This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Atmospheric mercury (Hg) usually tends to accumulate in the upper horizons of soils. However, the physico-chemical characteristics of some soils, as well as pedogenetic processes, past climate changes, or soil degradation processes, can lead to a redistribution of mercury through the soil profile. In this work, the presence and accumulation of mercury was studied in three deep polycyclic soils from a mountainous area in NW Iberia Peninsula. The highest total Hg values (HgT) were found in the organic matter-rich O and A horizons of FL and MF profiles (169 and 139 μg kg−1, respectively) and in the illuvial horizon of RV (129.2 μg kg−1), with the latter two samples showing the maximum Hg reservoirs (29.3 and 29.0 mg m−2, respectively). Despite finding the highest Hg content in the surface horizons, considerable Hg reservoirs were also observed in depths higher than 40–50 cm, indicating the importance of taking into account these soil layers when Hg pools are evaluated at a global scale. Based on the mass transfer coefficients, we can rule out the contribution of parent material to the Hg accumulation in most of the horizons, thus indicating that pedogenetic processes are responsible for the Hg redistribution observed along the soil profiles. Finally, by means of principal component analysis (PCA) and stepwise linear regression we could assess the main soil components involved in the Hg accumulation in each soil horizon. Therefore, PC1 (organic matter and low stability Al-hummus complexes) showed a higher influence on the surface horizons, whereas PC2 (reactive Al-Fe complexes and medium-high Al-hummus complexes) and PC4 (crystalline Fe compounds and pHw) were more relevant in the Hg distribution observed in the deepest soil layers.
Hg is a global pollutant of environmental concern, which can cause serious damage to the health of both humans and wildlife, especially in the methylated form (
Podzols and soils with podzolic characteristics can serve as examples of soils with a recognized ability to accumulate Hg in the deeper soil horizons. In Podzols, Hg can be mobilized through the soil profile together with organic matter and Al and Fe compounds and afterwards immobilized due to its adsorption to metal (Al, Fe)-humus complexes and Al and Fe oxyhydroxides, leading to the characteristic soil vertical pattern of Hg reported in different works (
Besides the particular chemical properties that determine Hg occurrence in soils such as Podzols, the distribution of Hg in other soils can also be defined by pedogenetic processes, past climate changes, or soil degradation processes (forest fires, deforestation, erosion) as reported by
The present study assess the occurrence, accumulation, and vertical distribution of Hg in three deep polycyclic soils from a mountain area in NW Iberia Peninsula. This study aims to evidence the relevance of deeper soil layers in terms of soil Hg pools, and attempts to further understand the main soil chemical characteristics and soil processes involved in the observed Hg patterns with soil depth.
This study was carried out in the Xistral Mountains, a medium altitude mountain range (maximum 1,060 m a.s.l.) located in the north of the province of Lugo (Galicia). This area is characterized by a mean annual temperature between 7°C and 10°C, and a moderate total annual rainfall (1,400–1800 mm) with a scarce rainfall seasonality and abundant fogs throughout the year above 600 m a.s.l. (
At each soil sampling site, samples were collected on a surface cut of a forest track, removing the first 30 cm-thickness of the profile in order to obtain a fresh soil surface and avoid potential anthropogenic disturbances. From each individual horizon identified in the field, soil samples were collected at an interval of 10–20 cm with a plastic garden trowel, which was rinsed twice between samples with a diluted HNO3 solution and then dried. The total number of soil samples collected was 45. In addition, a fresh rock sample representing the parent material for each sampled soil was collected. Soil and rock samples were stored in plastic bags and transported to the laboratory in a portable fridge at 4°C. Once in the laboratory, after plant debris and stone removal, soil samples were air-dried and sieved (2-mm mesh), whereas rock samples were washed with distilled water and any soil residue removed. Approximately 0.5 kg of sieved soil was quartered with a stainless steel riffle-splitter to obtain samples with enough representativeness and homogeneity for subsequent general chemical characterization and Hg analyses.
The physico-chemical characterization of soil samples was carried out in the fine earth fraction (<2 mm). Soil pH was determined in soil suspensions obtained after addition of distilled water (pHw) or saline solution (0.1 M KCl, pHK) to soil samples, maintaining a 1:2.5 soil:solution ratio. The total content of biophilic elements (i.e., C and N) were determined in an autoanalyzer after combustion of milled soil samples, with total C values assumed to be organic C due to the absence of inorganic carbonates in the studied soils. Neutral saline solutions, 1 M NH4Cl (
The fractionation of Al and Fe in soil samples was carried out following the procedures used by
For the Hg measurement, samples of soil and rock were milled in a mechanical agate mortar (Retsch RM100, Retsch RM200). About 100 mg of each sample was analyzed twice using a DMA-80 Hg analyzer (Milestone), which is based on thermal decomposition and atomic absorption spectroscopy. Measurements were repeated when the coefficient of variation was higher than 10%. In order to test the accuracy of the method (for quality assurance and quality control purposes), different standard reference materials were analyzed at the start of each sample run and every fifteen samples, obtaining recovery percentages of 91% for GBW 07402 (average 13.6 ± 0.7 μg kg−1;
For each sampled depth, the pool of Hg (HgRes_d) was calculated taking into account the corresponding thickness (Td), bulk density (Bd), coarse fragment proportion (C) (%), and total Hg content (HgT) as in
The Hg reservoir in each soil horizon identified (HgTres) was calculated as the sum of all HgRes_d included in the same horizon.
Mass transfer coefficients of Hg (τHg) were calculated following the open-system mass transport function (
The τHg for each soil horizon identified was calculated as the average of all Tau values included in the corresponding depths analysed. As
All the statistical analyses described in this section were done using SPSS version 25.0 software for Windows.
In order to reduce the number of soil variables studied into a few components, a principal component analysis (PCA) was conducted, applying varimax rotation that maximizes the sum of the variances of the square loadings. Each principal component consisted of variables with loadings higher than 0.50 (Abdi and Williams, 2010).
A principal component regression (PCR) analysis was carried out with the “new” variables obtained in the PCA as independent variables, and the principal components which are not correlated between them (orthogonal) and Hg as dependent variables. Using this method, we could predict Hg concentration and distinguish which soil properties are most involved in the Hg depth distribution observed. The weight of each component (wPC) was used to calculate its participation in the Hg prediction and it was estimated by multiplying the score of each component by the corresponding standardized regression coefficient (
The general chemical properties of the three soil profiles are shown in
Mean values per soil horizon of some chemical characteristics of FL, MF, and RV soils.
Profile | Horizon | Depth |
|
|
|
C |
|
N | S |
|
|
---|---|---|---|---|---|---|---|---|---|---|---|
cm | ------------ % ------------ | cmolc kg−1 | |||||||||
FL | O | 0–10 | 1 | 4.7 | 4.3 | 20.8 | 10.3 | 1.5 | 0.05 | 5.9 | 1.3 |
A | 10–30 | 2 | 4.8 | 3.9 | 11.3 | 7.1 | 0.7 | 0.05 | 1.3 | 6.5 | |
AC | 30–50 | 2 | 5.1 | 4.1 | 5.0 | 4.3 | 0.3 | 0.02 | 0.3 | 4.3 | |
2A | 50–60 | 1 | 4.8 | 4.2 | 5.7 | 5.1 | 0.3 | 0.01 | 0.3 | 3.7 | |
2AC | 60–74 | 1 | 4.7 | 4.3 | 4.5 | 2.2 | 0.2 | 0.01 | 0.2 | 4.0 | |
2BC | 74–90 | 1 | 4.7 | 4.4 | 3.2 | 2.9 | 0.1 | 0.01 | 0.2 | 1.9 | |
3BC | 90–150 | 3 | 4.6 | 4.5 | 2.4 | 1.5 | 0.1 | 0.01 | 0.2 | 1.8 | |
3C | 150–185 | 3 | 4.9 | 4.4 | 0.6 | 0.0 | 0.0 | 0.00 | 0.2 | 0.7 | |
MF | A | 0–38 | 3 | 4.8 | 3.6 | 10.9 | 5.9 | 0.6 | 0.03 | 1.6 | 5.2 |
Bhs | 38–48 | 1 | 5.0 | 4.3 | 3.0 | 1.7 | 0.1 | 0.01 | 0.2 | 2.1 | |
Bs | 48–68 | 1 | 5.0 | 4.4 | 3.9 | 3.2 | 0.2 | 0.01 | 0.3 | 2.1 | |
2CB | 68–100 | 2 | 5.1 | 4.5 | 1.6 | 1.0 | 0.1 | 0.00 | 0.2 | 1.3 | |
3C | 100–190 | 6 | 5.2 | 4.3 | 0.6 | 0.1 | 0.0 | 0.00 | 0.4 | 1.2 | |
RV | A | 0–43 | 4 | 4.8 | 4.2 | 7.7 | 5.7 | 0.5 | 0.03 | 0.6 | 4.2 |
B | 43–75 | 6 | 4.8 | 4.5 | 4.5 | 3.9 | 0.3 | 0.02 | 0.4 | 1.8 | |
2A | 75–95 | 3 | 4.7 | 4.4 | 3.1 | 2.2 | 0.2 | 0.01 | 0.4 | 1.7 | |
2AB | 95–110 | 2 | 4.7 | 4.5 | 1.2 | 0.9 | 0.1 | 0.01 | 0.4 | 1.1 | |
2B | 110–140 | 2 | 4.8 | 4.3 | 0.3 | 0.1 | 0.0 | 0.01 | 0.4 | 1.5 | |
2C | 140–155 | 1 | 4.8 | 4.0 | 0.1 | 0.0 | 0.0 | 0.01 | 0.3 | 1.8 |
n is the number of samples for each soil horizon.
pHw and pHK are soil pH in water and in saline solution.
Cp is the pyrophosphate-extracted C.
BS is the sum of base cations (Ca, Mg, Na, K) and AlK is the exchangeable Al.
The distribution of the Al and Fe compounds in the soil solid phase of the different horizons studied is summarized in
Mean contents per soil horizon of Al and Fe compounds and Hg (HgT), total reservoir of Hg (∑HgTres) and Tau.
Profile | Horizon | Depth |
|
|
|
|
|
|
|
|
|
|
---|---|---|---|---|---|---|---|---|---|---|---|---|
cm | g kg−1 | µg kg−1 | mg m−2 | |||||||||
FL | O | 0–10 | 1 | 4.7 | 4.8 | 5.2 | 4.9 | 5.8 | 7.1 | 169 | 8.6 | 460.4 |
A | 10–30 | 2 | 4.4 | 4.7 | 5.8 | 4.7 | 6.5 | 6.9 | 117 | 20.2 | 199.0 | |
AC | 30–50 | 2 | 4.6 | 5.5 | 5.7 | 4.9 | 4.9 | 6.3 | 44 | 7.0 | 72.5 | |
2A | 50–60 | 1 | 8.6 | 10.0 | 10.7 | 7.1 | 9.0 | 9.9 | 79 | 6.6 | 135.3 | |
2AC | 60–74 | 1 | 8.9 | 9.8 | 12.8 | 5.3 | 6.4 | 7.3 | 91 | 10.0 | 137.3 | |
2BC | 74–90 | 1 | 6.8 | 9.9 | 10.6 | 2.4 | 2.5 | 3.8 | 47 | 7.0 | 83.4 | |
3BC | 90–150 | 3 | 6.3 | 9.3 | 16.2 | 1.2 | 1.7 | 1.7 | 36 | 22.6 | 58.2 | |
3C | 150–185 | 3 | 2.2 | 3.0 | 17.5 | 0.6 | 0.8 | 2.7 | 15 | 5.7 | 20.4 | |
MF | A | 0–38 | 3 | 4.4 | 5.2 | 5.9 | 4.8 | 7.1 | 8.8 | 139 | 29.3 | 59.4 |
Bhs | 38–48 | 1 | 6.7 | 7.7 | 10.2 | 4.6 | 8.1 | 9.4 | 52 | 3.5 | 44.6 | |
Bs | 48–68 | 1 | 11.3 | 13.3 | 18.9 | 2.8 | 3.7 | 6.8 | 93 | 11.3 | 23.3 | |
2CB | 68–100 | 2 | 4.7 | 9.0 | 9.3 | 1.6 | 2.2 | 4.8 | 40 | 9.9 | 13.7 | |
3C | 100–190 | 6 | 2.1 | 4.1 | 18.3 | 1.5 | 1.9 | 6.0 | 22 | 12.5 | 9.4 | |
RV | A | 0–43 | 4 | 9.4 | 10.0 | 11.9 | 7.3 | 7.5 | 12.5 | 113 | 29.0 | 3.2 |
B | 43–75 | 6 | 9.0 | 15.0 | 16.2 | 7.2 | 8.6 | 14.2 | 129 | 32.5 | 5.0 | |
2A | 75–95 | 3 | 5.7 | 10.0 | 14.8 | 3.4 | 3.9 | 7.5 | 88 | 13.3 | 2.3 | |
2AB | 95–110 | 2 | 2.7 | 5.7 | 13.7 | 1.2 | 1.6 | 4.0 | 28 | 3.0 | 0.4 | |
2B | 110–140 | 2 | 1.3 | 2.8 | 16.7 | 0.8 | 1.2 | 3.1 | 15 | 4.1 | −0.5 | |
2C | 140–155 | 1 | 0.7 | 1.4 | 17.5 | 0.4 | 0.6 | 2.7 | 9 | 1.5 | −0.6 |
n is the number of samples for each soil horizon.
Alp (Fep), Alo (Feo), Aln and Fed are Al (Fe) extracted with Na-pyrophosphate (p), ammonium oxalate-oxalic acid (o), Al extracted with Na hydroxide (n) and Fe extracted with Na-dithionite-citrate (d), respectively.
HgT and HgTres are total mercury content and the mass of total Hg in areal basis for each whole horizon.
τHg are the mass transfer coefficients of Hg for each whole horizon.
Vertical patterns of Al-humus complexes of low (Alob), moderate (Alom) and high stability (Aloa) in the soils FL, MF, and RV.
In the case of the Fe distribution, organically-complexed Fe (Fep) dominate in the A horizons (4.7–7.3 g kg−1), whereas in the B and C horizons there is an equal partition between Fe-hummus complexes (Fep) and crystalline compounds (Fec) (
Considering the morphological features observed during field sampling such as the horizon diversity and discontinuities recognisable by the occurrence of stone lines and charcoal and the previous discussion of main soil physicochemical characteristics, the three soils studied are representative of complex polycyclic soils typically found in the mountain landscapes from NW Iberian Peninsula. According to this and following the IUSS-WRB (2014), although some classification requirements could be not wholly satisfied such as colour in the illuvial horizons, a tentative classification of the studied soil was Folic Umbrisol for soils FL and MF and Cambic Umbrisol in the case of RV soil.
The total mercury content (HgT) is shown in
Variation of observed (blue circles) vs. predicted Hg (red triangles) with soil depth in the soils FL, MF, and RV.
Although the HgT in the samples of the three soils studied were considerably higher than the values of HgT found in their parent material (0.7, 1.3, and 4.0 μg kg−1 for soils FL, MF and RV, respectively), most of the samples were below 130 μg kg−1, the critical load of Hg in soils (
In general, the values of HgT obtained in the present study are in the same order as those reported for acid forest soils not directly affected by Hg emission point sources worldwide (
The Hg reservoir (HgTres), shown in
The Hg reservoir (8.6 mg m−2) of the soil horizon with the highest HgT content (O horizon - FL soil profile) was higher than the range (0.02–7 mg m−2) found for this type of horizon worldwide (
The horizons with the highest HgTres were the B-horizon from RV soil (32.5 mg m−2), located at 43–75 cm depth followed by both the surface A horizons from RV and MF soil profiles (29.0 and 29.3 mg m−2, respectively). Similar thickness (32–41 cm depth), bulk density (0.81–0.99 g cm−3) and high total Hg contents (113–139 μg kg−1,
The lowest HgTres occurred in the deepest sampled horizons from FL and RV soil profiles (3C: 5.7 mg Hg m−2 and 2C: 1.5 mg Hg m−2, respectively) manly due to their low HgT contents and the Bhs horizon from MF soil (3.5 mg Hg m−2) because of its low thickness (10 cm) and high coarse fragment proportion (54%). Therefore, the commonly observed declining pattern of Hg concentrations with depth (
The complexity of polycyclic soils is also conspicuous when evaluating possible gains and losses of Hg regarding parent material. The highest τHg values in each soil profile (
The mass transfer coefficients of Hg calculated for the three soil profiles were of different orders of magnitude but all indicated a net gain of Hg in the soil horizons regarding their parent material. In fact, the contribution from parent material to soil Hg accumulation is very low in almost all of the horizons evaluated. Taking into account that the studied polycyclic soils are located relatively close to one another and therefore climate conditions are assumed to be the same, in addition to exogenic inputs associated to atmospheric Hg deposition, pedogenetic processes should be also involved in Hg distribution accounting for its gains and losses throughout the soil profiles as it was indicated by
In order to understand the involvement of soil processes in Hg distribution through soil profiles, in addition to common soil properties derived from the general characterization of soil samples, the different forms of Fe and Al associated to the soil solid phase together with Hg concentrations were included in the PCA analysis. The PCA extracted four components (
Loadings of the soil properties used in the principal component analysis.
PC1 | PC2 | PC3 | PC4 | Com | |
---|---|---|---|---|---|
N |
|
0.18 | 0.09 | 0.01 | 0.98 |
C |
|
0.21 | 0.14 | −0.01 | 0.98 |
eCEC |
|
0.01 | 0.33 | −0.07 | 0.96 |
S |
|
0.21 | 0.01 | 0.08 | 0.89 |
BS |
|
−0.18 | −0.15 | −0.01 | 0.86 |
AlK |
|
0.07 | 0.45 | −0.09 | 0.93 |
Cp |
|
0.43 | 0.26 | 0.08 | 0.96 |
AlLa |
|
0.22 | 0.53 | −0.12 | 0.96 |
Alol |
|
0.26 | 0.54 | −0.13 | 0.93 |
pHK |
|
0.33 | −0.53 | 0.02 | 0.87 |
Alo | −0.10 |
|
−0.19 | 0.11 | 0.93 |
Alp | 0.12 |
|
0.12 | 0.15 | 0.92 |
Alom | 0.15 |
|
0.11 | −0.06 | 0.86 |
Aloh | −0.09 |
|
−0.01 | 0.26 | 0.86 |
AlCu | 0.47 |
|
0.32 | −0.09 | 0.90 |
Alc | −0.39 |
|
−0.31 | 0.15 | 0.84 |
Fep | 0.30 |
|
0.30 | 0.49 | 0.94 |
Feo | 0.29 |
|
0.46 | 0.42 | 0.95 |
Fed | 0.16 |
|
0.17 | 0.72 | 0.97 |
Alia | −0.35 |
|
−0.48 | 0.01 | 0.62 |
Feia | 0.12 | 0.31 |
|
0.04 | 0.60 |
Aln | −0.56 | −0.07 |
|
0.31 | 0.75 |
Fec | −0.10 | 0.29 | −0.34 |
|
0.91 |
pHw | −0.28 | −0.45 | 0.08 |
|
0.53 |
Eigv | 8.6 | 7.1 | 3.0 | 2.2 | |
Var | 35.6 | 29.6 | 12.7 | 9.1 |
PC1—PC4: components.
Com: communality, proportion of the variance of each parameter explained by the extracted components.
Eigv: eigenvalue.
Var: percentage of variance explained by each component.
Soil properties determining each principal component (PC) appear in bold.
In order to elucidate the influence of the soil properties in the Hg distribution observed in the three soil profiles studied, we performed a stepwise regression analysis using the scores of the four PCs extracted by PCA, considering them as new soil variables. The best model obtained included PC1, PC2, and PC4 with and adjusted R2 of 0.893. The predicted Hg (Hgpred) for each soil sample was estimated through the following Eq.
The model is rather accurate with a root mean-square error (RMSE) of 18 μg kg−1 for the whole group of samples as it can be seen in
Contribution of each principal component (wPC) to the predicted HgT of each sample (weight: regression standardized coefficient for a given component multiplied by the score).
In the case of the FL soil, there is a clear effect of the PC2 (Al and Fe compounds), from the 2A horizon (50–60 cm) up to the 3C horizon in which, due to the low Hg values, no component showed influence. In the MF profile, of podzolic features, Al and Fe compounds, especially Al (Fe)-hummus complexes, showed the highest weights in Hg prediction. However, in the 3C horizon the crystalline Fe forms were the most important components in the estimation of the Hg content.
In the RV soil, the Hg content in the B and 2A horizons is mostly influenced by Al and Fe complexes (PC2) but also crystalline Fe (PC4) to a lesser extent. In this sense, secondary Fe minerals are recognized to promote Hg accumulation instead of organic compounds in those soil layers deeper than 1 m, as it was suggested by
The results of this work indicated a noticeable contribution of the Hg deposition from the atmosphere to the Hg concentrations found in the soils studied instead of the lithological source. This fact is supported by the highest Hg values in surface horizons that diminished with depth and peaks in illuvial horizons which were, in all cases, considerably higher than the Hg concentrations of the parent material samples. The Hg depth distribution of the three soils studied was determined by the presence of soil components such as organic matter and Al and Fe complexes. Despite the fact that the maximum Hg contents were observed in O and A horizons, the highest Hg pools appeared in the deepest soil layers (>40–50 cm), representing about 50%–65% of all the Hg stored in those soil profiles. For this reason, subsurface soil horizons should be systematically evaluated in works about Hg distribution in soils, since they could contribute substantially to the total Hg accumulated in soils worldwide. In addition, the biogeochemical stability of Hg accumulated in deep soil layers prevents its mobility to other components of terrestrial ecosystems, leaving aside toxicity risks to wildlife and human health as well as a decline in the quality of groundwater and surface waters.
The raw data supporting the conclusion of this article will be made available by the authors, without undue reservation.
All authors listed have made a substantial, direct, and intellectual contribution to the work and approved it for publication.
AG-A acknowledges to the Xunta de Galicia a predoctoral grant (ED481A-2016/220). MM-L acknowledges the predoctoral grant FPU of Ministerio de Educación y Formación Profesional (FPU17/05484). It is also recognized the financial support of the Consellería de Cultura, Educación e Universidade (Xunta de Galicia) through the contract ED431C2021/46-GRC granted to the research group BV1 of the University of Vigo and the research project ED431F2018/06-EXCELENCIA. The Spanish Ministry of Science and Innovation also supported this research through the funds provided to the project InMerForEcos (Ref. PID 2021-125114OB-I00).
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.