Nine AMS 14C ages are listed in Table 1. Calibrated 14C ages were calculated using Calib 7.0.2 with INTCAL13 model (Reimer et al., 2013). All calibrated ages are in range of 33–40 cal ka BP. OSL ages have been shown in Table 2. The OSL decay curves and regenerative-dose growth curves fitted by sums of exponential and linear functions are presented in Fig. 3. AMS 14C ages are obviously younger than OSL ages of sediments in the XT Section (Fig. 4). Due to the lack of plant remains and other more suitable radiocarbon dating material, bulk organic matter of the sediments was used for radiocarbon dating. 14C ages except the one from the section base are nearly in accordance with the stratigraphic order from top to the bottom (Fig. 4). However, two OSL ages of the sediments in XT Section are 57.9 ka (depth at 40 cm) and 69.1 ka (depth at 190 cm), showing 25–30 ka older than 14C ages from the same stratum.
Sample Depth (cm) Material δ13C (‰) 14C age (a BP) 14C calibrated age (2σ) (cal a BP)* XT0 0 TOC -22.7 29 050±25 33 289±274 (33 015–33 563) XT10 20 TOC -21.7 31 170±100 35 039±304 (34 735–35 343) XT30 60 TOC -22.0 29 100±80 33 329±312 (33 017–33 640) XT50 100 TOC -21.9 29 590±90 33 775±209 (33 566–33 984) XT65 130 TOC -22.6 32 980±110 37 065±550 (36 515–37 615) XT95 190 TOC -21.4 32 660±100 36 568±330 (36 238–36 898) XT105 210 TOC -23.3 34 050±45 38 556±202 (38 354–38 758) XT115 230 TOC -23.3 35 170±120 39 734±393 (39 341–40 127) XT120 240 TOC -23.4 30 110±90 34 149±258 (33 891–34 407) *Calibrated by software Calib 7.0.2 (Reimer et al., 2013).
Table 1. AMS 14C ages of the sediments in the XT Section
Sample Depth (cm) K (wt.%) Th (ppm) U (ppm) Water content (%) Dose rate (Gy/ka) De (Gy) OSL age (ka) XT1 40 1.55±0.06 6.35±0.37 6.92±0.33 10±5 3.55±0.27 205.92±12.99 57.9±5.7 XT6 190 1.63±0.07 8.11±0.50 3.03±0.26 10±5 2.83±0.22 195.42±7.50 69.1±6.0
Table 2. OSL ages of the sediments in the XT Section (quartz 90–125 μm)
Figure 3. OSL decay curves and regenerative-dose growth curves for samples XT1 (a), (b) and XT6 (c), (d). Lx. Regeneration dose; Tx. test dose.
The difference between radiocarbon dating and OSL dating is also found in other sites in the Qaidam Basin. The 14C ages (bulk organic matter used as the dating material as well), and OSL ages are typically different in the Shell Bar (36.5140°N, 96.2021°E) that is located in the southeastern part of the Qarhan playa. The uncalibrated 14C ages (TOC) are 22.11 ka BP (at depth 53 cm) and 30.51 ka BP (at depth 152 cm) (Zhang et al., 2008), whereas OSL ages (38–63 μm quartz as the dating mineral) are 99 ka (at depth 144 cm) and 113 ka (at depth 178 cm) (Lai et al., 2014). The younger 14C ages of the sediments in the ISL1A core (Fan et al., 2013; Long, 2011) are inconsistent with the older OSL ages (Long, 2011).
Fan et al. (2013) postulated that the organic carbon of the sediments in the ISL1A core may be contaminated by ground water that comes from the eastern Kunlun Mountains and surface water that is supplied to deep brines during the exploitation of deep brines for fertilizer. Long (2011) and Long and Shen (2015) concluded that the underestimated 14C ages for the ISL1A core are caused by contamination with modern carbon after deposition or during sampling and preparation for dating. The postulated contamination of the organic carbon in lacustrine sediments in the Qaidam Basin indicates that the source of organic carbon may be very complicated.
Based on comparative study between 14C ages and OSL ages for the Shell Bar in the Qaidam Basin, Lai et al. (2014) proposed that the 14C dates on the Shell Bar are severely underestimated by sample contamination during burial, diagenetic alteration, and even sample mixing during laboratory processing. Impact of 2% contamination with modern carbon for a 60-ka-old sample could yield an age estimate of about 30 ka BP (Pigati et al., 2007). Most OSL ages of paleoshorelines in northwestern China fall between 70–120 ka, whereas 14C ages for the same shorelines concentrate between 30 and 40 ka BP (Lai et al., 2014). The above differences between OSL ages and 14C ages indicate that the 14C ages are probably severe underestimates. Therefore, 14C ages older than 25 ka BP of lacustrine sediments in arid regions should be re-investigated (Lai et al., 2014).
Because bulk organic matter may be contaminated by modern carbon from the ground water that comes from the eastern Kunlun Mountains and surface water that is usually supplied to deep brines during the exploitation of the brines for fertilizer in the Xitaijinair Salt Lake region, 14C ages of sediments in the XT Section are probably underestimates. The result of this study supports the conclusion that the 14C ages larger than ~30 ka BP are severe underestimates in the Qarhan Salt Lake (Long and Shen, 2015; Lai et al., 2014; Fan et al., 2013; Long, 2011).
OSL dating method as a new approach to the salt lake deposits has been applied to date strata containing salts in the SG-1 core, western Qaidam Basin, and the reliability of OSL ages are confirmed by U-series dating results (Han et al., 2013). However, whether the OSL ages in sediments of the XT Section are more reliable than 14C ages in this section is unclear, because the bleaching of quartz during deposition and unstability of the dose rate in these sediments containing salts would affect the accuracy of the OSL ages.
Assuming the sedimentation rate is constant and the preliminary OSL ages are more reliable than 14C ages, the surface age of the XT Section is about 55 ka. This age indicates that the XT Section experienced erosion during the Late Pleistocene, in agreement with other studies (Lai et al., 2014; Han et al., 2013).
Nearly 80% of the surface area has been covered by playas in the Qaidam Basin (Fig. 1; Han et al., 2013). The U-series dating results of the Dafengshan (DFS) pit section, and the U-series dating and OSL dating results of the SG-1 core show that the surface of the playas has an age of about 100 ka (Han et al., 2013). U-series dating of gypsum from the D26 Section at Dalangtan playa plain placed the playa plain formation at about 100 ka (Ma et al., 2011). The extrapolation of magnetostratigrapy dating results of the sediments in the Liang ZK05 core also indicate the surface age at Dalangtan playa is about 111 ka (Shi et al., 2010). The OSL age of the surface sediment in the XT Section is about 50 ka younger than the above ages of the playa formation ages (Han et al., 2013; Ma et al., 2011; Shi et al., 2010). Both OSL and 14C ages of the XT Section indicate that the surface of the Xitaijinair region has been eroded. Additionally, some evidence shows that the lithologies of the yardangs in the western Qaidam are composed of Tertiary mudstone, siltstone and sandstone (Niu et al., 2011; Halimov and Fezer, 1989). Large areas of the yardangs distributed in the western Qadiam Basin imply that wind erosion is the main transport mode in the Qaidam Basin. Thus, the erosion of the surface of the Xitaijinair Salt Lake region was probably caused by wind erosion. Erosion of the Qaidam Basin has been concluded by Bowler et al. (1987), Kapp et al. (2011), Han et al. (2013) and Lai et al. (2014). Lai et al. (2014) re-interpreted the Shell Bar as a remnant of a river channel and proposed that the wind erosion has left the fluvial channel sediments topographically inverted. An et al. (2012) proposed that the possibility that the Qaidam Basin can contribute dust from the basin to the Chinese Loess Plateau, whereas Sun (2002) suggested that it was unlikely that the Qaidam Basin could be the dust source of the Chinese Loess Plateau. The chronology of the XT Section, which is close to the Shell Bar, also shows the erosive nature of the landscape during the Late Pleistocene.
The mineral compositions of the sediments in the XT Section are summarized in Fig. 5. The sediments in the whole XT Section are composed of quartz (volume%, 11%–44%, average 32.76%), muscovite (20%–40%, average 30.97%), clinochlore (4%–29%, average 11.25%), albite (0%–23%, average 4.84%), calcite (3%–13%, average 7.62%), dolomite (0%–6%, average 1.68%), ankerite (0%–4%, average 0.32%, occurred at depth 136–168 cm and 204 cm), halite (0%–15%, average 5.14%), gypsum (0%–29%, average 5.00%), and also aragonite, magnesite, bromargyrite and hexahydroborite.
Figure 5. Mineral compositions of the sediments in the XT Section (the shaded boxes show the coarse sediments).
Silicates (quartz, muscovite, clinochlore and albite) and carbonates (calcite, dolomite, ankerite, aragonite and magnesite) are in ranges of 60%–92% and 4%–18%, respectively (Fig. 5). The contents of muscovite and clinochlore are relatively lower in the coarse sediments (depth at 25–47 and 185–199 cm) than those in the other part of the section. However, the content of albite is relatively higher in the coarse part of the section (Fig. 5).
Feldspars are thought to be the most abundant of the labile minerals (Nesbitt and Young, 1982). As one of three endmembers of the feldspars, the albite (NaAlSi3O8) can be transported by rivers after the weathering and leaching of the rocks. The albite of the sediments in the XT Section is affected by distribution and characteristics of the source rocks, and climatic condition. The mountain ranges around Qaidam Basin are mainly composed of igneous and metamorphic rocks (Xu et al., 2006) which provide large amounts of silicates for the basin. The high content of the albite in the coarse sediments of the XT Section may be produced by suitable climatic condition having relatively more precipitation in the catchment of the Xitaijinair region. Meanwhile, more precipitation would strengthen the activity of the river transportation, so the coarse sediments were carried to the Xitaijinair Lake Basin. In sum, the mineral compositions of the sediments in the XT Section are probably related to past climatic conditions, and the albite is an especially sensitive and indicative mineral for past climate change.
Generally, for closed lakes on the QTP, exogenous carbonate delivery to the lake is relative weak during cold periods (Li and Kang, 2007). Interestingly, the coarse sediments in the XT Section contain more dolomite than the fine ones (Fig. 5). This feature may indicate that more dolomite enters the Xitaijinair Salt Lake during relative warm periods.