To explore geochemical factors affecting the chronological reconstruction of a historical arsenic pollution accident in Lake Qionghai at southwest of China, heavy metals and radionuclides (210Pbex and 137Cs) in two sediment cores were sampled and analyzed. Our study revealed that (1) industrialization and urbanization of Xichang since the mid-20th Century, aggravated soil erosion in the catchment and resulted in a significant and lasting increase in the sedimentation rate of Lake Qionghai; (2) sharp arsenic peaks in two sediment cores recorded an arsenic pollution accident in 2006 in which a large amount of arsenic was discharged into the lake from a chemical plant. Arsenic peaks in sediment cores A and B were dated to 1988 ± 3.5, 2006 ± 3.1, respectively. Only the arsenic peak time in core B: 2006 ± 3.1, agreed with the real pollution accident time of 2006. Linear correlation analysis suggested that arsenic sorption by Fe/Mn minerals and organic materials in sediment core A probably changed the depth distribution of the peak and affected the accuracy of chronological reconstruction. This case study suggested that sorption of heavy metals by Fe/Mn minerals and organic materials should be assessed before the chronological reconstruction of the metal pollution accident in the lake.
Heavy metal contamination in aqueous sediment is of concern for government and public due to its severe toxicity and potential risks in the future (Shi et al., 2015; Wang et al., 2014; Ho et al., 2013). Many studies reveal that heavy metal pollution is related to human activities (Garcia-Orellana et al., 2011; Abrahim and Parker, 2008; Cundy et al., 2003). Moreover, industrialization and urbanization are large-scale anthropogenic activities that usually included deforestation, desertification and pollution. Following these activities, sedimentation in lakes often accelerates and, even worse, the lakes become easily subject to the heavy metal pollution.
Identifying what kind of human activities historically caused severe heavy metal lake pollution is important for public and government in order to prevent more pollution in the future. Chronological reconstruction of metal pollution is an important means to accomplish this task (Garcia-Orellana et al., 2011; Cundy et al., 2003). Nowadays, sediment dating can be realized through calculation following measurement of natural and artificial radionuclides in sediment at different depths. One of the most important modern sediment dating methods is the 210Pb method. 226Ra finally decays to atmospheric 210Pb through 222Rn. Then atmospheric/exterior210Pb (210Pbex) is transported to the lake and deposited to the bottom sediment by precipitation and dry deposition. Two models, based on different assumption: CRS (constant rate of 210Pb supply) or CIC (constant initiate concentration of 210Pb), can be used to interpret the 210Pbex data and date the sediment. Detailed discussions about 210Pb dating have been documented in literature (McDonald and Urban, 2007; Alonso-Hernandez et al., 2006; Appleby and Oldfield, 1978). Another sediment dating method is based on the peaks of artificial radionuclides like 137Cs, 90Sr, 239-240-241Pu (Jeter, 2000) in sediment. This kind of radionuclide in sediment mainly originated from nuclear weapon tests. Enrichment of 137Cs in European sediment appeared in 1953, corresponding to the beginning of nuclear weapon testing (Magninu et al., 1990) and reached a maximum in 1959 and 1963 (Putyrskaya and Klemt, 2007). Moreover, the Chernobyl atomic plant power station accident in 1986 was the most severe nuclear accident in history and injected a large amount of artificial nuclides into the atmosphere (Cambray et al., 1987). Although there were no reports on atmospheric 137Cs fallout in China, many studies showed a 137Cs peak corresponding 1986 in some Chinese lakes due to Chernobyl, e.g. Lake Erhai, in southwest China (Wan et al., 2003) and Lake Guchenghu, in southeast China (Xiang et al., 2002). A sediment dating result using 210Pb can be used only when it is consistent with results from other methods.
The application of sediment dating methods include the chronological reconstruction of heavy metal pollution over time: e.g. sharp mercury peak in the sediment core of Banbury Reservoir in London, which was finally attributed to soil fertilization, road construction and land use for cattle in the catchment (Yang and Rose, 2005). However, these on-going activities appeared as difficult-to-interpret “sharp” peaks, which suggest that the source of the pollution should appear and disappear fast. Incomplete records about pollution incidents make it difficult to trace those pollution-incurring activities in history and to interpret these unusual peaks in sediment profiles. More importantly, whether those sharp heavy metal peaks in sediment profiles can be accurately dated is still unknown. To resolve this problem, Lake Qioinghai was chosen as a case study. An arsenic pollution accident happened in 2006 due to the illegal emission of a local chemical plant to the lake (Ji et al., 2007). The aims of the present study were as follows: (1) chronological reconstruction of this arsenic pollution accident; (2) exploration of the geochemical factors that can affect the dating result of the heavy metal peak in sediment.
Materials and methods
Lake Qionghai is located 5 km southeast of Xichang City of Sichuan Province of China (Figure 1). It has two inflowing rivers: Guanba River and Ezhanghe River, and one seasonal outflowing river: Haihe River (Figure 1). The Haihe River can flow backward into Lake Qionghai when the water level of the river it flows into, the Anning, increases above that of the lake. Most of the time, the water table of Lake Qionghai is higher than the Anning River.
Lake Qionghai is located in a subtropical plateau mountain area. It is a national scenic area and wetland park of China. This lake is located between 27°47′ and 27°52′N, 102°16′ and 102°21′E, with an altitude of 1510 m. Its area is approximately 31 km2. It has an average depth about 10.3 m and reaches 34 m in the deepest region. Its total volume is 320 million m3 and mean annual runoff is 147.3 million m3. Its watershed catchments area reaches 309 km2.
Sediment sampling and preparation
Two sediment cores were set in Lake Qionghai. Sediment core A was set in the position near the Xichang City (Figure 1) to investigate the chronic effect from potential heavy metal sources in the urban and industrial area, and Sediment core B was set in the centre of the lake (Figure 1) to investigate the overall status of heavy metal contamination in this lake. All the sediment samples were collected in November 2011. Sediment cores were collected using a flag-gravity corer (Serial No. XDB0205, China) that has inner and outer diameters of about 9.5 cm and 11 cm, respectively. The depth of sediment cores is shown in Figure 2. In addition, core sediments were undisturbed, as shown by the obvious water sediment interface and the preservation of good lamination in the sediment. The upper part of the sediment cores mainly consisted of dust-colour particles, while the lower part consisted of greyish yellow particles.
Sediment cores were taken out from the inner pipe of the flag-gravity corer. And two sediment cores were immediately sectioned in situ with a ceramic cutting knife at 1.0 cm intervals above 30 cm depth, and 2.0 cm intervals below 30 cm depth. In this process, the knife was frequently washed with pure water to prevent mutual contamination. All sediment sub-samples were put into Teflon packaging to prevent contamination during transportation to laboratory. All sediment sub-samples were freeze-dried in a refrigerator. When dry, all sediment samples were crushed and sieved with a 100-mesh sieve. Finally, fine sediment powder (<150 μm) was collected and weighed before chemical analysis.
Sediment samples were sent to the Nanjing Institute of Geography & Limnology, Chinese Academy of Sciences, to determine the radionuclide activity. Measurement of radioisotope concentration in sediment samples was performed using an ORTEC spectrometer with an HPGe detector of GWL-120-15 (AMETEK-AMT ORTEC Co.) with a resolution of 2.10 KeV at 1.33 MeV line of 60Co. Unlike other well detectors that have a hole all the way through the germanium crystal, GWL Series ORTEC's well detectors have a “blind hole” with at least 5 mm of active germanium at the bottom of the hole. This near 4 π geometry provides the maximum absolute counting efficiency available. The large well (1.55-cm diameter and 4.0-cm long) accommodates an extensive range of sample sizes. 210Pb was measured via gamma emission at 46.5 KeV with the relative intensity of 3.6%. Moreover, 226Ra was measured by 295 and 352 KeV γ-rays based on the assumption of secular equilibrium with its progenies with a relative intensity of 33.6%. Referring to the National Standard of China (GBT 11743–2013, 2013) regarding determination of radionuclides in soil/sediment by gamma spectrometry, the expanded uncertainty (with a confidence level of 95%) of determining the radionuclide activity of 226Ra and 210Pb, in soil/sediment is less than 10%. 210Pbex was calculated according to total 210Pb and 226Ra in the sediment.
where tx is sediment age at the depth x, λ is the decay constant of 210Pb, A0 (Bq cm−2) is the inventory of 210Pbex in the core, and Ax (Bq cm−2) is the inventory of 210Pbex below depth x.
Deionized water and analytical-grade chemicals were used in the analysis. Fine sediment samples (<150 µm) were digested with a mixture of HNO3, HClO4, and HF (USEPA Method 3052; USEPA, 1996a) and the diluted solutions were measured (USEPA Method 3050B; USEPA, 1996b). For each sediment core, one replicate sediment sub-sample was analyzed for each three to test whether there is significant error in sampling and analysis. Metals like Cd, Cr, Ni, Th, U, Pb, Zn, Cu, Tl, Mo, Mn, and Fe were measured using inductively coupled plasma mass spectrometry (ICP-MS); the mean recovery percentages ranged from 98 to 105%. Hg, As, and Se were measured using atomic fluorescence spectrometry (AFS); the mean recovery percentages ranged from 98% to 101%. The total organic carbon (TOC) in the sediment was determined using titration with Cr3+ after wet oxidation (Schumacher, 2002).
Results and Discussions
Profiles of 137Cs in sediment cores
The depth distribution of 137Cs for sediment cores A and B were shown in Figure 2. Lake Qionghai, Lake Erhai, and Lake Chenghai are all located at the southwest of China. Studies on Lake Chenghai (Wan et al., 2005) and Lake Erhai (Wan et al., 2003) both suggested the existence of a 137Cs peak attributed to the Chernobyl Nuclear accident at 1986. Therefore, we inferred that the largest peak of 137Cs in the two sediment cores of Lake Qionghai (in sediment cores A at 18.5 cm and B at 19 cm) were referenced to 1986, which was also proved by the independent dating result from CRS model using 210Pbex (Figure 2). In addition, some small 137Cs peaks distributed around the largest peak in cores A and B. Input of 137Cs into a lake mainly depends on the atmospheric deposition and the contribution from the catchment's area. Catchment-derived contribution was probably a second important source of 137Cs for this lake, because this lake is not an enclosed one but connected with two major inflowing rivers at the North and South shore and one seasonal out flowing river at the Northwest shore (Figure 1). Therefore, 137Cs contribution from inflowing rivers (Guanbahe River and Ezhang River) was more easily accumulated in core B than A (Figure 1). More 137Cs peaks in core B than core A supported this (Figure 2).
Sediment dating using CRS model
Following the procedure suggested by Appleby and Oldfield (1983), the 210Pb data has been assessed. The CRS model was considered suitable because the relationship between 210Pbex and depth is non-linear for cores A and B. Sediment dating results using the CRS model and 137Cs time tracer were presented in Figure 2. Result of sediment dating reflected that sediment core A and B were accumulated about from 1930 to 2011 and from 1864 to 2011, respectively. The result of sediment dating with the CRS model was also consistent with the 137Cs peak corresponding 1986 for both cores A and B (Figure 2), suggested that sediment dating using the CRS model is reliable.
The sedimentation rate of the two cores were calculated based on the ratio of sediment sub-sample dry mass with age and presented in Figure 3. At the position of core A, the sedimentation rate increased from 0.0558 ± 0.0808 g·cm−2·yr−1 (1956) and reached a maximum value of 0.3215 ± 0.0930 g·cm-2·yr-1 (1964). This kind of high sedimentation rate in waters at the position near Xichang city (Figure 1) was probably related to soil erosion in the urbanization of Xichang in the period from 1956 to 1964. This type of rapid sedimentation in the same period (1950 s to 1960 s) due to the large-scale deforestation and urbanization in China has also been found in the upper Yangtze River (Lu and Higgitt, 1998) and lakes (Xiang et al., 2002).
At the position of core B, the average sedimentation rate (ASR) was low before 1975 (ASR: 0.0936 ± 0.0978 g·cm-2·yr-1 during the period from 1864 to 1975) and increased after 1975 (ASR: 0.2657 g·± 0.0944 cm-2·yr-1 during the period from 1976 to 2011). Moreover, the first peak of the sedimentation rate at the position of core B appeared at 1975, which was the beginning of local urbanization in rural regions around the lake. In summary, the high sedimentation rate of Lake Qionghai currently in effect, has been occurring since 1990 for core A and 1975 for core B, respectively.
Geochemical record of arsenic pollution accident
Mean concentration of arsenic in sediment cores A and B was 7.24, 9.71 mg kg-1, respectively, both exceeding the arsenic abundance in the crust of China: 1.9 mg kg-1 (Lee and Yao, 1970). Two arsenic profiles showed unusual sharp peaks in the sediment cores (core A at 17.5 cm; core B at 5.5 cm), while sedimentation rate and other metals (Cd, Cr, Ni, Th, U, Pb, Zn, Cu, Tl, Mo, Mn, Fe and Hg) did not show similar peaks, suggesting there was an arsenic contamination event. This is consistent with previous record about an arsenic pollution accident at Lake Qionghai resulting from the Kangxi chemical plant at 2006 (Ji et al., 2007). A similar kind of mercury peak in lake sediment has been reported in cores from Loch Chon at Banbury Reservoir in London, UK, which was finally attributed to the anthropogenic input of heavy metals into the catchment (Yang and Rose, 2005). Whereas Banbury Reservoir had a mercury peak at only one core of six, both cores at Lake Qionghai showed unusual arsenic peaks (Figure 4).
Chronological reconstruction of heavy metal contamination in sediment can be used to trace the time mark of pollution activities in the past. In comparison to the real accident time of 2006, arsenic peaks in cores A and B corresponded to 1988 ± 3.5 and 2006 ± 3.1 according to the sediment dating result. Only the arsenic peak time in core B: 2006 ± 3.1 agreed perfectly with the real time at 2006.
Factors affecting arsenic peak dating
Linear correlation analyses between As and Mn, As and Fe, As and TOC in three sediment cores were performed and presented in Figure 5. Generally, Fe/Mn minerals and organic materials in sediment can adsorb arsenic in water, and their sorption degree mainly depended on the physical chemistry conditions (Deschamps et al., 2003; Catalano et al., 2011). Arsenic in core A showed significant and positive linear correlations with Mn/Fe/TOC and other trace heavy metals (e.g. Mo, Co), while As in sediment core B did not show any linear correlations with Mn/Fe/TOC (Figure 5) or other trace elements (not shown). Thus our results suggested that the arsenic in core A probably vertically redistributed, while the arsenic profile in core B has been preserved well. The larger width of the arsenic peak in core A than that of core B also supported this point (Figure 4). Arsenic sorption by Mn/Fe/Organic materials in sediment core A probably changed the depth distribution of arsenic peak. Therefore, chronological reconstruction of the metal pollution accident in the lake can be affected by the sorption of metals by Fe/Mn minerals and organic materials in the sediment.
Industrialization and urbanization in Xichang city aggravated soil erosion and resulted in a significant increase in the sedimentation rate of Lake Qionghai since 1950 s. The dividing time between the high and low sedimentation rate agreed with the outset of local urbanization. Sharp arsenic peaks in the two sediment cores recorded an arsenic pollution accident in 2006 (Ji et al., 2007). Deviation of the arsenic peak dated time of 1988 ± 3.5 for core A and real pollution time of 2006 was attributed to the redistribution of the arsenic peak due to the sorption of arsenic by Fe/Mn minerals and organic materials. Only the arsenic peak time in core B: 2006 ± 3.1 agreed perfectly with the real time of 2006. This case study suggested that sorption influence of Fe/Mn/Organic materials on heavy metals in sediment cores should be assessed before chronological reconstruction of metal pollution in history.
We are grateful for the aid of Huang Haofei, Yang Weihe, Liao Jingxing and Liao Chao in the field sampling.
The National Natural Science Foundation of China (41373120) and the Key Project of Natural Sciences of the Department of Sichuan Provincial Education (15ZA0077, 16ZA0098) financially supported this research