INTRODUCTION
Lake sediments from arid and semi-arid areas provide high-resolution records of climatic and environmental changes. Radiocarbon (14C) dating has been widely used to develop chronologies of lake sediments as old as ~50,000 yr (Martin Reference Martin1999). Terrestrial plant remains in lacustrine sediments are ideal materials for 14C dating, because their carbon is derived directly from atmospheric CO2 (Bertrand et al. Reference Bertrand, Araneda, Vargas, Jana, Fagel and Urrutia2012). However, in arid regions the sparse vegetation cover and rapid organic matter decomposition rate often result in the scarcity of terrestrial plant remains in the lake sediments (Zhang et al. Reference Zhang, Ming, Lei, Zhang, Fan, Chang, Wunnemann and Hartmann2006). Therefore, in most cases lake sediment chronologies must be based on the dating of bulk organic matter, aquatic plant remains, and/or carbonates (e.g. Shen et al. Reference Shen, Liu, Matsumoto, Wang and Yang2005a, Reference Shen, Liu, Wang and Matsumoto2005b; Liu et al. Reference Liu, Shen, Wang, Wang and Liu2007; Zhao et al. Reference Zhao, Yu, Zhao, Ito, Kodama and Chen2010; An et al. Reference An, Colman, Zhou, Li, Brown, Jull, Cai, Huang, Lu, Chang, Song, Sun, Xu, Liu, Jin, Liu, Cheng, Liu, Ai, Li, Liu, Yan, Shi, Wang, Wu, Qiang, Dong, Lu and Xu2012). However, these materials are usually composed of a complex mixture of carbon from various sources, and, depending on the magnitude of the reservoir effect of the lake, they typically yield 14C ages that are older than the true age (Fontes et al. Reference Fontes, Gasse and Gibert1996). Therefore, the chronological framework provided by the dating of these materials may be unreliable unless the reservoir effect is properly assessed (Stein et al. Reference Stein, Migowski, Bookman and Lazar2004; Zhou et al. Reference Zhou, Chen, Wang, Yang, Qiang and Zhang2009; Ascough et al. Reference Ascough, Cook, Church, Dunbar, Einarsson, McGovern, Dugmore, Perdikaris, Hastie, Friðricksson and Gestsdóttir2010; Hou et al. Reference Hou, D’Andrea and Liu2012).
The potential reservoir effect of a lake can be assessed using different methods. Assuming the sedimentation rate was constant during the recent past, the reservoir effect can be estimated from the intercept on the age axis using a linear regression between 14C age and depth (Fontes et al. Reference Fontes, Gasse and Gibert1996; Shen et al. Reference Shen, Liu, Matsumoto, Wang and Yang2005a, Reference Shen, Liu, Wang and Matsumoto2005b). However, this assumption is unlikely for most lakes, and therefore, polynomial regressions are often used when the sedimentation rate in a lake has changed through time (e.g. Li et al. Reference Li, Wang, Morrill, Anderson, Li, Zhang and Zhou2012; Zhou et al. Reference Zhou, Cheng, Jull, Lu, An, Wang, Zhu and Wu2014; Zhang et al. Reference Zhang, Ma, Qiang, Huang, Li, Guo, Henderson, Holmes and Chen2016a, Reference Zhang, Meyers, Liu, Wang, Ma, Li, Yuan and Wen2016b). In addition, the reservoir effect of lakes can also be determined by a mass balance model that allows for different sources of carbon transported to the lake water if their respective 14C activities can be quantified (Yu et al. Reference Yu, Shen and Coleman2007). Such a model requires a better understanding of modern carbon cycles in watershed ecosystems. However, the past carbon cycles of a lake are difficult to quantify; furthermore, the 14C ages of dissolved inorganic carbon of lake water (DICLW) and aquatic plant remains from surface sediments usually vary between different areas of a lake (e.g., Geyh et al. Reference Geyh, Schotterer and Grosjean1998; Mischke et al. Reference Mischke, Weynell, Zhang and Wiechert2013), indicating that the reservoir effects of sediment cores from different lake areas may be inconsistent; this further compounds the difficulty of assessing the potential reservoir effect. Therefore, thorough investigations of the variability of the reservoir effect and its potential determining factors are essential to provide insights into the assessment of the reservoir effects of sediment cores.
In general, the reservoir effect of aquatic plant remains is controlled by the age of the DICLW, which depends mainly on the input of exogenous old carbon and on CO2 exchange between the lake water and the atmosphere (Hatté and Jull Reference Hatté and Jull2007; Zhang et al. Reference Zhang, Ma, Qiang, Huang, Li, Guo, Henderson, Holmes and Chen2016a). In lakes with a relatively stable hydrological cycle and well-mixed lake water, the photosynthesis of aquatic plants will significantly affect lake-water pCO2 and CO2 exchange between the lake water and the atmosphere (Longhurst and Harrison Reference Longhurst and Harrison1989; Coletta et al. 2001), and therefore it is likely to impact the reservoir effect of DICLW. However, detailed studies of the influence of photosynthesis on reservoir effects are scarce. Here, we present the results of 14C dating of different carbon species from the Genggahai Lake system, a small, shallow lake in the Gonghe Basin on the northeastern Qinghai-Tibetan Plateau; submerged macrophytes currently flourish in the lake. Combined with the results of monitoring of the physical and chemical properties of the lake water from different parts of the lake with different aquatic plant communities, we attempt to investigate the variability of 14C ages of different carbon sources and their potential influencing factors.
STUDY AREA
Genggahai Lake (36°11′N, 100°06′E) is in the central Gonghe Basin (Figure 1a), ~50 km south of Qinghai Lake, at an altitude of 2860 m above sea level (asl). From the meteorological data from Gonghe Station from 1981 to 2010 AD, the mean annual temperature was ~4.6°C and the mean annual precipitation was ~325 mm; precipitation was mainly from May to September (Figure 1a). The basin was filled by fiuviolacustrine sediments (the Gonghe Formation), during the early- to mid-Pleistocene, and was subsequently tectonically up lifted after the mid-Pleistocene (Perrineau et al. Reference Perrineau, Van Der Woerd, Jing, Gaudemer, Pik, Tapponnier, Thuizat and Zheng2011). Aeolian activity is prevalent in the basin at the present (Qiang et al. Reference Qiang, Liu, Jin, Song, Huang and Chen2014).
Figure 1 Settings and location. (a) Physical environments of the Gonghe Basin. Insets: Major atmospheric circulation systems influencing the study area (the modern extent of the Asian summer monsoon is indicated by the grey dashed line; after Gao et al. Reference Gao, Xu, Guo and Zhang1962); and the monthly mean precipitation and mean temperature at Gonghe Meteorological Station from 1981 to 2010 AD (source: Chinese Meteorological Administration). (b) Topography of the area surrounding Genggahai Lake, lake bathymetry, and location of sampling sites. The filled triangles, filled circles, filled squares, and filled pentagrams represent the locations of the lake-water DIC samples, macrophyte-remain samples, the spring-water DIC samples, and the living P. pectinatus sample, respectively.
The lake contains two separate basins, Upper Genggahai and Lower Genggahai Lake (Figure 1b). The latter is currently almost dry, and in this study Genggahai Lake refers to Upper Genggahai Lake. It is a small shallow lake with an area of ~2 km2 and a maximum water depth of ~1.8 m. The lake is currently occupied by the dense growth of submerged macrophytes (e.g., Potamogeton pectinatus, Myriophyllum spicatum, and Chara spp.) and is surrounded by grassland. Currently, there are no large glaciers on the summits of the surrounding mountains and thus excess meltwater has little effect on the modern hydrology. The lake has no natural discharge outlets or direct surface inflows and is fed mainly by groundwater. Within the lake basin, springs emerge as artesian water. Small spring-water streams emanate from sediment outcrops in the northwest part of the catchment and feed Genggahai Lake and sustain the grassland (Figure 1b). The modern spatial distribution of aquatic plants in the lake depends on water depth (Figure 1b). Chara spp. occupy the depth range of 20–100 cm, while P. pectinatus and M. spicatum grow within the depth range of 50–180 cm (Qiang et al. Reference Qiang, Song, Chen, Li, Liu and Wang2013). Desert steppe mainly dominates the vegetation community in the terrestrial part of the catchment (Qiang et al. Reference Qiang, Liu, Jin, Song, Huang and Chen2014). Human activity is of low intensity around the lake, consisting of grazing by the animals of Tibetan herdsmen.
MATERIALS AND METHODS
In the summer seasons (May to September) from 2012 to 2013 we conducted a monitoring program of the physical and chemical properties of the lake, in areas with different aquatic plant communities (Figure 1b). Temperature, dissolved oxygen (DO) concentration, pH, and conductivity of lake water were measured using a portable water quality analyzer (Aquaread AP-1000). Macrophytes (Chara spp., P. pectinatus, and M. spicatum), lake water at the 20-cm depth of the water body from the monitoring sites, and spring water in the northwestern part of the catchment, were collected. Three sediment cores were recovered from different areas of the lake using a modified Livingstone piston corer. Large pieces of macrophyte remains (MR) from the uppermost 1-cm interval of the sediment cores were picked for AMS 14C dating. Samples MR-C1 and MR-C2 are the remains of Chara spp. Samples MR-P, MR-M1 and MR-M2 are the remains of vascular plants (P. pectinatus and M. spicatum) (Figure 1b).
Water samples were stabilized using 20 µL saturated mercuric chloride (HgCl2) solution and sealed in the field. Dissolved inorganic carbon (DIC) was precipitated as BaCO3 by adding 10 mL of saturated BaCl2 solution to 500 mL water before filtering with cellulose nitrate filter papers, followed by drying at 40°C. DIC concentration was measured by acid and alkali titration. Stable carbon isotopes of aquatic plants (δ13Corg) were measured using an on-line Conflo III-Delta Plus isotope ratio mass spectrometry, combined with a Flash EA1112 elemental analyzer. The results are reported in ‰ relative to VPDB. The precision was better than 0.1‰ from replicated measurements of standards. All the above experiments were carried out in the Key Laboratory of Western China’s Environmental Systems, Lanzhou University. Six DIC samples, five samples of macrophyte remains and one sample of living P. pectinatus (Figure 1b), were radiocarbon dated using accelerator mass spectrometry (AMS) by Beta Analytic Inc. (Miami, FL, USA). The fraction of modern carbon (pMC) of the sample is defined as (Donahue et al. Reference Donahue, Linick and Jull1990)
$${\rm pMC}{\equals}{\rm D}^{{14}} {\rm C}_{{\rm S}} \,/\,{\rm D}^{{14}} {\rm C}_{{1950}} $$
where D14CS is the measured ratio for the sample, blank-corrected and adjusted to δ13C values; and D14C1950 is the measured ratio of the standard, blank-corrected, adjusted to δ13C values, and recalculated to AD 1950. The AMS 14C dates are standardized using a consensus value for the National Institute of Standards and Technology (NIST) modern reference standard (SRM 4990C). The 14C age of a sample is defined as
$$^{{14}} {\rm C}_{{{\rm age}}} {\equals}-\!8033{\times}{\rm ln}\left( {{\rm pMC}} \right)$$
RESULTS
14C Ages
The sampling locations are shown in Figure 1b and sample details and the 14C ages of DIC and macrophyte remains are listed in Table 1. Two spring water DIC samples (DICSpring; i.e., DIC-S1 and DIC-S2) have significantly different 14C ages (7030 ± 30 14C yr BP and 5630 ± 30 14C yr BP) (Table 1). Compared to spring water, the 14C ages of lake-water DIC (DICLW) and macrophyte remains (MR) from the uppermost 1-cm interval of the sediments are much younger, ranging from 490 ± 30 to 1440 ± 30 14C yr BP (Table 1). It is noteworthy that the 14C ages of DICLW and MR from areas dominated by Chara spp. (DIC Chara and MR Chara ; i.e., DIC-C1, DIC-C2, MR-C1 and MR-C2) are younger than those from areas dominated by vascular plants (DICVascular and MRVascular; i.e., DIC-P, DIC-M, MR-P, MR-M1 and MR-M2) (Table 1). In addition, a living individual of P. pectinatus (Living-P) was dated to 1500 ± 30 14C yr BP.
Table 1 14C ages of spring water, lake water and plant macrophytes from core top samples.
Physical and Chemical Properties of the Lake Water and δ13Corg Values of Aquatic Plants
The pH values and conductivity of the lake water range from 8.1 to 9.5 and from 1.5 to 2.1 ms cm–1, respectively. The temperature of the lake water is high in June and July. In addition, the lake water from areas dominated by Chara spp. has a higher DO concentration (DO Chara ) and lower DIC concentration (DIC Chara ) than those from areas dominated by vascular plants (DOVascular and DICVascular) (Figure 2d–e). The concentrations of DO Chara and DOVascular range from 100–248% (mean of 188%) and 85–220% (mean of 157%), respectively. The ranges of DIC Chara and DICVascular concentrations are 8–14 mmol L–1 (mean of 10 mmol L–1) and 10–15 mmol L–1 (mean of 12 mmol L–1), respectively. In addition, Chara spp. has more negative δ13Corg values, with the range of –18.6 to –14‰ (mean of –16.1‰), than those of vascular plants, which have the range of –15.2 to –9.3‰ (mean of –12.5‰) (Figure 2f).
Figure 2 Physical and chemical properties of lake water in areas dominated by different submerged aquatic plants, and stable carbon isotope values of aquatic plants. (a–e) Temperature, conductivity, pH, DO, and DIC contents of the lake water, respectively. (f) δ13Corg of various aquatic plants.
DISCUSSION
The two DICSpring samples, collected in 2012 and 2016, have significantly different 14C ages: 7030 ± 30 14C yr BP and 5630 ± 30 14C yr BP (Table 1). The catchment substrate of Genggahai Lake consists mainly of surface fluvial-fan sediments and loose fluvio-lacustrine sediments of the Gonghe Formation (Xu et al. Reference Xu, Xu and Shi1984; Perrineau et al. Reference Perrineau, Van Der Woerd, Jing, Gaudemer, Pik, Tapponnier, Thuizat and Zheng2011), which contain abundant ancient carbonates. Dissolution of carbonates in topsoil or sediments within the catchment produces old DIC which infiltrates into the groundwater (Arslan et al. Reference Arslan, Kadir, Abdiglu and Kolayli2006; Olaaon Reference Olaaon2009). However, the infiltrating water also carries atmospheric CO2 and the CO2 generated by plant root respiration, which results in relatively younger 14C ages of the groundwater. Nevertheless, the distinctly different 14C ages of the two DICSpring samples may suggest that the 14C age of groundwater within the lake basin is quite sensitive to the meteorological conditions, which is probably related to the regional geological and geomorphic features. The porous nature of the sediments and the steep hydraulic gradient make the catchment of Genggahai Lake highly permeable (Qiang et al. Reference Qiang, Song, Jin, Li, Liu, Zhang and Zhao2017), which is conducive to the input of old DIC into the groundwater and may largely account for the significant difference in the 14C ages of the two DICSpring samples. In addition, variations in the regional meteorological conditions may also play an important role in determining the 14C ages of the groundwater, which need to be further investigated by long term, continuous monitoring of the 14C ages of spring water.
The DICLW and MR have much younger 14C ages compared to those of the groundwater, varying from 490 ± 30 to 1440 ± 30 14C yr BP. Further, the 14C ages of DICchara and MRchara are younger than those of DICVascular,MRVascular and Living-P (Table 1). The water in Genggahai Lake has a high alkalinity which could inhibit carbonate dissolution; therefore, the old carbon in the lake is mainly derived from dissolved lithogenic and biogenic sources due to inflowing groundwater (Abbott and Stafford Reference Abbott and Stafford1996; Geyh et al. Reference Geyh, Schotterer and Grosjean1998; Billett et al. Reference Billett, Garnett and Harvey2007). Consequently, it would be expected that the input of groundwater (spring water) into lakes may affect the 14C ages of DICLW and MR. Based on the dating of DICLW from Gahai Lake in the Qaidam Basin, Zhang et al. (Reference Zhang, Ma, Qiang, Huang, Li, Guo, Henderson, Holmes and Chen2016a) pointed out that the input of groundwater into the lake leads to site-specific 14C ages of the DICLW, which decreased rapidly with increasing distance from the springs that feed the lake. In the Genggahai Lake basin, the 14C ages of samples DIC-S1 and DIC-S2, collected at different times, differ by 1400 yr, while samples DIC-C1 and DIC-C2 have similar 14C ages, i.e., 720 ± 30 14C yr BP and 710 ± 30 14C yr BP. This reflects the fact that the input of exogenous old DIC dissolved in the groundwater has not significantly changed the 14C ages of DICLW. The springs flowing into Genggahai Lake emerge as artesian water, reflecting the lower position of the lake relative to the catchment water table (Qiang et al. Reference Qiang, Song, Jin, Li, Liu, Zhang and Zhao2017). Therefore, it is plausible that artesian water flowing into Genggahai Lake might also be found in other parts of the catchment and even in the lakebed. Nevertheless, the sub-surface artesian springs feeding lakes probably lead to efficient mixing of the lake water (Anderson et al. Reference Anderson, Abbott, Finney and Burns2005), and thus it would be expected that the ages of DICLW would be similar, even in different areas of the lake. However, the apparent difference between the ages of DICLW and MR from the areas dominated by different aquatic plants suggests that the input of exogenous old DIC transported by the spring-water streams cannot fully explain the differences in the dating results.
Besides the input of old carbon, CO2 exchange between the lake water and the atmosphere also has a large impact on the reservoir ages of lakes (Hatte and Jull Reference Hatté and Jull2007). A significant exchange of CO2 between them would impart a younger bias to the DICLW ages (Fontes et al. Reference Fontes, Gasse and Gibert1996). At Genggahai Lake, the much younger 14C ages of DICLW, compared to the ages of groundwater, suggest that the lake water is proceeding towards equilibrium with the atmosphere. In general, the CO2 exchange between lake water and the atmosphere is controlled by lake water stratification (i.e., water depth), wind regime, and ice-cover duration (Zhang et al. Reference Zhang, Ma, Qiang, Huang, Li, Guo, Henderson, Holmes and Chen2016a). There is no thermal stratification in such a shallow water body as Genggahai Lake and in addition wind regimes in the study area are generally strong; thus, these conditions favor a high rate of CO2 exchange between the lake water and the atmosphere. In addition, the photosynthesis of aquatic plants exerts a large influence on the rate of CO2 exchange between lake water and the atmosphere (Longhurst and Harrison Reference Longhurst and Harrison1989). Coletta et al. (Reference Coletta, Pentecost and Spiro2001) have demonstrated that that the intense photosynthesis of aquatic plants in the summer season consumes a large amount of DIC and decreases the lake-water pCO2, which effectively enhances the rate of CO2 exchange between lake water and the atmosphere, and results in the increased absorption of atmospheric CO2 by the lake water. Therefore, the overall young 14C ages of DICLW and MR in Genggahai Lake can be ascribed to the photosynthesis of aquatic plants.
The temperature, pH and conductivity of the lake water are consistent across the different lake areas (Figure 2a–c), which reflects the fact that the shallow water body and the strong wind regimes at Genggahai Lake promote thorough mixing. Nevertheless, the 14C ages of DICLW from areas of the lake with different aquatic plant communities are still variable. The lake-water samples were collected at 20-cm water depth, which excludes the influence of water depth on the variations in the 14C ages of DICLW. In fact, the photosynthesis rates of Chara spp. and vascular plants are distinctly different (Van den Berg et al. 2002), which is also revealed by the different DO and DIC concentrations in areas of the lake with different aquatic plant communities (Figure 2d–e). Aquatic plants absorb DIC and release oxygen through photosynthesis, resulting in high concentrations of DO and low concentrations of DIC in lake water. Therefore, the high concentrations of DO Chara and the low concentrations of DIC Chara reflect the higher photosynthesis rate of Chara spp. compared to vascular plants (Figure 2d–e). Moreover, P. pectinatus, M. spicatum, and Chara spp. preferentially utilize 12C for photosynthesis, giving rise to depleted 13C values in plant tissues (Hammarlund et al. Reference Hammarlund, Aravena, Barnekow and Possnert1997; Sand-Jensen Reference Sand-Jensen1983; Ray et al. Reference Ray, Klenell, Choo, Pedersén and Snoeijs2003). The more negative δ13C Chara is further evidence of the high photosynthesis rate of Chara spp. (Figure 2f). The photosynthesis of aquatic plants not only fixes carbon into plant tissues, but it also promotes carbonate precipitation (Apolinarska et al. Reference Apolinarska, Pełechaty and Pukacz2011). The high photosynthesis rate of Chara spp. can lead to heavy carbonate precipitation directly onto charophyte plant surfaces as mineral encrustations (Apolinarska et al. Reference Apolinarska, Pełechaty and Pukacz2011). Blindow (Reference Blindow1992) noted that encrustations account for up to 70% of the dry mass of Chara tomentosa. In fact, abundant encrustations are also found on Chara spp. from Genggahai Lake (Qiang et al. Reference Qiang, Song, Chen, Li, Liu and Wang2013). During the growing season, the rapid and high consumption of DIC in the lake due to the intense photosynthesis of Chara spp. significantly promotes CO2 exchange between the lake water and the atmosphere, thereby inducing a large bias in the ages of carbon species towards younger values in areas dominated by Chara spp. The 14C age of DIC-C1 is almost identical to that of DIC-C2, although the former was affected by the input of much older exogenous carbon by the inflowing groundwater (Table 1). The lake water in areas dominated by Chara spp. at the time of sampling of DIC-C1 has a higher concentration of DO Chara (165%) and a lower concentration of DIC Chara (9 mmol L–1), compared to DIC-C2 (135% and 13 mmol L–1, respectively), reflecting a higher photosynthesis rate. The higher photosynthesis rate of Chara spp. may have largely compensated for the influence of the input of exogenous old carbon and maintained the overall young age of DIC-C1. The difference in the ages of lake water and macrophyte remains in Genggahai Lake suggests that photosynthesis of aquatic plants plays an important role in determining changes in the reservoir effect of lake water.
CONCLUSIONS
In Genggahai Lake, the 14C ages of living plants, DICLW and MR in the uppermost sediments are much younger than those of groundwater. This can be ascribed to the exchange CO2 between the lake water and the atmosphere, in addition to the absence of stratification and the presence of a strong wind regime. The 14C ages of the lake-water DIC and macrophyte remains in the uppermost core sediments vary from site to site within the lake, which may be the result of differences between the photosynthesis rates of Chara spp. and vascular plants. Our results suggest that the photosynthesis intensity of aquatic plants, or the occurrence of succession within the aquatic community, should be considered when assessing the reservoir effects of a lake system, especially for lakes with a dense growth of aquatic plants.
ACKNOWLEDGMENTS
We thank the editor Prof. AJT Jull and two anonymous reviewers for their constructive comments and suggestions to improve the early version of this paper. We also thank Dr. J Bloemendal for his helpful comments and language improvements. This research was supported by the National Key R&D Program of China (Grant NO. 2017YFA0603402) and the National Science Foundation of China (Grant Nos. 41671190 and 41271219).
