3.1. Activity Concentrations and Correlation of Radionuclides
Out of all the sites, only one sample from S8 displayed a
137Cs activity concentration of 2.58 Bq/kg. This level was considerably lower than the soil radioactivity in the region affected by radiation fall-out from the nuclear weapon testing and lower than that caused by the Chernobyl NPP accident [
22,
23]. The
137Cs concentrations affected by radiation fall-out are normally in the range of 15–30 Bq/kg [
13]. For example, the reported mean
137Cs concentrations of soil range from 7.1 Bq/kg to 30.5 Bq/kg [
2,
4,
23,
24,
25,
26]. In Taiwan, the average detected soil
137Cs concentrations were less than 10 Bq/kg in the north east area [
16,
27], with the monitoring area located in the south tip of Taiwan which is less affected by radiation fall-out.
The activity concentrations of
40K,
226Ra,
228Ra, and
232Th of the soil samples are listed in
Table 3. In all soil samples tested, there were detectable concentrations of
40K,
226Ra,
228Ra, and
232Th. The activity concentrations of
228Ra and
232Th showed a strong positive correlation (r = 0.97,
p < 0.001, as seen in
Figure 2a). The average activity ratio of
228Ra/
232Th was 0.96 ± 0.10, with the ratio close to 1.0 indicating that the thorium series was in a state of secular equilibrium [
28,
29]. Since
228Ra is the progeny of the thorium decay series, the
232Th activity concentrations were used to represent the thorium series in this study. Additionally, as shown in
Figure 2a, there was a strong positive linear correlation between
232Th activity concentrations and
226Ra (r = 0.79,
p < 0.001).
Additionally,
Figure 2b shows that there was a strong positive correlation (r = 0.89,
p < 0.001) between the activity concentrations of
232Th and
40K, as well as a positive correlation (r = 0.66,
p < 0.001) between
226Ra and
40K. This correlation suggests that these radionuclides may have come from the same original sources and that they coexist at the sampling sites [
6,
23,
30]. It is also possible that these nuclides respond similarly to the soil’s geochemical behavior and to other environmental processes, as they are commonly distributed in the environment [
10,
14,
30]. Some studies indicate positive correlations among the nuclides
226Ra,
232Th, and
40K at each concentration [
10,
14,
23,
30].
The activity concentrations ranged from 77.2 Bq/kg to 517.7 Bq/kg for
40K, 3.9 Bq/kg to 31.6 Bq/kg for
226Ra, and 5.3 Bq/kg to 39.1 Bq/kg for
232Th. The mean and standard deviation values were 344.4 ± 116.1 Bq/kg, 18.6 ± 5.5 Bq/kg, and 26.5 ± 8.4 Bq/kg for
40K,
226Ra, and
232Th, respectively. The activity concentrations of the radionuclides in the present study were lower than the world soil average activity concentrations of 420 Bq/kg, 32 Bq/kg, and 45 Bq/kg for
40K,
226Ra, and
232Th, respectively [
14]. In these soil samples,
40K had significantly higher activity concentrations than
232Th and
226Ra, while
232Th had significantly higher activity concentrations than
226Ra (
p < 0.001). The abundance of potassium in the earth’s crust [
7,
31] means that
40K was responsible for the majority (82.2–90.4%) of the activity concentrations found in the soil samples [
1,
11,
18,
29,
32]. The levels of
232Th and
226Ra in the soil are impacted by the minerals present [
1,
3,
7,
32]. Some soil samples may have higher concentrations of
232Th, while others may have higher concentrations of
226Ra [
4,
18,
33,
34].
The levels of activity in the soil samples S14 and S15 were notably lower than those in the other samples. These two soil samples were collected near the coastline of the Bashi Channel, and their composition may be similar to that of the beach sand surrounding the Maanshan NPP, as shown in
Figure 1. The beach sand is made up of debris from shells and coral reefs, resulting in significantly lower activity levels than soil. The white beach sands contain predominantly a mix of coral reef, shell, and light minerals, hence natural nuclides have low activity concentrations. Activity concentrations ranging from 10.7 Bq/kg to 58.3 Bq/kg, from 2.7 Bq/kg to 7.1 Bq/kg, and from 2.8 Bq/kg to 11.0 Bq/kg for
40K,
226Ra, and
232Th, respectively, have been found in beach sands [
35,
36,
37]. When mixed with soil, the sand causes a considerable decrease in activity levels.
Soil activity concentrations vary across the world and are affected by factors such as mineral composition, particle size, geological conditions, and environmental processes specific to each region [
5,
29,
38,
39,
40]. The size of soil particles in the samples varied extensively, as presented in
Table 1. The average percentage of clay was 35.0 ± 14.5%, with a coefficient of variation (CV) of 41%. On the other hand, the mean percentage of sand was 47.4 ± 18.7%, with a CV of 39%. The sampled area was situated in the Maanshan Formation, which comprises sedimentary layers ranging from the Pliocene to the Pleistocene. These layers consist of limestone clasts, blue–gray shale, and gray sandstone, all of which contain diverse fossils such as shellfish and foraminifera [
10]. The sedimentary soil found in this area has low concentrations of natural radionuclides [
7,
11].
Maanshan NPP nears the coastline of the Bashi Channel, and the sampling area is relatively small, covering less than 10 km
2. Based on the variance in radioactivity, the sample seemed to contain a mix of soil, debris from shellfish, and foraminifera. Tsai et al. [
11] measured the radioactivity of soils near nuclear power plants and storage facilities in Taiwan. Specifically, two soil samples were obtained from the vicinity of the Maanshan NPP. The radioactivity of one sample was 21.1 Bq/kg, 25.8 Bq/kg, and 268.3 Bq/kg and the radioactivity of the other sample was 3.6 Bq/kg, 4.0 Bq/kg, and 73.6 Bq/kg for
238U,
232Th, and
40K, respectively. The high radioactivity of the soil sample was similar to the average radioactivity in the current study and the low radioactivity sample was similar to the radioactivity measured in samples S14 and S15 in the current study. Hence, the activity concentrations were lower than in those in the soil from other areas in Taiwan [
1,
4,
11,
27,
33,
41,
42,
43].
3.2. Descriptive Statistics and Frequency Distribution of Activity Concentrations
The descriptive statistics were used to infer the types of frequency distributions of the radioactivity [
6,
18,
43,
44]. The statistical parameters included the mean, standard deviation (SD), minimum, maximum, median, coefficients of variation (CV), skewness, kurtosis, and
p-values from a normality test. The computing methods for skewness (
G1) and kurtosis (
G2) followed Equations (5) and (6) [
45]:
where
g1 and
g2 are traditional measures of skewness and kurtosis, respectively, and
n is the sample size.
Table 4 lists the descriptive statistics of the activity concentrations for
40K,
226Ra, and
232Th. The mean and SD ranges are discussed in
Section 3.1. To determine the spread of a probability distribution, the CV percentage was obtained by dividing the standard deviation by the mean value [
6,
46]. A smaller coefficient of variation implies less deviation in the data, translating into a lower risk with respect to the data. In this study, the CV percentages of the samples ranged from 29.5% to 33.7%, indicating moderate deviation in the activity concentrations [
6]. This moderate deviation in radioactivity was due to certain samples, such as S14, S15, and S29, having significantly lower activity levels compared to the mean values.
To assess the degree of peakedness in activity concentrations compared to a normal distribution, the kurtosis parameter can be utilized [
4,
25,
47]. In this study, the kurtosis values of three types of radioactivity were analyzed, and they were all positive, indicating a higher peak than the normal curve. The kurtosis values for
40K and
232Th were close to 0, while the kurtosis value for
226Ra was 2.79 due to a concentration of activity levels within a relatively small range. Specifically, 80% (24/30) of the
226Ra activity concentrations was found to fall between 15.0 Bq/kg and 25.0 Bq/kg.
The skewness parameter determines whether the activity concentration is symmetrically spread around the mean value, with positive and negative values indicating right and left tails of activity concentrations, respectively [
4,
25,
47]. In the present study, the skewness values for the activity concentrations of three nuclides were negative (−0.80 to −1.04), indicating that the radioactivity was slightly skewed to the left of the mean.
Figure 3a–c shows the frequency distribution and the cumulative percentage curves of
40K,
232Th, and
226Ra activity concentrations. To test the normality of the activity concentration, the K–S method was employed. The results indicate that the activity concentrations were distributed normally, with significant values of normality test exceeding 0.05, as shown in
Table 4. The activity concentration for the three radionuclides was mainly concentrated in the middle data. Around 80% of the
226Ra samples had activity concentrations in the range of 15–25 Bq/kg, 83% of the
232Th samples had activity concentrations in the range of 20–36 Bq/kg, and 80% of
40K samples had activity concentrations in the range of 250–500 Bq/kg. The normality of the distribution of
40K,
232Th, and
226Ra was reflected by the small values of kurtosis and skewness (except for
226Ra, which had a relatively high kurtosis value,
Table 4) [
4,
25,
34].
3.5. Activity Ratios of Activity Concentration
The activity ratio of
232Th to
226Ra can indicate the origin and stability of the nuclides within the
232Th and
238U decay series [
9,
30,
42]. Furthermore, the activity ratios of
40K/
232Th and
40K/
226Ra can be used to evaluate the presence of heavy or light minerals in the sampling sites [
9,
54].
Figure 4a–c displays the activity ratios of
232Th/
226Ra,
40K/
232Th, and
40K/
226Ra, respectively. The activity ratios of
232Th/
226Ra ranged from 0.82 to 1.86, with an average value of 1.43 ± 0.22, as shown in
Figure 4a. The majority of the ratios (90%) were greater than 1.20, with the exception of the S10 sample which had a ratio of less than 1.0. The activity ratios had a CV value of 15.6%, which indicates a weak variation of ratios. The
232Th/
226Ra ratio was greater than 1.0, signifying that the thorium and uranium decay series were at disequilibrium and that Th-bearing minerals were higher than Ra-bearing minerals [
9,
54,
55]. The average activity ratio was close to the ratio of the world soil average of 1.41, with mean concentrations of 45 Bq/kg and 32 Bq/kg for
232Th and
226Ra, respectively [
14]. An average mean activity ratio close to 1.41 has been observed in China [
56] and Egypt [
30,
57]. The ratio depends on the area’s local geological and geochemical settings [
9]. Several studies have observed
232Th/
226Ra activity ratios higher than those found in the present study [
47,
52,
54,
55,
58]. In contrast, some studies have reported lower activity ratios than those found in this study [
9,
32].
The ratios of
40K/
232Th (
Figure 3b) varied from 8.47 to 15.67, with an average ratio of 13.1 ± 1.9. The CV value of 14.4% indicates low variation. The ratios of
40K/
226Ra (
Figure 3c) varied between 10.2 and 26.8, with an average ratio of 18.8 ± 4.1. The CV value of 22.1% indicates moderate variation. Additionally, the
40K/
232Th and
40K/
226Ra ratios in the soil samples were higher than the world average values of 9.3 and 13.1, respectively [
14]. The higher ratios suggest that the mineral composition of the soil samples contained light minerals such as sandstone and limestone [
9,
55]. This confirmed the relatively low activity concentrations of natural radionuclides in comparison to world average concentrations in the soil.
3.6. Radiological Hazard Indices
Figure 5a–d shows the radiological hazard indices, which include the external hazard index (H
ex), radium equivalent activity (Ra
eq), absorbed dose rate (D
Rex), and annual effective dose equivalent (AED
ex). These indices were calculated using the equations listed in
Table 2.
The external hazard index is an assessment of the hazard of the natural gamma radiation and is used to detect the radiological suitability of a material [
14,
19]. The H
ex values ranged from 0.05 to 0.31, with an average of 0.22 (
Figure 5a), all of which are less than the permissible value of unity.
The radium equivalent activity is widely used as a hazard index, with a formula for comparing the specific activity of materials containing different amounts of
226Ra,
232Th, and
40K [
14,
19]. It is defined as a single quantity that represents the combined specific activities of
226Ra,
232Th, and
40K, and provides a numerical indicator of an external dose to the public [
19]. The Ra
eq values varied from 17.5 Bq/kg to 116.3 Bq/kg, with a mean of 83.0 Bq/kg (
Figure 5b). These values are lower than the recommended limit of 370 Bq/kg [
19].
The outdoor absorbed dose rate ranged from 8.3 nGy/h to 54.4 nGy/h, with an average value of 39.0 nGy/h. The calculated annual effective dose equivalent due to gamma radiation ranged from 10.1 μSv/y to 66.7 μSv/y, with a mean of 47.8 μSv/y. All D
Rex and AED
ex values are also lower than the global population weight average of 59 nGy/h and 70 μSv/y, respectively [
14]. Samples S14, S15, and S29 had much lower radiological hazard indices due to their low radioactivity.
The radiological risk values in the present study compared to soils around NPPs and nuclear facilities in other countries are listed in
Table 7. The values from the present study are lower than those reported by previous studies across Taiwan [
11,
16], Thailand [
17], and India [
18] and are similar to those reported from Korea [
13]. The radiological hazard indices H
ex, Ra
eq, D
Rex, and AED
ex values indicate that the area does not pose a significant radiological health risk. Therefore, the naturally occurring radionuclides in the soil around the Maanshan NPP do not pose a significant radiological risk to residents and tourists. The data obtained in this study provide useful information on the background radioactivity of the study area.