatial temperature simulation. From the above-mentioned regional models I–III, the temperature at a regional scale at any location and depth of the investigation area is forecast. The numerical model is a finite element (FE) tetrahedron model (figure 2) in which selected structures are exactly integrated from the geological model. Although it might be advantageous to account for most geological structures, the CPU-intensive tetrahedron models had to compromise on the representation of layers at the scale considered here. The model fitting procedure (see section 4) could be performed with a maximum of 200 000 nodes. Since the focus of the investigations is on the deeper structures, the crystalline basement in particular was reproduced. Important criteria in the selection of the geological structures to be discretized were the possible variation of thermal conductivity and the complexity of layering that could possibly cause difficulties for the calculations of productivity models. From first assessments, a simplification into the following five layers has been decided: (1) deeper crystalline, (2) topmost 500 m thick crystalline, (3) PCT, (4) Jura Mountains and (5) Mesozoic/Tertiary/Quaternary. With the exception of the Jura Mountains, which have a complex structure, the sedimentary layers could be strongly merged since the thermal conductivity of the major layers does not vary strongly. A more refined treatment of the geological units using a tetrahedron model was in the past applied at a local scale, extending only a few kilometres around a borehole (Andenmatten and Kohl 2003). As the simplest starting point, only diffusive mechanisms that include effects from topography, thermal refraction, heat production or transient were considered in the full 3D simulations. By this, any data treatment or corrections become obsolete. In particular, effects from thermal refraction (Kohl and Rybach 1996), transient diffusion or topographic influence (Kohl et al2001) are already treated. The following basic diffusive thermal transport equation is assumed: with t being the time, λ the thermal conductivity and A the heat production. For the purpose of our temperature analysis, transient effects are treated sufficiently by accounting only for the effect of the last Ice Age. Earlier studies have identified this climatic effect as most prominent. In the subsequent thermal model, we characterize the effect of the Ice Age by a 10 K cooler surface temperature occurring in a period ranging between 100 000 and 10 000 years ago. A small variation of the assumed temperature step shows only little sensitivity to the greater temperature field. Disregarding this effect had in the past created, however, wrong extrapolations of the deeper thermal field in Central Europe (Clauser et al1997, Rybach 1992b). The thermal modelling had to also take into account hydraulic driven temperature signals. The discrimination of diffusive from advective heat transfer represented a major task of the thermal analysis. Advective heat transfer is expected to be a major origin for a misfit of the regional diffusive model. If borehole data apparently exhibit an advective component, the following treatment is performed. Considering a simple 1D case with vertical Darcy velocity, vD, the heat flow effect can be described by the Péclet analysis (e.g. Kohl et al (2003)) as where with j0 being the surface heat flow, jb the basal heat flow, λ the thermal conductivity of the rock matrix, (ρcP)f the heat capacity of fluid and Δz the vertical depth section. The same effect could be produced by an equivalent heat production, A, with and A can then be expressed after reformulating the upper two equations as function of a vertical flow velocity 1D test models easily show the validity of these considerations when comparing diffusive–advective heat transfer without heat production to pure diffusive heat transfer and equivalent heat production. Parallel to the Péclet analysis, this methodology only provides a rough estimation of the Darcy flow velocity for 3D assumptions. This method is advantageous when information on the hydraulic pressure or on transmissivity is missing for a borehole, as it is often the case for archived borehole data. Since the diffusive heat flow is defined from both temperature gradient and thermal conductivity, the uncertainty in the determination of thermal conductivity directly affects the modelled heat flow distribution. This one can be regarded as characteristic for the investigation area, but is of secondary importance to the specific goal that requires only the determination of the temperature field. However, the uncertainty of the base temperature data needs to be quantified since various data of different qualities are used. In northern Switzerland, temperature data from >50 boreholes are available with highly variable depth and measurement quality. In our procedure, the uncertainty of these data is mapped using a linear approach, similarly to evaluations of seismic hazard. Herein, the uncertainty analysis is restricted only to temperature data, even though more complex parameters could be taken, such as the steepness of geological layers or the fault zone density. For each borehole (figure 3), the classification of quality and depth was taken from Schärli and Kohl (2002) and binned into three classes: poor, good and very good. This attribute covers density, existence and quality of the temperature data. Figure 3. Open in new tabDownload slide Map of boreholes with available temperature data. The circles scale with data quality. The axes indicate the Swiss coordinate system (also on subsequent figures). Figure 3. Open in new tabDownload slide Map of boreholes with available temperature data. The circles scale with data quality. The axes indicate the Swiss coordinate system (also on subsequent figures). The areal distribution assumes that the quality of the temperature data at a selected depth interval can be extrapolated using a linear weight function. Typically, the calculated quality at any distance, Qi, decreases from the quality at the borehole, Q0, with increasing lateral distance, r, from a borehole until a cut-off radius, rmax, is reached. Given the temperature data density of the investigation area, a value of rmax = 20 km around a borehole seemed to be appropriate. At larger distances the information from the borehole is ignored. As a function of distance, the original quality value of each borehole drops to a local value by a maximum of one class (equation (11)). The analysis is computed for a quadratic grid size of 5 × 5 km2. The calculated value at a local grid cell corresponds to the maximum impact from a borehole and is not the result of a superposition from the impact of all boreholes. Thus, they cannot exceed the highest contribution from a single borehole. Currently, a linear weighting function is chosen, but the analyses could be extended to anisotropic, nonlinear cases that would also integrate the geologic impact. The present procedure allows us to illustrate the available data density at different depths. Two examples, z = 250 m and z = 1000 m, are shown in figure 4. The colour scaling ranges from dark grey (very good quality) to white (no information at all). In parallel to the data density, the quality distribution also reduces with depth. This is also illustrated by the fact that 40 out of 180 available boreholes in Switzerland reach a depth of 2000 m, with only three of them having sufficiently good data quality. This highlights, however, the need for a sophisticated data analysis. Figure 4. Open in new tabDownload slide Quality map at a depth of 250 m (top) and 1000 m (bottom). 3: very good, 2: good, 1: poor quality. Figure 4. Open in new tabDownload slide Quality map at a depth of 250 m (top) and 1000 m (bottom). 3: very good, 2: good, 1: poor quality. 4. Thermal modelling Thermal modelling evaluates the subsurface temperature as a key parameter for the resource assessment. Therefore, an analysis of existing borehole temperature measurements is performed to derive a 3D thermal ‘calibration model’. The thermal data basis for our simulation is taken from Schärli and Kohl (2002). The temperature compilation includes bottom hole temperature (BHT) values and high-resolution temperature (HRT) logging data. Two boreholes are deeper than 2 km and eight boreholes are deeper than 1 km in the investigation area. The temperature data generally show elevated temperature gradients with values between 30 and 40 K km-1. Near the surface, even higher values can be reached. The petrophysical compilation includes parameters such as thermal conductivity (partly anisotropic), porosity, density, heat capacity and heat production. During the simulation procedure, a model fit of all measured borehole temperature profiles is intended by combining petrophysical data as averaged values for each geological layer. In the first step of an iterative procedure, thermal parameters and boundary conditions were derived from temperature fitting at specific borehole locations that were considered as a representative for the local area. Next, the parameters and boundary conditions from all locations were compared and homogenized—with the exception of crystalline rock—for the total area of investigation. Maximum deviations of 10–20% for thermal conductivity from the averaged measurements were allowed. Especially for PCT and Mesozoic sediments this lateral homogenization is a necessary simplification since petrophysical measurements are scarce. In contrast, the petrophysical data from Schärli and Kohl (2002) exhibit a strong variation in crystalline rock properties. During the modelling, it became impossible to force the crystalline basement into a single homogeneous parameter range. Instead, existing thermal conductivity data from crystalline rock have been used for a regional lateral interpolation. This permitted us to account for possible internal composition of the rock. The fitting of a regional basal heat flow for a heterogeneous crystalline rock provided clearly improved adjustments of the temperature data. During the manual fitting procedure, a successful data interpretation is achieved when the selected distribution of thermal conductivity values of all units and the selected basal heat flow boundary condition could explain the temperature data. At the surface, an altitude-dependent surface temperature distribution with a lapse rate of 4.5 K km-1 was chosen. The following figures illustrate the procedure in the example of model I (models II and III were treated similarly). In figure 5, the interpolated distribution of the thermal conductivity in crystalline rock with the individual boreholes in model I is shown. Care has to be taken when evaluating this map, since the distribution is calculated only from five locations. No measurements are available for the southeastern part: here the mean value 3.1 W m-1 K-1 is used. As mentioned, crystalline rock is represented by the range of laboratory measurements (2.7–3.55 W m-1 K-1). The other material sets assume a uniform thermal conductivity of 2.9, and 1.7 W m-1 K-1 for the sedimentary cover and PCT, respectively. Heat production is kept uniform with values of 0.5, 2.5 and 3.6 µW m-3 attributed to the sedimentary cover, PCT and crystalline rock, respectively. Figure 5. Open in new tabDownload slide Thermal conductivity distribution in crystalline basement for model I. Borehole locations with thermal conductivity measurements are shown by black circles, other boreholes (squares) either do not have petrophysical data or do not penetrate into the crystalline rock. Figure 5. Open in new tabDownload slide Thermal conductivity distribution in crystalline basement for model I. Borehole locations with thermal conductivity measurements are shown by black circles, other boreholes (squares) either do not have petrophysical data or do not penetrate into the crystalline rock. Both the resulting heat flow and thermal conductivity distribution were finally accomplished in a thermal model, which is calibrated on the existing temperature data. The large-scale simulations provided good fits of the temperature data of most boreholes. As a confirmation of our procedure the local model of the Weiach borehole (shown in detail in figure 6) can be well explained by diffusive thermal transport in the top 2 km. Deeper in the crystalline basement, a possible advective influence becomes significant. Assuming vertical flow in the rock matrix, Kohl et al (2003) have evaluated a flow velocity of vD = 5 × 10-10 m s-1. Given the rough distribution of geological structures, the quality of the fit is very satisfying, for both the temperature and the temperature gradient. Figure 6. Open in new tabDownload slide Temperature fit of the borehole Weiach that penetrates the crystalline basement (dots: measurements, curve: modelling results). Temperature is shown on the left and the temperature gradient to the right. The abbreviations identify the local geological setting (k: crystalline, c: Carboniferous, p: Permian, s: Buntsandstein, mo: Upper Muschelkalk, ju: Lias, jm: Dogger, jo: Malm, OMM: Upper Marine Molasse). Figure 6. Open in new tabDownload slide Temperature fit of the borehole Weiach that penetrates the crystalline basement (dots: measurements, curve: modelling results). Temperature is shown on the left and the temperature gradient to the right. The abbreviations identify the local geological setting (k: crystalline, c: Carboniferous, p: Permian, s: Buntsandstein, mo: Upper Muschelkalk, ju: Lias, jm: Dogger, jo: Malm, OMM: Upper Marine Molasse). Temperature fits using the 3D thermal calibration model for boreholes with HRT data below 200 m depth are shown in figure 7. Again, most boreholes exhibit overwhelmingly a diffusive heat transport, with the exception of Riniken where an additional advective heat transport had to be assumed: a vertical flow velocity of vD = 5 × 10-10 m s-1 is indicated by assuming the 1D vertical advection (equation (10)). The borehole Ruchfelden exhibits temperature variations in the topmost sedimentary layers that are clearly related to advection in the OMM layer. Within the framework of the present resource analysis, these variations have not been treated since they do not provide different temperature patterns at greater depth. However, they illustrate the importance of the temperature data to highlight hydrogeological characteristics. Figure 7. Open in new tabDownload slide Temperature fit of selected boreholes. With the exception of Riniken, diffusive heat transport can explain the data (for notation, refer to figure 6). Figure 7. Open in new tabDownload slide Temperature fit of selected boreholes. With the exception of Riniken, diffusive heat transport can explain the data (for notation, refer to figure 6). The temperature distribution in this calibration model seems to be strongly reliable with maximum temperature differences of 5 K only at single individual borehole sections (see figure 7). Even data that have not been included in the definition procedure of the calibration model could be well matched (note, only deep HRT data are documented in figure 7; single BHT measurements or shallow boreholes have been omitted). In contrast to the well-fitted temperature field, the heat flow value is not well constrained given the uncertainties in the thermal conductivity. The fitting resulted in a basal heat flow distribution that did not drastically vary. The calibration model suggests, however, high basal heat flow values, mainly in the area of model I. At the bottom of the model in -10 km depth, basal heat flow reaches maximum values of >100 mW m-2 next to the northern border of the PCT. The fact that near surface heat flow may even reach >140 mW m-2 highlights the exceptional values of this area on a heat flow map of Europe! Definitely, the calibration model allows us to extract temperature at arbitrary depth or on arbitrary geological structure. As an example, figure 8 displays the temperature extracted along the top crystalline layer. It is obvious from figure 8 that the general dip of this layer has to be considered: it reaches the surface at the Black Forest in the north (very low temperatures) and dips towards greater depth in the southeast (higher temperatures). Furthermore, the PCT transects the crystalline rock in an E–W direction. It results in an elevated temperature at the greatest depth along the central trough. Temperatures therefore vary between 240 °C in the PCT and 10 °C in the outcrop of the crystalline rock in the Black Forest. This structure seems to be very promising for future geothermal production due to its hydrogeological characteristics. Figure 8. Open in new tabDownload slide Extracted temperature distribution from the thermal calibration model along the topmost 500 m crystalline layer. Figure 8. Open in new tabDownload slide Extracted temperature distribution from the thermal calibration model along the topmost 500 m crystalline layer. 5. Resource modelling The resource analysis is performed for resource classes 2 and 3. On the basis of our structural model of geological units, the key parameters of our assessment, temperature (from numerical modelling), hydrogeology (from available measurements) and reservoir geometry, have been established. Using the Gringarten (1978) solution, the geothermal productivity and extractable heat (equations (5) and (6)) can now be determined for various aquifer systems or arbitrary depth slices. Further assumptions have to be made, especially on the boundary conditions of surface utilization. The assumptions on the maximum density of doublet systems at the surface lead to a maximum recovery factor, Rmax, of 5%. Clearly, the realistic value will vary between 0% and Rmax. Strictly speaking, R depends on the hydraulic transmissivity, which is the most significant parameter for the determination of the distance between injection and production. The calculation procedure does not allow this distance to fall below 200 m. Furthermore, the operation time is set to t = 30 yr, during which temperature is not allowed to drop. A reinjection temperature of 15 °C is assumed, i.e. optimum cascade utilization at the surface. In the area of northern Switzerland, a total rock volume of 12 500 km3 is considered for resource class 2 and a volume of 42 500 km3 for resource class 3. When integrating the thermal calibration model, a total energy of EHIP = 1300 EJ (or 100 PJ km-3) for resource class 2 and EHIP = 13 000 EJ (or 300 PJ km-3) for resource class 3 would result. These numbers exceed by far the Swiss national energy need of 1.3 EJ yr-1. Since the extractable energy depends on the value of R, the producible geothermal energy is calculated for three major zones of interest: the crystalline rock, Upper Muschelkalk and Upper Marine Molasse. 5.1. Crystalline rock In places, the crystalline rock seems to be highly altered in the uppermost 500 m. Deeper crystalline layers have, however, rarely been drilled and no forecast of their properties can be given. Therefore, we restrict our analysis to a vertically 500 m thick layer in the topmost part of the crystalline rock. Fault zones with local transmissivity increase intersect it. There is even evidence from surface observations that these fault zones consist of multiple fracture sets and not of a single fracture. We assume a mean thickness of 500 m for each fracture zone. An ambient uniform hydraulic conductivity of K = 5 × 10-8 m s-1 for the crystalline rock and K = 2 × 10-7 m s-1 for the fault intersections has been taken. Due to the limited database, the use of uniform properties is suitable for a large-scale resource evaluation even though the detailed hydrologic situation is certainly strongly variable and far from being homogeneous. Under these premises, figure 9 illustrates the calculated geothermal productivity of this area. The central trough areas, the intersections with fault zones and the southeastern part seem to have the highest productivity. Along the reservoir thickness, a maximum productivity of Pg > 100 MWt is estimated from our evaluation for the most promising areas. These productivity values represent values for a single doublet system at an arbitrary location in the top crystalline layer. If production is, however, expanded over the total area, the system geometry has to be accounted for. With continued production over the assumed duration of 30 yr, Eg ≈ 40 000 PJ could be extracted. The volumetric average value of Eg = 13 PJ km-3 includes not only the deep, highly productive parts of the crystalline rock, but also the ‘cold’ near-surface parts in the NW. Figure 9. Open in new tabDownload slide Calculated thermal productivity for doublets in the topmost 500 m of the crystalline rock. Figure 9. Open in new tabDownload slide Calculated thermal productivity for doublets in the topmost 500 m of the crystalline rock. An appraisal of figure 9 must also consider the depth. Productivity is a function of temperature, which varies strongly with depth. In our model, the central trough has a depth of >4 km as well as the southeast part, where the crystalline rock is expected to be drilled only at >5 km depth. 5.2. Upper Muschelkalk The Upper Muschelkalk is treated similarly to the crystalline rock. The depth of the Upper Muschelkalk varies between 4.5 km in the southeast and reaches the surface close to the Swiss–German border. Formation temperature increases nearly linearly from the outcrop towards the SE. Maximum temperatures of 200 °C can be expected in the deepest parts. It also includes the same fault zones as the crystalline basement, which are probably the most promising structures in the Muschelkalk. In contrast to the crystalline rock where fractures generally form preferred flow paths, fault zones may represent flow barriers under particular conditions. We neglect this possibility and assume slightly elevated hydraulic conductivities of K = 3 × 10-7 m s-1 as a mean value for the undisturbed rock and K = 3 × 10-6 m s-1 for the intersections with the 500 m thick fault zones. The general NW–SE dipping trend is also clearly displayed in the productivity map (figure 10). In spite of the high K-values, the productivity is clearly lower than in the altered crystalline layer. This is mostly due to the small thickness of the Upper Muschelkalk layer (only 70 m) and to the lower temperature. Maximum productivity values of 20 MWt are reached at the greatest overburden and at the most permeable structures at the intersection with fault zones. When producing continuously over 30 yr, the Muschelkalk could provide an energy of Eg ≈ 3500 PJ. On average, Eg = 9 PJ could be produced per km3 of rock. Figure 10. Open in new tabDownload slide Estimated mean geothermal productivity of the Upper Muschelkalk. Figure 10. Open in new tabDownload slide Estimated mean geothermal productivity of the Upper Muschelkalk. 5.3. Upper Marine Molasse As noted earlier, the Upper Marine Molasse is the most used aquifer in Switzerland. Utilization is generally restricted to low-temperature systems with temperatures below 40 °C. At the greatest depth, temperatures of 60 °C can be expected in the SE part of the simulation area. We estimate hydrogeological conditions of K = 2 × 10-7 m s-1 for compact OMM rock. Due to their low hydrogeological importance, faults are not treated. In the productivity map, the fault zones could be more promising, especially in the deeper parts of the SE dipping structure. Maximum values of >15 MWt could be expected here (figure 11). Thickness increases towards the Alpine front where it exceeds 1000 m. A mean thickness of 250 m can be attributed to this SE dipping layer. The productivity increases up to 10 MWt as a function of thickness. The low reservoir temperature is also the reason why only a relatively small portion of the Swiss geothermal potential can be attributed to the OMM, in spite of the relatively high permeabilities and thicknesses. It is estimated that about 2700 PJ could be extracted after 30 yr of operation. Figure 11. Open in new tabDownload slide Estimated mean geothermal productivity of the Upper Marine Molasse. Figure 11. Open in new tabDownload slide Estimated mean geothermal productivity of the Upper Marine Molasse. 5.4. Surface utilization The various temperature or productivity values of the resources can be obtained either on a structural scale (isohypses of geological units), on a depth scale (for drilling purposes) or on a vertical slice relative to the model boundaries (relative to sea level). They represent the ideal position for combination with further data. As such, the productivity map can be connected to GIS data to link depth information to surface utilization. Multiple criteria can be conceived to identify optimum locations of geothermal production for individual utilization scenarios. For example, locations in close proximity to industrial or residential areas are of major interest for low-enthalpy systems since heat is most effectively produced close to consumers. For electricity production, mainly industrial areas are of interest, since here the large waste heat can potentially be used more effectively. In addition, the cooling of the power plant condenser needs to match different criteria: air cooling is less effective and is realized preferably under dry climatic conditions; however, water cooling might be the preferred solution in areas with sufficient precipitation. Here, areas in close proximity to surface waters are optimum locations. By such an approach, a new perspective for political and industrial decision-making is provided since both the consumption and the subsurface conditions can be displayed on one map. The best geothermal conditions are not necessarily at locations with an optimum consumer structure. Figure 12 illustrates a possible application of our approach. Here two constraints, industrial areas and usable surface waters, are shown together with the productivity map of the top crystalline layer (section 5.1). The map displays only the GIS information from northern Switzerland; the adjacent German utilization is blacked. Starting from a minimum distance of 500 m from both constraints the data start to grey until they are completely blacked at distances of 2 km. If a decision were required for optimum locations of geothermal production from the topmost crystalline layer, this map would indicate a favourable location next to the area of Brugg (Swiss coordinates x = 660/y = 260). Clearly, this figure identifies only one combination; more possibilities are conceivable such as depth (= drilling cost) and major river systems for pure EGS-type systems that would lead to different options. The more refined the GIS data provided by local authorities for urban and regional planning, the more precise the conclusions drawn from this procedure. Figure 12. Open in new tabDownload slide Combination of GIS data with a productivity map from the topmost crystalline. The coloured scheme exhibits optimum distances to industrial areas or to surface water systems. Areas without GIS data or without appropriate surface utilization are in black (see the text). Figure 12. Open in new tabDownload slide Combination of GIS data with a productivity map from the topmost crystalline. The coloured scheme exhibits optimum distances to industrial areas or to surface water systems. Areas without GIS data or without appropriate surface utilization are in black (see the text). 6. Conclusions In the geothermal community, different approaches have been taken to assess the subsurface geothermal potential. The approach presented here describes the important steps taken towards a dynamic, full 3D state-of-the-art evaluation. Of note, a first assessment for one of the most populated areas of Switzerland has been completed. The data compilation and treatment are performed in fully digital form and additional data are easily incorporated into this scheme. With the development of the Geoproductivity program, a single procedure can be applied to combine the subsurface geological/hydrogeological structures with temperature distributions to derive an areal, depth-dependent geothermal productivity distribution. Likewise, a future geothermal resource atlas can account for technical and geological aspects elaborated for individual areas of regional extent (∼50 km). Clearly, the analysis depends on different factors: the availability of hydraulic and thermal borehole data, a systematic geological examination and appropriate numerical interpretation tools. For the region considered, good data quality and geological reconnaissance were mostly provided from earlier investigations for nuclear waste disposal sites. The appraisal of subsurface temperature fields required, however, major effort. It turned out that the combination of 3D geological models and 3D numerical FE calculations is straightforward. In that way, a thermal calibration model of northern Switzerland was established. The quality of the fit with the borehole temperature data is generally excellent, even when geological information is simplified and homogenized material properties have been used partly. Uncertainty analyses of the temperature data show that numerical modelling is sometimes the only way of assessing the deeper temperature field. The results of our modelling indicate that diffusive transport dominates over advective heat transport in the study area, although at a local scale advective signatures can be identified. Three different geological structures have been considered in detail: the upper crystalline basement, Upper Muschelkalk and Upper Marine Molasse. The majority of the identified structures are ‘class 2’ systems for pure heat utilization. It may be estimated that 15 EJ could be extracted from ‘class 2’ systems over a time span of 30 yr. In addition, 50% of the topmost crystalline basement is linked to ‘class 3’ and could have favourable conditions for electricity production. The present assessment has determined that the Swiss primary energy consumption of 1.3 EJ yr-1 could be theoretically covered in large parts by utilization of these ‘class 3’ systems alone. Over a 30 yr period, a thermal energy of 40 EJ could be extracted from the most accessible part of the crystalline basement in northern Switzerland. It is obvious that this 500 m depth range represents only a small part of the total national geothermal resource, since deeper structures have not been integrated into this analysis due to insufficient data. It can even be assumed that the described fault zones penetrate deeper into the crystalline rock and may represent the objective of future geothermal production. The present assessment is conducted only for doublet systems. Although this is economically not always a viable perspective, it is a suitable approximation. In future, the same procedure will be applied to other areas of the Swiss Plateau. In contrast to this well-known geology, Alpine areas are more difficult to treat on similar regional scales due to possible bad coverage of geological and geophysical data. For such applications, the approach presented should be used in a flexible manner to include available data by local-scale 3D models. When using similar recovery factors, the results obtained are comparable to other geothermal resource assessments. However, our integrated procedure supplies more precise and extensive information. Evidently, one important aspect is the combination of the subsurface evaluation with GIS surface data that provides planning authorities with a powerful tool. Moreover, an extension of the methodology presented could easily include cost assessments of drilling and operational conditions. Herewith, an immediate forecast of geological or hydrogeological structures can be made available already at early stages of a geothermal project. The future will certainly bring along some new, innovative solutions for geothermal energy extraction and conversion. They can especially improve the recovery factors. Even with these improvements, the resource information represents the essential base for future energy scenarios. Acknowledgments The authors are very grateful to H Gorhan, M Geissmann and E Kissling for continuous and encouraging support. Financing is provided by the Swiss Geophysical Commission and the Swiss Federal Office of Energy. The thorough review of an anonymous reviewer is strongly appreciated. We would like to thank C Clauser for his editorial efforts. References Andenmatten Berthoud N , Kohl T . , 2003 Assessment and evaluation of geothermal potential in Switzerland (Atlas des Ressources Géothermiques Suisses), Swiss Geophysical Commission (Zürich: Office Fédérale de l'Energie) Andenmatten N , Kohl T . , 2003 Numerical simulations of 3-D thermal fields from GOCAD geological models 23rd GOCAD User Meeting (Nancy) Mallet J-L . (pg. 1 - 8 ) Bertani R . , 2005 World geothermal generation 2001–2005: state of the art Proc. World Geothermal Congress 2005 (Antalya, 24–29 April 2005) Clauser C . , 1997 The thermal regime of the crystalline continental crust; implications from the KTB , J. Geophys. Res. B , vol. 102 (pg. 18417 - 41 ) Google Scholar Crossref Search ADS WorldCat Gringarten A C . , 1978 Reservoir lifetime and heat recovery factor in geothermal aquifers used for urban heating , Pure Appl. Geophys. , vol. 117 (pg. 298 - 308 ) Google Scholar Crossref Search ADS WorldCat Hudson R B . , 1998 Electricity production from geothermal energy Proc. Int. Summer School ‘Electricity Production from Geothermal Energy’ Ponta Delgada, Azores Forjaz V H , Bicudo Ponte C A , Popovski K . (pg. 4.3 - 4.31 ) Jung R . , 2002 Abschätzung des technischen Potenzials der geothermischen Stromerzeugung und der geothermischen, Kraft-Wärmekopplung (KWK) in Deutschland (Hannover: BGR and Institut für Geowissenschaftliche Gemeinschaftsaufgaben) Kohl T , Andenmatten N , Rybach L . , 2003 Geothermal resource mapping—example from northern Switzerland , Geothermics , vol. 32 (pg. 721 - 32 ) Google Scholar Crossref Search ADS WorldCat Kohl T , Rybach L . , 1996 Thermal and hydraulic aspects of the KTB drill site , Geophys. J. Int. , vol. 124 (pg. 756 - 76 ) Google Scholar Crossref Search ADS WorldCat Kohl T , Signorelli S , Rybach L . , 2001 Three-dimensional thermal investigation below high Alpine topography , Phys. Earth Planet. Inter. , vol. 126 (pg. 195 - 210 ) Google Scholar Crossref Search ADS WorldCat Köhler S . , 2005 Geothermischer angetriebene Dampfkraftprozesse—Analyse und Prozessvergleich binärer Kraftwerke , PhD Thesis TU Berlin, Germany Lavigne J . , 1978 Les ressources géothermiques françaises—possibilités de mise en valeur , Ann. Mines , vol. 4 (pg. 57 - 72 ) OpenURL Placeholder Text WorldCat Lund J W , Freeston D H , Boyd T . , 2005 World-wide direct uses of geothermal energy 2005 Proc. World Geothermal Congress 2005 (Antalya, 24–29 April 2005) Mallet J-L . , 1992 Discrete smooth interpolation in geometrics modeling , Comput. Aided J. , vol. 24 (pg. 178 - 91 ) Google Scholar Crossref Search ADS WorldCat McKelvey V E . , 1972 Mineral resource estimate and public policy , Am. Sci. , vol. 60 (pg. 32 - 40 ) OpenURL Placeholder Text WorldCat Muffler P , Cataldi R . , 1978 Methods for regional assessment of geothermal resources , Geothermics , vol. 3 (pg. 53 - 89 ) Google Scholar Crossref Search ADS WorldCat Müller W H , Naef H , Graf H R . , 2001 Geologische Entwicklung der Nordschweiz, Neotektonik und Langzeitszenarien Zürcher Weinland NAGRA Techn. Bericht 99–08, Wettingen Rybach L . , 1982 Geothermal conditions in Switzerland , Bull. Vereins Schweizerischer Petr.-Geol. , vol. 48/115 (pg. 23 - 34 ) OpenURL Placeholder Text WorldCat Rybach L . , 1992a Geothermal potential of the Swiss Molasse Basin , Eclogae Geol. Helv. , vol. 85/3 (pg. 733 - 44 ) OpenURL Placeholder Text WorldCat Rybach L . , 1992b An attempt to interpret the temperature profile of the KTB pilot drill hole (Germany) by paleoclimatic considerations , Palaeogeogr. Palaeoclimatol. Palaeoecol. , vol. 98 (pg. 193 - 7 ) Google Scholar Crossref Search ADS WorldCat Rybach L , Kohl T . , 2004 Waste heat problems and solutions in geothermal energy , Energy, Waste and the Environment Gieré R , Stille P . London The Geological Society of London (pg. 369 - 80 ) Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Sanyal S K . , 2005 Classification of geothermal systems—a possible scheme Proc. 30th Workshop on Geothermal Reservoir, Engineering Stanford, 31 Jan —2 Feb 2005 (pg. 106 - 13 ) Sanyal S K , Butler S J . , 2004 National assessment of US Enhanced Geothermal Resource Base—a perspective Geothermal Resources Council Transactions, Palm Springs, 29 Aug—4 Sept 2004 , vol. 28 (pg. 233 - 8 ) Schärli U , Kohl T . , 2002 Archivierung und Kompilation geothermischer Daten der Schweiz und angrenzender Gebiete, Swiss Geophysical Commission (Beiträge zur Geologie der Schweiz: Geophysik, no. 36) Seitz A . , 2004 Aspekte der Hybridisierung geothermischer Energie zur Stromproduktion (München) pg. 128 Signorelli S , Andenmatten Berthoud N , Kohl T . , 2004a Geothermischer Ressourcenatlas der Schweiz Jahresbericht 2004 Swiss Geophys. Commission Signorelli S , Andenmatten Berthoud N , Kohl T , Rybach L . , 2004b Statistik Geothermische Nutzung der Schweiz für die Jahre 2002 und 2003 Swiss Federal Office for Energy Bern no 0417033 Swiss Federal Statistical Office . , 2003 , Statistical Yearbook of Switzerland—Country Portraits http://www.bfs.admin.ch/bfs/portal/en/index/international/laenderportraets.html Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Williams C F . , 2004 Development of revised techniques for assessing geothermal resources Proc. 29th Workshop on Geothermal Reservoir Engineering, Stanford, 26–28 Jan 2004 © 2005 Nanjing Institute of Geophysical Prospecting
TI - Development of a regional geothermal resource atlas
JF - Journal of Geophysics and Engineering
DO - 10.1088/1742-2132/2/4/S11
DA - 2005-12-01
UR - https://www.deepdyve.com/lp/oxford-university-press/development-of-a-regional-geothermal-resource-atlas-5hRhaUsKPo
SP - 372
VL - 2
IS - 4
DP - DeepDyve
ER -