Nature of Volatile Organic Matter in Lake Sediments as a Reflection of Paleoclimate Changes Occurring at 4 ka in the Central Qaidam Basin

This study explores the paleoclimate changes around the 4 ka BP period in the central Qaidam Basin (QB), assessing the differences in spectral characteristics and organic composition of salt lake sediments under different climate change conditions. Sediment samples (10-m-depth profile) were collected from the middle of dry salt flats in East Taijinar Lake (China). Sediment organic matter (SOM) was assessed by Fourier transform infrared spectroscopy (FTIR) and gas chromatography-mass spectrometry (GC-MS). Results showed a significant difference in the TOC content of sediments with different lithological characteristics. A lower TOC content in salt-bearing strata was attributed to the extreme sedimentary environment with minimal exogenous inputs. FTIR spectroscopy revealed that the SOM in sediments included aliphatic C, ketones and alcohols. Sediments of salt-bearing strata generally exhibited a rise in the content of ketone C=O groups and a decrease in aliphatic C, with an equal content of alcohols compared with silty sand. Therefore, exogenous OM and the content of TOC in silty sand strata are higher than in salt-bearing strata, indicating that the paleoclimate became warmer and wetter after 4 ka BP. GC-MS analysis showed a significant difference between the phenol and aldehyde content in different strata, further indicating that the paleoclimate changed from dry to relatively warm around 4 ka BP. Seven organic compound types were identified in SOM, including aldehydes, hydrocarbons, phenols, esters, ketones, alcohols, and furans. Different strata exhibited different distributions of organic compounds, with particularly high concentrations of aldehydes in salt-bearing strata and phenols in silty sand. Correlation analysis was performed between detrital minerals and OM types in all samples. Results showed a strong positive correlation between detrital minerals and phenols and a strong negative correlation between detrital minerals and aldehydes, with a negative correlation also identified between detrital minerals and ketones. Overall, the reduction in volatile organic compounds demonstrates that the paleoclimate changed from cooler and dry to wet and warm around the 4 ka BP period in the central QB, with the carbon preference index and n-alkane values further demonstrating these results. This study also confirms the importance of volatile organic compound monitoring to assess paleoclimate changes.

This study explores the paleoclimate changes around the 4 ka BP period in the central Qaidam Basin (QB), assessing the differences in spectral characteristics and organic composition of salt lake sediments under different climate change conditions. Sediment samples (10-m-depth profile) were collected from the middle of dry salt flats in East Taijinar Lake (China). Sediment organic matter (SOM) was assessed by Fourier transform infrared spectroscopy (FTIR) and gas chromatography-mass spectrometry (GC-MS). Results showed a significant difference in the TOC content of sediments with different lithological characteristics. A lower TOC content in salt-bearing strata was attributed to the extreme sedimentary environment with minimal exogenous inputs. FTIR spectroscopy revealed that the SOM in sediments included aliphatic C, ketones and alcohols. Sediments of salt-bearing strata generally exhibited a rise in the content of ketone C O groups and a decrease in aliphatic C, with an equal content of alcohols compared with silty sand. Therefore, exogenous OM and the content of TOC in silty sand strata are higher than in salt-bearing strata, indicating that the paleoclimate became warmer and wetter after 4 ka BP. GC-MS analysis showed a significant difference between the phenol and aldehyde content in different strata, further indicating that the paleoclimate changed from dry to relatively warm around 4 ka BP. Seven organic compound types were identified in SOM, including aldehydes, hydrocarbons, phenols, esters, ketones, alcohols, and furans. Different strata exhibited different distributions of organic compounds, with particularly high concentrations of aldehydes in salt-bearing strata and phenols in silty sand. Correlation analysis was performed between detrital minerals and OM types in all samples. Results showed a strong positive correlation between detrital minerals and phenols and a strong negative correlation between detrital minerals and aldehydes, with a negative correlation also identified between detrital minerals and ketones. Overall, the reduction in volatile organic compounds demonstrates that the paleoclimate changed from cooler and dry to wet and warm around the 4 ka BP period in the central QB, with the carbon preference index and n-alkane values further demonstrating these results. This

INTRODUCTION
The Qaidam Basin, located at the northern margin of the Tibetan Plateau (TP), is one of the highest and most arid regions worldwide (Yu and Lai, 2012). The QB is influenced by three major climatic systems, including the East Asian summer monsoon (EASM), the westerlies and the Indian monsoon (IM). Reconstructions of the paleoclimate in the QB have attracted much research attention due to its sensitive to climatic variability (Xiang et al., 2013;Stauch et al., 2017). Different geological archives, including ice cores (Thompson et al., 1997), eolian sand and loess sections (Sun et al., 2007;Stauch, 2015), and lake sediments Ma et al., 2019) have been used to reconstruct climate changes in the TP at different locations. In particular, the organic matter (OM) present in salt lake sediments, such as n-alkanes, fatty acids and lignin, can serve as a highly sensitive indicator for environmental reconstructions.
Salt lakes are saline environments containing relatively high salt concentrations, which widely occur worldwide (Zheng, 2011). Owing to the complexity and specific geotectonic-climatic conditions required for salt lake formation, the sedimentary characteristics of saline lakes differ from other lake environments. Saline lake sediments are a product of the physical-geographical and geological environment, with the development of sedimentary strata retaining information on the changing surrounding environment. Therefore, saline lake sediments provide a sensitive record of paleoclimatic and paleoenvironmental changes (Zheng et al., 2016). The OM accumulated in saline lake sediments is considered to reflect changes in terrestrial vegetation, as well as the hydrological conditions and aquatic plant productivity in lakes, providing a historical source of information on the regional environment (Prokopenko et al., 1999;Meyers and Teranes, 2001;Lücke and Brauer, 2004;Wischnewski et al., 2011;Aichner et al., 2012).
The sources of sediment organic matter (SOM) in salt lake sediments are complex and can include planktonic and benthic fauna and flora, in addition to terrestrial material from riverine and anthropogenic inputs. SOM can exist in anaerobic environments for a long period of time, being subjected to a variety of geological processes. The partition or distribution of OM in sediments mainly depends on the functional groups and molecular structures contained in OM and the surface properties of sediment mineral particles (Kleber et al., 2007;Liang et al., 2011). It has been reported that salinity may be an important factor in determining the distribution of OM (Means, 1995;Tremblay et al., 2005), with the sources and redox conditions of OM in sediments also known to affect its distribution behavior (Wilson and Xenopoulos, 2009;Riedel et al., 2013). For example, the high concentration of organic carbon and the C/N ratio in the surface sediments of Bled lake (Slovenia) has been shown to depend on the sediment oxidation environment (Muri and Wakeham, 2006). Therefore, the chemical and functional properties of OM in sediments depend on its source material and biogeochemical history (Tfaily et al., 2017). The assessment of OM in sediments allows a better understanding of the biogeochemical cycling of carbon in lake and sediment environments (Tuo et al., 2011;Fan et al., 2017). However, few studies have explored the distribution behavior of OM in salt lake sediments.
In the early stages of the diagenetic process, there is a close relationship between SOM and mineral particles (Keil et al., 1994). For example, early authigenic quartz fills primary pores and decreases interparticle porosity, controlling the distribution of migrated OM during early diagenesis, subsequently affecting the development and volume of OM pores in siliceous shale gas windows (Zhao et al., 2017a). Biomineralized calcite plays an important role in the function of many organisms and in turn, uses organic macromolecules to control crystal growth (Hakim et al., 2017). Furthermore, the surface properties of minerals may affect important interaction processes, such as organic matter interactions (Valdrè et al., 2012). In the later stages of sediment diagenesis, mineral grains provide a degree of protection and reduce the rate of OM decomposition in sediments (Wang and Lee, 1993). Despite the low level of OM reserves in sediments, the structure and strength of OM adsorption on mineral surfaces is relevant to various environmental applications, such as CO 2 storage and contamination remediation (Quicksall et al., 2008;Brown and Calas, 2012).
However, conflicting results have been obtained using different lake sediment proxies for moisture reconstruction, such as in Lake Qinghai (China) (Chen et al., 2016 and references therein). An et al. (2012) proposed that the northeastern Tibetan Plateau climate was wet during the early Holocene period, based on carbonate analysis. However, it has also been proposed that the early Holocene climate was relatively dry, based on analysis of pollen assemblages (Shen et al., 2005). In addition, previous studies on climate change in the QB have reported varying results for the late Holocene period around 4 ka BP. Sun et al. (2019) and Wu et al. (2018) proposed that a rapid cold and drought event occurred around 4 ka BP, leading to the gradual drought in the QB in the late Holocene period. Xiang et al. (2013) suggested that the climate was mainly warm and dry after 3830 BP, with a warm cool and extremely dry climate predicted during 3,040-2,600 ka BP based on both n-alkane and pollen proxies. In contrast, Chen et al. (2008) proposed that the topography of the TP and the adjacent Asian highlands led to a dry climate in the early Holocene period, while the climate became relatively humid (less dry) after 8 ka BP. It is of note, most studies have mainly been focused on the eastern or southern regions of the TP which are sensitive and ideally located to reflect the paleoclimate change of the whole TP region Cheng et al., 2013), while little research has been performed on the central QB region. Therefore, it is important to clarify the interpretations of environmental data in the central QB region around 4 ka BP. Also, few studies have assessed the chemical composition of SOM in these regions, especially in salt lake sediments. Due to the high salinity of salt lakes and the presence of various toxic ions (Co, Ni, and Mn), the composition of SOM in salt lake sediments is relatively unique (Liu et al., 2018;Isaji et al., 2019), making a precise understanding of SOM composition in salt lake sediments valuable information. In this study, samples were collected directly in the central QB region for analysis of the OM composition and distribution characteristics, to reflect the palaeoclimate changes at 4 ka BP. Meanwhile, some established palaeoclimate indicators, such as detrital mineral flux, nC 31 /nC 29 ratio, and CPI, were used to verify the results.
The aim of this study was to assess climate changes in the late Holocene period at 4 ka BP, in the middle of the QB using OM in lake sediments. First, Fourier transform infrared (FTIR) spectroscopy and gas chromatography-mass spectrometry (GC-MS) were used to investigate the composition of OM in salt lake sediments. FTIR spectroscopy provides insight into the chemical functions of OM based on its functional groups (Derenne and Quénéa, 2015), while GC-MS is an established analytical method for volatile organic compounds (Ene et al., 2012), providing a molecular fingerprint for OM based on its molecular weight and elemental composition (Stashenko and Martínez, 2014). Then, the paleoenvironmental changes in the middle of QB around the 4 ka BP period were reconstructed based on the n-alkane distribution and comparison with the OM in salt lake sediments.

Geological Background
The East Taijinar salt lake (37°21′54″-37°36′05″N, 93°45′33″-94°06′48″W) is located in the Tertiary anticline structural depression belt of the QB (Qinghai, China) ( Figure 1). The formation of this salt lake is associated with the uplift of the Qinghai-Tibet Plateau, the strong tectonic movement of the QB, and the migration of residual saltforming brine from western to central and eastern regions (Zhang et al., 1987). The lake is mainly supplied by the Nalinggele River from the Kunlun Mountains in the southwest. The brine subtype is magnesium sulfate with a pH of 7.9 (Zheng and Liu, 2009). The East Taijinar salt lake spans approximately 300 km 2 , which includes a lake area of 100 and 200 km 2 of dry salt flats. The dominant sediment component in the upper stratum of the salt lake area is halite, along with small amounts of mirabilite and gypsum and other relatively simple minerals such as quartz and feldspar.

Sample Collection
Sediment samples were collected from a depth profile (13.5 m) in the middle of the dry salt flats using a polyvinyl chloride corer (30 mm inner diameter and 250 mm length). All samples were collected in clean vertical sections. They were covered with a black cloth and sealed in black plastic bags, then wrapped with tape to avoid exposure to light or contamination.
The lithological profile of the core samples from top to bottom was as follows: silty sand strata (SS) from 0 to 8.0 m, consisted of loose silty sand; salt-bearing strata (SB) from 8.0 to 13.5 m, harder and denser than that of silty sand, composed mainly of white halite. The first sample was collected at a depth of 0.5 m, with subsequent samples collected from below this along a vertical depth profile at intervals of 0.5 m, resulting in 27 samples being collected for each profile. In this study, a total of ten samples were selected for analysis, including five from each of the two different lithologic strata zones, with the depth and age details for each sample shown in Figure 2.
All selected sediment samples were dated using optically stimulated luminescence in the Luminescence Dating Laboratory of the Qinghai Institute of Salt Lakes, Chinese Academy of Sciences. The dating of all samples have been done in our previous work, and the reliability and accuracy of the dating results have been discussed in the paper (Ma et al., 2021). The effective age of the salt-bearing strata could not be identified because of the low content of quartz and feldspar, although according to the lithology, this layer should be the product of an arid environment in the 4 ka BP period. The result of the OSL dating with their equivalent dose (De) values are given in Table 1. The remaining samples were stored in sterile polyethylene bags and maintained at 4°C. Prior to analysis, samples were freeze-dried, ground, and then passed through a 200-μm mesh sieve.

FTIR Analysis
FTIR spectroscopy was performed at room temperature, using the KBr pellet technique. The samples were dried immediately prior to analysis to minimize interference due to moisture. OM characterization by infrared spectroscopy involved the collection of FTIR absorbance spectra in the mid-infrared region from 4,000 to 400 cm −1 . Confirmation of the chemical bonds and functional groups attributed to absorption peaks in the FTIR spectra was performed using previously reported literature.

GC-MS Analysis
The OM content in all ten samples from 0 to 8.0 m and 8.5-13 m depths were analyzed using gas chromatography mass spectrometry (GC-MS), using a Thermo TRACE 1300 Series Gas Chromatograph combined with a TSQ 8000 Evo MS detector (Thermo Scientific, United States). Solid-phase microextraction (SPME) was performed using a 50/30 μm divinylbenzene/carboxen/polydimethylsiloxane (DVB/CAR/PDMS) fiber (Supelco, Bellefonte, PA, United States). GC analysis was carried out using a 30 cm × 0.25 mm DB-Wax column (0.25 µm film thickness; J&W Scientific Inc., United States). Prior to GC-MS analysis, the 2 g of samples were heated at 60°C for 60 min in an incubating box, then subjected to SPME for 30 min at a constant temperature of 60°C, with the injection volume set to 1 μl and an injector temperature of 250°C. The GC oven was maintained  at 50°C for 2 min and then increased to 220°C at 3°C/min and maintained for 1 min. Nitrogen (99.999% purity) was utilized as the carrier gas at a flow rate of 1 ml/min. The ion source temperature was maintained at 250°C and a 70 eV electrode was used for ionization. Finally, the compounds were identified by comparison with the National Institute of Standards and Technology (NIST) mass spectra library.

X-Ray Powder Diffraction Analysis
The mineral composition of sediments was determined through XRD analysis at the Qinghai Institute of Salt Lake, Chinese Academy of Sciences (Xining, China), using an X Pert-PRO diffractometer (BRUKER, Germany) with a CuKα radiation source and a scanning speed of 0.02°/s. The applied voltage and tube currents were 45 kV and 30 mA, respectively.

Quantification of Total Organic Carbon
Quantification of total organic carbon (TOC) was performed as per the procedure previously reported by Wu et al. (2012), with some modifications. Briefly, 2 g of sample was treated with 20 ml of 1 mol/L hydrochloric acid, then washed with deionized water five times until the acid was removed and dried in a crucible at 900°C. The sample was ground to a powder using an element analyzer (Elementar, Germany) and decomposed at 950°C. Finally, the carbon mass percentage was determined.

n-Alkane Analysis
The salt lake sediments were subjected to two cycles of extraction using an accelerated solvent extractor (Dionex ASE 350, Thermo Scientific, United States) with dichloromethane/methanol (93:7) at 100°C and 1,600 psi. The n-alkane was separated using a deactivated silica gel column eluted with n-hexane. The n-alkane concentration was quantified using the GC-MS system (7890B/5977B, Agilent technologies, United States). The capillary column was an HP-5MS silica capillary column (30 m × 250 μm × 0.25 μm), coated with 5% phenyl methyl silox. The GC-oven temperature was initially 50°C (held for 1 min), then increased to 315°C (held for 16 min) at 8°C/min. Individual n-alkanes were identified and quantified by comparing with the spectra obtained using an alkane-mixture standard (C 10 -C 40 , Fluka, United States).

Mineralogy and Total Organic Carbon
The XRD spectra show that the mineral components in all samples were similar. Results indicate that the sediment samples were comprised of halite, quartz, albite, muscovite, gypsum, chlorite, calcite, aragonite, and dolomite ( Table 2). However, the two different lithologic strata exhibited significant differences in the content of halite and detrital minerals (quartz, albite, and muscovite). SS strata samples exhibited the characteristics of low salinity and a high content of detrital minerals, while SB strata samples exhibited the opposite characteristics of high salinity and a low content of detrital minerals. Quartz was the most dominant detrital mineral in the SS strata (relative proportion of up to 41%), followed by muscovite and albite. In contrast, in the SB strata, quartz occurred at low concentrations and muscovite and albite were not detected. Other than detrital minerals, samples also contained carbonate (calcite), sulfate (gypsum), silicate, and halite minerals. The content of silicate ranged from 72 to 84%. Sulfate was identified in all sediments, although the content of carbonate in SB strata was only 1%. Generally, the mineral components were more abundant in the SS strata than in the SB strata.
The TOC content of lake sediment samples is also presented in Table 2, exhibiting differences in the samples from different sediment strata. The TOC values ranged from 0.02 to 0.17%, with the TOC content of SS strata samples ranging from 0.12 to 1.17% (mean 0.14%), while that of SB strata ranged from 0.02 to 0.04% (mean 0.03%). The observed variation between strata may be attributed to the differences in environment conditions and external inputs during the formation of salt lake sediment layers.

Sediment Organic Matter Functional Groups
The FTIR spectra of samples from the same stratum exhibit similar patterns (Figure 3). FTIR spectra of samples from the SB strata exhibit a more complex range of characteristic adsorption peaks than SS strata samples, especially at the range of 2,000-400 cm −1 . The determination of spectral peaks was based on the characteristic peaks of relevant organic and inorganic component functional groups, as reported in the literature ( Table 3).
The absorbance band at 3,700-3,618 cm −1 corresponding to the stretching of O−H bonds in kaolinite (Rahier et al., 2000), with this band is only detected in SS. The absorption peak at 3,554-3,409 cm −1 was assigned to the stretching vibrations of O−H bonded hydroxyl groups of alcohols, phenols, and water molecules (Ellerbrock et al., 1999), as well as NH functional groups (Szymański, 2017). The absorption peaks at 2,936-2,847 cm −1 were assigned to asymmetric and symmetric stretching vibrations from aliphatic CH 2 groups (Dutta et al., 2013). The absorbance peaks at 1,869-1,678 cm −1 were attributed to the weak C O stretching of conjugated ketones and anhydride (Coates, 2006), as well as to carboxylic acids and carbonyl bands in esters (Sarkhot et al., 2007), with these absorbance bands detected in all sediment samples from both lithologic strata. The bands in the region 1,645-1,620 cm −1 were attributed to the O−H bending modes of water (Ramasamy et al., 2009). The absorbance peak at 1,451-1,432 cm −1 was relatively complex, with Parolo et al. (2017) relating this peak to CO 3 2− vibrations in calcite and dolomite group minerals, or C−H bending vibrations of CH 3 and CH 2 groups. The band at 1,470 cm −1 was assigned as hydrogen-bonded O−C−O stretching of the carbonate ion (Oh et al., 2005). The peak in the range of 1,031-1,028 cm −1 may be related to the presence of OCH 3 or OH in alcohols (Rosa et al., 2016).
In addition to organic compounds, FTIR spectra also exhibited characteristic absorbance peaks for inorganic components. The absorbance band at 2,519-2,514 cm −1 was associated with the vibrations of CO 3 2− in calcite and dolomite group minerals. Generally, bands occurring below 1,000 cm −1 are related to a mixture of organic and inorganic compounds, such as quartz and clay minerals (Haberhauer et al., 2000;Calderón et al., 2013). The absorbance peak at 880-873 cm −1 was attributed to the same CO 3 2− vibrations as 2,519-2,514 cm −1 . The band in the range of 837-747 cm −1 , was assigned to Si−O−Si symmetrical stretching vibration in quartz (Lü et al., 2013). According to Ellerbrock et al. (1999), the band at 730-427 cm −1 was associated with inorganic components such as kaolinite, quartz, and gypsum.

Sediment Organic Matter Characterization by GC-MS
The characterization of SOM by GC-MS exhibited similar molecular compositions for both lithologic strata. The identified compounds included hydrocarbons, esters, aldehydes, ketones, furans, phenols, and alcohols. However, the proportions of these compounds in OM were significantly different. A total of twenty-five to thirty-four compound types were detected by GC-MS in SS, with phenols having the highest content, followed by hydrocarbons and aldehydes, while alcohols were only detected in sample SS02. The results are presented in detail in the supporting information (Supplementary Table S1).
Phenolic compounds are one of the most important secondary metabolites in plants (Karakaya, 2004), which are generally weakly acidic and easily oxidized in the environment. Four phenolic compounds were detected in SS strata samples, accounting for 25.09-55.73% of OM (mean proportion of 41.92%), most of which were diphenols.
Among the detected hydrocarbons, alkanes accounted for a distinctly larger proportion than other hydrocarbons. Alkanes are thought to be derived from lipid breakdown and are important biomarkers in geological analysis (Shanina, 2002;Fang et al., 2014). In total, between 10 and 17 hydrocarbons were identified in each SS strata sample, accounting for 29.33-56.02% of OM (mean proportion of 40.44%), with the second highest abundance after phenols. However, among all the identified compounds, hydrocarbons accounted for the most compound types.
The content of aldehyde compounds in SS strata samples ranged from 3.47 to 19.24% of OM (mean proportion of 10.19%). Among all the detected aldehydes, the average proportion of nonanal was highest. It is well established, that nonanal is an oxidative catabolite of fatty acids . Therefore, the high proportion of nonanal throughout the sediment depth profile indicates periods of intense vegetation (Dąbrowska et al., 2014).
The esters detected in SS strata samples are mainly methyl, butyl, and amyl esters, accounting for 3.34-11.41% of OM (mean proportion of 6.85%). Two ketones were identified, include 3heptanone and methylheptenone. Ketones primarily result from amino acid degradation (Zhuang et al., 2016), with amino acids being the main protein units present in all living organisms. 2pentylfuran was the only furanoic compound detected in all SS strata samples, accounting for 0.07-0.55% of OM (mean proportion of 0.18%). Alcohols were only detected in sample DP02, including 1-octanol and 2-ethyl-1-hexanol, with an abundance of 1.23%.

n-Alkanes in Sediments
n-Alkanes with carbon numbers between C 15 and C 36 were found in all samples (Figure 4). Short-chain n-alkanes (C 15 -C 20 ) were less abundant, with relative abundances varying from 1.15 to 8.61%, with an average content of 3.51%. The abundance of midchain n-alkanes (C 21 -C 25 ) varied from 9.37 to 35.99%, with an average content of 24.54%. In contrast, the long-chain n-alkanes (C 26 -C 36 ) were highly abundant in all sediments, ranging from 53.25 to 83.99%, with an average content of 69.83%. In SB sediments, the carbon maximum (C max ) of long-chain n-alkanes occurred at C 31 , while for SS sediments, the C max occurred at C 25 and C 27 . The carbon preference index (CPI) values in SB sediments varied from 3.79 to 11.02 (avg 6.72), the Pristane/Phytane (Pr/Ph) values varied from 0.23 to 0.83 (avg 0.41), and the Pr/nC 17 and Pr/nC 18 values varied from 1.03 to 1.62 (avg 1.35) and 0.98-1.17 (avg 1.11), respectively. In SS sediments, the CPI values varied from 1.22 to 1.53 (avg 1.31), the Pr/Ph values varied from 0.26 to 0.44 (avg 0.38), and the Pr/ nC17 and Pr/nC18 values varied from 0.70 to 1.54 (avg 1.16) and 1.20-1.87 (avg 1.46), respectively.

FTIR Analysis of Sediment Organic Matter
The identified FTIR absorbance bands from sediments in the same stratum exhibited identical peak positions and comparable absorbance intensities, indicating a similar molecular composition of SOM. This observation is in agreement with the previously reported results of Giovanela et al. (2010) and Pongpiachan et al. (2013) who found similar molecular compositions of SOM from within the same stratum or the same area. In the present study, several similar peak positions were found to exist in both strata, confirming that the same organic matter was present in different strata.  Despite the broad similarities in SOM chemical composition between the two different lithologic strata, there were significant differences observed in the relative abundance of certain C functional groups and distribution trends within strata. Compared with SS strata, five sediments of SB strata generally exhibited a higher content of ketone C O groups (1,869-1,678 cm −1 ) and a lower content of aliphatic C (2,926-2,847 cm −1 ), with an equal abundance of alcohols (1,168-1,078 cm −1 and 1,031-1,028 cm −1 ) (Figure 3). The enrichment of aliphatic C forms in SS strata may be associated with a more advanced stage of OM decomposition (Omar et al., 2020), with the persistence of aliphatic C groups during decomposition suggesting that it may represent relatively stable and recalcitrant forms of C in sediments (Lorenz et al., 2007;Calderón et al., 2011). Therefore, OM and the microorganisms that can decompose OM should be more abundant in SS than in SB strata. A lower abundance of ketone C O groups (1,869-1,678 cm −1 ) was observed in SS strata. According to Ndwigah et al. (2013), ketones are mainly produced by halophilic fungi, supporting the theory that the SB layer was deposited in an arid climate under increased salinity conditions. Alcohols exist widely in organisms and sediments, and the genesis of alcohols in sediments is extremely complex, which not only depends on the source of OM, but is also affected by many factors such as biochemistry and geochemistry (Duan et al., 1995). Previous studies on the genesis of alcohols have been based on characteristic organisms and different sedimentary environments, with the main influencing factors including either alga (Robinson et al., 1986;Nichols et al., 1990) or terrestrial organic matter (Volkman et al., 1987;Goossens et al., 1989). This result explains why the alcohol content of the different strata is almost the same, despite the SS sedimentary environment being different from that of SB strata.
Furthermore, the SS strata exhibited an abundance of characteristic peaks for carbonate (2,519-2,514 cm −1 , 1,451-1,432 cm −1 , 1,471 cm −1 , and 880-873 cm −1 ) and detrital minerals (730-427 cm −1 ). According to Keil et al. (1994), OM is associated with mineral grains, which subsequently reduces the rate of decomposition (Wang and Lee, 1993), with OM primarily filling the mineral pores of quartz aggregates, forming OM networks (Zhao et al., 2017b). Detrital minerals are mainly composed of quartz, sodium feldspar and muscovite, with these compounds serving a significant role in the diagenetic process and reflecting extrabasinal and intrabasinal sediment inputs (Zhao et al., 2017a). In general, detrital minerals come from peripheral areas of the lake and are supplied by processes including surface run-off, river transport, atmospheric rainfall, and lakeshore erosion (Shen et al., 2010). In humid climates, increased precipitation promotes surface run-off, carrying the insoluble detrital minerals into the salt lake and depositing them directly on the bottom of the lake bed. As a result, the paleoclimate appears to be warmer and wetter after 4 ka BP, with the exogenous OM and TOC content in SS strata being higher than in SB strata.

SOM Analysis by GC-MS
GC-MS analysis showed that the molecular types OM in different lithologic strata are relatively similar, with the identified substances classified into seven types, including aldehydes, hydrocarbons, phenols, esters, ketones, alcohols, and furans. The proportional abundances of different types of OM in different lithologic strata are shown in Figure 5.
Different strata have different distributions of organic molecular types. SS strata contain the highest content of phenolic substances, while the content of hydrocarbons is close to that of phenolic substances, with the total content of the two substances being more than 80%. The main sources of phenols in sediments are root exudates, microbial secretions, and plant tissue decomposition (Bao et al., 2013a(Bao et al., , 2013bVane et al., 2013), while hydrocarbons are affected by many factors such as organic source, microbial activity, and the redox environment (Wu et al., 2001). In view of the particular characteristics of salt lake sediments, exogenous OM and microbial communities significantly contribute to the OM content in SS strata. In SB strata, the content of aldehydes accounts for almost half of the OM. In high salinity environments, aldehydes are mainly produced by algae (such as diatoms) (; Chen et al., 2014a;Chen et al., 2014b) and therefore, the main contributor to OM in SB strata is likely to be algae.
In general, according to the sedimentary characteristics of the salt lake, the environmental conditions of SB strata during the sedimentary period are more extreme than for SS strata (Shen et al., 2010). However, more organic compounds were identified in SB strata (mean of 41 compounds) than in SS strata (mean of 30 compounds). This observation is partially supported by a previous study showing that the higher salinity of halite may affect the enrichment of some organic compounds (Strehse et al., 2018). Higher salinity also results in stratification of the water column, creating an anoxic sedimentary environment and inhibiting bacterial proliferation (Zhu et al., 2004), thereby reducing the consumption of OM. Klinkhammer and Lambert (1989) presented evidence for organic carbon preservation in high salinity sediments, while Jellison et al. (1996) reported a positive correlation between the accumulation of organic carbon and salinity levels. Overall, these phenomena may be attributed to the specific characteristics of sediment layers, such as salinity and biological composition.
The main differences in OM between the two strata layers are the proportions of phenols and aldehydes ( Figure 6). The proportion of phenols in the SS strata was much larger than in the SB strata, while the proportion of aldehydes in the SB strata was higher than in the SS strata. Many organisms have developed protective mechanisms allowing phenol to be used as their sole carbon and energy source (Schie and Young, 2000). Thus, the degradation of phenols by halophilic bacteria in high salinity environments is of relevance (Peyton et al., 2002;Li et al., 2019). Haddadi and Shavandi (2013) reported a Halomonas sp. strain that exhibited the ability to utilize phenol as the sole carbon and energy source in saline environments. Preliminary evidence has been reported that indicates that phenol degradation proceeds in some halophilic bacteria through a meta-cleavage pathway (Acikgoz and Ozcan, 2016). The survival and activity of halophilic bacteria are better suited to the environment of SB strata than SS strata, leading to more phenol degradation in SB strata. This phenomenon also results in a larger reduction in the Frontiers in Earth Science | www.frontiersin.org August 2021 | Volume 9 | Article 734458 8 proportion of phenols in SB strata. In addition, it further confirms that the paleoclimate of 4 ka BP appears to have been more extreme and that phenols may be a potential biomarker for reconstruction of the paleoclimate in this period. This is partially supported by a previous study showing that lignin phenol can be used as a biomarker for paleovegetation in soils (Kovaleva and Kovalev, 2015).
Algae are another main producer of OM in high salinity environments, with a large number of aldehydes produced at the same time. Especially in the later stage of diatom growth, the content of aldehydes continues to increase, with previous studies reporting that the content of aldehydes can reach up to 46.8% of the total volatile fraction (Chen et al., 2014a). In a previous study on nine kinds of diatoms, Chen et al. (2014a) found that nonanal was dominant among the volatile aldehydes of diatoms in the later stage of sedimentation. This result corresponds to the observed content of nonanal in the present study. Therefore, it may be inferred that the higher content of aldehydes in the SB strata occurs mainly due to the higher abundance of diatoms in salt lake water during the sedimentary period, as compared to SS. The aldehyde content also confirmed that the paleoclimate around 4 ka BP changed from cold and dry to relatively warm and wet.

Correlation Analysis Between Mineral Composition and Organic Matter
In this study, correlation analysis was performed between detrital minerals and various types of OM in all samples ( Table 4). Results show a strong positive correlation between detrital minerals and phenols, with a strong negative correlation between detrital minerals and aldehydes, which was consistent with the results of GC-MS and XRD analysis. However, a negative correlation between detrital minerals and ketones was observed. According to Schmidt et al. (2015), most volatile OM components are considered to be side products of primary and secondary metabolism, with ketones produced by bacteria or fungi (Piechulla and Degenhardt, 2014). These findings agree with the results of a previous study, showing that some halophilic fungal species in salt lakes were able to produce a wide range of secondary metabolites such as ketones (Ndwigah et al., 2013). Therefore, it can be preliminarily determined that the production of ketones in this study may be closely related to some halophilic fungi in lake sediments, indicating that ketones should be produced in large quantities in salt lakes under arid climate conditions.
OM is always associated with mineral grains (Keil et al., 1994) in lake sediments and due to the high salinity in salt lake sediments, OM is mainly formed by microorganisms (Han et al., 2020) and algae (Meyers and Ishiwatari, 1993). Detrital minerals serve a significant role in the diagenetic process and reflect extrabasinal and intrabasinal sediment inputs in salt lake environments (Zhao et al., 2017b). In general, detrital minerals originate from peripheral areas of the lake and are supplied by processes including surface run-off, river transport, atmospheric rainfall, and lakeshore erosion (Shen et al., 2010). In humid climates, increased precipitation levels promote higher levels of surface run-off, carrying insoluble detrital minerals into the lake and depositing them directly on the bottom of the salt lake. This spatial distribution of detrital mineral in sediments shows that the preferential feeding areas correspond to SS in our study. The high  abundances of quartz and feldspar in the sediments reveal a period of active detrital input. Wirrmann et al. (2001) demonstrated that detrital mineral flux can be used as a proxy indicator for the mean rainfall intensity and based on this, precipitation during SS deposition was higher than during SB deposition. Furthermore, the content of detrital minerals increases while salinity is simultaneously reduced, which is consistent with the results of XRD analysis in the present study. This conclusion is supported by Yang (2015) who reported the occurrence of frequent catastrophic floods by the Nalinggele River during the 4.7-3.5 ka BP period.
The correlation between mineral composition and OM in the sediments of East Taijinar salt lake, demonstrates that the concentrations of aldehydes and ketones were relatively high due to the dry climate during the period of deposition and the relative development of halophilic algae in the lake prior to 4 ka BP. However, with adaptation of the climate to warm and humid conditions, a large amount of exogenous OM was imported, resulting in a relative increase in the phenol content after 4 ka BP. Meanwhile, the detrital mineral flux in different strata also confirmed that the paleoclimate had changed from dry and cool to warm and wet in the 4 ka BP period.
Further Verification of Paleoclimate Changes Around 4 ka BP Using the Content of n-Alkanes n-Alkanes in sediments mainly originate from fatty acids and waxes in organism cells (Sachse et al., 2004), serving as widely studied biomarkers that are widely distributed in soils , loess deposit (Zeng et al., 2011), and lacustrine deposits. Bacteria and algae produce both odd and even carbon number short-chain n-alkanes (Meyers, 2003), while the abundance of odd long-chain n-alkanes (C 27 -C 31 ) has been extensively utilized as an indicator of terrestrial or land-derived OM (Pearson and Eglinton, 2000;Zhao et al., 2003). In the present study, the SS sediments samples exhibited an odd-over-even predominance and were mainly dominated by C 31 , while SB sediment samples were dominated by C 25 and C 27 without an odd-over-even predominance. This result is in agreement with a previous study showing that long-chain n-alkanes can also be produced by bacteria and algae, although these long-chain n-alkanes do not exhibit an odd-even dominance (Castañeda and Schouten, 2011).
The nC 31 /nC 29 ratio of n-alkanes in sediments is often used as an indicator of environmental and paleoclimate change (Meyers and Ishiwatari, 1993). It has been shown that nC 31 is the main n-alkanes peak in herbaceous plants, while nC 29 is the main n-alkanes peak in woody plants. When herbaceous plants are dominant, the climate tends to be warm and humid, resulting in nC 31 being the main n-alkane compound; when woody plants are dominant, the climate tends to be cold and dry, resulting in n-alkanes being dominated by nC 29 (Cranwell, 1973;De las Heras et al., 1989). Therefore, nC 31 /nC 29 > 1.0 indicates that the climatic environment in this period was relatively warm and humid, while nC 31 /nC 29 < 1.0 reflects that the climatic environment was cold and dry. The established nC 31 /nC 29 ratios are presented in Figure 2, with the ratio of nC 31 /nC 29 ranging from 0.64 to 0.79 in SB, while the values were >1.0 (1.41-2.31) in SS sediments. Therefore, the paleoclimate around 4 ka BP during the sedimentary period ranging from SB to SS, changed from dry and cool to warm and wet.
Furthermore, CPI values > 1.0 indicate terrestrial OM inputs (Bray and Evans, 1961). The CPI values in SS sediments ranged from 3.79 to 11.02 (avg 6.72), demonstrating that the main source of OM appears to be terrestrial in SS sediments. The Pr/Ph, Pr/C 17 , and Pr/ C 18 values are presented in Figure 2. The Pr/Ph ratio is used in a lacustrine setting to infer the oxic vs anoxic state during OM deposition, with Pr/Ph ratios <1.0 indicating anoxic conditions, whereas values > 1.0 reflect suboxic to oxic environments (Didyk et al., 1978). The Pr/Ph value of East Taijinar salt lake sediments from around the 4 ka BP period, suggest that deposition occurred under anoxic conditions due to the overlying lake water. It has been reported that Pr/nC 17 and Pr/nC 18 values of OM without degradation are very low (0.1-0.5), while higher values occur when OM is degraded by microorganisms (Zhao et al., 2016). In the present study, the ratio of Pr/nC17 (0.70-1.62 avg 1.26) and Pr/ nC18 (0.98-1.87 avg 1.28) indicate strong microbial activities in East Taijinar salt lake sediments around the 4 ka BP period.
These findings indicate that the OM in SS sediments was mainly terrestrial OM, while the OM in SB sediments could have a bacterial and algal source. The paleoclimate of the SB deposition period appears to be drier than that of SS deposition period, with the climate becoming warmer after the 4 ka BP period. Furthermore, the OM in all sediment samples was found to be strongly affected by microorganisms in anoxic environments, which is consistent with the results on OM distribution

CONCLUSION
This study preliminarily investigates the spectral characteristics and volatile organic composition of East Taijinar salt lake sediments around 4 ka BP. XRD, FTIR, and GC-MS analysis enabled the characterization of sediment OM. The TOC concentration was significantly lower in SB strata than SS strata, which may be attributed to environmental differences during the formation of salt lake sediments. FTIR spectroscopy showed a relative enrichment of ketone C O groups in SB strata, and a decrease in aliphatic C compared with the SS strata. These findings are confirmed by the results of GC-MS analyses and the relationship between OM and detrital minerals. The high content of phenols in the SS strata indicates that the organic sources and microorganisms contribute greatly to the OM content of SS, with the paleoenvironment likely being warm and humid. The high content of aldehydes in SB strata indicates a high abundance of diatoms in the salt lake water during the sedimentation period, with the climate likely to be relatively dry and cool at that time. Correlation analysis showed a strong positive correlation between detrital minerals and phenols, while a strong negative correlation was observed between detrital minerals and aldehydes, and a negative correlation was also observed between detrital minerals and ketones. The correlation analysis results indicate that OM in SS sediments is strongly affected by exogenous organic inputs, with the detrital mineral flux in different strata further confirming that the climate had changed from cool and dry, to wet and warm at 4 ka BP. The analysis of n-Alkanes in sediments, provides a tool to sensitively monitor paleoclimate changes in the middle area of the QB around the 4 ka BP period. Overall, the dominant carbon number, nC 31 /nC 29 ratio and CPI values all indicate that the climate changed from dry to warm, confirming that the spectral characteristics and content of VOCs in salt lake sediments are related to paleoclimate changes.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors.

AUTHOR CONTRIBUTIONS
ZM and XL performed the fieldwork; YZ and LY performed laboratory analysis; XL and LY prepared the figures and tables; XL, YZ, and FH wrote the manuscript with the help of all coauthors.

FUNDING
This work was supported by the Natural Science Foundation of Qinghai Province (2019-ZJ-911).