Reconstruction of atmospheric soot history in inland regions from lake sediments over the past 150 years

Historical reconstruction of atmospheric black carbon (BC, in the form of char and soot) is still constrained for inland areas. Here we determined and compared the past 150-yr records of BC and polycyclic aromatic compounds (PACs) in sediments from two representative lakes, Huguangyan (HGY) and Chaohu (CH), in eastern China. HGY only receives atmospheric deposition while CH is influenced by riverine input. BC, char, and soot have similar vertical concentration profiles as PACs in both lakes. Abrupt increases in concentrations and mass accumulation rates (MARs) of soot have mainly occurred since ~1950, the establishment of the People’s Republic of China, when energy usage changed to more fossil fuel contributions reflected by the variations in the concentration ratios of char/soot and individual PACs. In HGY, soot MARs increased by ~7.7 times in the period 1980–2012 relative to the period 1850–1950. Similar increases (~6.7 times) were observed in CH. The increase in soot MARs is also in line with the emission inventory records in the literature and the fact that the submicrometer-sized soot particles can be dispersed regionally. The study provides an alternative method to reconstruct the atmospheric soot history in populated inland areas.

Presently, historical reconstruction of BC variations in the atmosphere are primarily available from remote areas such as high-latitude ice-covered regions [15][16][17] , where ice cores can be used to record atmospheric BC, especially the resulting soot deposition 15 . In addition, statistical data on fuel usage along with their combustion emission factors have been utilized to reconstruct the BC emissions regionally and globally 18,19 to produce a rough indication of atmospheric BC history. However, the ice core reconstruction cannot reflect inland BC history, where BC emissions are dominated by sources related to urban and industrial activities. Moreover, the reconstructions of the long-term history from fuel usage are limited because of the unavailability of fuel usage data in the literature. Particularly, some important sources such as residential emissions from cooking and heating are difficult to estimate. Thus, more reliable measurements of atmospheric BC, especially soot emissions (and subsequent deposition to the land surface) and the reconstruction of its history are needed for inland areas [20][21][22] .
Lake sediments can record the long-term BC history from both atmospheric and river inputs in inland areas because BC is inert and resistant to degradation 1 ; however, it is hard to differentiate atmospheric deposition from the river input. Efforts to reconstruct the atmospheric BC history via sediment records were conducted in remote areas 22 , but these results cannot be easily extended to more densely populated regions. Soot and char as the two subtypes of BC have different size fractions and thus different transport pathways 2,14,23 ; the smaller soot fraction is atmospherically widely distributed, while the larger char fraction tends to be deposited close to the emission source. Consequently, soot can be far distributed in the atmosphere 2 and thus more reliably reflects the combustion history at a regional scale than char. Thus, we hypothesize that lake sedimentary soot may be used to reflect its atmospheric deposition history. To test the hypothesis, the comparison of soot history from a common lake that receives riverine inputs with a "standard" archive that is almost exclusively driven by the deposition from the atmosphere, is needed. Huguangyan Maar lake (HGY) is suggested to uniquely receive atmospheric deposition with very limited inputs from other sources such as tributary rivers and soil erosion 24 and thus can provide such "standard" archive for this study.
PACs are co-produced with BC from biomass and fossil fuel combustions 25 , and may even serve as precursors of soot formation in combustion processes 26,27 . Although some PACs such as polycyclic aromatic hydrocarbons (PAHs) have been extensively studied in lake sediments of China and worldwide [28][29][30] , very few studies have been extended to their derivatives such as oxygenated PAHs (OPAHs) and azaarenes (nitrogen heterocyclic PAHs, AZAs). Some of the derivatives have higher mutagenic and/or carcinogenic effects than their parent materials 31 . The polar PAC derivatives can be formed primarily by the combustion process itself or secondarily by reactions between parent-PAHs and oxidants such as ozone (O 3 ), the hydroxyl radical (OH·) or NO x , photolysis and microbial degradation of parent-PAHs 31 . Consequently, the composition of the polar PAC mixture can change during long-range transport. Combined analysis of PACs, BC, char, and soot in lake sediments facilitates our identification of their origins, sources, transport, and fates, and can confirm the reliability of their reconstructed emission histories.
In this study, sedimentary BC, char, soot, and PACs histories covering the past ca. 150 years were reconstructed from two lakes of eastern China, HGY and Chaohu lake (CH). HGY has only a minor local tributary while CH is fed by several rivers in the inland. Both of the lakes are located in the most industrialized region of contemporary China. We used concentration ratios of char/soot and individual PACs to infer the changes of energy usage. Moreover, we hypothesized that the sedimentary records of char, soot, and PACs could reflect the start of the fast industrialization in China after the 1950s and the changes in type of energy sources from mainly wood burning in the pre-1950s to fossil fuels in the post-1950s. By comparing soot records of the two lakes, in addition to BC emission inventories in the literature, we demonstrated that the sedimentary record can be used to reconstruct atmospheric soot history in populated areas.

Chronology dating
The constant rate of supply (CRS) 210 Pb model 32,33 was applied for age dating of the two studied cores (Fig. 1). Also, 137 Cs activities are presented as independent dates to confirm the CRS chronology (A3 and B3 in Fig. 1). The reconstructions with the CRS model for both cores precisely reflected the starting of global nuclear testing of ~1952-53 and the global nuclear fallout of ~1963-64, which are in good agreement with the 137 Cs data. Furthermore, the small peak of the 137 Cs activity at 10.75 cm in the CH core corresponds with the Chernobyl nuclear accident of 1986. We observed substantially lower mass accumulation rates (MARs) in HGY than in CH, which supports the dominant atmospheric deposition of particles in HGY. The 137 Cs activities of the HGY core, however, did not decrease after the 1964 peak. A similar finding was reported for the Imitavika Lake in the Arctic 34 and Sombre Lake in the Antarctic 35 and attributed to a constant delivery of 137 Cs to the lake sediments via atmospheric deposition and some suspended matters from glacier melt water. We therefore suggest that 137 Cs is being continuously delivered to the HGY sediments by the atmospheric deposition and some soil erosion. As 210 Pb has a half-life of 22.26 years, we can only produce reliable chronologies of 150-170 years. The dating of sediment layers older than 1850 were inferred from average accumulation rates of the upper layers. Consequently, our dating of the sediments older than 1850 is less reliable. Therefore, we refrained from discussing the older sediments in this paper.  38 ; however, the strong increase in the 1950s had not been reported previously.

Historical variations in concentrations and
From pre-1850 to ~1930-40s, the overall trends of BC, char, and especially soot concentrations and MARs varied little in spite of a few peaks, which are comparable with those in 1950-1980 (Fig. 2). The lowest concentrations of BC and char occurred, although not notable, for the two cores in the same time interval (~1930s) when China fell into warfare.
Similar vertical concentration trends of PACs with BC, char, and soot. The vertical trends of the concentrations of parent-PAHs (∑28PAHs, excluding perylene) and OPAHs (∑15OPAHs) paralleled those of BC (and their sub-fractions) in both lakes, indicating dominant human emissions for these pollutants and their co-emission from combustion sources. In a previous study, we found that PAHs concentrations are more closely correlated with those of soot in sediments than in soils, indicating that soot is a major sorbent of PAHs in sediments 39 . Our finding of strong correlations (r = 0.76-0.88, p < 0.001) of the concentrations of ∑28PAHs and ∑15OPAHs with those of soot in the HGY and CH sediment cores (Table S2) is in line with the previous observation. The correlations between PACs concentrations and those of BC were weaker in CH than in HGY, which may be attributable to a dilution effect from the river inputs into the lake.
However, there were also asynchronous developments of the concentrations of ∑28PAHs and BC fractions (Fig. 2). In the HGY core, the lowest ∑28PAHs occurred around the 1860s, the beginning of the Westernization Movement (WM), the pioneer industrialization period in China; however, the lowest soot concentrations were found around the 1930s. Moreover, in recent years (after the economic reforms starting in 1978), there were rapid increases in concentrations of ∑28PAHs, ∑15OPAHs, and ∑4AZAs (sum of the concentrations of four azaarenes) in HGY, whereas the concentrations of BC decreased after 2004-06 (Fig. 2). This indicates that the dominant factor determining the historical variations of PACs are their emission strengths, although their relationships with BC are enhanced by their sorption to co-emitted carbon fractions. Even though BC fractions and PACs are co-emitted from all combustion processes, each combustion source can generate different characteristic mixtures of carbonaceous matters and PACs 40 . PAHs are generally produced at relatively lower temperatures than soot and can also act as precursors of soot 26 . The dissimilar distribution patterns of ∑28PAHs, ∑15OPAHs, and ∑4AZAs in HGY and CH (Figs S2-S4) suggest that the two regions were influenced by different BC and PACs sources.
AZAs showed different vertical concentration distributions in the two lakes (Fig. 2). In HGY, the vertical distribution of ∑4AZAs paralleled those of the concentrations of BC fractions, ∑28PAHs and ∑15OPAHs, with two sharp increases at the beginning of the 1950s and in the late 1970s, respectively. However, in CH, ∑4AZAs concentrations increased from the pre-1850s to 1950 but thereafter decreased. We attribute the differences to the river inputs to CH and thus a specific local source enriched in AZAs. Our results demonstrate that HGY mainly received BC and PACs from the atmospheric input with some minor input from soil erosion immediately adjacent to the maar lake, while the input from the tributary rivers was a significant source of BC and PACs to the sediments of CH.
Previous studies have suggested that perylene in pristine sediments originates mainly from in-situ biogenic diagenesis under anoxic conditions 41,42 . In line with this, perylene in HGY had higher concentrations in the lower parts of the core (Fig. S5), which is opposite to other PACs. Relatively lower concentration and different vertical distribution pattern of perylene was found in CH compared to HGY (Fig. S5). We suggest that this is associated with the different oxygen concentrations in the two lakes. HGY with a water depth of ~20 m likely shows anaerobic conditions near the ground. For CH, the water depth is only ~2.7 m, and the water currents from the tributary and leaving rivers likely create aerobic conditions near the ground.
Changing composition of fossil energy sources inferred from char/soot and PAC concentration ratios. Industrialization in the Western developed countries began in ~1750, the start of the Industrial Revolution. Although pioneer industrialization in China can be traced back to the 1860s (the beginning of the WM), accelerated industrialization occurred after the establishment of the PRC (1949-50). The char/soot ratio has been suggested to be indicative of the type of fuel combusted in urban areas. Biomass burning usually produces higher char/soot concentrations ratios while fossil fuel combustions, especially motor vehicle emissions, produce lower ratios 43,44 . There was a sharp decrease in char/soot ratios in both lakes in ~1950 (A3 and B3 in Fig. 1), supporting the increase in fossil fuel combustion contributions after the establishment of the PRC. However, in both lakes a small increasing trend of the char/soot ratio was observed since 1950. This is very likely due to the rapid increase in the population of eastern China in this time period (Fig. S6). The increasing population probably resulted in significantly increasing deposition of char into these lakes from multiple local sources related to the increased human activities. Abrupt increases in the MARs of retene further suggest increases in biomass burning (in both regions), which will also lead to increasing char/root ratios (Fig. S7).
Trends of selected concentrations ratios of individual PACs may assist in tracking PACs sources 45,46 , although the composition of the PAH mixtures are potentially altered during transport in the atmosphere and after deposition to soils before being eroded into lakes. The concentration ratio of low-molecular-weight (LMW) to high-molecular-weight (HMW) PAHs (Fig. 3, see explanations in the footnote, similar as below) is the most common tool to reflect the relative contribution of biomass to fossil fuel 46 . This ratio had an overall decreasing trend from bottom to top of the CH core while the lowest value was found in the 1950s (Fig. 3). This is consistent with the char/soot ratio that proves the increase in the contribution of fossil fuel combustion in the 1950s. In HGY, the LMW-/HMW-PAHs concentration ratios were higher in the bottom than the top of the core; however, the lowest ratio was seen in the beginning of the WM (around 1860s) because HGY is located much closer to the southern countries. This finding suggests that the pioneer industrialization by the WM had released more toxic PAHs (HMW-PAHs) into the atmosphere than the later development, which is reflected by the low LMW-/ HMW-PAHs concentration ratios.
Similar to the trends of LMW-/HMW-PAHs concentration ratios, ∑15OPAHs/∑28PAHs also decreased since ~1950 (Fig. 3). This further supports our conclusion that the combustion of fossil fuel which emits lower OPAHs-to-parent-PAH amounts (compared to combustion of biomass 47,48 ) increased during and after the 1950s. Finally, the decreases in the concentration ratios of IcdP/(IcdP+ BghiP), FLUA/(FLUA+ PYR), and ANT/ (ANT+ PHE) (see explanations in footnote of Fig. 3) in the two lakes, more obviously in HGY, are in line with the significant increase in petroleum combustion 29,45 . Atmospheric vs. river debris inputs of pollutants. The concentration ratios BeP/BaP and 9-FLUO/ FLUO were higher in HGY than in CH sediments (Fig. 4). Because BaP is more susceptible to atmospheric degradation than BeP 47,49 and 9-FLUO is one of the oxidation products of FLUO 50,51 , long-term exposure to atmospheric ultra-violet (UV) radiation would lead to higher BeP/BaP and 9-FLUO/FLUO concentration ratios. Therefore, we interpret the differences in BeP/BaP and 9-FLUO/FLUO concentrations ratios between the two lakes as an indication of the dominance of atmospheric deposition to HGY, while river input of PACs plays an important role in CH. In addition, as population and human activities in eastern China increased, it is reasonable that the contribution from local emissions in the HGY region has increased in recent years. This also explains the decreases in BeP/BaP and 9-FLUO/FLUO concentration ratios (Fig. 4).
In HGY, the highest MARs of soot, which occurred in recent years, is ~21.3 times of the lowest MARs of soot in the 1930s, and the MAR of soot in the period 1980-2012 is ~7.7 times that of the period 1850-1950 (Table 1). However, the average MAR of char in the period 1980-2012 is only ~3.6 times that of the period 1850-1950. These observed differences in the temporal trends of char and soot are an indication of a higher contribution of local sources to char than soot in HGY.
The different transport distances of char and soot can also explain their historical variations in CH. The average MAR of soot in the period 1980-2012 is 6.7 times that of the period 1850-1950 (Table 1), which represents only a slightly smaller ratio than in HGY. We attribute the lower MARs of soot to the dilution by soot-poor clastic material deposited into CH by the riverine inputs. It has been shown that the current soot concentrations in the atmosphere are similar regionally 14,23 and thus presumably the current soot deposition to the lake sediments are similar at both locations as well. The consistent increases of MARs of soot in the two lakes between 1980-2012 and 1850-1950 suggest that soot MARs in sediments reflect the variations in atmospheric soot concentrations, which provides a new pathway to reconstruct atmospheric soot history for inland areas using lake sediments. However, in CH there was only a small increase of the char MARs in 1980-2012 compared with that in 1850-1950 (~1.8 times on average), supporting that local river inputs can dilute the concentrations of char. Husain et al. 22 have also reconstructed the atmospheric BC history using a lake sediment record at Whiteface Mountain, New York. They determined BC concentrations with the thermal/optical NIOSH method, which has been suggested to measures BC fractions tending to more resemble refractory BC such as soot than the IMPROVE method used in our work 38,52 .
Because sedimentary soot originates mainly from regional atmospheric deposition, the ratio of ∑28PAHs concentrations to soot can be used to reflect the local-to-regional influences. Small variations in ∑28PAHs/soot in HGY confirm the mainly atmospheric deposition of PAHs and soot to the Maar lake, while the increasing trend of ∑28PAHs/soot ratio in CH after 1850 is likely associated with increasing contributions from the local human activities via riverine inputs (Fig. S8). The far higher ratios of ∑4AZAs and retene to soot in CH than in HGY further support this conclusion.

Comparison of the soot deposition history with BC emission inventories in China.
The sedimentary record of soot concentrations in HGY and CH parallels the BC inventories for China 18,19,53,54 (Fig. 5). The slight deviations between our lake soot records and the emission inventories of different groups 18 that there may be local modifications of the national deposition patterns. Thus, we suggest that lake sediments are a reliable archive of the atmospheric soot history.
The available studies using ice cores to reconstruct the soot deposition history in China demonstrated a ~2.5 times increase in soot deposition between the pre-1950s and the post-1980s 55,56 , which is far lower than that of our sediment data (~6.7-7.7 times). This suggests that the long-range transport of soot in remote areas may decrease the influence of human activities, and thus ice core records cannot document the atmospheric soot history in populated areas.
The concentrations of PAHs, OPAHs, and AZAs showed similar vertical distributions as the MARs of soot (Fig. S9). However, the rates of increase in MARs of PACs and soot in the two regions were different (Table 1). It is likely that PACs in the two lakes originated mainly from local/regional emission sources, which explains the differences between the two lakes (Figs S2-S4). The higher increase rates of PACs concentrations in CH than in HGY sediments may indicate the significant additional input of PACs into CH by the riverine inputs. In contrast, the HGY catchment receives PACs only from atmospheric input. The atmospheric deposition of PAHs in HGY parallels their emission inventories 57,58 (Fig. S9), further supporting our suggestion that the Maar lake provides a robust PACs history. Because there is a lack of long-term OPAHs and AZAs inventories in China, our results provide the first historical record of atmospheric OPAHs and AZAs concentration variations at inland regions of China.
Implications of the reconstruction of the history of atmospheric pollutants. The differentiation between char and soot provides a new way to reconstruct the atmospheric soot history. This may be suitable for much longer-term atmospheric soot reconstruction such as for the past 2000 years, the Holocene, and the glacial-interglacial cycles. The atmospheric soot reconstruction in this study provides a robust record, which can be used to test BC emission inventories obtained from fuel usages. An abrupt increase in soot concentrations occurred in ~1950 in China, when more fossil fuels were used (Fig. S6). The timing is far later than that in the western developed countries, where the peak in atmospheric BC was found in the 1910s 16    Although both char and soot are light-absorbing, they have different optical properties. Char absorbs visible light weakly and tends to absorb short-wavelength light more strongly, while the light absorption capacity of soot is strong 44 . The increase in char and especially more soot emissions to the atmosphere associated with the change to fossil fuel usage may result in greater climatic warming effects. However, the climate system is complicated. For example, the composition of carbonaceous aerosol as characterized by the contributions of organic carbon (OC) and BC may vary among different energy source emissions. Because the OC/BC ratios have been suggested to indicate the warming or cooling effects of carbonaceous aerosols 60 , the change from biomass burning to fossil fuel combustion such as that occurring in 1950 seems likely to lead to more warming effects. However, more aerosol emissions from fossil fuel consumption, especially those of light-scattering particles such as sulfate would result in cooling effects. Therefore, understanding the effects of the various compounds emitted to the atmosphere by fossil fuel combustion requires knowledge of the histories of different atmospheric chemical species including trace gases and aerosols and their direct and indirect climatic radiative forcing. Our suggested new method for the reconstruction of the atmospheric soot history for inland regions contributes to closing the knowledge gap.

Method
We selected Huguangyan Maar lake (HGY) and Chaohu lake (CH) in the developed regions of eastern and southeastern China, respectively, for this study (Fig. S1). HGY is located in Guangdong province and has a water area of ~2.5 km 2 with an average water depth of ~20 m, and there are no river inputs and outlets. CH is located in Anhui Province and has a water area of ~775 km 2 with a water depth of ~2.7 m, and there are eight larger tributary rivers and one outlet to Yuxi River, which is a tributary of the Yangtze River. Two sediment cores were obtained using a gravity corer from HGY (HGY12-2) and CH (CH12-3), respectively, in January 2012. The cores were sectioned into 0.5-cm intervals of the upper 20 cm and 1.0-cm intervals of the lower part.
All samples were freeze-dried. Radionuclide concentrations of 137 Cs, 210 Pb and 226 Ra were analyzed for dating by direct gamma counting of 3-6 g of dried sediments using a Canberra HpGe well detector model GCW2022. BC, char, and soot concentrations were quantified using the thermal/optical method with the IMPROVE protocol after acid pretreatment 12,61 . Furthermore, the concentrations of 29 parent-and alkyl-PAHs (including perylene), 15 OPAHs, and 4 AZAs 47,62,63 were measured in the samples. Details of the locations, sampling, dating, analytical procedures and quality assurance/control are explained in the Supplementary Online Materials (SI). A summary of the concentrations of carbon fractions and PACs as well as their abbreviations are presented in Table S1.