Introduction

Radiation naturally occurs and has been part of Earth’s environment since its inception. It is found in Earth’s crust, building materials, food, drinking water, and even the air we breathe. Humans have always been exposed to natural radiation sources such as Earth itself and cosmic radiation from space. Additionally, human activities also contribute to radiation exposure through X-rays for medical applications, the fallout from nuclear weapon testing, and the release of small amounts of radioactive materials from coal and nuclear power plants (NPPs) [1]. However, during the routine operation of a NPP, minimal amount of radioactivity, within regulatory limits, is discharged into the environment. Public exposure can occur internally through inhalation or consumption of radionuclides and externally from deposited activity or effluent plumes.

The safety of both occupational workers and the public is given the top priority in nuclear industries in India. The Kaiga Generating Station (KGS) in Kaiga, Karnataka, operates four PHWR (pressurized heavy water reactor) type nuclear reactors, each with a capacity of 220 MWe, where rigorous protocols limit the discharge of radioactive materials. Liquid effluents are treated to significantly reduce contamination levels, ensuring they are well below permissible limits. Gaseous effluents are filtered through HEPA (high-efficiency particulate air) filters before being discharged through a stack of height 100 m. Continuous emission monitoring from the facility ensures compliance with regulatory standards [2].

However, environmental radiological monitoring around NPPs is crucial to ensuring public and environmental safety. The Indian nuclear industry follows a systematic monitoring program, beginning with a comprehensive pre-operational survey around the NPP site. At KGS, a detailed pre-operational survey was conducted over a 30 km radius of the site to set baseline radioactivity levels [3]. Once operational, continuous environmental monitoring is conducted. The Environmental Survey Laboratory (ESL) equipped with highly sensitive instruments, operated by the Environmental Monitoring and Assessment Division, Bhabha Atomic Research Centre (BARC), oversees these efforts to ensure public radiation exposure remains within safe limits as set by national and international guidelines.

During normal or accidental operations, NPPs can release radioactive effluents in gaseous or liquid states. Tritium (3H; t1/2 = 12.3 years) is a major by-product of nuclear reactions in PHWR-type nuclear reactors at KGS, Kaiga, which use heavy water (D2O) as its coolant and neutron moderator [4, 5]. It can enter into the atmosphere and water bodies, making its monitoring crucial for assessing environmental impacts and public radiation doses. Other significant radionuclides, like 137Cs (t1/2 = 30.2 years) and 90Sr (t1/2 = 28.7 years), are also rigorously monitored due to their high fission yields and long physical half-lives, which allow them to remain active in the environment for several decades [6]. Additionally, 137Cs and 90Sr possess significant radiotoxicity and can accumulate in muscle tissues and bones of the organisms upon ingestion [7]. Consequently, their presence in the environment is of critical concern, and monitoring their levels is vital for assessing radiation risks to the public and the environment [6, 7].

The monitoring covers various environmental matrices such as air, water, sediment, soil, vegetation, and food samples, to detect any radioactivity levels exceeding natural background levels due to KGS operations. The measurement of gross-alpha and gross-beta activity in the air primarily serves screening purposes and helps in the long-term monitoring of radionuclide behaviors [8]. Samples from terrestrial environments like soil, grass and other vegetation act as trend indicators and markers of events near a nuclear installation. Vegetation and food samples including crops, vegetables, milk, and meat are the main dose-contributing samples of terrestrial origin. Meanwhile, samples of aquatic origin such as surface water, fish, and sediments cover the aquatic aspects and assess the impact on water bodies due to discharges from the NPP. These indicators are selected based on their relevance in assessing radiation exposure and environmental contamination stemming from NPP operations. This holistic approach to environmental monitoring and radiation assessment helps enhance the understanding of the environmental implications of nuclear energy production.

Globally, extensive research has been conducted to examine the environmental radiological impacts of NPPs, with numerous studies examining the long-term trends in environmental radioactivity levels [9,10,11,12,13,14,15,16,17]. For instance, a study at Qinshan NPP in China monitored the radioactivity in drinking water and the surrounding environment, as well as cancer incidence rates among the residents, over 9 years from 2012 to 2020 [10]. Similarly, research conducted at the Wolsong nuclear power plant in Korea analysed environmental radioactivity data collected from 1998 to 2010 to understand long-term accumulation trends [17]. These long-term assessments often employ trend analysis methods to evaluate the ongoing trends in environmental monitoring data. Trend detection techniques can be categorized as parametric and non-parametric. Parametric tests assume a normal distribution of data and require a large sample size [18]. In cases where the underlying distribution is unknown, non-parametric methods are more appropriate [19]. The present study utilized the non-parametric Mann–Kendall test, considering the inadequate knowledge of the parent distribution, which is widely used in analyzing the time series data related to environmental radioactivity [10, 17, 20,21,22].

Site description

The KGS is situated in Kaiga village within the Indian state of Karnataka, at Latitude 14.86° N and Longitude 74.44° E. On India’s southwest coast, this location is approximately 58 km east of the coastal town of Karwar. Kaiga is encircled by tropical forests, ranging in elevation from + 20 to + 700 m above mean sea level, and is home to a wide diversity of flora and animals. Rainfall in the area averages 3700 mm during 4 months from June to September. At Kaiga, the ambient temperature ranges from 13.3 to 41.0 °C, and the relative humidity varies between 17.7 and 99.9% [23]. Figure 1 shows the location of KGS, Kaiga.

Fig. 1
figure 1

Location of the KGS, Kaiga

Full size image

Environmental monitoring programme

Monitoring environmental radioactivity poses significant challenges due to the spatial and temporal variations of radioactivity levels within environmental samples. Therefore, it is imperative to analyze numerous samples from various locations and time periods to establish a statistically significant concentration profile of radioactivity. The collected environmental samples were categorized into three groups:

  1. 1.

    Samples such as food/crops, vegetables, drinking water, air, etc., relevant for the estimation of impact to the members of the public.

  2. 2.

    Samples such as soil, grass, leaf, and sediment, serve as trend indicators, and

  3. 3.

    Samples such as goat thyroid, accumulate specific radionuclides to a greater extent (e.g., 131I), serving as a sensitive indicator.

The sampling locations around KGS were identified based on considerations such as prominent wind direction, atmospheric dilution factor, proximity to the site, population centres, availability of samples, and accessibility. The region around KGS was divided into four different circular zones (beyond the exclusion zone of 2.3 km) namely zone-1 (Z1: 2.3–05 km), zone-2 (Z2: 05–10 km), zone-3 (Z3: 10–15 km) and zone-4 (Z4: 15–30 km). The samples were collected from various locations within these zones, covering a radial distance of up to 30 km from KGS. The zone-wise sampling map around KGS is shown in Fig. 2. All samples were collected and analyzed using standardized procedures [24].

Fig. 2
figure 2

Zone-wise sampling map around KGS, Kaiga

Full size image

For air sampling, monitoring included gross-alpha, gross-beta, 3H, and gamma-emitting radionuclides like 137Cs and 131I activity. Particulate samples were collected on the glass-fibre filter paper through continuous air sampling at a flow rate of 50 L per minute, utilizing the air sampler installed in the pre-dominant downwind sector (WNW) at ESL, Kaiga (12.3 km, in Z3). Activated-charcoal impregnated filter papers were specifically utilized for detecting 131I in air samples. Weekly cumulative air particulate samples were examined for long-lived gross-alpha and gross-beta activity using low-level counting systems. Additionally, gamma spectrometry was employed to identify and quantify gamma-emitting radionuclides such as 137Cs, 131I, etc. These measures could be especially effective in detecting short-term releases of particulates from the operational NPP [11]. Tritium levels in the air were measured by condensing ambient air moisture using the cold finger method, followed by analysis with a liquid scintillation analyzer (LSA) [4, 24].

In the aquatic environment, the monitoring program included sampling of water, freshwater fish, and sediment samples from the Kadra reservoir (or downstream), a liquid effluent discharge body for KGS. Fish samples were collected monthly from the local vendors and sediment samples were taken from Hartuga village (a nearby population centre) in zone-1. Water samples from all the zones were analyzed for 137Cs, 90Sr and 3H activities, with tritium concentrations determined via LSA following standard procedures [24].

Samples from terrestrial matrices including cereals, grass, soil, leaves, and various dietary items like milk, meat, leafy vegetables, diet, and eggs were collected from the surrounding villages (based on accessibility and availability, and therefore zone-wise sampling was not applied for the samples of plant and animal origin). Subsequently, these samples were analyzed for 137Cs and 90Sr activities. Additionally, goat thyroid samples were collected from local shops at Kadra village, near Kaiga Township in zone-3, and analyzed for 131I concentration using gamma spectrometry.

During the study period, all the gamma-emitting radionuclides were assessed using a gamma spectrometric system, and 90Sr was measured using low-background beta counters, following radiochemical separation. Gamma-emitting radionuclides in aquatic and terrestrial matrices were analyzed utilizing HPGe detector with 15% relative efficiency coupled to a PC-aided 8K MCA. The laboratory currently utilizes a 50% relative efficiency HPGe gamma spectrometer, improving the detection limits over the years.

Trend analysis method

The study utilized the Mann–Kendall (MK) statistical test, a non-parametric method widely used to detect monotonic trends in environmental data series [10, 17, 22, 25,26,27,28]. The MK test involves two hypotheses; the null hypothesis (Ho), which suggests no trend, and the alternate hypothesis (H1), which indicates the presence of a significant increasing or decreasing trend. The fundamental principle of the MK test involves assigning specific values of 1 for an increase, − 1 for a decrease and 0 for equality while computing the difference between later and former measurements, (Xj − Xi), where j > i [29, 30]. This can be expressed in Eq. (1) below,

$$sgn\left( {X_{j} - X_{i} } \right) = \left\{ {\begin{array}{*{20}l} 1 \hfill & {if\; \;X_{j} - X_{i} > 0} \hfill \\ 0 \hfill & {if\, \;X_{j} - X_{i} = 0} \hfill \\ { - 1} \hfill & {if \;\;X_{j} - X_{i} < 0} \hfill \\ \end{array} } \right.$$
(1)

The MK test statistic (S) is calculated as the sum of integers with Eq. (2):

$$S = \mathop \sum \limits_{i = 1}^{n - 1} \mathop \sum \limits_{j = i + 1}^{n} sgn\left( {X_{j} - X_{i} } \right)$$
(2)

where n is the total number of data points, while Xi and Xj correspond to i = 1, 2,…, n − 1, and j = i + 1,…, n. A large positive value of S suggests that the measurements taken at later times are greater than those taken earlier, indicating an upward trend. Conversely, a significantly negative S value indicates that later values are lower than earlier ones, suggesting a downward trend. Instead, a small absolute value of S does not signify a noticeable trend. When n > 8, the statistic S approximates a normal distribution, and under the condition that the mean of S is 0, the variance V(S) of the statistic S is obtained by Eq. (3) [31, 32]:

$$V\left(S\right)= \frac{n\left(n-1\right)\left(2n+5\right)- {\sum }_{k=1}^{m}\left({t}_{k}-1\right)\left({2t}_{k}+5\right)}{18}$$
(3)

where m represents the number of tied groups, and tk signifies the number of ties for the kth value. The standardized normal test statistic Zs (z) is computed with Eq. (4),

$$Z_{S} = \left\{ {\begin{array}{*{20}l} {\frac{S - 1}{{\sqrt {V\left( S \right)} }} } \hfill & {if\; \;S > 0} \hfill \\ 0 \hfill & {if\;\; S = 0} \hfill \\ {\frac{S + 1}{{\sqrt {V\left( S \right)} }}} \hfill & {if\; \;S < 0} \hfill \\ \end{array} } \right.$$
(4)

The z-values that are positive or negative indicate an increasing or decreasing trend, respectively. Further, the statistic S is closely associated with Kendall’s coefficient τ, as expressed by Eq. (5),

$$\tau = \frac{S}{D}$$
(5)

where,

$$D= {\left[\frac{1}{2}n\left(n-1\right)-\frac{1}{2}\sum_{k=1}^{m}{t}_{m}\left({t}_{m}-1\right)\right]}^\frac{1}{2}{\left[\frac{1}{2}n\left(n-1\right)\right]}^\frac{1}{2}$$
(6)

The τ coefficient ranges from − 1 to + 1 and is similar to the correlation coefficient used in regression analysis. To determine the presence of a significant increasing or decreasing trend at a significance level α, the test compares |z| to the critical value z(1 − α/2). For this study, α is set to 0.05. The p-value lower than 0.05 (p < 0.05) affirms the presence of a statistically meaningful trend [17]. At α = 0.05, z > 1.96 implies a significant increasing trend, while z < − 1.96 denotes a significant decreasing trend in the monitoring data.

To determine the magnitude (change per unit time) of the trend, Sen’s slope estimator was used [33, 34]. The slope of a time series with N pairs of data, Xi = x1, x2,..., xn, is estimated using Eq. (7). The median of N values of βi yields the Sen’s estimator of slope β.

$${\beta }_{i}= \frac{{x}_{j}- {x}_{i}}{j- i} , \, for \, \, all \, i <j \, and \, i=\text{1,2},\dots , N$$
(7)

The study utilized the ‘R’ library package “trend” to perform these statistical tests [35]. The trends were categorized as ‘increase’, ‘decrease’, or ‘no trend’, where ‘no trend’ indicates no statistically significant consistent direction [17].

Results and discussion

The MK test was exclusively applied to cases where the activity concentration exceeded the minimum detectable activity (MDA), and the trend analysis was conducted using the geometric mean of sample results. However, in certain samples, including diet, meat, egg, fruit, vegetables, and indicator organisms such as goat thyroid, the concentration levels were below MDA. Consequently, these samples were excluded from the trend analysis.

Trends of radioactivity in air

During the study period, the annual average concentrations of gross-alpha and gross-beta activity in airborne particulate samples varied from 0.02 to 0.3 mBq m−3 and 0.3 to 2.6 mBq m−3, respectively. The observed activity levels were comparable to the pre-operational values [3]. Typically, the sources of gross-alpha and -beta activities in the air are primarily the long-lived daughters of gaseous radon (222Rn). In particular, gross-beta activity is mainly attributed to 210Pb (t1/2 = 22.2 years) and 210Bi (t1/2 = 5.01 days), while gross-alpha activity largely arises from 210Po (t1/2 = 138.4 days) [8].

During the mid-2000s, the elevated concentrations of gross-alpha and gross-beta activity were likely due to increased dust load in the air, resulting from the ongoing construction activities in the surrounding areas of the sampling location. The trend analysis results for these activities showed a decreasing trend as presented in Table 1 and Fig. 3. However, activity levels of 137Cs, 90Sr, and 131I were below MDA in all air particulate samples, and hence no trend analysis was performed for these radionuclides.

Table 1 Results of MK test of gross-alpha(1) and gross-beta(2) activity
Full size table
Fig. 3
figure 3

Gross-alpha and gross-beta activity in ambient air (WNW, Z3)

Full size image

The trend analysis results for 3H-in-air activity, outlined in Table 2, indicated a statistically significant decreasing trend within the monitoring zones Z2 and Z3 (z < − 1.96 and p < 0.05), while no discernible trend (z > − 1.96 and p > 0.05) was observed in zones Z1 and Z4. The trend graphs for 3H activity in the air are shown in Fig. 4. Over the period from 2005 to 2020, the annual mean activity concentration of 3H in zone-1 (Z1) ranged from 0.2 to 1.2 Bq m−3, which was significantly below the regulatory limit of 6036 Bq m−3 for the annual average derived air concentration (DAC) of tritium at the site boundary (2.3 km) of KGS [36]. Furthermore, there was a noticeable decrease in activity levels with distance from the plant throughout the study period.

Table 2 Results of MK test of 3H-in-air activity
Full size table
Fig. 4
figure 4

Trends of the annual average concentration of 3H-in-air activity around KGS

Full size image

Trends of radioactivity in water samples

The trend analysis results for 3H activity in water samples of downstream liquid effluent discharge water bodies across various zones are presented in Table 3. The results (z < − 1.96 and p < 0.05) indicated a statistically significant decreasing trend across the monitoring zones, Z3, and Z4. Conversely, no trend (z > − 1.96 and p > 0.05) was observed in zones Z1 and Z2. The corresponding trend graphs of the annual mean activity concentrations are shown in Fig. 5. During the period from 2005 to 2020, the activity levels of 3H in downstream water in Z1 ranged from 10.2 to 23.6 Bq l−1, significantly lower than the acceptable limit of 880 Bq l−1 set by the regulatory authorities and far below the World Health Organization’s (WHO) guideline of 10,000 Bq l−1 for 3H in drinking water [36, 37]. These findings indicate a significant decrease in 3H activity over the years of continuous operation of KGS.

Table 3 Results of MK test of 3H activity in water
Full size table
Fig. 5
figure 5

Zone-wise trends of annual mean activity concentration of 3H in downstream water around KGS

Full size image

Furthermore, water samples in various zones were analyzed for 137Cs and 90Sr activities, which were found to be below the detection limits of 2.0 mBq l−1. Consequently, no statistical analysis was conducted for 137Cs and 90Sr activities in water. It indicates that radioactivity levels in liquid effluents discharged from the KGS are negligible.

Trends of radioactivity in dietary matrices

137Cs and 90Sr, both radionuclides of significant biological concern due to their long half-lives and biological affinity, were monitored in various dietary items around the KGS. The activity of 90Sr in all dietary matrices and 137Cs concentrations in the samples of diet, meat, fruit, eggs, and vegetables were all found to be below the MDA (0.05 Bq kg−1 edible weight). As a result, the MK test was not performed for these samples. However, 137Cs activity was detected in the cereals, milk, and fish samples. The occurrence of radiocesium (134Cs and 137Cs) in the environment is primarily due to fallout from nuclear weapons testing beginning in 1945, major nuclear accidents like Chernobyl (1986) and Fukushima (2011), and potentially small amounts released during normal NPP operations [38, 39]. The shorter half-life of 134Cs (t1/2 = 2.06 years), compared to 137Cs (t1/2 = 30.2 years), results in its relatively lower persistence in the environment, making it difficult to detect after a few decades [38]. Since 134Cs activity was not detected in these samples, the source of 137Cs activity in these matrices is solely from historical fallout.

The trend analysis results of 137Cs activity concentrations in these dietary items, as presented in Table 4, showed either no trend or a decreasing trend in radioactivity levels.

Table 4 Results of MK test of 137Cs activity in dietary items
Full size table

Figure 6 presents the annual mean concentrations of 137Cs activity in dietary items, indicating no increase in radioactivity levels attributable to plant operation. However, no statistically significant trends were observed in the 137Cs activity in milk and fish samples (z > − 1.96, p > 0.05). The variations in radioactivity levels in milk could be attributed to several factors affecting its environmental transfer through the soil-grass-cow-milk pathway [40]. In the case of fish samples, variability may occur due to the analysis of different fish species, having distinct habitats and feeding behaviors [41].

Fig. 6
figure 6

Trends of annual mean activity concentration of 137Cs activity in dietary matrices around KGS

Full size image

Analysis of trend indicator samples

The results of the MK test for 137Cs activity (Bq kg−1 dry wt.) in trend indicator samples such as grass, leaves, soil, and sediment are outlined in Table 5. The detailed analysis revealed a decreasing trend in grass samples (z < − 1.96 and p < 0.05), whereas no significant trend was detected in leaf, soil, and sediment matrices (|z|< 1.96 and p > 0.05). The plots of annual average values of 137Cs activity concentration are shown in Fig. 7.

Table 5 Results of MK test for 137Cs activity in trend indicating samples
Full size table
Fig. 7
figure 7

Annual mean activity concentration of 137Cs in trend indicator samples

Full size image

The levels of 137Cs activity in sediment samples exhibited higher values in recent years. As 134Cs was not observed in downstream water bodies, the source of 137Cs activity in sediment is fallout origin only. Kaiga being a tropical forest area getting heavy rainfall with an annual mean of 3700 mm, physical processes, such as erosion could be the major cause of the redistribution of 137Cs in soils and its movement from soil to reservoir sediment. The hilly topography of this region may also contribute to the soil erosion process. Therefore, the observed activity buildup in sediment or redistribution of activity in soil is likely due to natural phenomena, specifically soil erosion. It is concluded that this trend has occurred due to natural phenomena altering the natural levels of radioactivity, rather than the effect of KGS operations. The 90Sr activity was found to be below MDA (0.5 Bq kg−1) in the grass, leaf, soil, and sediment samples and as a result, the trend analysis was not performed for 90Sr activity in these matrices.

Conclusions

In conclusion, the analysis of environmental radioactivity data (2005–2020) collected around KGS revealed significant insights. The outcomes from the long-term trend analysis, statistically carried out using the Mann–Kendall test, indicated either a decreasing trend or no trend in activity levels across aquatic, atmospheric, and terrestrial matrices. Trace levels of 3H were observed in the atmospheric and aquatic environment. The observed activity levels were much less than the acceptable derived limits prescribed by the atomic energy regulatory body. According to MK test, there was evidence suggesting a decreasing trend in the various zones at 5% significance level.

In the case of 137Cs activity, it was detected in some dietary components. However, the observed levels were solely attributed to natural fallout origins. Notably, 137Cs activity exhibited a decreasing trend in the cereals and grass samples. However, no significant trend was detected for 137Cs activity in milk, fish, leaf, soil, and sediment samples. Furthermore, the 90Sr activity remained below detection levels across all the matrices studied. These comprehensive findings confirm that no significant radioactivity material was released from KGS during the study period and there is no accumulation of radioactivity in the surrounding environment. The study provides substantial evidence affirming the safety of both the local public and the surrounding ecosystem.