Assessment of physicochemical and radon-attributable radiological parameters of drinking water samples of Pithoragarh district, Uttarakhand

This study evaluates the quality of drinking water samples (sample size = 52) taken from various locations of Pithoragarh district, Uttarakhand. The parameters include physiochemical properties viz. total dissolved solids (TDS in mg/L), electrical conductivity/salinity (µS/cm), pH and radiological dose attributable to radon in water (µSv/y). TDS values for the tested samples varied within the range of 18–434 mg/L with average value of 148 mg/L. Electrical conductivity and pH for these samples was measured as 36–868 µS/cm (average: 296 µS/cm) and 6.8–8.2 (average: 7.2), respectively. Radon activity concentration for these water samples was measured using scintillation-based radon monitor, immediately after sampling at the location site. Radon activity concentration was measured as 0.6–81.9 Bq/L with an average value of 17.8 Bq/L. The paper also estimates the annual effective ingestion dose (µSv/y), annual effective inhalation dose (µSv/y) and total effective dose (µSv/y) attributable to radon in drinking water samples.


Introduction
Water quality is one of the important parameters determining the health risks to humans. Measurement of various characteristics of water, associated interpretations and hazard estimations have been performed worldwide at different places of relevance. It is a fact that numerous contaminants pollute water sources and this impacts human health risks. The source of contaminants could be natural or anthropogenic and the level of degradation of water quality depends on their source-term. Radioactive contamination is also a serious issue around the globe. While the mineralogy acts as the source term for physicochemical parameters, natural radio-nuclides present in the earth crust acts as the main source for radiological impurity of ground and surface water [1]. Total dissolved solids (TDS), salinity/electrical conductivity and pH are three vital physicochemical parameters which vary for different water sources and also get affected due to human activities. Similar to the air matrix, radon/ thoron gases and its decay products primarily released from the natural radio-active precursors (decay chains) determines the radioactivity concentration profile of water. Rn 222 and Rn 220 are two isotopes of radon with different half-life (Rn 222 has a half-life of 3.82 days and Rn 220 has a half-life of 55.5 s) [2]. Due to its gaseous nature, it travels through cracks and pores between the grain in the soil and easily escape into the ground and surface water. Previous studies in other parts of Uttarakhand Himalaya indicate the presence of a significant amount of natural radioactivity in the air, soil, and water [3][4][5][6][7][8][9]. Pithoragarh is a high-altitude terrestrial region and groundwater is the main source of water to be used for domestic as well as industrial purposes. Most of the population in the region depends on natural spring sources for drinking water [10]. Rn 222 emits alpha particles and its short-lived progenies (Po 218 , Po 214 , Pb 214 and Bi 214 ) interact with atmospheric particles [11]. Previous studies in Kumaun Himalayas reported radiological dose levels due to Rn 222 activity concentration [12][13][14][15]. Such measurements, their usage in the estimations of radiological doses and the interpretations are highlighted as an important data-base by standard international commissions [16,17]. Therefore, measurement of radioactivity concentration and the estimation of radiological doses due to the ingestion of Rn 222 has been done by several researchers in past as well [18][19][20].
This study focuses on the measurement of physicochemical properties and Rn 222 concentration in drinking water samples of Pithoragarh district, Uttarakhand, India. Rn 222 concentration was measured using a portable Rn 222 monitor on the spot of sampling right after the sample is collected. The physicochemical properties (TDS, Electrical Conductivity, and pH) were measured in the laboratory. Statistical analysis of the obtained data is discussed in detail in this work.

Geology of study area
Pithoragarh is the easternmost district of Uttarakhand state of India. It confines higher Himalayan Mountain ranges, passes, valleys, alpine meadows, forests, waterfalls, rivers, glaciers, springs, and snow-capped peaks. The geographical area of Pithoragarh is 7,110 square Km. The district is located in the Kumaun division of Uttarakhand. The northern part of the district is attached to the Tibetan plateau and Nepal is on its eastern border that is separated by the Kali river also known as Sharda river. The elevation from sea level ranges between 500 and 6400 m. Variation of temperature is considerable from area to area depending upon the altitude. The lower Himalaya is confined by the Main Boundary Thrust (MBT) in the southern part and Main Central Thrust in the Northern part. It consists of the late Proterozoic to early Cambrian sediments intruded by some granites and volcanic rocks [21]. The rock types in this land are quartzites, siltstone, carbonates, Phyllite, schist with subordinate impure marbles, metamorphic rocks, and orthogneisses [22]. Almost 50% of the area is occupied by snowy Himalayan Mountain ranges in the north forming several glaciers in the region. Therefore, at lower reaches, numerous natural springs are available for domestic and irrigation purposes. The district receives a good amount of rainfall, which infiltrates into the ground through the soil matrix and recharges the groundwater level [23]. During the postmonsoon period, the water level reduces gradually affecting running water inventory for the pre-monsoon phase. The radon exhalation and water discharge rate are largely influenced by the meteorological parameters like pressure, temperature, relative humidity etc. Radon level in water was found to be higher in monsoon and post-monsoon season in comparison to pre-monsoon season in a study conducted in Garhwal region [24]. The concentration of radon is linked to water quality in terms of radiological considerations. This study was performed for the post-monsoon season only and the results are interpreted in terms of water quality characterization using the results obtained from the measurements made in this season. Whereas physicochemical properties represent the parameters mainly for post-monsoon season, annual inferences are also made for radiological dose estimations. As the radon concentration values remains in midlevel (maximum in pre-monsoon and minimum in monsoon period) range, these can be assumed as representing yearly averaged values for calculations.

Experimental techniques and methodology
Sample collection 52 water samples of drinking water were collected from different villages of the study area from various sources like natural springs (49) and hand pumps (3) in the month of December. Radon concentration profile follows a seasonal pattern due to the change in the water level as the function of rainfall. Radon concentration is expected to be at maximum level in the summer season while it reduces in monsoon period due to the increase in water level. The water samples that have been taken from the locations represent water which is used by the local population for domestic purpose as well as for drinking. Water samples were collected in polypropylene double capped sampling bottles. The bottles were washed thoroughly prior to sampling to reduce external contamination. Before pouring the sample into the bottles, a container is used to collect the water. The polypropylene sample bottle was then immersed into the container. One end of a polypropylene pipe is immersed into the opening of water source. After ensuring that no air bubble is forming inside the pipe, the other end of the pipe is purged into the sampling bottle, replacing the water in the sample bottle. To collect water sample from the hand-pump, the pipe was plunged inside the pumping cylinder of hand-pump up to the water level and hand-pump lever was operated to pump out water. After ensuring that there are not any bubbles forming inside the pipe, the other end of pipe then plunged inside the sample bottle that was immersed inside the water in large container. The sample bottle was capped inside the container to prevent the atmospheric contaminants. This sampling methodology was used to reduce the chance of instant degassing of radon while pumping the water out from the handpump, and to reduce external contamination from the atmosphere. After sampling, these water samples were analyzed for Rn 222 on the spot with a portable Rn 222 monitor.
After the Rn 222 measurement, water samples were taken to the laboratory for physiochemical analysis (Fig. 1).

Measurement of Rn 222 concentration
Water samples were analysed for Rn 222 concentration with Smart RnDuo using a bubbler kit. Smart RnDuo is an online active Rn 222 monitor that is developed by Bhabha Atomic Research centre, Mumbai. The working of Smart RnDuo is based on the alpha scintillation method. An Ag:ZnS coated cell (Lucas cell) is connected to a photomultiplier tube that records scintillation counts when an alpha particle hits the wall of the lucas cell [25]. The basic set-up for the measurement of radon concentration in water involves 50 ml glass sample bottle attached to a bubbler. The bubbler is connected to the radon monitor using polypropylene pipes and a particle filter. The air pump is used to circulate the air within the sample through the Lucas cell to obtain escaping Rn 222 from the sample inside the Lucas cell. For 15 min of the measurement cycle, a pump is set to on for 5 min and for the rest 10 min, the alpha count continues. After completing one cycle, average Rn 222 activity concentration is displayed on the monitor. Figure 2 shows the schematic diagram for the setup for the measurement of Rn 222 activity concentration using the bubbler method [26]. This instrument gives Rn 222 concentration in the air volume having unit Bq/m 3 . To calculate Rn 222 activity concentration in water volume (C liq ) one can use the following Eq. 1 for the setup [27].
where C air is the Rn 222 activity concentration of air given by the Rn 222 monitor, K is the partition coefficient (i.e. K = 0.25 for water and air), V air and V liq is the volume of air present in the setup and volume of liquid sample present in the sampling bottle, respectively. (1)

Measurement of Physiochemical parameters
The measured physiological parameters for the water samples i.e., Total Dissolved Solids (TDS), pH, and electrical conductivity due to salinity were measured using standard analytical methods available at Nuclear Research Laboratory, USERC, Dehradun. The Oakton CON 550 Benchtop Conductivity/TDS meter kit was used for the measurement of TDS and Electrical Conductivity of the water samples. The instrument measure electrical conductivity/TDS though a conducting probe, and has conductivity measuring range of 0-200 mS/cm with an accuracy of ± 1.0% full-scale ± 1 digit, and TDS measuring range of 0-100 g/L with a resolution of 0.1 mg/L. TDS is the measurement of the total weight of solids dissolved in the water sample volume that is expressed in parts per million or milligrams per unit volume. pH shows whether the sample is alkaline or acidic. It is a negative logarithm of hydrogen ion concentration. For pH measurement, GeNai pH digital meter-117800 GB is used in the laboratory. It has pH range of 1-14 with 0.5 ± 1 digit accuracy pH [28][29][30].

Estimation of radiological dose
Radon dose received by the population through ingestion of water can be changed with some factors as time delay between pour and drink, pour method, boiling of water before drink. In this study the dose was calculated by assuming that water is consuming as such it pours from the source. The annual effective dose due to ingestion and inhalation of Rn 222 has been calculated using the parameters established in the reference [31,32].
Annual effective inhalation dose can be calculated by the Eq. 2: where, E w.Ih is the annual effective inhalation dose, C R n w is the Rn 222 activity concentration in water (Bq/L),R a.w is the ratio of Rn 222 in air and water (0.1 Bq.m −3 /Bq.L −1 ) [33], F is the equilibrium factor between Rn 222 and its progenies (0.4), O is average annual indoor occupancy time per person (7000 h/y) and DCF is the dose conversion factor due to Rn 222 exposure i.e. 9 nSv/(h.(Bq/m3))[34].
Age-dependent annual effective ingestion dose was calculated using Eq. 3 as follows [35]: where, AED ing is the age-dependent annual effective ingestion dose(µSv/y), A is the activity concentration of Rn 222 in water (Bq/L),w ing is age-dependent annual average water intake by the local population (L.y −1 ) and DCF ing is the agedependent dose conversion factor (nSv/Bq), and their values are given in the references [36,37].

Result and discussion
The statistical parameters (i.e. minima, maxima, mean, standard deviation, skewness, and kurtosis) calculated for Rn 222 activity concentration (Bq/L), TDS (mg/L), pH, Electrical conductivity (µS/cm), age-dependent annual effective ingestion dose (µSv/y), Annual effective ingestion dose (µSv/y), Annual effective inhalation dose (µSv/y) and Total annual effective dose (µSv/y) in Table 1. The frequency plot regarding the data obtained for Rn 222 activity concentration in collected water samples is shown in Fig. 3a. The frequency plot indicates that more than half (54%) of samples have Rn 222 activity level below 10 Bq/L and in 56% samples radon level is below 11.1 Bq/L (recommended level by USEPA) [40]. All of samples have Rn 222 levels within the safety limit recommended by WHO (i.e. 100 Bq/L). To obtain the best fit distribution among Lognormal, Exponential, Gamma and Weibull, K-S (modified) test and A-D test has been performed on radon dataset, and goodness-of-fit statistical parameters are shown in Table 2. Results shown in Table 2 and Fig. 3 suggest that except exponential distribution, all other distributions (Lognormal, Gamma and Weibull) captured the trend of the variation of the data [41,42]. Figure 3b demonstrate the probability plot generated for the radon dataset. Lognormal, Gamma and Weibull dataset lie over the fitted reference line, indicating good fit for the data. However, in the Exponential probability plot, data points outreach the reference fitted line which indicates rejection of Exponential distribution over the radon dataset.

Physicochemical parameters
Total dissolved solids (mg/L) values of the samples occur in the range between 18 and 434 mg/L with an average value of 148 mg/L. From frequency plot (Fig. 4a), it can be seen that; the TDS concentration was below 50 mg/L in 37% samples and in 97% samples its value was below 300 mg/L. Also shown in Table 1, this dataset is not uniform with − 0.2 kurtosis and lightly-tailed or non-uniformly random. Goodnessof-fit statistics using K-S (modified) and A-D test has been performed over TDS dataset, and calculated parameters are shown in the Table 2. The goodness-of-fit statistics suggest that Lognormal, Gamma and Exponential distribution are not followed by the TDS dataset. However, Weibull distribution is accepted as best fit among assumed Distributions. Figure 4b demonstrate the probability plot of the TDS dataset is rejected for Lognormal, Gamma and Exponential distribution. However, it is observed that the K-S (modified) test signifies that data-points are within the reference fitted line which confirms that Weibull distribution is best fit for the dataset (Table 2), meanwhile Lognormal, Gamma and Exponential distributions are rejected for the TDS dataset.

3
The pH values of collected samples range from 6.8 to 8.2 with a mean value of 7.2. Thus, the samples are slightly alkaline but safe for drinking and other domestic purposes. Skewness and kurtosis values show asymmetry of the dataset distribution which indicates a lack of external parameters to affect the pH of the water samples. Electrical conductivity varies from 36 to 868 µS/cm, with a mean value of 296 µS/ cm.

Radiological Dose assessment due to Rn 222 exposure
The age-dependent annual effective ingestion dose (AEI g D) estimated due to Rn 222 exposure to the local population is shown in Table 1. For infants (age below 1 year), the annual ingestion dose (µSv/y) varies from 2.7 to 376.6 µSv/y with an average value of 81.7 µSv/y. These values estimated for the children aged between 1 and 3 years vary within the range of 0.9-125 µSv/y with an average value of 27.2 µSv/y. For children aged 4-8 years, the value of AEI g D ranges between 1.0 and 144.9µSv/y with an average value of 31.4µSv/y. For toddlers aged 9-13 years, the average value of AEI g D is 21.8µSv/y within the range of 0.7-100.3µSv/y. For teenagers who come in the age group of 14 to 18 years, the range of AEI g D occurs between 1.2 and 171.9 µSv/y with the mean value of 37.3 µSv/y. For the adults, the AEI g D ranges from 1.5 to 209 µSv/y with an average value of 45.4µSv/y. The AEI g D values for infants and children are higher than for teenagers and adults because of the high value of dose conversion factors for infants and children. High dose conversion factors for infants and children indicate that this age group is at relatively higher risk due to similar exposure of Rn 222 compared to other age groups. For the study area, these dose values are well within the safe limits prescribed by various agencies [41]. Therefore, water samples analyzed for this region are safe for drinking for all age groups. Figure 5 shows the violin plot for annual effective ingestion dose for the various age groups. The violin-like graph shows the kernel densities of the dataset of age-dependent annual ingestion dose. The white dots between the densities are the median value. It can be seen that for infants the median as well as the span of the dataset is highest.
The annual effective Ingestion dose (µSv/y) due to Rn 222 exposure through drinking water, annual effective inhalation dose (µSv/y) due to inhalation of Rn 222 escapes from the water and total annual effective dose (µSv/y) are also represented in Table 1. The annual effective ingestion dose ranges between 1.48 and 209.20 µSv/y with the mean value 45.37 µSv/y. The annual effective inhalation dose occurs in the investigation within the range of 1.46-206.33 µSv/y and the average value is 44.76 µSv/y. The total annual effective dose due to Rn 222 exposure ranges between 2.94 and 415.54 µSv/y with a mean value of 90.14 µSv/y. Figure 6 illustrates the violin plot of annual effective ingestion dose (µSv/y), annual effective inhalation dose (µSv/y) and total annual effective dose (µSv/y), respectively. The white dots in between quartiles indicate the mean value of the data and the red line demonstrates the median value. The mean value is in the second quartile for all the doses which usually happens when most of the data occur at low values. The total dose increases when the Rn 222 concentration increases. It is also demonstrated in the study that although children drink less water than adults yet they receive more doses due to high dose conversion factor proposed for infants. Similar conclusions are also made by other studies in past [32,36].

Spatial analysis
This study is a field survey; hence, it is imperative to analyze the dataset over terrestrial land, with available data points. QGIS 3.4 LTS software is used for spatial analysis of Rn 222 activity concentration (Bq/L) and TDS (mg/L). It is to be noted that the sampling of water in the study was carried out on point locations of the map. The latitude and longitude were also taken by GARMIN-GPS during the measurement of the location of sampling. The pointwise data is not enough to illustrate areal distribution. Therefore, the inverse distance weighting (IDW) interpolation method is used to calculate the approximate values on the neighbouring locations where the data was not taken during fieldwork. It is an adaptive interpolation method that distributes the data values over a spatial area. The simplest weighting function that is used to interpolate data over the area is the inverse of the distance between two points with measured data on the location of sampling. This is a good method when need to apply for fewer location points otherwise kriging should be used for interpolation [43]. However, the accuracy of the predicted values for regions other than the sample nodes increases with the increase in sample size. The study represents the spatial distribution based on chosen data nodes only for indicative purposes.  Figure 7b demonstrates the predicted Rn 222 activity concentration values in water samples for the study region. It is observed that the northern area shows high values of Rn 222 activity concentration in water samples although TDS is comparatively lower than 200 mg/L in the water samples taken from those nearby locations. The northern area of Pithoragarh is high altitude Terrain. More studies are needed to investigate higher Rn 222 concentration values observed for high altitude regions. The observance of higher Rn 222 activity concentration in low altitude regions (e.g. sampling node close to Dharchula) indicate the possibility of transportation of minerals by natural spring water, aquifers and rivers.

Conclusion
In this study, 52 samples of drinking water were collected from Pithoragarh district of Uttarakhand, India. The samples were investigated for Rn 222 activity concentration (Bq/L) level, TDS concentration (mg/L), Electrical Conductivity (µS/cm) and pH. Age-dependent radiological doses to local population due to radon ingestion and inhalation due to water intake were also calculated and are compared with reference parameters recommended by ICRP (2002). All of the water samples have lower Rn 222 activity concentration level than the reference limit recommended by European Commission (2001) and WHO (2008). The Rn 222 activity concentration dataset follows Lognormal, Weibull and Gamma frequency distribution when tested with K-S (modified) and A-D tests for goodness-of-fit. Most of the samples have TDS below 250 mg/L which indicates that the samples are of good quality (mineral-wise). For the case of frequency distribution of TDS, Weibull distribution was found to be best fitted as per the tests for goodness-of-fit. On average, the tested water samples were found to be slightly alkaline. Higher agedependent annual effective ingestion dose values for the infants reflect the impact of higher dose conversion factors. In 97% samples the estimated dose values were found to be lower than the recommended levels, wherever applicable. The spatial distribution analysis demonstrated that the water samples from the southern region have higher TDS (mg/L) values and the Northern region have higher Rn 222 activity concentration (Bq/L) levels.