Biomarkers of Modern Plants and Soils from Xinglong Mountain in the Transitional Area Between the Tibetan and Loess Plateaus

See discussions, stats, and author profiles for this publication at: http://www.researchgate.net/publication/223698344

Biomarkers of modern plants and soils from Xinglong Mountain in the transitional area between the Tibetan and Loess Plateaus

ARTICLE in QUATERNARY INTERNATIONAL · MAY 2010 Impact Factor: 2.06 · DOI: 10.1016/j.quaint.2009.12.009

CITATIONS READS 11 42

7 AUTHORS, INCLUDING:

Hucai Zhang Yang Pu Yunnan Normal University Nanjing University of Information Science …

63 PUBLICATIONS 915 CITATIONS 11 PUBLICATIONS 82 CITATIONS

SEE PROFILE SEE PROFILE

Wenxiang Zhang Yunnan Normal University

17 PUBLICATIONS 94 CITATIONS

SEE PROFILE

All in-text references underlined in blue are linked to publications on ResearchGate, Available from: Hucai Zhang letting you access and read them immediately. Retrieved on: 04 December 2015 Quaternary International 218 (2010) 143–150

Contents lists available at ScienceDirect

Quaternary International

journal homepage: www.elsevier.com/locate/quaint

Biomarkers of modern plants and soils from Xinglong Mountain in the transitional area between the Tibetan and Loess Plateaus

GuoLiang Lei a,b, HuCai Zhang a,b,*, FengQin Chang b, Yang Pu b, Yun Zhu b, MingSheng Yang c, WenXiang Zhang a a Key Laboratory of Western China’s Environments, MOE, College of Earth Sciences and Environments, Lanzhou University, Lanzhou 730000, PR China b State Key Laboratory of Lake Science and Environment, Nanjing Institute of Geography and Limnology, Chinese Academy of Sciences, Nanjing 210008, PR China c Key Laboratory of Boyang Lake, MOE, School of Environmental and Chemical Engineering, Nanchang University. Nanchang 33003, PR China article info abstract

Article history: Organic compounds found in living plants and modern soils in different bioclimatic areas are important Available online 16 December 2009 in understanding the paleoenvironmental implications of the organic matter signals from geological records. Using GC–MS techniques, a series of biomarkers, mainly including n-alkanes and n-alkan-2- ones, were identiﬁed from the typical plants (e.g. broadleaf-deciduous trees and conifers), and the soils collected from Xinglong Mountain, a transitional area between Tibetan and Loess Plateaus. The C15–C33 n-alkane homologues were present in all samples. Typically, C27 or C29 n-alkanes were most abundant in tree samples and C31 was most abundant in grasses samples. Meanwhile, the soil samples were mainly dominated by C29 and C31 homologues. The n-alkan-2-ones showed a strong odd-over-even predominance of carbon numbers. C23,C25 or C29 n-alkan-2-one homologues were most abundant in plant samples, while C27 or C29 was abundant in soils samples. The lipid biomarker distribution patterns of all the plants and soil samples suggest that the vegetation is a primary source of organic matter for the soils, and that reprocessing of microbes and physicochemical reactions in the soil played an important role in the degradation of organic matter. Research has advanced knowledge about transformation of biomarkers in the plants (organic matter)–microbes–forest system. It also can help reconstruct the process of plant successions recorded by the lipids. Ó 2009 Elsevier Ltd and INQUA. All rights reserved.

1. Introduction It is notable that the lipid fractions of n-alkanes and n-alkan-2- ones can provide reliable information about the sources and Lipid biomarkers can be identiﬁed from various environments, diagenetic alteration of organic matter. The n-alkanes can preserve such as marine sediments, lacustrine deposits and modern soils primary sources of organic matter because of their low suscepti- (Meyers and Ishiwatari, 1993; Ohkouchi et al., 1997; Sheng et al., bility to microbial degradation in comparison with other types 1999; Feakins et al., 2007; Wang et al., 2007; Zhang et al., 2007). of organic matter (Meyers, 2003). For instance, the n-alkanes and Such records not only contain biological and geological informa- n-alkan-2-ones distribution patterns, together with the individual tion, but also can provide climate-related information based on the hydrocarbon isotopic composition, can indicate the source of the content, the relative abundance (mainly biomarker ratios), the organic matter in a variety of sediments (e.g. Rielley et al., 1991; Xie carbon number distribution and the isotope compositions. Previous et al., 2002; Zhang et al., 2004; Rommerskirchen et al., 2006; Bai studies have proven that lipid biomarkers in sedimentary records et al., 2009). However, the n-alkan-2-ones signals from sediments can be sensitive indicators for the past environmental conditions are often not fully utilized in comparison with n-alkanes and n- (e.g. vegetation, temperature and precipitation) (Meyers and Ishi- alkanoic acids, which can be directly derived from higher plants. watari, 1993; Zhang et al., 2006; Eglinton and Eglinton, 2008). This is mainly due to the fact that the alkan-2-ones are often absent or low in abundance in living organisms, though they can be found in a variety of sediments in the Earth geohistory. The long-chain n-alkanes (C27,C29,C31 and C33) in modern soil usually represent the higher plant input (Bull et al., 2000; Zhang * Corresponding author at: State Key Laboratory of Lake Science and Environ- et al., 2003). It was found that the n-alkanes extracted from woody ment, Nanjing Institute of Geography and Limnology, Chinese Academy of Sciences, 73 East Beijing Road, Nanjing 210008, PR China. plants were dominated by C27 or C29, whereas C31 was more abun- E-mail address: [email protected] (H.C. Zhang). dant in grasses (Schwark et al., 2002). Therefore, the ratios of C27/C31

Fig. 1. A DEM map of the locality of the study area.

or (C27 þ C29)/C31 can be used to evaluate the relative abundance of periods, the herbs were more abundant than woody plants in Loess forest and grass coverage, and this relationship had been confirmed Plateau, especially in the Luochuan area, where there was not by lipid biomarker records in lake sediments (Hanisch et al., 2003). development of forest during the MIS11 (Zhang et al., 2008). This There is a close relationship between the organic matter in the raises an important question: are the biomarker indices unsuitable modern soil and in living vegetation, where the organic matter of for differentiating the vegetation sources, or are paleosols such as surface soil comes largely from the local vegetation but can be the S4 misclassified? Therefore, it is crucial to understand the reworked microbially and by early diagenesis (Dinel et al., 1998; processes that control biomarker distribution in plants and soils in Howard et al., 1998). Study of modern soils underlying different the modern environment, which has been believed to be climati- vegetation types has shown that the lipid fractions in soil can vary cally analogous to conditions during the formation of the S4 in the spatially depending on the vegetation type (Wang et al., 2003, 2007). Chinese Loess Plateau. This will provide a solid background in The soil organic matter in forested areas can be a mixture of all reconstructing the past climate history and vegetation evolution vegetation types (e.g. trees, shrubs and grasses). As a result the lipid using the biomarker records from geological archives. fractions extracted from organic matter in soil are mixed signals of This paper reports the lipid fractions extracted from the organic trees and grasses (Zhang et al., 2008). Partitioning of these mixed matter of various plants and modern soils from Xinglong Mountain signals is crucial in understanding the dynamics of the plant–soil– in Northwest China. The aim is to understand the relationship environment system using the molecular index. For example, the between plants and underlying soils in affecting biomarker signals well-developed paleosol S4 in the Luochuan loess profile in the through a large database on the n-alkanes and n-alkan-2-ones as Loess Plateau, corresponding to the MIS11 in age, was classified as well as the lipid information of plants and soils. This is one of the Leached cinnamon soil, the soil developed under warm-humid first studies showing that components of biomarkers are strongly climate under forest cover. However, based on biomarker indices it related to vegetation types. Therefore, this study provides impor- was found that during the formation of paleosol S4, the n-alkanes tant information on the modern processes in driving the biomarker were dominated by C31 homologue, indicating that the primary dynamics and can be readily applied in reconstructing paleo- organic input originated from herbs. During the warm and humid environments using fossil biomarkers.

Table 1 Concentration and proxies of n-alkane and n-alkan-2-one from plant and soil samples.

No. Species Description n-Alkane n-Alkan-2-one

Concentration (mg/g) Proxy Concentration (mg/g) Proxy

nC27 nC29 nC31 CPIh ACL Rh/l nC25 nC27 nC29 CPI P1 Picea wolsonii Tree trunk 3.9 1.2 0.6 2.2 23.9 2.3 0.2 0.1 0.3 3.4 P2 Populus davidialga Tree branch 5.1 2.4 0.6 1.5 22.6 1.6 0.3 0.2 0.1 2.1 P3 Populus davidialga Tree leaf 156.0 58.8 11.3 11.5 26.4 28.5 5.5 2.9 7.0 17.1 P4 Picea wolsonii Tree leaf 19.3 31.5 13.5 7.1 27.1 8.8 1.3 0.9 0.9 0.8 P5 Picea crassifolia Tree leaf 10.6 15.4 10.6 3.7 25.1 3.3 1.4 0.6 1.1 4.0 P6 Entodon sp., Hygrohypnum sp. Moss 23.8 21.5 11.9 3.4 24.0 2.1 3.5 3.5 3.5 3.4 P7 Stipa bungeuna, Stipa bungeunsis Grass 33.2 95.3 286.0 12.3 29.9 56.2 3.1 16.9 20.3 10.2 S1 Surface soil 12.8 11.7 11.0 8.4 27.8 17.9 1.9 3.4 3.6 3.6 S2 Surface soil 6.0 9.1 7.1 5.4 27.0 11.9 1.4 1.5 0.9 4.4 S3 Soil at 20 cm depth 6.1 8.6 9.8 3.8 27.0 6.4 1.8 2.4 2.2 2.6 S4 Soil at 40 cm depth 8.5 8.8 9.3 5.2 27.0 11.5 2.4 3.1 2.6 2.9 S5 Soil at 70 cm depth 10.1 10.1 10.4 7.0 27.2 12.5 1.8 2.6 2.0 3.2

Note: only the concentration of total samples of S3, S4 and S5 are listed here. G.L. Lei et al. / Quaternary International 218 (2010) 143–150 145

Fig. 2. Carbon number distribution of the n-alkanes of all 18 samples.

2. Sampling and analytical methods vegetation samples were cut into small pieces and ground to 178 mm before 20 g soil samples and 5 g plant samples were prepared for Xinglong Mountain (35470N; 104040E, 2150 m to 3435 m a.s.l., lipid extraction. The extracting method involved two steps: first the Fig. 1) is a unique island-like area where primitive forest is well ground samples were put in the solvent of dichloromethane and preserved at the western edge of the Chinese Loess Plateau. The methanol in a volume ratio of 93:7 for 56 h, then the solution was mean annual temperature is 4.1 C, with a mean annual precipita- ultrasonic-extracted with an ultrasonic extractor twice for 20 min tion of 625 mm. The forests are mainly distributed between 2500– each time. The extracted material was not further separated into 3000 m a.s.l. and they are dominated by broadleaf-deciduous trees family components using a silica gel–alumina chromatographic (i.e. Populus davidiana and Betula platyphyla) and mixed with column. This procedure reduces loss of some main components various conifers (i.e. Picea crassifolia (Qinghai spruce) and Picea such as micro-saturated hydrocarbon and oxygen compounds, and wilsonii). The vegetation under the trees is dominated by bushes provides data that can be readily compared with the previous (i.e. Fargesii nitida, Sorbus koeheana, Crataegus kansuensis and studies (e.g. loess samples). Finally, the extracts were dried and re- R. xanthina), grasses (i.e. Gymnocarpium disjunctum, Philomis dissolved in chloroform prior to the GC–MS analysis. A known umbrasa, Pyrola rotundifolia and Thalictrum) and mosses (i.e. amount of n-eicosane-d42 was added before GC–MS analysis. n- Hygrohypmn sp. and Entodon sp.). Leached brown soil covers the Alkane and n-alkan-2-one concentrations were calculated based on area, and the surface of the soil is covered with a thin watery the peak area comparison with the added standard. defoliation layer mixed with bryophyte/moss (Liu et al., 1990). Plant and soil samples were collected in the forested area at about 2600 m a.s.l. Soil samples were collected from the surface 2 Soil layer and also along a profile at depths of 20, 40 and 70 cm. Seven vegetation samples from different parts of the trees and plants in Surface soil the study area were collected around the soil sampling sites to 1 represent the most popular three typical plants, i.e. trees, grass and lower plant (see Table 1 for details). All the samples were collected P6 0 in late autumn in order to avoid seasonal variations of the P2 P5 biomarker memories (Maffei et al., 1993). For each of leaves, grass 2CP and moss samples, an equal amount of samples from at least three -1 locations were collected and well mixed as one composite sample P1 P4 to ensure the representativeness of the samples. All samples were P3 carefully collected to avoid any potential contamination and were Plant -2 dried immediately for further analysis. For biogeochemcial analysis, P7 the soil samples taken from the profile were further separated into two portions using 350-mm mesh in order to assess the impact of -3 grain size on lipid compositions. In total 18 samples were prepared -2-10123 for GC–MS analyses. PC1 To minimize the potential contamination, all the rotten part of the vegetation samples was completely removed mechanically and Fig. 3. PCA ordination plot showing the long-chain n-alkanes of C27–C31 for different then cleaned by distilled water 5 times before being dried. All the types of samples. B: soil; C: plant; >: surface soil. 146 G.L. Lei et al. / Quaternary International 218 (2010) 143–150

Table 2 been cleaned with distilled water and the rotten parts were Eigenvalues and variances of factor extraction by PCA. completely removed to ensure no contamination. This may be PC Rime eigenvalues caused by the relatively low proportion of long-chain n-alkanes,

Total % of Variance Cumulative % because tree trunk and branch were made up of pith and xylem. Previous studies have revealed long-chain n-alkanes of plants 1 2.657 44.278 44.278 2 2.178 36.297 80.575 originating from the epicuticular waxes (Eglinton and Hamilton, 1967; Cranwell et al., 1987; Hanisch et al., 2003). There is a difference in maximum carbons between the surface soil samples, which The GC–MS analyses were performed using an HP 5973 MSD, may be caused by the effect of partly decomposed plants, implying whichisinterfacedtoanHP6890gaschromatographandfitted that the organic matter of the surface soil might be not well mixed. with a fused silica capillary column (DB5, 30 m 0.25 mm i.d.). It is notable that the relative intensity of C31 decreased from the top The oven temperature was set to 80 C for 1 min before being to the lower parts on the profiles, while C27 increased concurrently increased to 300 Cat3C/min for 30 min. Helium was used as (Fig. 2). In bulk samples there is no significant difference in relative carrier gas at a linear velocity of 32 cm/s, with the injector intensity of n-alkanes in both <350 mm and >350 mm portions of operating at a constant flow of 0.9 ml/min. The MS was operated soils, indicating the limited impact of grain size of plant residues on with ionization energy of 70 eV, a source temperature of 230 C the n-alkanes signals in soils. The high relative intensity of long- and an electron multiplier voltage of 1900 V over the range of 35– chain n-alkanes showed that the organic matter originated from 550 Da. the higher plants, while the difference of major peaks in the Blank samples were also concurrently run. The GC–MS spec- different layers reflected the complexity in the sources of organic trum results for blank samples showed that the original data were compounds. free of artificial contamination. Principal component analysis (PCA) was conducted to assess the role of different sample types in affecting the composition of the organic matter across samples (Fig. 3). Only long-chain n-alkanes 3. Results and discussion with odd carbon numbers (C23–C33) are included in PCA. The first PCA axis accounted for nearly 45% of the total variance, while 36% of 3.1. The n-alkanes of plants and soils the total variance was explained by the second PCA axis. In order to reduce the variables, their eigenvalue was set to >1, which means n-Alkane and n-alkan-2-one concentrations are listed in Table 1. that the variance of each PC is at least as great as the variance of The carbon numbers of n-alkanes from the plants and soils mainly a prime variable. Thus, two factors (PC) were extracted, which range from C15 to C33 (Fig. 2). The long chain n-alkanes, such as C27, altogether cover 80.6% of the total variance (Table 2). C29 and C31, were more abundant in most samples (Table 1). Grasses Although the significance of the extracted factors is still unclear, had a maximum peak at C31, leaves of conifer had a maximum peak PC1 and PC2 contain the most information on the biomarker (this is at C29, and trunks, branches and leaves of Populus davidialga and what ordination analysis does, though the axes represent the latent moss were dominated by C27. The difference was visible among the variables, which may not be measured). In Fig. 3, the soil samples plant samples, so it is creditable to use the results as typical of concentrate in the top left quadrant, whereas the plant samples plants of the Xinglong Mountain, although here a small number of span more widely in the ordination space. The dispersion pattern is each species was used to represent the typical plants. Surface consistent with the large difference of the long-chain alkanes modern soils were dominated by C27 and C29, homologous with among vegetation samples. Notably, the surface soil sample sites high odd-over-even predominance. However, the bulk samples between plant samples and deep soils indicate that vegetation collected from upper and middle parts on the soil profile, in both signals are well recorded by the n-alkanes in the underlying soils portions of different sizes (e.g. <350 and >350 mm), were domi- through surface samples. nated by C31, while the samples from the lower part had a main Other commonly used proxies of organic geochemistry were carbon peak at C29 (Fig. 2). also calculated in this study using the following equations. These Tree trunk and branch from the Xinglong Mountain are char- indicators include the odd-over-even carbon predominance (CPI), acterized by main peaks at C27 or C29, a typical range for woody the ratio between long-chain and short-chain n-alkanes (Rh/l) and plants. However, the low-number-carbon n-alkanes contained in the average carbon length (ACL) (Cranwell et al., 1987; Xie et al., them had relatively high intensity, even though the samples had 1999; Jeng and Kao, 2002; Wang et al., 2003) (see Table 1).

Fig. 4. The indices of n-alkanes and their changes from soil-upper to soil-lower. >: >350 mm portion; ,: <350 mm portion; C: bulk samples. G.L. Lei et al. / Quaternary International 218 (2010) 143–150 147

Fig. 5. The correlation between CPIh, ACL and Rh/l. B: soil; C: plant; >: surface soil.

X h X However, odd-numbered carbon short-chained n-alkanes may CPIh ¼ 2 odd ðC23 C33Þ= even ðC22 C32Þ X i derive from microbial organisms (Cranwell et al., 1987; Dinel et al., 1990). Compared with the surface soil, the higher value of R from þ even ðC24 C34Þ (1) h/l soil samples in the section (9.40) may therefore indicate that X X microbes and bacteria contribute some short-chained n-alkanes to CPIl ¼ odd ðC15 C21Þ=even ðC16 C22Þ (2) the organic matter in the soil. Following the increase of short- chained n-alkanes, there was a lower R in soil from the section X . X h/l compared with surface soil (Fig. 4). Correlation analysis shows ACL ¼ Ci i Ci ð15 i 34Þ (3) there is a strong association among CPIh, ACL and Rh/l across all X X samples (Fig. 5), indicating they are sensitive indices for different þ Rh=l ¼ C22= C21 (4) vegetation types. Translation of organic matter from plants to soils can be involved with complicated processes. The CPI , ACL and R proxies The odd-over-even carbon predominance both with high carbon h h/l all show a sharp decrease from the top to deep soils (Fig. 4), indi- number (CPIh) of the grasses (12.3) and the broadleaves (11.5) taken cating that the organic matter from upper to lower part in soil from Xinglong Mountain is larger than 10. The CPIh of coniferous profile experienced a degradation or the processing of microbes. leaves and moss are between 3.4 and 7.1, but tree trunk (2.2) and This is further supported by the presence of diploptene (hop- branch (1.5) have the lower values. The CPIh of surface soils ranges 22(29)-ene) in the soils (Fig. 6), as diploptene was considered to be from 5.4 to 8.4, while the soil samples from the lower part have CPIh an indicator of bacterial sources (Wakeham, 1990) and microbial from 3.8 to 7.7. The higher CPIh of the soil samples indicates a higher activity in diverse environments, including soils, microbial plant input. The odd-over-even carbon predominance values with mats, and oceanic and lacustrine sediments (Volkman et al., 1986; low carbon number (CPIl) are close to 1 in most plants, while leaves Venkatesan, 1988; Elvert et al., 2001). In particular, the reprocessing of Picea crassifolia, picea wolsonii and Populus davidialga show of microbes is important for the degradation in the soil formation values higher than 1. The CPIl of soil samples range from 0.6 to 0.7, (Chen, 1979). The fact that diploptene was identified in all soil which are lower than that of plant samples, indicating the samples indicates that microbes contribute organic matter directly reprocessing of organic matter in the soil. The ACL of trees are in the or indirectly to the soil and play an important role in affecting the range from 22.6 to 27.1, which is lower than that of grasses (29.9) composition of organic matter in forested areas. The difference but are generally higher than that in the moss (24). The ACL ranges between the upper soil layer and lower soil layer indicates that the of soils (26.3–27.8) lies well in the range of the values for higher degradation of organic matter becomes stronger with the increase plants. All these three proxies indicate the organic matter in soils is in depth, but it could also be a result of varying vegetation types in closely related with the overlying vegetation. Rh/l from high plants the past. in leaves of conifer (8.8–3.3), trunks (2.3), and branches (1.6) are The CPIh, ACL and Rh/l in the soil profile show an increasing trend lower than Rh/l from soils (average value 9.4). Lower plant in moss from 20 cm depth in the profile to the lower parts, and the CPI ,ACL also has a low value. But grasses and leaves of Populus davidialga as h and R at the depth of 70 cm are close to the values of the surface the primary source of organic matter in soil have higher values of h/l soil samples (Fig. 4). But CPIl shows a different pattern in that all soil Rh/l (28.5–56.2). The distribution patterns of Rh/l indicate the samples have almost similar values. The changes of CPI , ACL and organic matter of soils may derive from the overlying vegetation. h

Fig. 6. (a) Typical m/z 191 mass chromatography of soil; #1 refers to diploptene. (b) Mass spectra of diploptene. 148 G.L. Lei et al. / Quaternary International 218 (2010) 143–150

indicating an integrated composition of the main vegetation types. P7 30 It is interesting that soil samples are closer to the area of conifer in Grass composition than the broadleaf ones, as coniferous trees are the dominant vegetation in the forest zone between 2500 and 3000 m Soil a.s.l. at Xinglong Mountain. However, previous studies have shown that some species of conifer vegetation can also produce the 27 P4 n-alkanes distribution peak at C31 (Schwark et al., 2002). In this P3 LCA study, the n-alkanes extracted from coniferous trees are dominated Conifer by C29. Zhong et al. (2007) also reported a preferential C29-alkanes P5 peak in modern pine leaves on the western Chinese Loess Plateau, P1 af which is near the study area. 24 le ad ro P6 B 3.2. The n-alkan-2-ones from plants and soils P2 The n-alkan-2-ones have been considered askinds of biomarker 21 compounds with environmental and biogenesis signiﬁcance in 0.2 0.3 0.4 0.5 modern soils (Xie et al., 1999). In this study, the n-alkan-2-ones

C29/(C27+C29+C31) were identifiable in all samples. The most striking characteristics of the n-alkan-2-ones are that they appeared with a distribution of Fig. 7. Scatter plot showing the comparison between the vegetal signals extracted carbon numbers from C17 to C33 and had a strong odd-over-even B C > from ACL and C29/(C27 þ C29 þ C31). : soil; : plant; : surface soil. carbon number predominance (Fig. 8). The n-alkan-2-ones of high plants are dominated with peaks at C23,C25 and C29, while the soil Rh/l may be caused by the activity of microbes. Microbes usually are samples are dominated with the peaks at C27 and C29. more active at the upper part of soil, and less so at the lower part of Previous studies revealed that the n-alkan-2-ones are common the soil profile because of lack of oxygen (Chen, 1979). Therefore, in various sediments. For example, n-alkan-2-ones were detected in the increase of CPIh, ACL and Rh/l values from upper-middle to the the organic matter in atmospheric aerosol and showed an odd lower part in the profile is possibly caused by weakening of carbon predominance (Xie et al., 1999). They possessed a bimodal microbial degradation. The relative abundance of long-chain distribution pattern with the carbon number peak at C19 (or C21) and alkanes is relatively higher in the lower part of the soil profile, C29 (or C31). These n-alkan-2-one homologues were considered to possibly due to both the decreased degradation intensity and the be derived from the leaf wax on the plant surface, while the origin of quantitative decrease in microbial activity down the profile. n-alkan-2-ones

Fig. 8. The distribution pattern of n-alkan-2-ones in soil and vegetal samples. G.L. Lei et al. / Quaternary International 218 (2010) 143–150 149

Fig. 9. The relation of CPI and Rh/l between the n-alkanes and n-alkan-2-ones. vermicular red earth and snow samples (Xie et al., 1999), in which 4. Conclusions the carbon number ranged from C21 (or C24)toC31 with a peak at C29, with odd carbon predominance. This was believed to be derived Abundant lipid fractions were identified in the typical vegeta- from lipids and carbohydrates of high plants. tion and soils from the transitional belt between the Loess and In this study, the n-alkan-2-ones show odd carbon predomi- Tibetan Plateaus. This study reveals that the n-alkanes extracted nance with carbon number peaks at C27 and C29 and a similar from vegetation are variable even though the trees have carbon distribution pattern to the n-alkanes (Fig. 8). Cranwell et al. (1987) numbers peaking at C27 and C29, whereas grasses peak at C31. The reported that the n-alkanes were degraded to alkanone through the carbon numbers in the soil developed under the forests showed not chemical or microbial processing in lakes, while Nichols and Huang only the C27 and C29 peaks, but also a strong peak at C31. The first (2007) suggested that they were derived from sphagnum in two axes of PCA analysis extracted from odd carbon numbers of freshwater peatlands, but the origin of n-alkan-2-ones in a forest C23–C33 showed the transfer pathway of organic matter in the ecosystem is unclear. The CPI and Rh/l of n-alkan-2-ones are forest system from local vegetation to soils. The CPIh,Rh/l and ACL calculated following equations (1) and (3). CPI and Rh/l between the values showed strong sensitivity to the change in vegetation types. n-alkanes and n-alkan-2-ones both show a high correlation coef- The presence of diploptene in soils indicates that microbes ficient in this study (Fig. 9). The relatively high correlation between contribute organic matter to soil directly or indirectly. Biomarker CPIs of both compounds may indicate a common origin, and CPIs signals also will change with the microbial degradation in a soil can be used as a proxy for the source of organic matter. Soil samples system. In this case, n-alkanes of coniferous trees have a peak at C29. show small variation in both CPI and Rh/l values, and are strongly Using C29/(C27 þ C29 þ C31) and ACL, it can serve as a reliable proxy concentrated within a small area (Fig. 10). When CPI values of the to distinguish the shift between coniferous tree and broadleaf in soil samples increase, it indicates more grass and broadleaves. Xinglong Mountain, and the vegetation features overlying the soil When Rh/l value decreases, it indicates an increased contribution might be reconstructed in the transitional area between the Tibetan from conifers. and Loess plateaus. Abundant n-alkan-2-ones, which are characterized by prefer- ence for odd carbon numbers, were identified from all samples. They were dominated by the high carbon numbers with peaks at 16 C27 and C29, indicating the n-alkan-2-ones in soil samples origi- P7 nated from the high plants, while the appearance of the n-alkan-2- Grass ones with low carbon numbers indicated that microorganisms may play an important role in the soil formation. The distribution 12 af patterns of n-alkan-2-ones were comparable with those of the le ad n-alkanes, implying they were derived from a common origin. ro B P3 Soil The CPI and Rh/l of n-alkan-2-ones show a consistent variation l/h in plant samples. The CPI and Rh/l can be used to investigate the 8 R source of organic matter. Lipids of typical plants and the soils in modern forests at Xinglong Mountain demonstrated that plants were the main Con contributors to the organic matter in soil. The changes of the 4 ifer P1 P4 organic matter from vegetation to soils were closely related to the microbial activity. Although the lipids of soil were affected by P2 P5 various factors, this study identified the sources of the organic P6 0 matter of soil, and further study on the relationships in the plants– 05101520 soil–microorganism system may help us to understand the plant CPI record by lipid fractions and how to interpret the lipid signals in soils and paleosols.

Fig. 10. The CPI and Rh/l distribution pattern of n-alkan-2-ones extracted different This work provides a basic framework for the use of biomarkers samples. B: soil; C: plant; >: surface soil. on modern plants and soils derived from the forest area under arid– 150 G.L. Lei et al. / Quaternary International 218 (2010) 143–150 semiarid climate conditions. Much work needs to be done in order Meyers, P.A., 2003. Applications of organic geochemistry to paleolimnological to understand how biomarker signals of organic matter are trans- reconstructions: a summary of examples from the Laurentian Great Lakes. Organic Geochemistry 34, 261–289. formed from the forest to the soil, lacustrine deposits and paleosol, Meyers, P.A., Ishiwatari, R., 1993. Lacustrine organic geochemistry–an overview of e.g. paleosol sequences in loess sections. Such kinds of works will indicators of organic matter sources and diagenesis in lake sediments. Organic be useful and important in identifying sources of the organic matter Geochemistry 20, 867–900. Nichols, J.E., Huang, Y.S., 2007. C-23-C-31 n-alkan-2-ones are biomarkers for the in a variety of geological archives and in paleoenvironmental genus Sphagnum in freshwater peatlands. Organic Geochemistry 38, 1972–1976. reconstruction. Ohkouchi, N., Kawamura, K., Taira, A., 1997. Molecular paleoclimatology: reconstruction of climate variabilities in the late Quaternary. Organic Geochemistry 27, 173–183. Acknowledgements Rielley, G., Collier, R.J., Jones, D.M., Eglinton, G., 1991. The biogeochemistry of Ellesmere Lake, UKdI: source correlation of leaf wax inputs to the sedimentary This study is supported by NSFC (no. 40871096) and the lipid record. Organic geochemistry 17, 901–912. Rommerskirchen, F., Plader, A., Eglinton, G., Chikaraishi, Y., Rullko¨tter, J., 2006. Hundred Talent project of the Chinese Academy of Science. We are Chemotaxonomic significance of distribution and stable carbon isotopic grateful to two anonymous reviewers, and Prof. Dr. Zhaohui Zhang composition of long-chain alkanes and alkan-1-ols in C4 grass waxes. Organic and Dr. Yongli Wang for their very constructive suggestions and Geochemistry 37, 1303–1332. comments on the earlier version of draft that has improved the Schwark, L., Zink, K., Lechterbeck, J., 2002. Reconstruction of Postglacial to Early Holocene vegetation history in terrestrial Mid-Europe via cuticular lipid paper greatly. We also thank Prof. Jay Quade for his efforts to biomarkers and pollen records from lake sediments. Geology 30, 463–466. improve the English language. Sheng, G.Y., Cai, K.Q., Yang, X.X., Lu, J., Jia, G., Peng, P., Fu, J.M., 1999. Long-chain alkenones in Hotong Qagan Nur Lake sediments and its paleoclimatic implications. Chinese Science Bulletin 44, 259–263. References Simoneit, B.R.T., 1985. Application of molecular marker analysis to vehicular exhaust for source reconciliation. International Journal of Environmental Bai, Y., Fang, X.M., Nie, J.S., Wang, Y.L., Wu, F.L., 2009. A preliminary reconstruction Analytical Chemistry 22, 203–233. of the paleoecological and paleoclimatic history of the Chinese Loess Plateau Simoneit, B.R.T., Cox, R.E., Standley, L.J., 1988. Organic matter of the tropospheredIV: from the application of biomarkers. Palaeogeography, Palaeoclimatology, Lipids in Harmattan aerosols of Nigeria. Atmospheric Environment 22, 983–1004. Palaeoecology 271, 161–169. Venkatesan, M.I., 1988. Diploptene in Antarctic sediments. Geochimica et Cosmo- Bull, I.D., Bergen, P.F.V., Nott, C.J., Poultonb, P.R., Evershed, R.P., 2000. Organic chimica Acta 52, 217–222. geochemical studies of soils from the Rothamsted classical experimentsdV. The Volkman, J.K., Allen, D.I., Stevenson, P.L., Burton, H.R., 1986. Bacterial and algal fate of lipids in different long-term experiments. Organic Geochemistry 31, hydrocarbons in sediments from a saline Antarctic lake, Ace lake. Organic 389–408. Geochemistry 10, 671–681. Cranwell, P.A., Eglinton, G., Robinson, N., 1987. Lipids of aquatic organisms as potential Wakeham, S.G., 1990. Algal and bacterial hydrocarbons in particulate matter and contributors to lacustrine sediments. Organic Geochemistry 11, 513–527. interfacial sediment of the Cariaco Trench. Geochimica et Cosmochimica Acta Chen, H.K. (Ed.), 1979. Soil Microbiology. Shanghai Scientific & Technical Publishers, 54, 1325–1336. Shanghai, pp. 311–314 (in Chinese). Wang, Z.Y., Liu, Z.H., Yi, Y., Xie, S.C., 2003. Features of lipids and their implications in Dinel, H., Schnitzer, M., Mehuys, G.R., 1990. Soil lipids: origin, nature, contents, modern soils from various climate vegetation zones. Acta Pedologica Sinica 40, decomposition and effect on soil physical properties. In: Bollag, J.M., Stotzky, G. 967–970 (in Chinese). (Eds.), Soil Biochemistry. Marcel Dekker, New York, pp. 397–427. Wang, Y.L., Fang, X.M., Bai, Y., Xi, X.X., Zhang, X.Z., Wang, Y.X., 2007. Distribution of Dinel, H., Monreal, C.M., Schnitzer, M., 1998. Extractable lipids and organic matter lipids in modern soils from various regions with continuous climate. Science in status in two soil catenas as influenced by tillage. Geoderma 86, 279–293. China Series D (Earth Sciences) 50, 600–612. Eglinton, T.I., Eglinton, G., 2008. Molecular proxies for paleoclimatology. Earth and Xie, S.C., Yao, T.D., Kang, S.C., Duan, K., Xu, B., Thompton, L.G., 1999. Climatic and Planetary Science Letters 275, 1–16. environmental implications from organic matter in Dasuopu glacier in Xix- Eglinton, G., Hamilton, R.J., 1967. Leaf epicuticular waxes. Science 156, 1322–1334. iabangma in Qinghai-Tibetan Plateau. Science in China Series D (Earth Sciences) Elvert, M., Whiticar, M.J., Suess, E., 2001. Diploptene in varved sediments of Saanich 42, 383–391. Inlet: indicator of increasing bacterial activity under anaerobic conditions Xie, S.C., Wang, Z.Y., Wang, H.M., Chen, F.H., An, C.B., 2002. The occurrence of a grassy during the Holocene. Marine Geology 174, 371–383. vegetation over the Chinese Loess Plateau since the last interglacial: The Feakins, S.J., Eglinton, T.I., Demenocal, P.B., 2007. A comparison of biomarker records molecular fossil record. Science in China Series D (Earth Sciences) 45, 53–62. of northeast African vegetation from lacustrine and marine sediments (ca. 3.40 Zhang, Z.H., Zhao, M.X., Lu, H.Y., Faiia, A.M., 2003. Lower temperature as the main Ma). Organic Geochemistry 38, 1607–1624. cause of C4 plant declines during the glacial periods on the Chinese Loess Gonzalez-Vila, F.J., Polvillo, O., Boski, T., Moura, D., De Andres, J.R., 2003. Biomarker Plateau. Earth and Planetary Science Letters 214, 467–481. patterns in a time-resolved Holocene/terminal Pleistocene sedimentary Zhang, Z.H., Zhao, M.X., Yang, X.D., Wang, S.M., Jiang, X.Z., Oldfield, F., Eglinton, G., sequence from the Guadiana river estuarine area (SW Portugal/Spain border). 2004. A hydrocarbon biomarker record for the last 40 kyr of plant input to Lake Organic Geochemistry 34, 1601–1613. Heqing, southwestern China. Organic Geochemistry 35, 595–613. Hanisch, S., Ariztegui, D., Pu¨ ttmann, W., 2003. The biomarker record of Lake Albano, Zhang, Z.H., Zhao, M.M., Eglinton, G., Lu, H., Huang, C.Y., 2006. Leaf wax lipids as central ItalydImplications for Holocene aquatic system response to environ- paleovegetational and paleoenvironmental proxies for the Chinese Loess mental change. Organic Geochemistry 34, 1223–1235. Plateau over the last 170 kyr. Quaternary Science Reviews 25, 575–594. Howard, P.J.A., Howard, D.M., Lowe, L.E., 1998. Effects of tree species and soil Zhang, H.C., Chang, F.Q., Li, B., Lei, G.L., Chen, Y., Zhang, W.X., Niu, J., Fan, H.F., physico-chemical conditions on the nature of soil organic matter. Soil Biology Yang, M.S., 2007. Branched aliphatic alkanes of shell bar section in Qarhan Lake, and Biochemistry 30, 285–297. Qaidam Basin and their paleoclimate significance. Chinese Science Bulletin 52, Jeng, W.L., Kao, S.J., 2002. Lipids in suspended matter from the human-disturbed 1248–1256. Lanyang River, northeastern Taiwan. Environmental Geology 43, 138–144. Zhang, H.C., Yang, M.S., Zhang, W.X., Lei, G.L., Chang, F.Q., Pu, Y., Fan, H.F., 2008. Liu, G., Song, B.J., Liu, Q.H., Li, Z.Q., Li, F., Qi, Y.Z., 1990. Soil formation and distribution Molecular fossil and paleovegetation records of paleosol S4 and adjacent loess of natural preservation area in Xinglong Mountains. Journal of Gansu Agricul- layers in the Luochuan loess section, NW China. Science in China Series D (Earth tural University 25. 405–410 (in Chinese). Sciences) 51, 321–330. Maffei, M., Mucciarelli, M., Scannerini, S., 1993. Environmental factors affecting the Zhong, Y.X., Chen, F.H., An, C.B., Xie, S.C., Huang, X.Y., 2007. Holocene vegetation lipid metabolism in Rosmarinus officinalis L. Biochemical Systematics and covers in Qin’an area of western Chinese Loess Plateau revealed by n-alkane. Ecology 21, 765–784. Chinese Science Bulletin 52, 1692–1698.