A radon survey was selected to detect the presence of permeable zones (faults and fractures) in the geothermal field (Gingrich, 1984), that may act as pathways for fluid movement in the study area. This methodology has been adopted for geothermal exploration in many countries such as New Zealand (Whitehead, 1984), USA (Fleischer, 1988), Mexico (Lopez et al., 1987; Balcazar et al., 1993), and Indonesia (Phuong et al., 2012; Haerudin et al., 2013). A comprehensive methodology was adopted to carry out radon measurements in soil and water.

For soil-gas radon measurements, a sampling grid covering an area of 1km2 was designed and plotted around the hot springs (Fig. 2). In total, 36 equally spacing (200m) sampling points were contemplated to fulfil the purposes of sampling. RAD-7 Electronic Radon Detector (DURRIDGE Co., USA) was used for the measurement of radon concentration in the Tattapani area. A stainless soilgas probe was used for in-soil radon measurement. Alpha particle emission activity in the soil-air sample was recorded using RAD-7. Air was pumped into the instrument and allowed to circulate for 20min to record the alpha particle emission measurement. Measurement for radon in water was made on-site for samples collected from an open well, hot springs and cold springs. The RAD H2O technique of RAD-7 was used to calculate the concentration of radon levels in water samples. This took about 30min.

Pre-set grab samples from RAD-7 were taken in single runs with four cycles of 5min each. Mean radon concentration was calculated averaging the radon concentration measured in all four cycles using the following equation

where C1, C2, C3 and C4 are the radon concentrations of the respective cycles.

For the chemical and isotope analysis, eight hot springs were selected along with water samples from the nearby river, dug well and fresh water springs. Water samples were collected in high quality Polyethylene (PET) plastic bottles: 200ml for cations, anions, silica (SiO2), and isotopes (2H, 18O), and 1L for 3H. Samples collected for cation analysis were acidified with 1% HNO3 (yielding pH<2) to avoid precipitation upon storage. Samples collected for SiO2 were diluted (1:1) with deionized water to prevent silica polymerization (Arnórsson, 2000). Samples of river water were filtered using filter paper (0.45μm).

Physico-chemical parameters such as temperature, pH, flow rate, Total Dissolved Solid (TDS) and Electrical Conductivity (EC) were determined in the field. Temperature and pH measurements were made using Crison pH-meter PH25+, which was calibrated by using three buffer solutions of pH 4.00, pH 7.00 and pH 9.00. Electrode was placed into the water sample and left until thermal equilibrium was established. Readings were recorded up to 0.01 pH-units together with the temperature of the water. For EC and TDS, a multimeter was used. Electrode was washed with deionized water before and after the measurement to ensure the precision.

Chemical analyses were carried out through Atomic Absorption Spectrometry (AAS) using a Varian Flame Atomic Absorption Spectrometer. A direct read out is given for concentrations of Na, K, Ca and Mg (Giggenbach and Goguel, 1988) in mg/l. Concentration of silica (dissolved Si in water) was measured using a UV–visible spectrophotometry. For concentrations of major anions (e.g. HCO3, Cl and SO4), the water samples were analysed using the methods described by APHA (2005) and Watson (1978).

The stable isotope compositions (2H and 18O) in water samples were measured by mass spectrometry. Using the CO2 equilibration method (Epstein and Mayeda, 1953), 18O values of water were analysed relative to V-SMOW with a standard error of ±0.1‰. The tritium content of the samples was determined by liquid scintillation counting after electrolytic enrichment of the water samples with a standard error of ±1TU (Hussain and Asghar, 1982).

Out of 36 radon sampling locations, 6 points were located over the loose gravel (fluvial), 7 on the shaley rocks, 7 on the sandstone of the Murree Formation, 4 on the carbonates of the Muzaffarabad Formation and 12 on the thick soil covering the underlying rocks (Table I). Radon concentration varied between 2.1 and 29.5KBqm-3 with a mean value of 7.62KBqm-3 and standard deviation of 6.23KBqm-3 (Table I). Concentration of radon was highest at sampling point A3 and lowest at C5. Sampling points A3, B3, F3, and F4 showed high anomalous values relative to those of other locations. Variations in radon emission from different rocks and soils depend upon the permeability and structural components (faults / fractures) along with various climatic aspects (Mogro-Campero and Fleischer, 1977). Studies conducted worldwide have revealed a persistent low radon level emission from Earth’s crust to atmosphere except for sites of geological fault, uranium deposits, volcanoes and geothermal sources (Nishimura and Katsura, 1990). In the present study, the radon concentration was relatively low at sample points located on or near the river bank. Gravel deposited along the river bank was not consolidated, so the large interconnected pore spaces between the sediments possibly provide pathways for surface air to dilute the gases escaping from the subsurface (Table I).

Soil gas radon data have been statistically distributed into three populations: low (less than 6.23KBqm-3), high (between 6.23 and 13.84KBqm-3) and anomalous concentrations (between 13.84 and 21.46KBqm-3; Table I). Spatially, two zones with anomalous radon concentrations were identified quite opposite to each other on the contour map, i.e. one in the South and the other in the North (Fig. 3). However, a N-S trending zone can be assumed if both anomalies are joined together. The N-S trending zone with high and anomalous radon concentrations happened to be located just above the contact between the Muzaffarabad and Patala formations where an igneous intrusion was observed near the secondary site (Fig. 4). Generally, the process of out-gassing through crust is a fracture controlled phenomenon, which varies according to the fracture spatial distribution (Sac et al., 2011). Once released from source, radon tends to migrate upwards. The movement of radon and its concentration in soil is influenced by a number of factors including the permeability of rocks, porosity and humidity, atmospheric pressure, micro-cracks, surface winds and granulations (Toutain and Baubron, 1999). A possible interpretation for these anomalous zones in the study area could be related to the presence of pathways along the contact allow that the movement of fluids from depth to surface.

Data from six traverse lines of the grid (A, B, C, D, E and F) that run across the Tattapani anticline was plotted to evaluate the effect of lithological changes over radon concentration as shown in Figure 4. Radon releasing capacity of rocks and soil (Morawska and Phillips, 1993), host rock permeability and flow of carrier gases (CO2, CH4, N2 and He) were found to be the influencing factors for in-soil radon concentration, analysed at surface conditions (Ball et al., 1991). The increase in radon concentration is associated with the escape of soluble radon from water at high temperatures. The shales of the Patala Formation have prominent radon concentration due to their high permeability. However, radon concentration is higher towards the shales exposed on the southwestern limb of the anticline (Fig. 4), indicating the presence of a subsurface source along the shale and carbonate contact, from which radon and other carrier gases (CO2, CH4, N2 and He) are following a direct escape route.

FIGURE 3

Contour map showing the spatial variability of radon concentration over the study area. The empty diamond symbols show the location of the radon sampling points.

Radon levels in the water samples were relatively low as compared to the ones in the soil. Thermal water had higher radon concentration than cold water samples. The highest concentration of radon was 9.1KBqm-3 for THS- 6 (hot spring) and the lowest was 5.89KBqm-3 for river water (Table IV). Solubility of radon in water depends upon temperature, decreasing by increasing temperature (Ball et al., 1991). Nevertheless, despite of its lower solubility in thermal water at higher temperature, the concentration of radon in the hot springs was higher than in the fresh water because it has been migrating to surface via thermal water.

The results of the physical analyses of water samples from the study area are shown in Table II. Hot spring waters were slightly acidic with pH ranging 6.62 to 6.69, and temperature ranged from 54.8ºC (THS-8) to 60.8ºC (THS- 6). Flow rates were variable from one to another among hot springs, between 3.6 and 15.6Ls-1. Hot springs with larger discharge showed the highest temperatures, EC and TDS values, ranging from 1115 to 1260μS/cm and 730 to 810ppm, respectively. Water samples other than hot springs were slightly alkaline with lower values of EC, TDS and temperature. Dug well and fresh water samples from cold spring had values of EC and TDS relatively higher than river samples because of their interaction with rocks. Dug well sample (TDW-1) showed abnormal values (Table II) which could be caused by thermal water recharge.

Analysis of major anions showed that the hot springs have relatively higher proportions of bicarbonate than sulphate and chloride. Ternary plot based upon concentration of major anions (Cl-, SO4 -2 and HCO3-) shows that the Tattapani hot springs are of bicarbonate type (Fig.5). Piper plot shows that major cations like alkali metals (Na and K) dominate the composition of the hot springs (Fig. 6). Water samples other than hot springs have relative high concentrations of Ca and Mg. Based on Piper plot, the Tattapani hot springs can be classified as Na-KHCO3, while the water from the open well can be classified as Ca-Mg-SO4-Cl (Fig. 6). All other water samples are alkali bicarbonate type.

FIGURE 4A

A) Variations in soil gas radon concentration along 6 traverses of grid moving across the Tattapani anticline cross-section.

FIGURE 4B
B) Index map showing geology as well as location of traverses A to F over Tattapani anticline.

Geo-thermometers based on total dissolved silica (e.g. quartz, amorphous silica and chalcedony) or on the relative concentrations of cations (e.g. K-Mg, Na-K, Na-K-Ca) in the thermal water can be used to calculate subsurface geothermal reservoir temperatures (Fournier, 1981). The equations of the geothermometers used in this study are as it follows:

Quartz (Truesdell, 1976)

(2)

FIGURE 5

Cl-SO4-HCO3 ternary diagram (after Giggenbach and Goguel, 1989) showing that Tattapani hot springs are bicarbonate type.

Quartz no steam loss (Fournier, 1977)

(3)

Quartz max steam loss (Fournier, 1977)

(4)

Na-K (Giggenbach, 1988)

(5)

K-Mg (Giggenbach, 1988)

(6)

K-Ca (Tonani, 1980)

(7)

Na-K-Ca-Mg (Giggenbach, 1988)

(8)

FIGURE 6

Piper plot showing the distinctive chemical nature of thermal water of Tattapani area.

Silica geo-thermometer is based on the solubility of quartz in the solution, which is highly dependent upon the temperature of fluid. According to Fournier (1989), the assumptions of silica geo-thermometer are based upon equilibrium between fluid/quartz and lack of fluid mixing during the ascending of fluid to the surface.

Reservoir temperatures based on quartz (conductive) geo-thermometer using equation (3) range between 96.01 and 122.53ºC, and between 97.54, and 120.31ºC based on quartz (adiabatic) geo-thermometer using equation (4) (Table III). The calculated reservoir temperatures suggest a near equilibrium with respect to quartz and undersaturation with respect to amorphous silica, indicating that concentration of quartz in the fluid is controlled either by adiabatic (max steam loss) or conductive (no steam loss) cooling processes.

The cation geo-thermometers assume an equilibrium state between the fluid and the minerals of the host rock. Results of the Na-K geo-thermometers using equation (5) show that the temperature ranges from 462.11 to 520.90ºC which could not be possible in this system. Such high temperatures might imply that equilibrium has not yet attained in case of Na-K exchange reactions at reservoir conditions, indicating the occurrence of dissolution of rocks (Truesdell, 1976; Fournier, 1979; Tonani, 1980; Arnórsson et al., 1983; Neiva et al., 1987). K-Mg requires less time than Na-K to reach equilibrium, making it more suitable for the current study. Temperatures calculated using equation (6) for K-Mg geo-thermometer, range between 111.21 and 115.16ºC (Table III). Reservoir temperatures calculated using the K-Ca geo-thermometer equation (7) are higher, ranging between 179.34 and 184.23ºC, while Na-K-Mg- Ca geo-thermometer equation (8) resulted in temperature ranging between 90.37 and 93.18ºC (Table III). Figure 7 shows a Na-K-Mg ternary diagram (Giggenbach, 1988) indicating that the thermal, cold and river waters fall into the category of immature waters. Therefore, re-equilibrium, mixing/dilution and/or dissolution of minerals associated with these cations could be the governing processes in such a system due to water-rock interaction (Reyes et al., 2010), as discussed in the next section.

Waters of Tattapani hot springs show a negative correlation between δ2H and chloride (Fig. 8A) indicating a mixing process between thermal water with low deuterium and high chloride concentration and shallow groundwater with high deuterium and low chloride concentration. Low deuterium ratio in the majority of thermal springs indicates conductive heat loss, whereas a relatively high deuterium ratio in THS-1, THS-6 and TDW-1 indicates heat loss by steam (Truesdell et al., 1977). Figure 8B shows a negative correlation between δ18O and chloride, while the δ18O vs temperature plot (Fig. 8C) indicates mixing of 18O enriched shallow groundwater with thermal water. The fact that the majority of samples lie above the trend line indicates that conductive heat loss is due to waterrock interaction (Ahmad et al., 2002). Figure 9 shows an inverse relationship between tritium and electrical conductivity which also indicates that a conductive cooling process dominates over the mixing one.

FIGURE 7

Na-K-Mg Ternary plot used to determine reservoir temperature and equilibrium between cations for geo-thermometry.

Indication of mixing is provided by plots of concentrations of tritium and chloride versus temperature (Fig. 10), and silica and sulphate concentrations as a function of chloride concentration (Fig. 11) which show a good linear relationship (Ahmad et al., 2002). Tritium has a negative correlation with both the surface temperature and chloride (Fig. 10) which indicates a mixing of thermal water containing high chloride and low tritium concentration with groundwater with low chloride and high tritium concentrations. The correlation between silica and chloride concentrations (Fig. 11) suggests neither precipitation nor dissolution of quartz after mixing (Arnórsson, 1985). This hypothesis was confirmed by the absence of secondary silica precipitation around the hot springs. Figure 11 also shows a positive correlation between concentrations of sulphate and chloride, which indicates the absence of precipitation and/or dissolution of minerals containing these species during the flow of the thermal water from the reservoir up to the surface.

The extent of mixing between thermal and groundwater is calculated by assuming that the tritium concentration in the thermal waters before mixing is zero. The mixing percentage is then estimated using the following expression (Sammel and Craig, 1981; Giggenbach et al., 1983; Chatterjee et al., 2016):

(9)

where X is the percentage of the mixing, Ths is the tritium concentration (in TU) of the thermal water and Tgw is the tritium concentration (in TU) of the local ground water. Table IV shows the concentration of tritium in the thermal, ground and river water of Tattapani area. Hot springs THS-7 and THS-8 show maximum mixing (8.68%) and THS-1, THS-5 and THS-6 show no mixing, as concentration of tritium in the latter samples was below detection limit. This indicates mixing of thermal water with different proportions of shallow groundwater before reaching the surface. Fraction of mixing can vary depending upon the season and subsurface geology of the area (Ahmad et al., 2007). It is inferred from the values in Table IV that the mixing of thermal water with groundwater cannot be the only phenomenon for the thermal water cooling.

A silica concentration versus temperature diagram (Fig. 12) shows that thermal waters lie between the solubility curves of amorphous silica and quartz suggesting that thermal waters are supersaturated with respect to quartz. Quartz supersaturation in thermal waters strongly indicates conductive cooling during up flow from reservoir to the surface (Ahmed et al., 2000).

FIGURE 8A

A) Relationship between chloride and δ2H in the thermal water and shallow groundwater.

FIGURE 8B
B) Relationship between chloride and δ18O in the thermal water and shallow groundwater.

FIGURE 8C
C) Relationship between temperature and δ18O in the thermal water and shallow groundwater.

Conductive cooling can be estimated using the following expression (Levet et al., 2006):

(10)

where Tadiabatic is water temperature of adiabatic mixing, Trt is the estimated reservoir temperature, Tgw is temperature of shallow groundwater and x is the percentage fraction of groundwater in thermal water. Then the extent of conductive cooling is the difference between Tadiabatic and Ths the measured surface temperature of the hot spring (Table IV). The results show that conductive cooling range between 33.21 and 58.9ºC, which is a significant portion of heat loss in the Tattapani hot springs. Thus, it is deduced that surface temperature of thermal water is mainly controlled by conductive cooling along with some mixing of shallow groundwater controlled by seasonal fluctuations.

FIGURE 9

Relationship between electrical conductivity and tritium content of thermal water and shallow groundwater.

Isotopic signatures are used to differentiate the possible origin (magmatic, oceanic and meteoric) of thermal water (Faure, 1991). δ18O and δ2H values of all water samples from the Tattapanihot springs lie between -6.72 to -6.14‰ and -39.37 to -36.43‰, respectively (Table IV). This isotopic data does not indicate any significant contribution of magmatic water which generally has δ18O from +6 to +9‰ and δ2H from -40 to -80‰ (Pearson and Rightmire, 1980; Giggenbach, 1992). Normally, ocean water has δ18O and δ2H about 0‰ (V-SMOW). Depleted values of δ2H and δ18O rule out the contribution of oceanic water. These values lie close to the Local Meteoric Water Line (LMWL) of this area (Hussain et al., 1995), indicating a possible recharge from meteoric water. The δ2H and δ18O values of the open well water lie close to those of thermal spring water (Fig. 13), which is possibly being recharged by the seepage of deep water of hot springs. The possible source of recharge of thermal and river water could be similar (e.g. meteoric) due to their δ2H values (Fig. 13). However, 18O concentration shifts due to rock-water exchange at higher temperatures as a result of which water of hot springs seems slightly less depleted in 18O (Fig. 13). The source of geothermal springs is probably meteoric water, feed by rain and/or snow melts from northern high altitude areas.

FIGURE 10

Relationship between temperature, chloride and tritium concentrations for thermal and groundwater.

FIGURE 11

Relationship between chloride, silica and sulphate concentrations for thermal and groundwater.

Tritium (3H) with half-life of 12.4 years can be used for dating and tracing young groundwater that has been recharged within the last 50 years. Tritium content in the thermal waters of Tattapani area is found to be below detection limits (i.e. 0.3TU; Table IV) indicating a recharge period of more than 50 years. Tritium content in the water sample from the dug well in the area is also below detection limit, suggesting a recharge from thermal water. Instead, cold spring water has considerable tritium content (7.95TU), which indicates significant contribution of young water in the recharge. Although the H and O isotopic data indicate a meteoric water recharge, the absence (or negligible levels) of tritium in the Tattapani geothermal springs suggests an age of more than 50 years for the thermal water (Ahmad et al., 2001).

FIGURE 12

Silica concentration as a function of temperature. The plot shows heat loss of Tattapani hot springs (solid symbols) due to conductive cooling.

FIGURE 13

δ2H versus δ18O plot using delta values of oxygen and hydrogen relative to LMWL (Hussain et al., 1995) and GMWL (Craig, 1961).