Evaluation of the Theoretical Geothermal Potential of Inferred Geothermal Reservoirs within the Vicano – Cimino and the Sabatini Volcanic Districts ( Central Italy ) by the Application of the Volume Method

The evaluation of the theoretical geothermal potential of identified unexploited hydrothermal reservoirs within the Vicano–Cimino and Sabatini volcanic districts (Latium region, Italy) has been made on the basis of a revised version of the classical volume method. This method is based on the distribution of the partial pressure of CO2 (pCO2) in shallow and deep aquifers to delimit areas of geothermal interest, according to the hypothesis that zones of high CO2 flux, either from soil degassing and dissolved into aquifers, are spatially related to deep hydrothermal reservoirs. On the whole, 664 fluid discharges (cold waters, thermal waters, and bubbling pools) have been collected from shallow and deep aquifers in the Vicano–Cimino Volcanic District and the Sabatini Volcanic District for chemical and isotopic composition, in an area of approximately 2800 km2. From this large hydro-geochemical dataset the pCO2 values have been computed and then processed to obtain a contour map of its spatial distribution by using geostatistical techniques (kriging). The map of pCO2 has been used to draw up the boundaries of potentially exploitable geothermal systems within the two volcanic districts, corresponding to the areas where endogenous CO2 raise up to the surface from the deep hydrothermal reservoirs. The overall estimated potential productivities and theoretical minimum and maximum thermal power of the two volcanic districts are of about 45 × 103 t/h and 3681–5594 MWt, respectively. This makes the Vicano–Cimino Volcanic District and the Sabatini Volcanic District very suitable for both direct and indirect exploitation of the geothermal resources, in view of the target to reduce electricity generation from conventional and poorly sustainable energy sources.


Introduction
In recent decades, the worldwide demand for energy has increased far beyond a critical threshold, with an associated rise in CO 2 emissions being observed.In such conditions, the development of a plan focused on non-carbon or reduced-carbon sources of energy involves the evaluation of the subsurface energy potential and the development of such technologies as the geothermal energy [1].The peri-Tyrrhenian sector of central Italy hosts a large thermally anomalous area that comprises southern Tuscany, the Latium region, and the Campanian volcanic areas of the Phlegraean Fields and Vesuvius, where exploration for high and medium enthalpy fluids has concentrated in areas of recent magmatism [2,3].Starting in late 1960s, a massive program of geothermal prospections has been conducted in large areas of Latium by ENEL (National Electric Energy Agency) and AGIP (National Oil Company) companies, in order to quantify the potential resources suitable for electricity generation [4][5][6][7][8][9].Several tens of 1-5 km-deep boreholes and about 180 test-holes have been drilled until the early 1990s [10] (and references therein), providing maximum bottom-hole temperatures exceeding 300 • C. Despite the significant potential highlighted by these preliminary explorations, the geothermal resources of Latium remain so far unexploited.The reasons for this failure have mostly been related to the low permeability of the reservoirs [11], resulting in dry and unproductive boreholes for almost half of the deep wells drilled in the whole region, whereas potentially productive wells have regularly suffered the presence of hot brines (up to 350 g/L at Cesano; [5]) and/or corrosive gases (H 2 S), so that operations always stopped at the preliminary phases.Notwithstanding these negative attempts, Italy is experiencing a renewed interest in geothermal energy favored by recent technological advances in exploration and exploitation, which have extended the potential of geothermal reservoirs to lower temperatures and greater depths, and encouraged by the growth of energy demand [1].In recent years, several exploration permits have been requested by private companies in Italy, many of these are in the Latium region [12], indicating the significant interest of industry for the development of this renewable resource.
A classical method proposed to estimate the theoretical geothermal potential of a given area is the volume method [13], which is based on the calculation of the heat stored in a certain volume of rock and requires information on the depth of the top of the geothermal reservoir and both an average porosity and an average fluid temperature of the reservoir itself.In more recent years, the old approach has been reviewed with the introduction of the CO 2 surveyed in the regional aquifers as a tracer for reservoir productivity [14], as typical geothermal systems exhibit an anomalous CO 2 degassing of deep provenance in areas of high heat flux [15][16][17][18][19].In this paper, a revised version of the volume method has been applied to estimate the theoretical geothermal potential [20] of inferred geothermal reservoirs within the Vicano-Cimino Volcanic District (VCVD) and the Sabatini Volcanic District (SVD) in the Latium region.Based on a large and very detailed geochemical dataset, the aim of this study is to identify potential areas for high-to-low temperature resource exploitation and refine the estimation of the theoretical geothermal potential.

Geological and Hydrogeological Background
The study area comprises the VCVD and the SVD, two of the four large (more than 1000 km 2 each) Quaternary volcanic districts of the Roman Magmatic province [21], and the adjacent Tolfa mountains, which formed from the emplacement of an intrusive body pertaining to the Tuscan Magmatic Province (i.e., the Tolfa Dome Complex [22]) on a sedimentary basement (Figure 1).Magmatism in the Tyrrhenian sector of central Italy was generated as the result of a post-collisional crustal extension which occurred at the back of the eastward-migrating Apennine fold-and-thrust belt.This extensional system also led to the development of dominantly NW-and minor NE-striking extensional fault sets arranged in a horst-graben pattern [10,23], and produced a strong crustal thinning (<25 km; [24]) and high heat flow (locally > 200 mW/m 2 ; [25]).Volcanic complexes grew up on buried horst-graben structures, as shown by gravimetric anomalies [10], whilst marine clastic sediments filled the structural lows.
The hydrogeological setting is dominated by a regional hydrothermal reservoir hosted in the carbonate-evaporite units and a shallow, mainly unconfined, regional aquifer within the volcanic rocks [32].Low-permeability Plio-Pleistocene deposits and/or the Ligurian s.l.rocks generally act as an efficient hydraulic barrier between the shallow and the deep aquifer.Locally, permeable layers within the low-permeability sedimentary deposits host perched aquifers that feed numerous springs of limited and discontinuous extent.Thermal and mineral springs, representing clear examples of localized rising waters from the deeper regional hydrothermal reservoir, abundantly emerge from the volcanic and sedimentary deposits in the presence of tectonic disturbance, since fractures in fault zones act as preferential paths for the fast upwelling of deep-originated fluids to the surface [33,34].The pre-volcanic basement of the SVD and VCVD comprises, from bottom to top (Figure 1): (1) Mesozoic carbonates overlying Triassic evaporitic facies (Burano Fm.); (2) a Cretaceous-Paleogene arenaceous-clayey-carbonate allochthonous flyschoid complex (Ligurian s.l.); and (3) a Miocene-Plio-Pleistocene autochthonous complex made of continental marls, sands, clays, and conglomerates.The youngest formations are Quaternary continental clastic sediments associated with travertines and diatomites [10,31].
The hydrogeological setting is dominated by a regional hydrothermal reservoir hosted in the carbonate-evaporite units and a shallow, mainly unconfined, regional aquifer within the volcanic rocks [32].Low-permeability Plio-Pleistocene deposits and/or the Ligurian s.l.rocks generally act as an efficient hydraulic barrier between the shallow and the deep aquifer.Locally, permeable layers within the low-permeability sedimentary deposits host perched aquifers that feed numerous springs of limited and discontinuous extent.Thermal and mineral springs, representing clear examples of localized rising waters from the deeper regional hydrothermal reservoir, abundantly emerge from the volcanic and sedimentary deposits in the presence of tectonic disturbance, since fractures in fault zones act as preferential paths for the fast upwelling of deep-originated fluids to the surface [33,34].

Basic Statistics and Geostatistical Analysis
Basic statistics (Table 1) show that the whole population of samples (664) is positively skewed with respect to pCO 2 (i.e., the mean is higher than the median value) and, consequently, characterized by non-normal distribution.The pCO 2 varies from 0.001 to 0.98 bar with the median value, which represents a more robust statistic parameter for non-normal distributions, of 0.017 bar and the inter-quartile range, representative of the distribution dispersion, of 0.046 bar (0.008-0.055 bar).In order to improve the symmetry of the highly-skewed normal distribution and obtain a distribution as close as possible to a Gaussian-type, a log 10 -transformation of the variable has been applied.As a result, the frequency distribution of pCO 2 (Figure 2) and the Quantile-Quantile (QQ) plot (Figure 3), both derived from the log-normal transformation of the variable, approach a symmetrical distribution although they do not follow a normal distribution.The histogram (Figure 2) shows a unimodal shape with a long right tail; more in detail, it highlights the presence of one maximum in the interval from −2.6 to −1.0 log bar and a uniform trend from −1.0 to 0 log bar.In the QQ plot (Figure 3) the pCO 2 values plot along a curve which can be modeled as the combination of four different log-normal populations [36], which reflect the complexity of the CO 2 degassing process from different sources.

Basic Statistics and Geostatistical Analysis
Basic statistics (Table 1) show that the whole population of samples (664) is positively skewed with respect to pCO2 (i.e., the mean is higher than the median value) and, consequently, characterized by non-normal distribution.The pCO2 varies from 0.001 to 0.98 bar with the median value, which represents a more robust statistic parameter for non-normal distributions, of 0.017 bar and the interquartile range, representative of the distribution dispersion, of 0.046 bar (0.008-0.055 bar).In order to improve the symmetry of the highly-skewed normal distribution and obtain a distribution as close as possible to a Gaussian-type, a log10-transformation of the variable has been applied.As a result, the frequency distribution of pCO2 (Figure 2) and the Quantile-Quantile (QQ) plot (Figure 3), both derived from the log-normal transformation of the variable, approach a symmetrical distribution although they do not follow a normal distribution.The histogram (Figure 2) shows a unimodal shape with a long right tail; more in detail, it highlights the presence of one maximum in the interval from −2.6 to −1.0 log bar and a uniform trend from −1.0 to 0 log bar.In the QQ plot (Figure 3) the pCO2 values plot along a curve which can be modeled as the combination of four different log-normal populations [36], which reflect the complexity of the CO2 degassing process from different sources.After basic statistics, the pCO2 values have been processed to obtain a map of its spatial distribution.The geostatistical technique applied herein involves the use of kriging [37,38] through semi-variogram computation and modeling.A directional semi-variogram of the log-transformed variable has been calculated to estimate the spatial variation of pCO2 by applying a lag distance of 2000 m (tolerance ± 1000 m) (Figure 4).This distance is comparable to the average minimum distance among pairs of samples.This representation takes into account the possible spatial anisotropy of the variable and allows verification of the presence of spatial autocorrelation among the experimental data.The directional semi-variogram shows a general growth of the semi-variance γ(h) with the distance, within a radius of influence (range), while beyond that distance the γ(h) remains approximately constant around the value (sill) of 0.46.This means that a spatial correlation among observations exists within the range distance, over which no spatial correlation between data is apparent.Moreover, for h = 0 the semivariance γ(h) > 0 due to the short scale variations enclosed within the first lag (nugget effect).The After basic statistics, the pCO 2 values have been processed to obtain a map of its spatial distribution.The geostatistical technique applied herein involves the use of kriging [37,38] through semi-variogram computation and modeling.A directional semi-variogram of the log-transformed variable has been calculated to estimate the spatial variation of pCO 2 by applying a lag distance of 2000 m (tolerance ± 1000 m) (Figure 4).This distance is comparable to the average minimum distance among pairs of samples.After basic statistics, the pCO2 values have been processed to obtain a map of its spatial distribution.The geostatistical technique applied herein involves the use of kriging [37,38] through semi-variogram computation and modeling.A directional semi-variogram of the log-transformed variable has been calculated to estimate the spatial variation of pCO2 by applying a lag distance of 2000 m (tolerance ± 1000 m) (Figure 4).This distance is comparable to the average minimum distance among pairs of samples.This representation takes into account the possible spatial anisotropy of the variable and allows verification of the presence of spatial autocorrelation among the experimental data.The directional semi-variogram shows a general growth of the semi-variance γ(h) with the distance, within a radius of influence (range), while beyond that distance the γ(h) remains approximately constant around the value (sill) of 0.46.This means that a spatial correlation among observations exists within the range distance, over which no spatial correlation between data is apparent.Moreover, for h = 0 the semivariance γ(h) > 0 due to the short scale variations enclosed within the first lag (nugget effect).The This representation takes into account the possible spatial anisotropy of the variable and allows verification of the presence of spatial autocorrelation among the experimental data.The directional semi-variogram shows a general growth of the semi-variance γ(h) with the distance, within a radius of influence (range), while beyond that distance the γ(h) remains approximately constant around the value (sill) of 0.46.This means that a spatial correlation among observations exists within the range distance, over which no spatial correlation between data is apparent.Moreover, for h = 0 the semi-variance γ(h) > 0 due to the short scale variations enclosed within the first lag (nugget effect).The geometric anisotropy for pCO 2 is defined along the N10 • E and N100 • E directions (Figure 4), which have been identified as representative of major (U direction) and minor (V direction) axes of the anisotropy ellipse, respectively.In particular, the major axis is the direction of maximum spatial continuity of the variable (maximum range: 12,000 m) and the minor axis is the direction of maximum spatial variability (minimum range: 8500 m).Ordinary kriging has been applied to produce the estimation map of pCO 2 (Figure 5) that, at a preliminary phase, has been back-transformed into the original variable values.
geometric anisotropy for pCO2 is defined along the N10° E and N100° E directions (Figure 4), which have been identified as representative of major (U direction) and minor (V direction) axes of the anisotropy ellipse, respectively.In particular, the major axis is the direction of maximum spatial continuity of the variable (maximum range: 12,000 m) and the minor axis is the direction of maximum spatial variability (minimum range: 8500 m).Ordinary kriging has been applied to produce the estimation map of pCO2 (Figure 5) that, at a preliminary phase, has been back-transformed into the original variable values.The assumption is that areas with anomalous pCO2 values represent the potentially exploitable geothermal reservoirs in the VCVD and the SVD (for any detail relative to the numbered areas see Table 2).Test-holes and geothermal wells are also reported in the map (see also Figure 1b).

Application of the Volume Method
The volume (V) of the unexplored geothermal reservoir is computed by multiplying the areal extension (A) and the average thickness (h) of the reservoir itself, according to the equation

V = A × h
where A is evaluated on the basis of the pCO2 distribution of emerging waters (Figure 5) and h is calculated from the difference between the depth of the potential reservoir top and the depth of 3 km, which is usually considered the maximum depth for an economically exploitable geothermal resource (Table 2) [11,14].All available information on: (1) the geological structure at depth and ( 2) the depth of the potential reservoir top have been acquired from the Ministry of Economic Development website [12].The assumption is that areas with anomalous pCO 2 values represent the potentially exploitable geothermal reservoirs in the VCVD and the SVD (for any detail relative to the numbered areas see Table 2).Test-holes and geothermal wells are also reported in the map (see also Figure 1b).

Application of the Volume Method
The volume (V) of the unexplored geothermal reservoir is computed by multiplying the areal extension (A) and the average thickness (h) of the reservoir itself, according to the equation where A is evaluated on the basis of the pCO 2 distribution of emerging waters (Figure 5) and h is calculated from the difference between the depth of the potential reservoir top and the depth of 3 km, which is usually considered the maximum depth for an economically exploitable geothermal resource (Table 2) [11,14].All available information on: (1) the geological structure at depth and ( 2) the depth of the potential reservoir top have been acquired from the Ministry of Economic Development website [12].The theoretical geothermal potential (W) of the identified geothermal reservoirs (Table 2), which is comparable to the thermal energy in [42], has been computed according to the equation [14] where C w (in J/kg K) is the specific heat capacity of fluids contained in the geothermal reservoir (data from [43]), (T) is the reservoir temperature (in K), Q (in t/h) is the hourly production rate, and 1000/3600 is a conversion factor (to convert Q from t/h to kg/s).The hourly production rate (Q) of the potential geothermal systems has been computed multiplying the volume of each geothermal reservoir for its specific productivity.The average specific productivity of 40 t/h km 3 , calculated for the liquid-dominated geothermal systems of Latium by [14] has been used for calculations.Reservoir temperatures have been derived taking into account, hierarchically, data from: (1) deep drilling (i.e., bottom-hole temperatures); (2) geothermometric evaluations based on chemical equilibria of gas species [19,[39][40][41]; and (3) the map of temperature of the potential reservoir top [12].Minimum and maximum temperatures (Table 2) have been considered to obtain the minimum and the maximum theoretical thermal power, respectively.

Discussion
The shape of the frequency distribution of logpCO 2 (Figure 2) highlights the presence of two distinct populations of: (1) low pCO 2 values (≤0.045 bar), which can generally be interpreted as the normal enrichment in soil derived-CO 2 of infiltrating waters from zones of low CO 2 flux, and (2) medium-to-high pCO 2 values (>0.045 bar), reflecting the input of deeply derived CO 2 .This hypothesis is confirmed by the δ 13 C-CO 2 values of the sampled waters, ranging from −27.6 to +2.3‰ vs. VPDB [19,39,40], which suggest that CO 2 has a two-fold origin: relatively negative δ 13 C-CO 2 values of the low pCO 2 waters imply dominant CO 2 contribution from soil respiration and aerobic decay of organic matter [44]; conversely, less negative δ 13 C-CO 2 values of the medium-to-high pCO 2 waters point to CO 2 production from thermo-metamorphic reactions involving carbonate formations (δ 13 C-CO 2 values from −2.0 to +2.3‰ vs. VPDB [45]) and minor contribution from mantle degassing (δ 13 C-CO 2 values from −7.0 to −3.0‰ vs. VPDB [46]).Consistent with this hypothesis, the different populations highlighted in the QQ plot (Figure 3) can be interpreted as representative of pCO 2 values fed by both biological and endogenous sources.In our interpretation, the populations with the lowest values (A,B) represent the background values (organic CO 2 ) while the other populations (C,D) reflect the input of CO 2 derived from depth.In spite of possible different interpretations of the origin of the populations, the pCO 2 value of 0.045 bar has been selected as the possible threshold value for the background biological CO 2 .
The geographic distribution of pCO 2 values in the shallow aquifers of the VCVD and the SVD (Figure 5) highlight the presence of areas characterized by medium-to-high pCO 2 values, i.e., those reflecting the input of endogenous CO 2 , whose distribution is not homogeneously affected from the complexity of the structural setting and the fracture-related upwelling of deep-originated fluids from the active hydrothermal system.Those areas, whose size is reported in Table 2, are characterized by pCO 2 values > 0.045 bar and are used to define the areal extension of potentially exploitable geothermal systems, as they are interpreted as the surface expressions of geothermal reservoirs located at depth.Fifteen areas with size comprised from 2 to 104 km 2 are highlighted in the map of Figure 5.They are both associated with (1) thermal waters upwelling from the deep reservoir, where the CO 2 -rich gas phase is originally dissolved, and (2) cold waters from the shallow aquifers, which receive the input of endogenous CO 2 that separates from the parent deep fluid and upwells as a single gas phase from the deep reservoir.These areas are generally located in correspondence with positive gravity zones, which have been interpreted as buried structural highs of the carbonate basement [19] and represent the sectors where the top of the geothermal reservoir is located at shallower depths.On the other hand, pCO 2 values < 0.045 bar, i.e., those reflecting the shallow production of biological CO 2 , are associated with gravity lows and are typical of cold waters circulating within the shallow aquifers, which have no relation with the tectonic framework and do not receive contributions from the geothermal reservoir.As for the lack of high values of pCO 2 in correspondence with the geothermal field of Cesano, one possible explanation is that the Cesano field is a hot brine [5], so self-sealing phenomena of fractures and faults by the highly saline fluids could occur and hinder their upwelling to the surface.
Calculated volumes, discharge rates, and theoretical geothermal potential estimates for the 15 areas are reported in Table 2.However, it is worth noting that the specific productivities obtained in this work represent rough approximations and their use can lead to significant uncertainties, as the factors controlling the productivity of geothermal reservoirs, such as permeability, vary from reservoir to reservoir.The following discussion is a brief description of each area.The Civitavecchia reservoir can be considered as an economically valuable, low-enthalpy geothermal resource, due to the low temperatures (50-85 • C), and the small depth of the reservoir top (~400 m).Both the high expected productivity (~8840 t/h) and the considerable geothermal potential that may be extracted (257-619 MWt), and the presence of potential users (e.g., in the town of Civitavecchia, at the distance of few km) make this site very attractive.The boreholes drilled in the past for geothermal exploitation by ENEL are currently used for greenhouse space heating and public spas.The reservoirs of Monte Romano-Blera and Tuscania have moderate production rates (2280-3192 t/h) and considerable expected theoretical geothermal potentials (up to 527 and 700 MWt, respectively).Based on the depth of the reservoir top (>1000 m) and the temperature (150-200 • C), they can be classified as useful accessible medium-to-high enthalpy geothermal resources for electric production.The Viterbo-Grotte S. Stefano thermal basin shows considerable expected productivity (7216 t/h) and theoretical geothermal potential (310-497 MWt) in an area with low temperatures (60-80 • C) at relatively shallow depths (~400-800 m), suggesting good perspectives for direct geothermal uses, also due to the presence of many potential users (e.g., in the town of Viterbo, at the distance of few km).About Grotte S. Stefano, it is worth noting that wells drilled in the past for geothermal resource exploitation have been soon abandoned after CO 2 -rich gas phase eruption, with no associated thermal water, at the interception of the reservoir top.The Canale Monterano-Mt.Solferata reservoir can be considered an economically valuable, due to the high temperatures (150-200 • C) and the moderate depth of the reservoir top (~600 m).Both high expected productivity (~9984 t/h) and geothermal potential (1498-2189 MWt) make this site a useful accessible resource both for power production and direct uses.The small Capranica reservoir has temperatures in the range 100-150 • C at a depth of ~500 m, a moderate expected productivity (~600 t/h), but low theoretical geothermal potential (53-90 MWt), suggesting good perspectives for both direct uses of the geothermal resources and power production.Similar to the Capranica reservoir are those of Manziana, Trevignano, and Castel Campanile, in terms of both areal extension (2-9 km 2 ) and temperatures (100-165 • C).Smaller expected productivities (168-360 t/h) and theoretical geothermal potentials (19-43 MW) make those sites useful accessible resources for direct use and power production.The Nepi, Isola Farnese, and Castelnuovo di Porto-Sacrofano reservoirs are characterized by considerable expected productivities (792-1700 t/h) and low-to-moderate geothermal potentials (70-316 MWt).Even though these systems have high temperatures (100-200 • C), based on the high depth at their top (1200-1500 m) they can be classified as useful accessible medium-to-high enthalpy geothermal resources for electric production.Moving eastward, the Orte-Montecchie, Ponzano Romano, and Fiano Romano reservoirs are efficiently cooled by the meteoric water recharge of the hydrothermal from the nearby Apennine range.This explains the low temperatures (25-100 • C) at the reservoir top, even if at a relatively shallow depth (200-400 m).On the basis of their considerable expected productivity (728-5200 t/h) and low theoretical geothermal potential (up to 163 MWt), these systems can be classified as useful accessible resources for direct uses.

Materials and Methods
Up to 664 water samples have been collected over an area of ~2800 km 2 from springs and wells fed by (1) the cold and shallow aquifers hosted in the volcanic and sedimentary rocks and (2) the deep hydrothermal reservoir.The sampling sites have been homogeneously distributed all over the investigated area, and the sampling density is high (Figure 1).The pCO 2 of the sampled waters has been calculated with the PHREEQC code [47], operating with the Lawrence Livermore National Laboratory database and using as input data the groundwater physico-chemical parameters, whose full dataset is reported elsewhere [19,39,40].
Descriptive statistics and graphical representations have been carried out to characterize the population of water samples with respect to pCO 2 .After that, experimental data have been processed to produce a pCO 2 -contour map by applying the techniques of geostatistics (i.e., variogram analysis and kriging estimation; [37,38]).Experimental directional variograms have been constructed to (1) investigate the spatial dependence of the pCO 2 values by calculating the variogram parameters (i.e., nugget, range and sill) and (2) determine the directional differences (anisotropy).Kriging has been applied to provide the best local estimate of the mean value of a regionalized variable (i.e., a certain property that varies in the geographic space) by using the measured values and a semi-variogram to determine the scale of variance and estimate the unknown values.
According to [14], the revised volume method has been applied on the produced contour map to compute the volumes of the deep geothermal reservoirs identified on the basis of the anomalies of pCO 2 .

Conclusions
The revised volume method has been used to evaluate the potential productivity and the theoretical geothermal potential of unexploited geothermal reservoirs of the VCVD and the SVD, as identified by the distribution of pCO 2 in shallow and deep aquifers.This parameter has allowed the delimitation of areas of potential geothermal interest which have been computed through a geostatistical approach (kriging).
By assuming a specific productivity of 40 t/h km 3 , a potential productivity of ~45 × 10 3 t/h and a total theoretical geothermal potential of 3682-5595 MWt has been estimated.This makes the exploitation of the identified geothermal resources in the VCVD and the SVD very suitable for both generation of electric power and direct uses that, due to the presence of many potential users (municipalities, industrial sites, agricultural, and touristic infrastructures), can play a significant role in the reduction of CO 2 emissions.

Figure 1 .Figure 1 .
Figure 1.(a) Geological sketch map of the Vicano-Cimino Volcanic District (VCVD) and Sabatini Volcanic District (SVD) showing the location of the fluid sampling sites; (b) location of test-holes, geothermal wells, and exploration permits (from [12,35]); (c) simplified geological model for the Figure 1.(a) Geological sketch map of the Vicano-Cimino Volcanic District (VCVD) and Sabatini Volcanic District (SVD) showing the location of the fluid sampling sites; (b) location of test-holes, geothermal wells, and exploration permits (from [12,35]); (c) simplified geological model for the VCVD and SVD geothermal systems (from [1]).

Figure 2 .
Figure 2. Histogram of logpCO 2 values of the collected waters.

Figure 2 .
Figure 2. Histogram of logpCO2 values of the collected waters.

Figure 3 .
Figure 3. QQ plot of logpCO2 values of the collected waters.The black arrow at the logpCO2 value of −1.347 (0.045 bar) separate the background populations (A,B) from those (C,D) where CO2 partially derives from deep sources.

Figure 3 .
Figure 3. QQ plot of logpCO 2 values of the collected waters.The black arrow at the logpCO 2 value of −1.347 (0.045 bar) separate the background populations (A,B) from those (C,D) where CO 2 partially derives from deep sources.

Figure 2 .
Figure 2. Histogram of logpCO2 values of the collected waters.

Figure 3 .
Figure 3. QQ plot of logpCO2 values of the collected waters.The black arrow at the logpCO2 value of −1.347 (0.045 bar) separate the background populations (A,B) from those (C,D) where CO2 partially derives from deep sources.

Figure 5 .
Figure 5. Map of pCO2 distribution as obtained from ordinary kriging after back-transformation.The assumption is that areas with anomalous pCO2 values represent the potentially exploitable geothermal reservoirs in the VCVD and the SVD (for any detail relative to the numbered areas see Table2).Test-holes and geothermal wells are also reported in the map (see also Figure1b).

Figure 5 .
Figure 5. Map of pCO 2 distribution as obtained from ordinary kriging after back-transformation.The assumption is that areas with anomalous pCO 2 values represent the potentially exploitable geothermal reservoirs in the VCVD and the SVD (for any detail relative to the numbered areas see Table2).Test-holes and geothermal wells are also reported in the map (see also Figure1b).

Table 1 .
Basic statistics (quantitative data) relative to the whole population of pCO 2 values.

Table 1 .
Basic statistics (quantitative data) relative to the whole population of pCO2 values

Table 2 .
[41]20]ion of fluid production rate (t/h) and theoretical geothermal potential (MWt)[13,20]calculated through the revised volume method of [14] for the VCVD and the SVD.= area of the geothermal reservoir calculated by kriging; h = average thickness of the geothermal reservoir; V = volume of the geothermal reservoir; Q = discharge rate of the geothermal fluid; T min-max = minimum and maximum temperatures of the geothermal reservoir; W = theoretical thermal power. 1 2[19];3 [39]; 4[40]; 5[41]. A