Analysis of Fluid Flow Pathways in the Mount Meager Volcanic Complex, Southwestern Canada, Utilizing AMT and Petrophysical Data

Defining the spatial distribution of geological structures and rock properties is important for understanding how fluid flow is controlled in a geothermal reservoir. Here, we present a procedure to examine the potential fluid pathways. By combining 3‐D resistivity models derived from audio‐magnetotelluric (AMT) data with available rock properties (porosity and permeability) and fluid sample data (fluid resistivity, salinity, and temperature), we investigated the relationship between electrical resistivity and fluid flow in an active volcanic system. Different petrophysical models and empirical relations are evaluated to determine the relationship between the fluid flow system at Mount Meager, British Columbia, and the resistivity model. In addition, we utilized porosity and permeability measured in the laboratory to define the porosity‐permeability relationship. The porosity of the volcanic core samples showed a range of 2.6%–23.2% and the permeability was in a range of 0.001–5,186.57 mD. The results showed the potential of 3‐D inversion of AMT data to map the fluid pathways at Mount Meager. These pathways are correlated with loss circulation zones in boreholes and can account for porosity up to 8.5%, which using the porosity‐permeability relationship translates to permeability of the order 0.249 mD. Not only are the fault and fracture zones important for reservoir exploitation, but they also provide permeability for the circulation of meteoric water. Our studies suggest that a set of fractures with 0.1 m spacing and 20 mm aperture can keep 40% fluid in pores and transmit fluid with possible permeability of 666 mD.


10.1029/2022GC010814
2 of 20 Different methods have been used by researchers to evaluate fluid flow pathways and their effective parameters in various geothermal reservoirs around the world. Numerical modeling (Gunnarsson & Aradóttir, 2015;Strehlow et al., 2015), structural and laboratory experiments (Eggertsson et al., 2020;J. Farquharson et al., 2015;Kendrick et al., 2013;Liotta et al., 2020), geochemical studies (Ásmundsson et al., 2014;Darling & Armannsson, 1989;Libbey & Williams-Jones, 2016), and geophysical methods (Aizawa et al., 2009;Bertrand et al., 2013Bertrand et al., , 2022Cordell et al., 2019;Hacıoğlu et al., 2021;Heise et al., 2016;Marwan et al., 2021;Miller et al., 2022;Pavez et al., 2020;Tang et al., 2008) have been performed to estimate the potential flow routes in volcanic zones, with different depths of investigation. Most studies on porosity estimation have considered either laboratory measurements or numerical models; however, utilizing geophysical imaging together with laboratory data provides detailed and reservoir-scale estimation of physical properties with depths of investigation down to hundreds of kilometers when utilizing a broad range of frequencies.
The Mount Meager Volcanic Complex (MMVC), in southwest British Columbia, has one of the highest geothermal potentials in Canada (Grasby et al., 2012). Abundant thermal springs and high enthalpy geothermal potential based on borehole temperature data suggest the presence of permeable conduits channeling fluid beneath the MMVC . The MMVC has been studied since the 1970s, but drilling from the 1980s to the 2000s did not reach reliable permeable zones with adequate flows for economical production . Studies at Mount Meager have included geochemical methods (Clark et al., 1982;Ghomshei & Clark, 1993;Ghomshei et al., 1986;Huang, 2019;Proenza, 2012) and geophysical methods (Candy, 2001;Flores-Luna, 1986;Jones & Dumas, 1993;Nevin Sadlier-Brown Goodbrand Ltd. [NSBG], 1975, 1981b, hydraulic conductivity (Jamieson, 1981), and surface geological mapping (Read, 1977(Read, , 1990 to evaluate the capability of the geothermal systems to produce energy at the MMVC. A multidisciplinary project including gravity, passive seismic, deep and shallow magnetotelluric (MT), and bedrock mapping commenced in 2019 to increase the knowledge of the potential reservoir .
Electrical resistivity is a physical property which depends on pore geometry, rock composition, fluid content, and temperature. In geothermal systems, fault and fracture zones that contribute to fluid circulation (as a function of porosity and permeability) may exhibit electrical resistivity differences (Muñoz, 2014). Therefore, resistivity measurements provide a valuable tool for reservoir evaluation. Resistivity models derived from MT data can define subsurface structures based on surface measurements of natural perturbations in the Earth's electric and magnetic fields. Previous MT studies at Mount Meager defined a deep magma body beneath Mount Meager and its relationship to the potential geothermal system (Candy, 2001;Hanneson & Unsworth, 2022;Jones & Dumas, 1993). A more comprehensive audio-magnetotelluric (AMT) data set was collected in 2019 (Craven et al., 2020) to study the structures in the top few hundred meters below the surface of the MMVC. Our recent analysis (Hormozzade Ghalati et al., 2021) correlated a model derived from the AMT data set to the lithology and hydrothermal alteration based on temperature and geological logs. This model was also similar to that of Hanneson and Unsworth (2022) in the depth range where they overlapped. Hormozzade Ghalati et al. (2022) also provided possible interpretations of potential reservoir and caprock zones and suggested conduits toward thermal springs that could represent fracture networks.
Building upon our previous study (Hormozzade Ghalati et al., 2022), here, we establish empirical relationships between petrophysical data from the laboratory, that are incorporated into our AMT model to evaluate the spatial relationship between the main geological structures and fluid circulation in the potential geothermal reservoir. In addition to the 3-D electrical resistivity model, we have utilized laboratory core measurements of porosity and permeability and fluid chemistry data. We developed new petrophysical models relating porosity to permeability and salinity to electrical resistivity. We calibrated known petrophysical models based on the Mount Meager well log and laboratory-based rock physical data. Our new petrophysical models were then applied to the resistivity model to derive porosity and permeability values. Possible fracture zones were also characterized for the study area, which had not been performed to date. This study utilizes reservoir parameters such as porosity, permeability, electrical resistivity, fluid salinity, and temperature to define the permeable pathways that control the upflow of thermal fluids.

Regional Geology
Volcanic belts of western Canada are controlled by the motions between Pacific, North American, Juan de Fuca (JDF), and Explorer plates. The Pacific plate is moving away from, and the JDF plate is moving toward, the North American plate at a rate of 4-5 cm/year (Jones & Dumas, 1993;Riddihough, 1984). At the Cascadia subduction zone, which extends along western North America, the JDF Plate goes underneath the North American plate in a north-easterly direction (with an 8°-16° dip under Vancouver Island) (Jones & Dumas, 1993;Spence et al., 1985). This motion caused the creation of the Cascade Volcanic Arc, which constitutes a chain of volcanoes from BC to California. The Cascade Volcanic Arc is segmented into the Quaternary-aged Garibaldi Volcanic Belt (GVB) and the High Cascades in the north and south, respectively (Figure 1) (A. S. Wilson et al., 2018). The JDF Plate is almost 5 Ma beneath the GVB and 10 Ma beneath the High Cascades (Venugopal et al., 2020;D. S. Wilson, 2002). The division of this arc is also related to the left lateral strike-slip Nootka Fault (Figure 1) (Hyndman et al., 1979). The GVB is a glaciovolcanic arc that formed from the late Tertiary to Quaternary (Jessop, 2008;Souther, 1977). Mount Meager, the focus of this study with an approximate 2 million years of eruptive history, is one of the major volcanic structures within the GVB (Warwick et al., 2019).

Mount Meager Volcanic Complex Geology and Tectonic Setting
The MMVC, situated approximately 150 km north of Vancouver, British Columbia, Canada, is characterized by steep topography due to high uplift rates and erosion over the past 2.5 Ma (Huang, 2019). The geology mainly consists of Mesozoic fractured crystalline and metamorphic rocks (e.g., quartz diorite, granodiorite, dacite, and gneiss) (Moore et al., 1983). Therefore, the hydraulic conductivity and permeability at Mount Meager are mainly controlled by fracture porosity (Jamieson, 1981). When drilling through the basement rock, fractures were detected by an observed high amount of fluid loss. Adams and Moore (1987) proposed that the distribution of the Figure 1. A simplified map of the Juan de Fuca (JDF) Plate subducting beneath the North American Plate along the Cascadia subduction zone. The JDF Plate moves under the less dense North American plate. The black triangles show the volcanoes, and the red triangle shows Mount Meager with yellow arrows as the regional stress direction (after Venugopal et al. (2020) andUSGS (1905)). upward fluid flow was controlled by fault and fracture zones, dikes, and breccias related to volcanic activities. They noted that carbonates from hydrothermal brecciation might have filled fractures in the shallow reservoir, which could cause reduced permeability.
Generally, fault zones can act as conduits, barriers, or combined conduit-barrier systems to fluid flow (Caine et al., 1996). Significant faults at the MMVC include the Meager Creek Fault (MCF), No-Good Fault (NGF), Camp Fault, and Carbonate Fault ( Figure 2) (Fairbank et al., 1981). The MCF is an east-west striking normal fault that dips 45°-50° northward. It is the key structure controlling the geothermal system in the region, but its exact geometry is unidentified (Huang, 2019). A fractured damage zone, 100 m wide at the MCF, provides a pathway for fluid and is parallel to an impermeable core plane of the fault (Mak, 2014). The MCF seems to act as a seal preventing fluid flow across the fault and is considered as the southern boundary of the reservoir (Huang, 2019). The NGF, Camp, and Carbonate faults are perpendicular to the MCF. The NGF, with an approximately 90° dip and north-south strike direction, is the oldest structure in the area, and it is intersected by boreholes where vertical fractures were observed (Huang, 2019). The left lateral strike-slip Camp fault, with a 35° strike and an approximate vertical dip, intersects the eastern section of the MCF. Further, the Carbonate fault, with a 130° strike and 50°-60° SW dip in the southwest of Pylon Peak, is supposed to provide a flow path to the Carbonate cold spring (Proenza, 2012).
There are steeply dipping regional fractures with a northwesterly direction and local fractures that follow local earth stresses and faults in MMVC's granodiorite (Figure 2; Balfour et al., 2011). Although the main attitude of joints to the south of Mount Meager (the South Reservoir) has a 130°/60° SW orientation, several different attitudes exist within this area. East-west trending fractures dipping to the north (40°) exist along Meager Creek between the NGF and Angle Creek, and north-south vertical fractures exist along the NGF (Fairbank et al., 1981).
The distribution of thermal springs in southwestern British Columbia is primarily controlled by the main crustal-scale faults that provide a permeable pathway for hot water seepage to the surface (Grasby & Hutcheon, 2001). Springs and vents at the MMVC trend spatially northward, where rocks are fractured and dissected by faults (Bernard, 2020). The discharge of hot water at the Meager Creek hot springs is caused by a topographic rise in the bedrock (Hormozzade Ghalati et al., 2022;Jamieson, 1981). Warm springs are located along the trace of MCF, which shows the presence of permeable zones allowing the outflow of thermal fluid to the surface. Also, a few cold springs are at higher elevations and are controlled by the geological contacts between volcanic layers, basement rocks, and unconsolidated deposits (Mak, 2014). The locations of thermal springs and faults are shown on the geological map in Figure 2.

Geochemistry and Petrophysical Data
The ability of fluid to flow in a reservoir is affected by the physical properties of the rocks, including porosity, permeability, and pore connectedness. Porosity is the amount of pore space in a medium. Permeability is a measure of the porous medium's capacity to pass the fluid, and normally correlates positively with pore connectedness (Bernabé et al., 2010). Considering these essential properties will help in evaluating a reservoir's hydraulic conductivity and flow characteristics. This paper investigated 21 core samples (3.8 cm in diameter and 1.5-7.84 cm in length), which comprise volcanic and basement rocks from sites within the MMVC (shown in Figure 2). The samples were dried for at least 48 hr prior to taking measurements. Subsequently, the porosity and permeability of these samples were measured using the Core Test Systems AP-608 Gas combined permeameter and porosimeter in the Institut National de la Recherche Scientifique (INRS). Porosity was measured by applying Boyle's law to the samples under various pressures of helium. Permeability values were measured by applying Darcy's law under different confining pressures and considering the Klinkenberg correction for gas slippage (Klinkenberg, 1941).
In addition to the permeability due to rock matrix porosity, fracture porosity exists in the rocks studied. Quantitative knowledge about the effects of fractures on porosity and permeability can be obtained by applying the theoretical models to the available fracture sets at the MMVC. Geometrical properties of available fractures and the relationships between them are important factors in evaluating the characteristics of fracture networks (e.g., Seifollahi et al., 2014). Therefore, the measureable features (dip angle, dip direction, orientation, length, spacing, and aperture) are utilized in this paper. Fracture data were collected from surface outcrops in the study area, where fresh bedrock was exposed (Chen et al., 2020;Jamieson, 1981). Data analysis methods were used to study the spatial distribution of fracture properties for further interpretation of petrophysical and AMT models. Fracture porosity ( ) is calculated as: where is fracture aperture (m), and is spacing (m) in i and j directions (CMG, 2016). Moreover, permeability through a set of fractures ( in mD ) can be expressed using the following equation, where the fracture aperture is in μm, and porosity is in % (Tiab & Donaldson, 2016).
Porosity involves information such as pore connectedness and electrical resistivity of bulk rock and fluid. This interrelatedness of porosity and electrical conductivity ( = 1 where is electrical resistivity) can be defined using petrophysical models, such as Archie's law (Archie, 1942), Modified Archie's law (MAL; Glover et al., 2000), Hashin Shtrikman (HS) model (Hashin & Shtrikman, 1962), and Waxman Smits model (Waxman & Smits, 1968). Archie's law is an empirically derived relationship between porosity and electrical conductivity which is expressed as follows: where is porosity, m is an empirical constant for the cementation factor, is the bulk electrical conductivity of the host rock, and is the fluid electrical conductivity (S/m). MAL defines the relationship between electrical conductivity and porosity considering two-phased electrical conductivity as follows: where is the solid phase conductivity of the material (S/m).
The HS model can define the upper ( + ) and lower ( − ) bounds of electrical conductivity in a two-phase medium (solid and fluid): As previously mentioned, the next parameter that plays an important role in porosity estimation is fluid conductivity. Experimental models show complexity in estimating fluid conductivity ; therefore, we need to study fluid conductivity before modeling (Sinmyo & Keppler, 2016). Various studies have shown the importance of temperature and salinity in estimating fluid conductivity (Peacock et al., 2020;Watanabe et al., 2022). The resistivity of water can change depending on temperature, salinity, and the chemical origin of the water in a reservoir (Tiab & Donaldson, 2016). Various methods exist to define electrical resistivity in water samples, including direct measurement of fluid resistivity, electrical resistivity well logs, and chemical analysis.
In 2022, Meager Creek Development Corporation collected eight groundwater samples (from locations shown in Figure 2) (MCDC, 2022). Chemical analyses including groundwater temperature, pH, major ions, total dissolved solids, and direct electrical resistivity measurements were performed on the samples. Since we had some direct electrical resistivity measurements from the reservoir water in the survey area, we studied a total of 171 legacy fluid salinity data from boreholes (Wells MC-1, MC-2, and MC-3) and thermal springs (Proenza, 2012 and references therein). The measured ionic concentration of available salts in these samples is converted to the total salinity ( sp in ppm) using Equation 6 (Tiab & Donaldson, 2016). Total salinity can then be used to define the fluid's electrical resistivity by utilizing Equation 7: where sp is equivalent NaCl concentration, sii is the concentration of each ion, n is the number of ions in the solution, is the weighting factors for each ion, and 75 is water resistivity at 75°F. This equation will be used as a reference to be compared with the new relationship calibrated for the Mount Meager data set in this study.
The electrical resistivity of fluid varies with temperature (Hersir & Árnason, 2009;Ussher et al., 2000). The relationship between electrical resistivity and temperature of the fluid can be expressed using Hilchie's equation as follows: where 1 and 2 are the electrical resistivity of the fluid at temperatures of 1 and 2 in °C (Tiab & Donaldson, 2016).

Well Logs
Geophysical well logging provides information about the composition and physical properties of rocks and fluids around a borehole. Temperature, pressure, and geological logs are available for all boreholes. A Neutron-Gamma ray survey was performed on well M9-80D using Schlumberger equipment in November 1980 and was digitized and used in this study ( Figure 3) (NSBG, 1981a). The high energy neutrons emitted from the probe can be reduced by collisions with the hydrogen available in borehole fluids and rocks. Therefore, a neutron log is sensitive to the amount of hydrogen in a formation and can be used in porosity determination (Tiab & Donaldson, 2016). A permeable zone was identified based on the graphic geological logs and negative spikes in the neutron log, which represents a water saturated fractured rock. The simultaneous increase in gamma ray emissions, higher than quartz diorite host rock, was interpreted as a fractured volcanic rock. The well logs in permeable zones, which are fractured volcanic rocks, are used to calculate the cementation factor in this study using Equations 3 and 4.

Magnetotellurics
To complement the petrophysical studies at Mount Meager, surface geophysical methods were used to obtain a larger-scale estimation of the rock properties. Magnetotelluric (MT) methods utilize the Earth's natural electromagnetic field variations to determine the subsurface electrical resistivity structure. The MT transfer function, also known as the impedance tensor is a complex tensor that defines the relationship between the horizontal components of the Earth's electric ( and ) and magnetic ( and ) fields considering the angular frequency ( ) so that: Likewise, the geomagnetic transfer function, or tipper, is a complex vector that defines the relationship between the horizontal ( , ) and vertical ( ) components of the Earth's magnetic field (Chave & Jones, 2012;Simpson & Bahr, 2005): The real portion of the tipper function can be interpreted graphically using induction arrows, where lateral variations in resistivity can be inferred by induction arrows that point toward (Parkinson's convention) or away from (Wiese's convention) the conductive anomaly (Jones, 1986;Parkinson, 1959;Wiese, 1962). The induction arrows, here using Parkinson's convention, are plotted in Figure 4. Induction arrows are plotted on top of the determinant mode of apparent resistivity to give information regarding the structures that can be anticipated from the inversion. Figure 4 shows that the main axes of the induction arrows of the stations around the NGF point toward north-northwest. However, induction arrows of sites around the south of the section point toward the conductivity changes in the north or the shallower conductor (above the MCF).
In order to better define the fluid flow correlation with fault/fracture networks and permeable zones from modeled electric resistivity, we inverted the AMT data with faults as tear surfaces and compared the new 3-D model and with the original (default) regularized inversion model of the area (Hormozzade Ghalati et al., 2022). A number of inversions with tear surfaces were tested with different data, model and regularization parameters utilizing RLM3D. We built a new 3-D electrical resistivity model that utilized CGG's RLM3D, which employs a finite difference solver to calculate forward models and non-linear conjugate gradients in the 3-D inversion engine. 3-D orthogonal Cartesian cells are defined based on the survey geometry and topography. Topography is handled by incorporating the DEM horizon grid and elevation data in each EDI file. Triangulated surface mesh files are fixed in the model in the form of vertices and triangles that define the 3D shape and location of fault surfaces (Rodi & Mackie, 2001;Soyer et al., 2018Soyer et al., , 2020. The triangulated surface mesh of each fault was built in Leapfrog Geothermal using available structural information of faults from surface mapping and drillholes. We then assigned these triangulated mesh files as tear surfaces within the inversion mesh. This procedure of introducing tear surfaces allows for interruptions in regularization across the surfaces and sharp changes in electrical resistivity around the faults. In this new model, we used a subset of 70 AMT stations to focus on the area of faults, boreholes, and core sample data. The data were rotated counter-clockwise by 29.4° to optimize the rectilinear mesh discretization. To model the topography, the first 63 vertical cells were 30 m thick and subsequently increased by a factor of 1.06 for each layer thereafter. The core part of the mesh consisted of 94 (i) × 83 (j) cells of 75 m height and width (total area of 7,050 m × 6,225 m) ( Figure S2 in Supporting Information S1).
To fulfill the boundary conditions, we assigned horizontal paddings on all four sides of the core part to increase cell dimensions laterally by a factor of 1.5. The corresponding mesh consisted of 120 (

Results
As shown in Figures 5 and 6 Sensitivity tests were carried out to assess the robustness of these conductors in the new electrical resistivity model (Figures S10 and S11 in Supporting Information S1). Both conductors were replaced with structures of 100 and 300 Ωm resistivity to test these features. Depth assessment tests were performed to show the sensitivity of data to solve for the depth of the discussed structures. These tests show an increase in overall RMS of forward models and normalized determinant phase after introducing features of 30, 300, and 1,000 Ωm resistivity to model. Therefore, the AMT data set across the survey area shows an effective investigation depth of 3 km below the sea level, which is discussed here ( Figure S9 in Supporting Information S1).
To evaluate the fluid pathways revealed in the AMT models, we started by calculating the electrical resistivity of the fluid samples. To assess the reliability of the legacy fluid samples and chemical models, we used newly collected fluid samples. Fluids are usually classified based on the major anions. Therefore, we plotted the thermal fluid samples from hot springs and boreholes based on the major ions in the Cl-SO4-HCO3 ternary diagram in Figure 7 (Giggenbach, 1988).
As shown in Figure 7, there is consistency between new and legacy hot spring samples. This plot shows that the majority of water samples from boreholes, which are close to mature geothermal waters, are within the deep chloride classification. It is notable that samples that have a high concentration of Cl are fed from a reservoir and can be used to identify permeable zones (Shah et al., 2019).
A total number of 635 measured fracture planes focused on the study area are categorized into four different strike clusters, with 32.5% in the NE-SW direction, 41% in the NW-SE direction, 20% in the NS direction, and 6.5% in the EW direction. The distribution of strike direction of the measured fractures is shown in rose diagrams in Figure 8. Two histograms are plotted for each fracture cluster, which show the statistical distribution of dips and spacings. The dip angle and spacing followed normal and log normal distributions, respectively.

Discussion
Plan view slices through the constrained final resistivity model are presented in Figure 5. They can be compared with the default model presented in Figure S3 in Supporting Information S1. Major features in the default model are C1 and C2 with average resistivity of approximately 15 Ωm hosted in a background resistivity of approximately 100 Ωm. Conductors in the constrained model are at the locations the of previously defined C1 and C2 with approximately the same extension and direction. This conductivity is expected to map the low permeability clay-rich layers with the measured temperatures of 70°C-160°C from different boreholes, acting as caprocks sealing deeper hot fluids. The shallow zone in the boreholes is related to the argillic alteration minerals characterized by the presence of smectite, illite and rare kaolinite (Proenza, 2012; and reports listed therein). Besides the amount and type of clay minerals, another possibility for conductive features is the existence of high salinity fluid. Given the temperature and reported salinity of near surface and hot spring fluids, a porosity factor of over 50% is needed (which is a high porosity for our study area) to decrease resistivity to about 20 Ωm. Highly conductive features could have a remnant of thermal saline fluid from the reservoir region. As the temperature rises to above 200°C in the reservoir region, the resistivity in the AMT model increases, which is followed by the formation of chlorite and epidote minerals in the propylitic alteration zone detected in mudlogs (Proenza, 2012; and reports listed therein; Ussher et al., 2000).
The major structures in the Mount Meager area that control thermal manifestations are MCF and NGF with east-west and north-south directions, respectively. Based on the default resistivity model, surfaces with dips of 50°-60° and 80°-90° can be considered for MCF and NGF, respectively, which are correlated with the structural geology data. The intersection of NGF and MCF in our model is related to conductive zones and possible pathways for fluid flow. This relationship is supported by a shift in the electrical resistivity at the location of the faults. This shift was shown in the western and southern margins of the model, which coincided with the existence of faults shown by the plane and lines in Figure 6. The constrained model shows that the MCF bound the conductors above the fault. Both models showed lateral changes in electrical resistivity below Meager Creek and revealed a moderately resistive edge at the footwall of the NGF and a highly resistive edge at the footwall of the MCF. By increasing the depth, the contrast in resistivity disappears and based on the resistivity values, both sides of these faults consist of quartz diorite at depth. There was no change in electrical resistivity associated with the borders of the Camp and Carbonate faults in both models, with the models showing mainly a sequence of layers of increasing resistivity. Therefore, the Camp and Carbonate faults cannot be considered as reservoir boundaries.
In the new model with tear surfaces, the northward conductor along the NGF (black circle) is narrower and appears thinner compared to the default model in the same profile direction ( Figure 5). Since conductors in the model are constrained to the surface of faults, the damaged fracture zones around the hanging walls of MCF and NGF act as permeable paths for fluid transmission (yellow arrows). Furthermore, the NGF and MCF seem to act as barriers that prevent lateral and vertical flow, respectively. Sensitivity analyses are provided in the Supporting Information S1. These tests show that the faults' inclinations and conductors are well defined ( Figures S5-S8 in Supporting Information S1). We calibrated known petrophysical models based on the Mount Meager well log and laboratory-based rock physical data. Our new petrophysical models were then applied to the resistivity model to derive porosity and permeability values. Possible fracture zones were also characterized for the study area, which had not been performed to date. This study utilizes reservoir parameters such as porosity, permeability, electrical resistivity, fluid salinity, and temperature to define the permeable pathways that control the upflow of thermal fluids.
Fractures are often consequences of magma intrusions and faulting. When they are filled with hot water, they are typically observable in electrical resistivity models as low resistivity features. Therefore, we can evaluate thermal fluid pathway zones by integrating the AMT model features at the location   of fractures with porosity-permeability values. The loss circulation zones were encountered during drilling at various depths (Proenza, 2012, and references therein). These zones can represent the intersection of boreholes with high permeability faults and fracture networks in that depth range. The loss circulation zones are within the moderate resistivity zone, correlated with the 40-300 Ωm resistivity range.
Chemical models exist to determine the electrical resistivity from a fluid's total salinity. We used the newly collected samples (which have direct electrical conductivity measurements from the laboratory) to test the reliability of available chemical models. Therefore, the total salinity for all fluid samples (legacy and new) was calculated. Subsequently, the chemical model (Equation 7) was applied to the fluid salinity data (Table 1). To evaluate the chemical models, we applied the model to the newly collected fluid samples to compare the results with the direct electrical resistivity measurements from the laboratory. Moreover, we used the laboratory data to develop a relationship between a fluid's electrical resistivity and salinity for the Mount Meager area, and calibrated Equation 7 for our field area ( Figure 9). Figure 9 illustrates a trend between salinity and electrical resistivity of fluid samples measured in the laboratory at room temperature (which is corrected further in Table 1 for reservoir temperature) and the known relationship (Equation 7) as a reference.
The salinity in Figure 9 shows the total dissolved salts in the fluid (Tiab & Donaldson, 2016). The dissolution of these salts can provide free ions acting as charge carriers, increasing conductivity (Chave & Jones, 2012;Simpson & Bahr, 2005). Therefore, it is expected that a fluid's electrical resistivity decreases as salinity increases. This new relationship (Equation 11) was also applied to the legacy data (Table 1) Because of the effects of temperature on boreholes' electrical resistivity, we adjusted the results to the reservoir temperature. These electrical resistivity values were then corrected for temperature and used in further porosity-permeability calculations. A comparison of the calculated electrical resistivity from these two models (Equations 7 and 11) is shown in Table 1.
For most rocks, the value of the cementation factor lies between m = 1.5 and 2.5 (Cordell et al., 2018). To have reliable modeling of porosity based on the AMT electrical resistivity model, we estimated the cementation factor for our survey area. Bulk electrical conductivity in granitic rocks is mainly controlled by conduction through the pore fluid (Glover & Vine, 1994). Therefore, the neutron porosity log was calculated for the permeable zone of well M9-80D (Figure 3). This porosity log was then used to calculate the electrical conductivity based on Archie's law and MAL. We considered the electrical conductivity of granitic rocks to be 0.0007 S/m and the fluid electrical conductivity to be 2 S/m (based on the borehole temperature of around 70°C in that depth range). The calculated electrical conductivity of the permeable zone was correlated with the electrical conductivity from the AMT model at a cementation factor of 1.6.
The expected amount of fluid within the Mount Meager rocks can be estimated using Figure 10. We calculated the porosity based on the MAL (Figure 10a) and HS upper bounds (Figure 10b) for the potential fluid pathways. This figure shows the variation of fluid fraction with respect to the change in bulk resistivity of rocks at Mount Meager. Fluid resistivity calculations show a range of values from 0.15 to 0.8 Ωm for the borehole samples at the reservoir temperature. As previously mentioned, the average electrical resistivity of the pathways in the AMT model is between 40 and 300 Ωm. Considering these models, a porosity range of 0.1%-8.5% is expected for this range of bulk electrical resistivity.
We estimated permeability based on porosity. Generally, permeability and porosity have a positive correlation (Revil et al., 2020). Figure 11 shows the  positive correlation between porosity (%) and permeability (mD) in the data, using the following equations for the basement rock (Equation 12) and the volcanic rocks (Equation 13). basement = 0.0015 × 0.3181 (12) volcanic = 8 × 10 −9 × 8.0624 Considering Equations 12 and 13, we expect the permeability for the estimated range of porosity ( Figure 10) to be 0-0.249 mD ( Table 2). The permeability of the order 1.01 mD is required for natural geothermal convection at 200°C-250°C reservoir temperature (Straus & Schubert, 1977). However, by increasing the temperature, which reduces fluid viscosity, the necessary level of permeability for fluid flow can be reduced to 0.1 mD (at reservoir temperature in the order of 350°C) (e.g., Hanano, 2000Hanano, , 2004. Our data show that the permeability of the basement rock samples is less than 0.002 mD and its porosity values lie between 0.2% and 3.2%, whereas the volcanic samples demonstrate a wide range of porosity (2.6%-23.2%) and permeability (0.001-5,186.57 mD) values. This variation can be related to thermal, mechanical, or chemical sources of micro-structures, such as fractures and pore networks within the rock (J. Farquharson et al., 2015). Of note, permeability measurements of the core samples depend on the physical sample size (Clauser, 1992;J. I. Farquharson et al., 2016;Heap et al., 2017). Therefore, macroscopic fractures might not be captured and considered in these core samples. Because the major permeability in the MMVC is due to fracturing, we must consider the fracture information.
The fracture orientations of NW-SE (which matches the local maximum principal stress) and NE-SW are dominant in the study area. Spacing and aperture measurements demonstrate a wide range (Figure 8), which can lead to a wide range of fracture porosity and permeability (Equations 1 and 2).  Table 2, the surrounding rocks in this area show low permeability: less than 0.249 mD, which corresponds to the rock porosity estimated using either HS + or MAL. Considering the spacing and aperture information for the fracture  porosity-permeability calculations (utilizing Equations 1 and 2), we estimate that the porosity and permeability of the fracture zones can increase up to 40% and 666 mD, respectively (Table 2). This is consistent with other studies that investigate the dependence of fluid pathways in volcanic systems on fractures (Heap et al., 2017;Nara et al., 2011;Pavez et al., 2020;Walker et al., 2013).
Comparing the modeling results to the results of the laboratory measurements shows that the resistivity model can reasonably map the subsurface resistivity structures and can be used as a tool to estimate the porosity of geothermal reservoirs. Figure 12 illustrates the results of estimating the porosity-permeability from the 3-D AMT model. It is worth noting that the temperature dependency of pore fluid resistivity was considered in the porosity modeling ( Figure 12). Temperature is assigned to the model using 21 borehole temperature logs throughout the study area and extrapolated between boreholes. Fractures are highlighted in the models to show the regions with potential enhanced permeability. Modeling based on electrical resistivity overestimates the porosity and permeability in the caprock because it assumes that the high conductivity is related to the pore fluid rather than the clay surface resistivity. Therefore, we have highlighted contour features in this figure for a better understanding. Sensitivity tests on the effect of cementation factor on porosity are provided in the Figure S12 in Supporting Information S1.
Fluid injections have not shown flow between wells MC-6 and MC-7 the (Locations shown in Figure 2; GeothermEx Inc., 2005). Based on the direction of these boreholes, the fracture sets with an east-west direction may not contribute to the fluid flow and might have been filled with alteration minerals. Another set of fractures with strike direction perpendicular to the flow direction, which are not productive, exists between wells MC-6 and MC-7. In contrast, fluid injection tests show a permeable zone between wells MC-6 and MC-8 (GeothermEx Inc., 2005). This zone is correlated with fracture set 2 with a major northwest-southeast fracture direction. This direction is parallel to the regional maximum horizontal stress direction shown in Figure 1. This information is plotted on a conceptual model in Figure 13.  The temperature scale in Figure 13 is based on the interpolation and extrapolation of temperature logs. These data in the southern boreholes showed that temperature exceeded 200°C within the first 1 km below the surface. Meteoric water which penetrates and circulates within this hot rock zone is one of the sources of warm water in the thermal fluid manifestations. This water is then mixed with deeper older thermal fluids (Huang, 2019). Water samples from MC1 indicate that cold water is mixed with thermal water (Clark & Phillips, 2000). This was supported by chemistry and isotope data of discharged water in boreholes and thermal springs in the study area (Ghomshei & Clark, 1993;Huang, 2019).

Conclusions
In this paper, we have presented combined geophysical and petrophysical modeling to define the permeability of the MMVC. A new 3-D AMT resistivity model within which the faults are assigned as regularization tear surfaces was generated and the results, along with those from a previous unconstrained model, were studied together with available core and fluid samples and petrophysical models. Both models showed similar structures, but the constrained model demonstrated thinner conductors and sharper changes in electrical resistivity near the MCF and NGF. Both models showed that conductive features are bounded by two main local faults and that conductors extend downdip along the NGF and above the MCF. This finding suggests that (a) the damaged zone of NGF and MCF can act as local pathways transmitting hydrothermal fluids, and (b) an impermeable fault core surface restricts flow and exists along the southern and western boundaries of the reservoir. Both resistivity models show that the conductive structures on the west and east sides of the Camp fault are  (Jamieson, 1981), possible fractures, temperature distribution on two planes that intersect boreholes, and extrusive and intrusive breccia (not to scale) (after Ghomshei et al. (2013) connected. This connectivity is not continuous because of the northwest tendency of the conductors. The results show that the Camp fault has a permeable core surface or a largely damaged zone with fractures and cannot be interpreted as the eastern boundary of the South reservoir. The fault's damaged zone and fractures are shown in drilling and surface structural geology studies, which should be considered in the fluid flow modeling.
Borehole fluid samples with high content of Cl can represent their reservoir source and were utilized for electrical resistivity calculations. Taking into consideration the relationship between fluid salinity and electrical resistivity, the components of Archie's law are defined for the study area. Utilizing these components and core data, we optimized petrophysical models for the survey area and estimated the porosity and permeability using the 3-D electrical resistivity model presented earlier by Hormozzade Ghalati et al. (2022). We estimate that the reservoir related to the electrical resistivity range of 40-300 Ωm has 0.1%-8.5% of fluid content with a permeability of 0.00-0.249 mD.
Cold water springs exist in volcanic rocks at higher elevations. The volcanic rocks showed higher permeability in laboratory measurements. It can be interpreted that the fluid is slowly transmitted through higher permeability rocks to the surface. However, warm and hot springs are located along Meager Creek in the basement rocks, close to the location of local faults. Their spatial location is probably related to fractures and a fault-damaged zone.
Low permeability values measured in the lab suggest the importance of macrofractures in increased fluid flow to hot springs in the south reservoir.
Based on our interpretation, fractures with a northwest-southeast direction can be related to flow. This direction correlates with the extension of the moderate resistivity pathway zone of the AMT model. Considering the pattern of this fracture set for drilling purposes will increase the permeability up to 666 mD. It should be noted that because of the high degree of alteration shown in the AMT model and based on the geological and temperature logs, fractures might be filled with alteration minerals at locations with higher temperatures. However, fracture strike direction and dip information should be considered when drilling directional wells. Furthermore, permeability depends on the pore shape and may vary in the horizontal and vertical directions. Knowledge of the physical properties of rocks that contribute to permeability is essential when estimating the long-term viability of geothermal resources.