Descriptive statistics
Microsoft Excel 2013 was used to calculate descriptive statistics. The descriptive statistics of the concentrations of 10 heavy metal elements, Sc, V, Cr, Mn, Fe, Co, Zn, As, Hf, and Ta, in the moss samples are presented in Table 1. The relative errors of the elemental concentrations for all analyzed elements are less than 15%. Concentrations are given in mg/kg. The coefficient of variation (CV) in percent was calculated as the ratio between the standard deviation and the mean concentration value. The normality of the concentration distributions for the 21 elements was tested using the Shapiro-Wilk test. The probability values (p-value) obtained for the element distributions are given in Table 1. The hypothesis of normality will be rejected (with 95% confidence) if the corresponding p-value is less than 0.05. It can be clearly seen in Table 1 that among the 10 analyzed elements, only three had a normal distribution: V, Fe, and Ta.
Table 1
El. | Min | Max | Mean | Median | STDEV | CV (%) | p-value |
Sc | 0.08 | 3.55 | 1.32 | 1.23 | 0.87 | 66 | 0.157 |
V | 0.98 | 18.53 | 7.75 | 7.53 | 4.56 | 59 | 0.098 |
Cr | 1.11 | 19.91 | 6.93 | 6.09 | 4.52 | 65 | 0.04 |
Mn | 50 | 207.4 | 102.28 | 89.59 | 38.46 | 38 | 0.018 |
Fe | 284 | 6542 | 2503 | 2213 | 1615 | 65 | 0.052 |
Co | 0.25 | 3.39 | 1.46 | 1.35 | 0.80 | 55 | 0.148 |
Zn | 43 | 1298 | 285.47 | 175.5 | 282.7 | 99 | < 0.001 |
As | 0.49 | 17.07 | 3.93 | 2.40 | 4.06 | 103 | < 0.001 |
Hf | 0.04 | 1.31 | 0.48 | 0.37 | 0.32 | 66 | 0.020 |
Ta | 0.02 | 0.41 | 0.15 | 0.14 | 0.10 | 66 | 0.068 |
The coefficient of variation for the analyzed elements varied from 38 to 103%. The maximum coefficient of variation for the concentrations was obtained for As (103%) and the minimum coefficient of variation was observed for Mn (38%). The concentrations of the 10 heavy metal elements in the moss samples decrease in the following order: Fe > Zn > Mn > V > Cr > As > Co > Sc > Hf > Ta.
The mean elemental concentrations (mg/kg) of the moss samples collected in Lamdong, other cities in Vietnam (Hanoi, Thainguyen), and Europe (Tver, Yaroslav and Tula regions of Russia, Silesia-Kraków and Legnica-Głogów Copper Basin of Poland, Prut river catchment region of Romania, Norway, and Moldova) are presented (Table 2). It should be noted that the same analytical technique (neutron activation analysis) was used to obtain the elemental concentration data for all regions mentioned above.
Table 2
The mean elemental concentrations (mg/kg) in moss obtained in the present work and in other cities in Vietnam and Europe
El. | Lamdong (Present work) | Hanoi (Nguyet et al. 2010) | Thainguyen (Nguyet et al. 2010) | Tver and Yaroslav regions, Russia (Ermakova et al. 2004a) | Tula region, Russia (Ermakova et al. 2004b) | Silesia-Kraków and Legnica-Głogów Copper Basin, Poland (Grodzinska et al. 2003) | Prut river catchment, Romania (Lucaciu et al. 2004) | Norway (Steinnes et al. 2016) | Moldova (Zinicovscaia et al. 2017) |
Sc | 1.32 | 1.16 | 1.3 | 0.14 | 0.57 | 0.19 | 0.48 | 0.1 | 1.1 |
V | 7.75 | 17.28 | 53 | 3.2 | 8 | 3.9 | 7.0 | 1.6 | 9.6 |
Cr | 6.93 | 20.81 | 16 | 1.5 | 5 | 6.2 | 5.4 | 1.1 | 10.3 |
Mn | 102.28 | 147.8 | 153 | 400 | 300 | 145 | 180 | 450 | 132 |
Fe | 2503 | 4400 | 4700 | 550 | 2200 | 1226 | 1900 | 490 | 3300 |
Co | 1.46 | 1.93 | 1.5 | 0.41 | 0.63 | 0.3 | 0.57 | 0.5 | 1.4 |
Zn | 285.47 | 457.7 | 121 | 34 | 54 | 150 | 40 | 36 | 39 |
As | 3.93 | 3.19 | 9.9 | 0.22 | 0.5 | 0.43 | 0.73 | 0.17 | 1.2 |
Hf | 0.48 | 0.68 | 1.8 | 0.18 | 0.82 | 0.17 | 0.53 | - | - |
Ta | 0.15 | 0.22 | 0.2 | 0.020 | 0.078 | 0.02 | 0.056 | - | - |
Comparing the concentrations of the elements in the moss samples from Hanoi and Lamdong, we found that except for Sc and As, the concentrations of the other metal elements, V, Cr, Mn, Fe, Co, Zn, Hf, and Ta, are from 1.32 to 3 times higher in Hanoi than in Lamdong. This is understandable because there are many more sources of pollution in Hanoi than in Lamdong. Hanoi is now considered one of the two most polluted cities, in Vietnam. In the case of Thainguyen, the concentrations of all elements except Zn are higher than in Lamdong. In particular, the concentrations of some elements in the samples of Thainguyen are very high in comparison with those in Lamdong, namely, 6.84 times for V, 3.75 times for Hf, and 2.52 times for As. Thainguyen is also the major center of the country and there are many ore mines, especially iron ore. Therefore, air pollution in this area is expected to be much higher than in Lamdong.
Looking the element concentrations in the moss from European countries listed (Table 2), it is clearly seen that heavy metal air pollution in these European countries, especially Norway, is much lower than in Lamdong, Hanoi, and Thainguyen.
To visualize more clearly the differences in the mean concentrations of 10 metal elements in the mosses of Lamdong, Hanoi, Thainguyen, and European cities, the ratios of the elemental concentrations in the mosses from cities in Vietnam and Europe to those of Lamdong are shown (Fig. 2).
Correlation analysis
The correlation coefficients of the elemental concentrations in the moss samples can give some information about their origin. The Pearson correlation coefficients with significance level p = 0.05 are presented in Table 3. Several pairs of elements are strongly correlated, such as Sc and Hf (r = 0.85), Sc and Ta (r = 0.787), Sc and As (r = 0.699), Mn and Co (r = 0.624), Fe and As (r = 0.627), Fe and Ta (r = 0.739), As and Ta (r = 0.773), and Hf and Ta (r = 0.745). It was found that natural soils having high levels of Co always have the presence of Mn and Fe (Kosiorek & Wyszkowski 2019). It can be confirmed from Table 3 that there are quite strong correlations between Mn and Co (r = 0.624) and Mn and Fe (0.544). It is observed that some elements, including V, Cr, Co, and Zn, are very weakly correlated with other elements (Table 3).
Table 3
| Sc | V | Cr | Mn | Fe | Co | Zn | As | Hf | Ta |
Sc | 1 | | | | | | | | | |
V | -0.029 | 1 | | | | | | | | |
Cr | 0.153 | 0.25 | 1 | | | | | | | |
Mn | 0.488 | 0.286 | 0.309 | 1 | | | | | | |
Fe | 0.591 | 0.321 | 0.456 | 0.544 | 1 | | | | | |
Co | 0.393 | -0.122 | 0.323 | 0.624 | 0.259 | 1 | | | | |
Zn | 0.095 | -0.142 | 0.116 | 0.041 | 0.108 | 0.294 | 1 | | | |
As | 0.699 | -0.005 | 0.032 | 0.452 | 0.627 | 0.102 | 0.142 | 1 | | |
Hf | 0.85 | 0.019 | 0.017 | 0.297 | 0.441 | 0.267 | -0.003 | 0.507 | 1 | |
Ta | 0.787 | -0.05 | 0.189 | 0.509 | 0.739 | 0.426 | 0.133 | 0.773 | 0.745 | 1 |
Contamination factor
The contamination factor (CF) is a quantity that can be used to assess the pollution level of an element at the surveyed location. According to Fernandez and Cảrballeira (2000), the elemental contamination factor can be evaluated by the following equation:
$${CF}_{El}=\frac{{C}_{El}}{{BG}_{El}}$$
where CEl is the mean concentration of the element of interest in all investigated moss samples in the region under study, and BGEl is the background concentration. Since the background concentrations of all elements of interest in Lamdong province are not available, the minimum values of the elemental concentration listed in Table 2 were chosen as background values. Therefore, for the background values of Sc, V, Cr, Fe, Zn, Co, and As, the data of Norway (Steinnes et al. 2016) were used, while for Hf and Ta, the data of Silesia-Kraków and Legnica-Głogów Copper Basin, Poland (Grodzinska et al. 2003) were used. Finally, the mean value of the 3 smallest concentrations of Mn among all moss samples collected in Lamdong province in this work was used as the background value for Mn.
Based on the value of the elemental contamination factor, pollution levels can be divided into 6 categories ranging from C1 to C6, as follows: category C1 (unpolluted) if CF ≤ 1; category C2 (may be polluted) if 1 < CF ≤ 2: C2; category C3 (polluted at low level) if 2 < CF ≤ 3.5: category C4 (moderately polluted) if 3.5 < CF ≤ 8; category C5 (polluted at high level) if 8 < CF ≤ 27; and C6 (extremely polluted) if CF > 27. The calculated values of the elemental contamination factors for Lamdong province are listed in Table 4. It can be seen from this table that the air in Lamdong province might be polluted by Mn (C2), is polluted at a low level by Co and Hf (C3), is moderately polluted by V, Cr, Fe, Zn, and Ta (C4), and is polluted at a high level by Sc and As (C5).
Table 4
Element | Sc | V | Cr | Mn | Fe | Co | Zn | As | Hf | Ta |
Background | 0.1 | 1.6 | 1.1 | 56.37 | 490 | 0.5 | 36 | 0.17 | 0.17 | 0.02 |
Mean | 1.32 | 7.75 | 6.93 | 102.28 | 2503 | 1.46 | 285.47 | 3.93 | 0.48 | 0.15 |
CF | 13.20 | 4.84 | 6.30 | 1.81 | 5.11 | 2.92 | 7.93 | 23.12 | 2.82 | 7.50 |
Category | C5 | C4 | C4 | C2 | C4 | C3 | C4 | C5 | C3 | C4 |
Factor analysis
Factor analysis is a very suitable tool to find the sources of elemental air pollution when using the moss biomonitoring technique. This method has been used by previous researchers (Schaug et al. 1990) to analyze the concentrations of elements in moss samples and identify possible sources of pollution. In this work, IBM SPSS software version 20 was used to analyze the concentration data. The results presented in Table 5 include the factor loadings of the elements as well as the eigenvalues, the explained variance, and the cumulative explained variance of the extracted factors.
Table 5
Factor analysis of the elemental concentrations in the moss samples
Element | Factor-1 | Factor-2 | Factor-3 | Factor-4 | Factor-5 | Factor-6 |
Sc | 0.45 | 0.80 | 0.24 | 0.07 | -0.03 | 0.03 |
V | 0.03 | -0.02 | 0.01 | 0.13 | 0.98 | -0.08 |
Cr | 0.07 | 0.00 | 0.18 | 0.95 | 0.12 | 0.05 |
Mn | 0.44 | 0.11 | 0.81 | 0.08 | 0.28 | -0.07 |
Fe | 0.71 | 0.30 | 0.15 | 0.43 | 0.27 | 0.04 |
Co | -0.04 | 0.24 | 0.88 | 0.21 | -0.17 | 0.21 |
Zn | 0.08 | 0.00 | 0.10 | 0.05 | -0.07 | 0.99 |
As | 0.90 | 0.32 | 0.06 | -0.08 | -0.03 | 0.08 |
Hf | 0.21 | 0.96 | 0.09 | -0.03 | 0.04 | -0.03 |
Ta | 0.66 | 0.61 | 0.26 | 0.14 | -0.10 | 0.05 |
Eigenvalue | 4.395 | 1.551 | 1.338 | 0.882 | 0.677 | 0.584 |
Expl. Variance (%) | 43.95 | 15.51 | 13.38 | 8.82 | 6.77 | 5.84 |
Cumulative (%) | 43.95 | 59.46 | 72.83 | 81.65 | 88.42 | 94.26 |
Six factors have been extracted that can explain 94.26% of the total variance. The explained variances of Factor-1, Factor-2, Factor-3, Factor-4, Factor-5, and Factor-6 are 43.95%, 15.51%, 13.38%, 8.82%, 6.77%, and 5.84%, respectively. If the value of any factor loading is greater than 0.6, then it is written in bold (Table 5).
Before discussing the factors that have been extracted, it should be emphasized that the Central Highlands, including Lamdong province, has the largest amount of aluminum bauxite in Vietnam. Moreover, Vietnam has been estimated to hold the third largest reserves of bauxite in the world. Currently, there are two large bauxite refineries operating in Lamdong province. Aluminum and ferric oxides are the main components of bauxite (Nechitailov et al. 2008). However, other toxic metals may also contaminate the surrounding environment, depending on the characteristics of the land and the land use activities (Abdullah et al. 2016). Therefore, these refineries are expected to be large atmospheric pollution sources of heavy metal and metalloid elements.
Factor-1 and Factor-2 explain 43.95% and 15.51% of the total variance, respectively. Factor-1 is heavily loaded by the elements As (0.90), Fe (0.71), and Ta (0.66), while Factor-2 is mainly loaded by the elements Hf (0.96), Sc (0.80), and Ta (0.61). All are elements in the Earth’s crust, and Fe, especially, is the fourth-most abundant element in the Earth’s crust. Therefore, we can conclude at a glance that Factor-1 and Factor-2 reflect contamination with soil dust. The presence of As in Factor-1 can also be explained by the presence of As in the agricultural soil of Lamdong. This province is very famous in Vietnam for growing vegetables, coffee, and flowers. These products are supplied to the whole country, especially the southern region. To grow these products, farmers must use a lot of chemical fertilizer and pesticides with a high concentration of As, so that the soil in Lamdong might be contaminated with As. The elements present in Factor-1 and Factor-2 can also be released from other industries and human activities. As highlighted above, there are two large bauxite refineries in Lamdong, and the presence of these elements in Factor-1 and Factor-2 might be a consequence of bauxite mining activities in Lamdong province. Therefore, it is possible to say that Factor-1 and Factor-2 represent pollution sources from agriculture and the bauxite refining industry.
Factor-3 has high values for Co (0.88) and Mn (0.81), and it accounts for 13.38% of the total variance. The presence of Co and Mn in the air can be caused by both natural and man-made sources. Some of the natural sources that emit Co into the air include weathering and erosion of rocks and soil, forest fires, and evaporation of seawater, etc. Crystalline rock is the strongest natural source of Mn in the air. The other natural sources of Mn in the air are sea spray, forest fires, and vegetation activity (Schroeder et al. 1987; Stokes et al. 1988).
Several man-made sources of cobalt air pollution are reported (ATSDR 2004), namely, coal-fired plants, emissions from vehicles, mining and processing of ores containing Co, utilization of chemical supplies containing cobalt, etc. The main anthropogenic sources of Mn released to the air are industrial emissions (such as ferroalloy production and iron and steel foundries, power plants, and coke ovens), combustion of fossil fuels, and re-entrainment of manganese-containing soils (Lioy 1983; Ruijten et al. 1994).
Factor-4 contains Cr (0.95) only and explains 8.82% of the total variance. Cr rarely occurs in nature (Barałkiewicz & Siepak 1999) so Factor-4 can be related to human activities. It was suggested by Cheng that the combustion of coal and oil is the most important emission source of Cr in China (Cheng et al. 2014). In Lamdong province, coal is still the main fuel used for cooking by many families, and oil is used to pump water from underground wells to irrigate fields of vegetables, flowers, and coffee plants.
Factor-5 is heavily loaded on V (0.98) and explains 6.77% of the total variance. It has been found that combustion of fossil fuels and oil is the major source of V in the atmosphere (Kousehlar & Widom 2019). In Lamdong province, the use of fossil fuels for cooking and people's daily activities is still common. Furthermore, farmers in Lamdong province regularly use diesel-powered engines to irrigate coffee and other industrial crops. These activities may be the main sources of vanadium emission into the atmosphere.
Factor-6 contains only Zn (0.99) and explains 5.84% of the total variance. An investigation of air quality in Asian countries conducted by Hopke et al. (2008) shows that Zn is emitted into the atmosphere by two-stroke vehicles, which are a very popular means of transportation in Asia, including Vietnam. In addition, Zn can be emitted from tire wear (Blok 2005; Longhin et al. 2016). Thus, Factor-6 may be related to two-stroke motor vehicles and tire wear.