Oscillation of mineral compositions in Core SG-1b, western Qaidam Basin, NE Tibetan Plateau

Uplift of the Tibetan Plateau since the Late Miocene has greatly affected the nature of sediments deposited in the Qaidam Basin. However, due to the scarcity of continuously dated sediment records, we know little about how minerals responded to this uplift. In order to understand this response, we here present results from the high-resolution mineral profile from a borehole (7.3–1.6 Ma) in the Basin, which shows systematic oscillations of various evaporite and clay minerals that can be linked to the variation of regional climate and tectonic history. In particular, x-ray diffraction (XRD) analyses show that carbonate minerals consist mainly of calcite and aragonite, with minor ankerite and dolomite. Evaporates consist of gypsum, celesite and halite. Clay minerals are principally Fe-Mg illite, mixed layers of illite/smectite and chlorite, with minor kaolinite and smectite. Following implications can be drawn from the oscillations of these minerals phases: (a) the paleolake was brackish with high salinity after 7.3 Ma, while an abrupt change in the chemical composition of paleolake water (e.g. Mg/Ca ratio, SO42− concentration, salinity) occurred at 3.3 Ma; (b) the three changes at ~6.0 Ma, 4.5–4.1 Ma and 3.3 Ma were in response to rapid erosions/uplift of the basin; (c) pore water or fluid was Fe/Mg-rich in 7.3–6.0 Ma, Mg-rich in 6.0–4.5 Ma, and K-rich in 4.1–1.6 Ma; and (d) evaporation rates were high, but weaker than today’s.

sequence and its location in the transitional zone between the arid Asian interior and the East Asian Monsoon regions not only make it an ideal place to study the high-resolution records of long-term climate and erosion histories 16 , but also provide important information on tectonic deformation processes and the potential linkage of tectonic uplift to climate change.
As part of a joint Sino-German project, a 723m-deep core, named SG-1b, was recovered from the top of the Jianshan Anticline (38°21′ 9.46″ N, 92°16′ 24.72″ E, Fig. 1) with an average recovery rate of ~93%. Paleomagnetic dating of the core yielded an age range of ~7.3-1.6 Ma, with a sediment accumulation rate (SAR) of 6.5 to 30.4 cm/ka 16 . According to Lu 17 , the lower section of the core (723-245 m, ca. 7.3-3.6 Ma) is characterized by gray to dark-gray fine siltstones with striking horizontal millimeter-scale laminae; the middle section (245-200 m, ca. 3.6-3.3 Ma) is characterized by massive bedding and occasionally weak laminations accompanied by coarse siltstones and thin bands of sandstones; and the upper section (200-0 m, ca. 3.3-1.6 Ma) contains abundant millimeter-to centimeter-scale interbedded sandstones and scattered gypsum crystals. Ooid carbonates begin to occur at 81 m (2.6 Ma) and occur more frequently upwards.
Here we present a 6-Ma (i.e. 7.3-1.6 Ma) long high-resolution profile of the mineral compositions (including evaporates, carbonates and clay minerals) recorded in Core SG-1b. We show that the Core's minerals have indeed documented important climate events in the region's paleoenvironmental history.
Calcite is the major carbonate phase present throughout the profile. The crystallinity of calcite was recorded using the full-width half-maximum (FWHM) function, and XRD of 3.03 Å (104) gave values ranging from 0.165 to 0.306. Small amounts of Mg are often integrated into the calcite lattice. The MgCO 3 content of calcite in the profile ranges from 0 to 3.32% (Fig. 2), suggesting that these calcites are Mg-poor. Most halite appears from 585 m (6.6 Ma) upward, with content ranging between 1-10%, gradually increasing upward (Fig. 2). Halite crystallinity (HC), represented by the integral breadth of the 2.82 Å peak, ranges from 0.22 to 0.39, with a mean value of 0.27 (Δ °2θ , Fig. 2).
XRD analyses suggest mixed-layer illite/smectite minerals (a thermodynamically metastable intermediate product between smectite and illite) 22 in the Core exhibit disorder (R 0 I/S) and order (R 3 I/S) (Fig. 3). Because there is no distinct boundary in the XRD spectrum between smectite and I/S, or between I/S and illite, most R 3 I/S identified by XRD could also be R 1 I/S 9 . Moreover, the R 1 I/S is more stable than any other mixed layer I/S phase because it has a unique structure, composition, and ordering pattern 9,23,24 . Thus, any clay mineral identified as R 0 I/S by XRD is more likely to be a mixture of discrete smectite and R 1 I/S, i.e. mixed-layer I/S with R > 1 is a mixture of R1 I/S and illite 9 .

Discussion
Origins. Carbonate minerals. The precipitation of carbonate minerals in typical modern lakes follows a low-Mg calcite → high-Mg calcite → aragonite sequence, accompanied by an increased Mg/Ca ratio in lake water. Most aragonite and some calcite should have been directly deposited by the paleolake water, e.g. when the salinity of a solution reached ~1.8 times the concentration of modern seawater 2 . However, in the Core, some calcite and trace amounts of dolomite are not authigenic, because suspended carbonates were found to contain calcite (5-21%) and trace amounts of dolomite (SI Table), presumably originating from the physical weathering of carbonate rocks and older carbonate sediments in the source region.
Dolomite can form as a primary precipitate, a diagenetic replacement, or out of hydrothermal fluid. With high permeability, a high fluid flow rate is possible. This facilitates the supply of Mg, essential to the formation of dolomite 3 . Bacterial metabolism may aid dolomite precipitation in settings where sulfate-reducing bacteria flourish 3 . The chemical composition of dolomite can be further modified during burial and metamorphism, in response to changing environmental parameters, including temperature, pressure, burial history, chemistry of pore fluid, and co-precipitating mineral phases. Because of its complex origins and formation conditions, it is difficult to identify climatic information based on dolomite alone. The formation of ankerite (Ca(Fe, Mg, Mn) (CO 3 ) 2 ) has been related to the Fe-Mg rich interstitial water/pore water present during the transformation of clay minerals. The varying abundance of ankerite can therefore reflect the variable availabilities of Fe, Mg and Ca in fluids (interstitial water/pore water).
Sulfates and halite. When the salinity of a solution reaches approximately five times the concentration of modern seawater, gypsum precipitates from brine 2 . The Sr in the celestite could have originated from three possible sources: 1) the weathering of rocks; 2) Sr-rich hydrothermal fluids; and 3) pore fluids. We may assume this because a celestite deposit which has been proposed to be of hydrothermal origin 25 occurs near the Core, in the QB (Fig. 1). During the process of calcite dissolution-reprecipitation, there is a net input of strontium into pore fluids that are oversaturated by celestite 26 . As one of the most common evaporative minerals, halite crystals should directly precipitate from brine and pore water when its salinity reaches 11 or 12 times that of modern seawater 2 .
Clay minerals. Generally, clay minerals in lake sediments are lithogenous as by-products of the physical and chemical weathering of rocks, e.g. chlorite and illite form under severe physical weathering conditions, and kaolinite and smectite form when chemical weathering conditions predominate. However, significant quantities of kaolinite and smectite appear in the modern sediments of Lake Nella and soils in Antarctica, a very cold region 27,28 , which would suggest that the presence of kaolinite and smectite is not always indicative of a warm climate and/or chemical weathering conditions. A weak correlation between abundance of kaolinite and smectite (R 2 = 0.23, Fig. 4a) also suggests their origins to be different. Despite that the sediments deposited in the Jianshan Anticline area were mainly transported there by fluvial process and wind action from the Altyn Mountains to the northwest 17 , most of the clay minerals in the Core would have then undergone thermal metamorphism, with a few that have experienced diagenesis (Fig. 3). Therefore, there could be other origins, such as an authigenic precipitation, and a transformation between the clay minerals. The formation of kaolinite possibly depends on the contents of Mg and Ca, while the formation of smectite relies on the Si, Ca, Mg and Fe contents in the lake/pore water 28 . The significant correlation between illite and chlorite (R 2 = 0.62, Fig. 4b), and between illite and I/S (R 2 = − 0.88, Fig. 4c), would suggest that most illite was transformed from I/S, and some from physical weathering with chlorite. Similarly, chlorite can also be transformed from kaolinite as well as from weathering processes. In the Core, there are likely to be a couple of explanations for this: (1) there was a significant relation between chlorite and kaolinite (R 2 = 0.56) (Fig. 4d); and (2) kaolinite content decreases as burial depth increases. The formation of authigenic chlorite is related to the presence of Fe-Mg rich interstitial water/pore water, while K-rich water favors the formation of illite and I/S mixed minerals 29,30 . The ions required for the formation of these minerals were perhaps derived from the breakdown of detrital Fe-Mg minerals and lacustrine sediments during deposition or early burial 5,31 . Despite that original clay mineral assemblages resulting from weathering have also been preserved in the Core 31 , it was difficult to quantify the proportion of original clays.
On the other hand, temperature is an important control of the crystallinity and chemistry of illite and chlorite, the formation of I/S mixed minerals, and the switch from disorder I/S to order I/S 5-9,32-34 . For example, low IC is preserved in the cold climate sections. IC in the anchizone of the Core indicates that the low-grade thermal metamorphic temperature ranged from 280-360 °C 19,21 (Fig. 3). The S-I reaction is an abiotic dissolution re-crystallization process that typically requires temperature conditions of 300-350 °C 9 . The transformation temperature from disorder R 0 I/S to order R 1 I/S is ~100-110 °C 32,33 , and R 1 I/S is stable between ~150 °C and ~225 °C 34 . The Fe-Mg contents of illite and chlorite will increase with increasing temperature 6 . Mg-chlorite occurs at temperatures of > 100 °C 5,7 . Implications for environmental changes. The  The presence of gypsum and celestite indicates that the paleolake water was brackish, with high concentrations of SO 4 2− after 7.3 Ma. The gradual increase in halite content after 6.6 Ma suggests an increasing salinity in the paleolake water, probably caused by a dry regional climate and high rates of evaporation (Fig. 3). Possible dry-cold events happened at 6.6 Ma and 2.9 Ma as the lower IC and variability of evaporates (Fig. 3). Grain sizes and lithofacies in the Core SG-1b and fossils in QB also suggest that the climate was dry but much wetter than the Quaternary during 7.3-3.6 Ma 17 . Climatic proxies suggest that rapid drying began principally from 8 Ma in the eastern QB (Huaitoutala section, Fig. 1) 35 , and the climate had already been dry after 5.3 Ma in the central QB (Yahu Section, Fig. 1) 36 with a drought event at 3.6 Ma in the western and central QB (Core SG-1b and Yahu Section, Fig. 1) 17,36 and a long-term stepwise drying trend in the western QB since 3.1 Ma (Core SG-1, SG-3 Fig. 1) 37,38 . The widespread distribution of evaporative minerals in the western QB also suggested drought climate since late Pliocene (SI Figure) 45 , the category of I/S mixed minerals turns from disorder to order (Fig. 3), and the value of (illite + chlorite)/(kaolinite + smectite) decreases dramatically. These changes indicate that the chemistry of the fluid or pore water was different between 4.1-1.6 Ma from between 7.3-4.5 Ma (Fig. 3). At ~3.3 Ma, the significant change in the relative abundance among calcite, aragonite, and gypsum, and a weak increase in halite abundance signify an abrupt change in the chemical composition of the paleolake water (e.g. the Mg/Ca ratio, SO 4 2− concentration, salinity). This agrees with the variations in granulometric parameters and lithofacies features, which suggest that the paleolake was shallow at ~3.3 Ma as a result of the rapid and continuous uplift of the Jianshan Anticline 17 .
The three observed changes at ~6.0 Ma, ~4.5-4.1 Ma and ~3.3 Ma are likely to be related with the rapid erosion and high SAR values noted in Core SG-1b at > 7.3-6.0 Ma, 5.2-4.2 Ma and 3.6-2.6 Ma (Fig. 3). These high SAR values can be linked directly to the tectonic evolution of the Qaidam Basin 16 . The tectonic deformation and uplift event at 8~6 Ma has been reported from the QB and from various parts of the TP [45][46][47] . The main and intense uplift of the northwestern QB happened at 5.3-4.5 Ma 45,47 , and rapid uplift happened at ~3.6 Ma 12 . Tectonic activity not only triggered a drought climate, but also resulted in the transportation of materials into the Basin 35,48 , a great inflow of hydrothermal fluid, and therefore high SAR rates. The similar variation of SAR among the cores in the western QB (SI Figure) also provided an evidence for the tectonic activities. Therefore, climatic change and uplift history of QB controlled the depositional environment and oscillations of minerals in the Core SG-1b.

Materials and Methods
Mineral compositions in the Core were examined at 1 m intervals using an X-ray diffractometer (XRD) (a Rigaku D/MAX-2000: Cu, Ka1, 1.5406 Ǻ, 40 kV, 40 mA). A subsection of these samples were treated with ultrapure water at 2 m intervals, then diluted with HCl and H 2 O 2 to remove dissolvable ions, calcium carbonate and organic matter; the clay mineral fraction (< 2 μm) was then separated in a centrifuge. For each sample, this fraction was then scanned using XRD, to include the sequential collection of three XRD patterns, i.e. under natural (air-dried) conditions (N), after saturation with ethylene-glycol for 24 hrs in desiccators (EG), and after being heated in a muffle furnace at 490 °C for 2 hrs (H). Diffraction patterns (2θ ) were scanned from 3° to 30°, with a step size of 0.07°. Minerals were identified by peak area. The measurements were conducted at the Micro Structure Analytical Laboratory (MSAL), Peking University, in compliance with Chinese oil and gas industry standard SY/T 5163-2010.
The MgCO 3 content in the calcite was deduced from the deviation of the XRD (104) using the following calibration 18  Si (111)