Concentration, spatial distribution, and source apportionment of heavy metals in agricultural soils from the Yarlung Zangbo River Basin, Tibetan Plateau

In this study, 112 agricultural topsoil samples were collected from the Yarlung Zangbo River Basin (YZRB) and its two tributaries, and the concentrations, spatial distribution, source apportionment, and potential ecological risk of eight heavy metals (Cd, Hg, Pb, Cr, Cu, Zn, Ni and As) were analyzed. The mean concentrations of Cu, Pb, Zn, Cr, Ni, Cd, As, Hg and Pb were 33.91, 28.21, 86.93, 109.04, 73.58, 0.12, 21.49 and 0.05 mg/kg, respectively. Except for Pb, the concentration of heavy metals in soil exceeded the corresponding local background values, but did not exceed the national soil pollution risk screening values. In terms of spatial distribution, it showed that the heavy metals were enriched in the Nianchu River, and followed by Lhasa River and YZRB mainstream. Furthermore, the ecological risks were ordered as Hg > Cd > As > Ni > Cu > Pb > Cr > Zn, and the potential ecological risk index of the YZRB fell in the range of moderate ecological risk. The high-risk areas were in the facility agriculture fields of the Nianchu River, which was mainly related to agricultural activities. In addition, the positive matrix factorization results showed traffic, agricultural activities, natural weathering of the parent minerals, and atmospheric transport sources were the four sources in soils of the YZRB. This work will provide a scientific reference to develop appropriate pollution control measures and facility farmlands management in the YZRB of Tibetan Plateau.


Introduction
As a sink for nutrients and pollutants, soil plays a vital role in the stability of ecosystems and human survival (Wu et al. 2018;Yan et al. 2023).Recently, with the rapid development of agricultural production, industrialization and urbanization, soil faced a variety of pollutants.Among them, heavy metals (HMs, including arsenic and mercury because of similar characteristics, hereafter) are especially dangerous due to their high toxicity and persistence (Chen et al. 2008).HMs not only affect the quality of soil and crop yield, but also could accumulate in crops and increase the risk of human consumption, which has aroused great concern.
As the water tower in Asia, the Qinghai-Tibet Plateau supplies a substantial part of water demands in Asia, in terms of its water-supplying function and the dependence of ecosystem on downstream, so its importance is self-evident, while its vulnerability deserves more attention (Immerzeel et al. 2020).Generally, the Tibetan Plateau (TP) is considered to be a primitive region where human influence is limited.Regions in the TP with low levels of human activity 577 Page 2 of 11 are often selected as background values for research.Nevertheless, scientists claimed that the TP has recently been under environmental menace by soil and sediment pollution especially in densely populated areas in the middle section of the Yarlung Zangbo River Basin (YZRB) (Wu et al. 2018;Li et al. 2020a, b;Bai et al. 2014;An et al. 2022).Most of the investigations on soil pollution in the TP by HMs were focused on concentration level, spatial distribution and pollution status.Research on the concentration and sources of soil elements in the TP mainly focuses on anthropogenic sources, such as mining exploration (Li et al. 2018), railway construction along the Qinghai-Tibet Railway (Zhang et al. 2022) or traffic emissions along national highways (An et al. 2022).However, little attention was paid to the source analysis of HMs in the soil of facility farmland distributed along the YZRB (Wang et al. 2023).Specifically, with the intensification of human activities, especially the extensive use of cultivated land, HMs cause potential threat consequences to the local ecosystem (Liu et al. 2020).
Considering that the middle section of the YZRB is an important part of supporting local food security and ensuring stable social development, and are heavily influenced by human activities.Correspondingly, multivariate human activities especially for facility farmland use would be the dominant contributors to the concentrated HMs in the YZRB.Naturally, investigating the pollution status and identification the major sources of HMs in this region is particularly important.Previous studies on soil HMs in the YZRB showed that parent material was the main source of HMs in farmland soil.Vegetable fields were more severely polluted by HMs than grain fields, and the accumulation of HMs caused by the extensive use of agricultural chemicals in vegetable production should be of concern (Wang et al. 2023).Thus, accurately define the sources of HMs in agricultural soils of YZRB is a focus of local crops productivity, due to its self-evident food production.As the increase of human activities and regional changes on the plateau, it is necessary to conduct long-term monitoring the topsoil HMs in this region, which could help improve farmland management and develop appropriate pollution control measures.However, previous studies on HMs in soils of the TP were limited, and the status of the HMs still needs to be explored.
The main objectives of this research were to: (1) investigate the concentration of eight HMs (Cd, Hg, Pb, Cr, Cu, Zn, Ni and As) and characterize their spatial distribution; (2) evaluate their potential ecological risk; (3) accurately define the potential source and quantify their contributions, so that the results could provide a scientific basis for environmental management and HMs pollution control in the YZRB.

Study area
As shown in Fig. 1, the study area (28°20′-30°20′ N, 87°00′-92°35′ E) of YZRB was located in the southern of the TP of China, including the middle of the YZRB and two tributaries (Lhasa River and Nianchu River), a well-known area for self-evident food production.One tributary of the YZRB flowing through the capital of Tibet (Lhasa), which is the most densely populated area.Duilongqu flows into the right bank of Lhasa River Basin, and the largest geothermal power station (Yangbajing) is distributed in the upstream.
Copper mining exploration and the leakage of geothermal water are important sources of Cu, Zn, and As enrichment in the soil environment in the Lhasa River Basin, respectively.The other tributary (Nianchu River) is distributed with the most intensive facility farmland, and agricultural activities are an important source of soil HMs enrichment.YZRB has deposits of rich and fertile soil, forming a major agricultural development, also is a major grain and vegetable production area in Tibet.The main crops of farmland are highland barley and potatoes, and the main crops of facility agriculture are vegetables, fruits and nurseries.The study area belongs to a semi-arid plateau temperate zone with annual precipitation ranging from 200 to 600 mm (Zhan et al. 2021).The key soil type in the middle of the YZRB is subalpine grassland soil zone, with loose sandy soil structure and easy to be eroded (Zhan et al. 2021).The chemical weathering process in the YZRB is relatively weak, and physical weathering is the main process of soil production (Li et al. 2009).In addition, the Gangdise polymetallic metallogenic belt is also located in the YZRB, with huge potential for Cu, Pb and Zn deposits (Huang et al. 2023).The YZRB valley is an important granary and oasis in Tibet and is most likely to be affected by HMs using agriculture and other activities.

Sample collection and analysis
In this study, 112 topsoil (0-20 cm) samples were collected from farmlands and facility farmlands after the autumn harvest in September 2020.The sample sites were arranged along the river and covered the agricultural land around the watershed.Each sample was at combination of five random subsamples within 20 m from the sampling sites.After mixed, 1 kg of soil samples was taken in polyethylene bags and brought back to the laboratory.Soil samples were dried in shade at room temperature and sieved through a 100-mesh sieve for element analysis.After microwave digestion procedure, specifically, approximately, 100 mg of sieved soil was digested after adding 3 mL concentrated HNO 3 and 1 mL HF in PTFE sealed digestion tanks for 30 min at 180 ℃ (Li et al. 2022).After that, the sample solutions were evaporated to nearly dry, and transferred to a volumetric flask for constant volume to 50 mL.The concentrations of Pb, Cu, Ni, Zn, Cd and Cr were detected using ELANDRC-e inductively coupled plasma mass spectrometer (ICP-MS, PE, USA).As and Hg concentrations were determined by AFS-820 and AFS-830 atomic fluorescence spectrometers (AFS, Beijing Titan Instruments Co., Ltd., China), respectively.The optimized operation conditions of the instruments are listed in Table S1 (Supporting Information).All sample concentrations were the average of three measurements, with a standard deviation of less than 5%.

Nemerow pollution index method
The monomial pollution index (P i ) can reflect the pollution degree of a certain heavy metal, when combined with Nemerow pollution index (P N ) method, can comprehensively reflect the heavy metal pollution status of the study area.The calculation formula (1-2) were as follows: where C i and S i are the measured and the background concentration of HMs in Tibetan soil (mg/kg), respectively.The background values of eight HMs in Tibetan soil are listed in Table 1.P avg and P max are the average and the maximum value of P i , respectively.The pollution index for P i and P N for pollution evaluation criteria are shown in Table S2 (Supporting Information).

Potential ecological risk index
The potential ecological risk index (RI) assessment organically combines the toxic effects and environmental effects of HMs, which is conducive to the ecological risk assessment of a individual heavy metal and the comprehensive ecological risk assessment of multiple HMs.RI is calculated using the following Eq.(3): where T r i is the toxic response factor of individual HMs (1, 2, 5, 5, 5, 10, 30 and 40 for Zn, Cr, Cu, Pb, Ni, As, Cd and Hg) (Hakanson 1980).E r i is the ecological risk index of monomial heavy metal.Risk classification and evaluation criteria by potential ecological risk index assessment are shown in Table S3 (Supporting Information).

Positive matrix factorization model
Positive matrix factorization (PMF) model is effective for source apportionment, and widely used to identify the source of HMs in agricultural soils (Li et al. 2022;Zhang et al. 2018).PMF model is a method based on a receptor model and uses sample composition to identify pollution sources (Wang et al. 2021).PMF model does not rely on emission (1) conditions, meteorological and topographical data of the emission source, nor does it need to trace the migration process of particulate matter, and it is easy to operate with reliable results.The basic principle of the PMF model is as follows: assuming that matrix X is the measurement sample data, matrix X can be decomposed into a fraction matrix G, a loading matrix F and a residual matrix E, which is calculated as follows: where X ij is the content of element j in sample i, G ik is the contribution of source factor k to sample i; F kj is the content of element j in factor k; E ij is the residual matrix, p is the number of factors.
The matrix X is decomposed by the PMF model to obtain the optimal matrices G and F, which leads to the minimum objective function Q: where U ij is the uncertainty of element j in sample i.When the concentration of an element is below or above the (4) detection limit, the equation for the concentration and uncertainty of the element is as follows: When the concentration is below the detection limit: When the concentration is above the detection limit: where x ij and d ij are the input concentration and detection limit, σ is the uncertainty, and c ij is the measured concentration (mg/kg).

Statistical analysis
Data analysis were performed using SPSS 24.0, the mean value, standard deviation, maximum value, minimum value and coefficient of variation (CV) were calculated and used to describe the concentration characteristics of HMs in soil.Considering the relationship between variables, after the maximum rotation of variance, factors with eigenvalues greater than 1 were extracted for principal component analysis (PCA).ArcGIS 10.5 software was used to map (6) the sampling sites within the study area and to display the spatial distributions and potential ecological risk assessment of HMs.ArcGIS 10.5 software was used to map the distribution of sampling sites in the study area, present the spatial distribution characteristics of HMs concentrations and potential ecological risk assessment values.Graphs such as correlation heat maps were obtained using Origin 2018 software.Source apportionment of HMs was resolved using EPA PMF 5.0 software.The PMF model was used to quantify the main sources of HMs in agricultural soil.By setting different factor numbers (3-7), running times (100 times) and minimizing Q value, the optimal number of sources and factor contribution rate can be obtained.

Descriptive statistics of soil heavy metals
The descriptive statistics for HMs concentration in the topsoil of agricultural land in the YZRB is presented in

Spatial distribution
To better analyze the spatial changes of HMs in soils and identify the possible sources, the spatial distribution patterns are presented in Fig. 2. Obviously, the spatial distribution of HMs in agricultural soil of the YZRB was highly variable, and the concentration of HMs in soil of Nianchu River was higher than that of Lhasa River and YZRB.The rapid development of agriculture utilization may be the causes attributed to the soil HMs accumulation (Wang et al. 2019;Yang et al. 2021a;Baltaset al. 2020).Specifically, high value of Cu (> 44.03 mg/kg) was relatively scattered spatial distribution, mainly concentrated in the Lhasa River Basin, which may be related to the upstream copper mining.The distribution of Pb was similar to that of Cu, but the difference was that the enrichment trend of Pb around the city was more obvious, indicating that Pb was closely related to anthropogenic activities.The concentration of Zn was generally high, which may be associated with the minerals.Spatial distribution for Cr and Ni in soils was highly similar.The high concentrations areas for Cr (higher than 152.56 mg/kg) and Ni (higher than 85.58 mg/kg) were mainly found in the facility farmlands around Shigatse city, which was presumably related to agricultural activities.The concentrations of Cd in the Nianchu River and Lhasa River were high, and the spatial distribution was similar to that of Pb, the concentration decreases with the distance from the city, indicating that they have the same source.As was high (> 30.49mg/kg) in the tributary of Lhasa River and Nianchu River, more evidence shows that As was a geothermal source.Hg exhibited a higher abundance in the Nianchu River Basin compared to the other regions, and gradually increased with anthropogenic activities.

Heavy metal pollution and risk assessment
To assess the HMs pollution, the values for P i were first summarized.The proportion of pollution degree of sampling sites based on the obtained values of P i is listed in Table 2. Correspondingly, all the eight HMs present a contamination risk.Fortunately, Pb and Zn showed no-pollution or slight pollution levels.However, the P i indices of Cr, Ni, and Hg were relatively high, indicating a severe contamination risk.Hg pollution was the most serious of all elements.Among them, clean and severe pollution accounted for the same proportion, with the main pollution concentrated in light, followed by moderate.Moreover, Nemerow pollution index based on P N values showed that the pollution of HMs in the whole basin was relatively serious, half of the sample sites showed moderate to severe pollution, and the remaining sample sites showed light pollution.The assessment results of different watersheds showed that the P N values was in the order of Nianchu River > Lhasa River > YZRB mainstream.
In the Nianchu River, the highest value of P N was observed in the facility farmlands in the Sangzhuzi district of Shigatse (15.82).Consequently, most of the agricultural lands in the study area were polluted, while the contamination level was generally low.
The mean E r i value for the eight HMs in the soils of the study area was 45.77 (Cd), 81.73 (Hg), 4.88 (Pb), 2.82 (Cr), 7.74 (Cu), 1.18 (Zn), 11.46 (Ni) and 11.49 (As), respectively.The proportion of pollution degree of sampling sites based on the obtained values of E r i is listed in Table 3.Fortunately, it can be clearly seen that Cu, Pb, Zn, Cr, and As posed a low-potential ecological risk.For Ni, 92.86% of the sampling sites were low risk, and only a few sampling sites in the facility farmlands area of Shigatse have slightly higher risk.Cd mainly showed moderate ecological risk, accounted for 57.14%, and there were 5 sample sites with E r i value of Cd greater than 160.Seriously, the E r i value for Hg ranged from 7.89 to 293.85, and 41.07% of the sites showed moderate-to-serious ecological risk levels.Conclusively, Hg and Cd were important contributors with high ecological risk assessment values.Clearly, the spatial distribution for RI values of soils in the study area was visualized using GIS method.The RI values in Fig. 3 ranged from 55.41 to 400.03, with an average of 167.07.Specifically, the RI values in the Nianchu River Basin mostly showed a moderate risk level, and gradually increased from southeast to northwest, in line with the direction of the river.Most of the Lhasa River Basin and the YZRB mainstream were at a low ecological risk level, and individual sites showed a considerable ecological risk level.The RI values in the Lhasa River Basin and the YZRB mainstream gradually decreased with increasing distance from the city.In conclusion, the level of risk in the study area was closely related to anthropogenic source.

Correlation analysis
Analyzing correlations between HMs can provide a better understanding of their sources, and strong affinity between them indicates that the sources are likely to be the same (Li et al. 2012).Due to the distance and possible spatial variability, spearman correlation was conducted separately in the three areas.As shown in Fig. 4a  Inconsistent with the previous, the three elements Cd, As, and Hg seem to have a high correlation with all elements, due to their low detection levels.

Source analysis of PMF model
The possible sources of HMs were quantitatively analyzed by the PMF model analysis.The factor profiles and corresponding contributions of HMs are presented in Fig. 4a.Four contribution factors were found in the Nianchu River Basin.Factor 1 (F1) was mainly loaded on Cu (33.2%),Pb (39.3%),Zn (30.7%) and Cd (49.9%).As shown in the previous correlation results, these elements have a significant positive correlation (p < 0.01) and coincide with spatial distribution patterns.Meanwhile, the spatial distribution also revealed that its concentration change decreased with the increase of urban distance, while the agricultural land in the study area was distributed along the road with heavy traffic.Generally, Pb, Cd and Zn are often used as vehicle tail gas sources, and accumulated in soil through atmospheric deposition, dust adsorption, etc. (Facchinelli et al. 2001;Huang et al. 2018;Liu et al. 2014;Pardyjak et al. 2008).Therefore, F1 was considered as a source of traffic pollution.In factor 2 (F2), Cr (63.3%) and Ni (82.5%) were the dominant elements with high variability and influenced by anthropogenic activities.HMs in agricultural soils may come from various anthropogenic sources, such as fertilizers, pesticides, and sewage irrigation, especially for seasonally planted vegetable fields with continuous human interference.Previous studies have shown that the concentration of Cr in fertilizers is generally high.Cr-containing fertilizers were important Cr sources for agricultural soils in China (Li et al. 2020a, b).Cr and Ni were highly correlated and had higher concentrations in the facility agriculture fields than that of farmland soils, indicating that these two elements are mainly input from agricultural sources.Factor 3 (F3) was predominantly characterized by As (71.9%), as mentioned previously, As seems to have little change and weak correlation with other metals.
Considering that TP is an As-rich region (Li et al. 2014;Sun et al. 2020;Yang et al. 2021b), suggesting that it originated predominantly from the sources rocks and were barely influenced by human activities.Hg (64.6%) was the dominant element in factor 4 (F4).Human sources are the main source of Hg pollution (Cai et al. 2015), and Hg is also the main component of pesticides and fertilizers (Giersz et al. 2017;Men et al. 2019).Moreover, the enhanced accumulation of Hg in the environmental medium of the TP is more commonly understood as atmospheric deposition (Yang et al. 2010;An et al. 2022;Du et al. 2022).Therefore, F4 can be identified as a mixed source of agricultural and atmospheric deposition.
The factor profiles in the Lhasa River Basin are shown in Fig. 4b.In F1, the contribution rate of Hg (74.8%) was high, and the Hg in the Lhasa River Basin varies greatly.According to E r i value, most sampling sites were polluted, so it was speculated that Hg pollution was related to human activities.Therefore, F1 can be attributed to the source of agricultural activities.Pb (73.4%),Zn (67.6%) and As (79.6%) had a higher contribution to F2. Irrigation with the water from Lhasa River is one of the possible ways for local farmland soils HMs enrichment, as the discharge of mining wastewater and geothermal water from the Duilongqu tributary are the causes of HMs (such as Zn, As, and Pb) pollution in the Lhasa River (Mao et al. 2019;Xiao et al. 2023).Therefore, F2 was inferred as mining exploration.In F3, Cu (47.2%),Cr (45.3%), and Ni (52.4%) were the dominant elements.The concentrations of Cu, Cr, and Ni in the Lhasa River Basin were below the local soil background values, indicating that these HMs were less affected by human activities.With the natural weathering of parent minerals, these HMs are released into the soil.In conclusion, F3 can be identified as the source of parent minerals, which has been proved in the previous studies.
The factor profiles in the YZRB mainstream are shown in Fig. 4c.Cr (52.5%) and Ni (70.2%) were the dominant elements in F1.Zn (75.3%) and Pb (76.7%) were the leading contributions in F2.Hg (73.2%) was the dominant element in F3.Although three contribution sources were obtained in the PMF model, the spatial distribution patterns of these elements were very similar.Based on the E r i values, these elements in the YZRB mainstream were at a clean level, with their concentrations were lower than the local soil background values.Correlation analysis also showed a high correlation between them, intensively indicated that these elements cannot be contributed by human activities.With the natural weathering of the parent minerals, these elements are released into the soils of the YZRB mainstream.

Conclusions
In conclusion, the concentrations, spatial distributions, sources, and potential ecological risk of eight selected HMs in the topsoil of agricultural land in the YZRB were comprehensively investigated.The concentration of HMs in soil exceeded the corresponding local background values, but did not exceed the national soil pollution risk screening values, implying that the study area was considered to be a clean state with limited human impact.Spatially, the high value areas of Pb, Cu and Cd appeared around the mining exploration area, while Cr and Ni appeared around the facility agriculture fields, implying that human activities have a contribution.The ecological risk index of different Basins based on RI values showed that the order of Nianchu River > Lhasa River > YZRB mainstream, was consistent with that of distribution intensity of facility agriculture.Hg and Cd were the main contributors to the results of high ecological risk assessment.PMF source analysis shows that the HMs in the soil of the YZRB were affected by traffic, agricultural activities and soil parent minerals.The research results will have important reference significance for identifying the sources of HMs in agricultural soils, for agricultural safety production, and soil pollution prevention in the YZRB.

Fig. 1
Fig. 1 Location map of the study area and the sampling sites of agricultural soils in the YZRB

577 Page 6 of 11 Fig. 2
Fig. 2 Spatial distribution of HMs in agricultural soils

Fig. 3
Fig. 3 Spatial distribution of RI values of HMs in agricultural soils

Fig. 4
Fig. 4 Spearman correlation and factor profiles of HMs in agricultural soils: a Nianchu River; b Lhasa River; and c YZRB mainstream

Table 1 .
(Huang et al. 2023;An et al. 2022;Bai et al. 2014;Li et al. 2009;Tian et al. 2022)s Cr > Zn > Ni > Cu > Pb > As > Cd > Hg, with their average values (ranges) are 109.04,86.93,73.58,33.91,28.21,21.49, 0.12, and 0.05 mg/kg, respectively.For comparison, the HMs (average concentration) measured in this work were found not to be significantly different from previously published data on these regions(Huang et al. 2023;An et al. 2022;Bai et al. 2014;Li et al. 2009;Tian et al. 2022), and the concentration levels were observed to be at almost the same level.In order to present the evolution of HMs in the continental crust more clearly, soil background were also listed in Table 1(Taylor and

Table 2
Statistical analysis of monomial pollution index (Pi) of heavy metals in soils