The Qaidam Basin is located in the northeast corner of the Qinghai-Tibet Plateau and has an irregular rhomboid shape. It has a length of 850 km from east to west, a width of 150–300 km from north to south, and an area of about 12.1 × 10⁴ km². Tectonically, the basin belongs to the northeast part of the Tethyan domain, with the Qilian orogenic belt to the north, the East Kunlun orogenic belt to the south, and the Altyn Tagh orogenic belt to the west. These orogenic belts have undergone long-term and complicated evolution and formed the complex tectonic-magmatic- hydrothermal framework of the surrounding mountain system (Figure 1; Dai and Cao, 2000; Tang et al., 2000; Tan et al., 2012).

During the Cambrian-Devonian period, the Qaidam Basin was a relatively independent Terrene (Tang et al., 2000). After the Indosinian tectonic event, the Qaidam terrene and its surrounding blocks gradually merged and transitioned to the tectonic stage of intra-continental basin development (Tang et al., 2000). Since the Cenozoic, the basin has widely received sediments and formed a set of overlying strata with a thickness of about 6,000–7,000 m and up to 10,000 m in certain areas (Zhao et al., 2013). Because of the influence of the “far-field effect” after the collision between the Indian plate and the Eurasian plate since the Cenozoic, the fault system at the margin of the basin was reactivated. Further, the Altyn Tagh fault to the western margin of the Qaidam Basin also began to be reactivated, and a series of NW-SE trending anticline and syncline structures developed in the west of the Qaidam Basin (Yin et al., 2007). These anticline and syncline structures can be identified, as the anticline structures are narrow, while the syncline structures are broad. This reflects the typical structural shape of the detachment system, indicating the existence of detachment in the deep part of the basin (Yin et al., 2007). These anticline structures in the western basin are rich in Ca-Cl brine resources (Li et al., 2015). Moreover, the fault system at the margin of the basin and part of the inner basement were also reactivated, and four types of fault systems were developed: the northern margin fault system of the basin, the northern margin fault system of the East Kunlun Mountain, the Altyn Tagh fault system, and the intra-basin fault system. Together, these faults controlled the Mesozoic and Cenozoic sedimentation of the basin (Luo and Pang, 2003).

The depositional strata since the Cenozoic can be divided into seven units from old to new: the Paleocene-Eocene Lulehe Formation (E₁₊₂₁), the Oligocene Lower Ganchaigou Formation (E₃g), the Miocene Upper Ganchaigou Formation (N₁g), the Lower Youshashan Formation (N₁y), the Pliocene Upper Youshashan Formation (N₂g), the Shizigou Formation (N₂S), and the Quaternary Qigequan Formation (Qp¹⁺²) (Table 1). In recent years, with the accumulation of chronology data, especially high-precision paleomagnetic data, accurate ages of basin strata and chronology frameworks have been established (Table 1; Sun et al., 2004; Nie et al., 2019).

After the formation of the basin, widely distributed Cenozoic strata have occurred in Qaidam Basin. From the perspective of sedimentary sequences in the western Qaidam Basin, at the end of the Paleogene period, the Qaidam Basin has entered a stage of subsidence and the paleontological sedimentary environment and experienced processes of water body development from shallow to deep (Su et al., 2014). This means that from the Lulehe Formation to Xiaganchaigou Formation, the sedimentary facies are mainly developed from coastal-shallow to semi-deep lake subfacies (Su et al., 2014). The Neogene strata are mainly developed as lake facies, and the lithology is mainly alluvial rock, gray mudstone, and silty mudstone (Zeng et al., 2021). Shallow coastal lake facies primarily dominate the Quaternary strata, and semi-deep lake facies appeared in the middle part during the late stage (Dang et al., 2004).

According to drilling core data, a large number of evaporite deposits, represented by carbonate, sulfate, and chloride, have been developed in the western basin since the Eocene (Zeng et al., 2021). Among them, the oldest evaporate deposits are developed in the late Eocene Upper Xiaganchaigou Formation in the Shizigou area (Guo et al., 2018). These evaporate deposits can provide an abundant source for Ca-Cl brine during their migration. Moreover, very thick piedmont pluvial deposits occur at the end of the Paleogene, with the thickest strata exceeding 2,600 m (Zhang et al., 1987). Thus, the Ca-Cl brines (oil-field Ca-Cl brines) are produced in Paleocene–Neogene strata. These Ca-Cl brines have been found in many tectonically controlled units (e.g., anticlines, fractures, faults) and have high total dissolved salts (TDS) that coexist with petroleum in clastic rocks (Li et al., 2015). Commonly, the Ca-Cl brines in Paleocene formations are >1,000 min-depth (ranging from 1,000 to 1,500 m) and Neogene formation Ca-Cl brines are <1,000 m in-depth (ranging from 500 to 1,000 m). All Ca-Cl brines from petroleum wells have very high pressure resulting in constant self-eruption for several decades, and they have temperatures close to or exceeding 100°C and TDS >100 g/L. Minerals salts precipitate on the surface once the Ca-Cl brines discharge from production tubing.

Most of the samples used for this study were collected from oilfield wells. A small number of samples (i.e., JS1, L2, E2, HZK01and YC2) from the middle of the basin were collected from Ca-Cl brine exploration wells (Figure 1). A final volume of each sample of 500 ml was ensured after measuring the specific gravity on site and sealing the samples for laboratory analysis.

Ninety-five samples from Nanyishan, Shizigou and Dafengshan areas as well as the central part of the basin were selected for hydrochemical characteristics analysis (Figure 1; sampling location). These samples basically cover the whole western region of the Qaidam Basin and are representative of the basic characteristics of the Ca-Cl brine in this basin. Considering the differences in the origin of different types of water bodies, only Ca-Cl water bodies in the study area were retained in this study. Firstly, the water types were defined according to the principle of published by Hardie and Eugster (1970) to eliminate non-Ca-Cl type hydrochemical data. Further, 63 sets of published hydrochemistry data of Ca-Cl brines from Nanyishan, Shizigou, Youdunzi, Youshashan, Xiaoliangshan and Lenghu areas were also compiled (Figure 1; compiled location; Liu et al., 2019; Tan et al., 2007; Li, 2007; Li, 2013; Li et al., 2015). Together, these hydrochemical data can reflect the hydrochemical characteristics of the Ca-Cl brine in the whole western Qaidam Basin; thirteen Ca-Cl brine samples from Shizigou (S38, S41H1-1-411, S41H2-1-411, S41H1-2-413, S41H1-1-511 and S49H1-3-512), Nanyishan (N6 and N13) and central basin (JS1, L2, E2, HZK01 and YC2) were subjected to B isotope analysis, these samples were distributed throughout the western basin and have a good representativeness. Considering the latest published Li-isotope data of Ca-Cl type brine in the western basin and the lack of related research of Ca-Cl brine in the central part of the basin, only five Ca-Cl type samples from the central basin were selected for analysis (Figure 1; JS1, L2, E2, HZK01 and YC2). The above data were integrated to explore the origin of Ca-Cl brine in the western Qaidam Basin.

The major element data are shown in Supplementary Appendix SA. Ca-Cl brine has high ion concentrations of Cl⁻, Na⁺, Ca²⁺, SO₄ ²⁻, Mg²⁺and HCO₃ ⁻. The K⁺, B₂O₃, and Li+ contents in this Ca-Cl brine are relatively high and have potential development value.

To clarify the relationship between the different major elements, the Pearson correlation coefficients between the 12 major elements were calculated (Table 2).

The correlation coefficients (Table 2) between Na⁺ and Cl⁻, TDS and Na⁺, as well as TDS and Cl⁻ in Ca-Cl brine are 0.96, 0.94, and 0.93, respectively. These high coefficients not only imply that the dissolution of NaCl provides the main source of Na⁺ and Cl⁻ in Ca-Cl brines, but also that these two elements are the main ionic components causing the high TDS in Ca-Cl brines. The correlation coefficients between K⁺ and Na⁺, K and Ca²⁺, K⁺ and Cl⁻, K⁺ and B₂O₃, and K⁺ and Li⁺ in Ca-Cl brines are 0.54, 0.81, 0.70, 0.53, and 0.78, respectively. In combination with the high correlations between Na⁺ and Cl⁻, TDS and Na⁺, as well as TDS and Cl-, these correlation coefficients suggest that the K⁺ in Ca-Cl brines may originate from both the dissolution of existing evaporites in the basin and other sources. The correlation coefficients between Ca²⁺ and Li⁺, as well as Ca²⁺ and Sr²⁺ in Ca-Cl brines are 0.90 and 0.77, respectively. To a certain extent, this result suggests that the sources of Li⁺ and Sr²⁺ in Ca-Cl brines are most likely originate from the same source area and are related to minerals or rocks with high Ca²⁺ levels. The correlation coefficients between Mg²⁺ and SO₄ ²⁻ as well as Mg²⁺ and HCO₃ ⁻ in Ca-Cl brines are 0.62 and 0.49, respectively, and the correlations between Ca²⁺ and SO₄ ²⁻, as well as Ca²⁺ and HCO₃ ⁻ are low. This result may be related to the incomplete dissolution of dolomite during the migration of Ca-Cl brine. In addition, the correlation between B₂O₃ and Li⁺ in Ca-Cl brines is 0.51. Some of the B³⁺ and Li⁺ elements in Ca-Cl brines may originate from bedrock weathering or water-rock interaction under hydrothermal conditions. Further, B₂O₃ correlates well with TDS, B₂O₃ correlates well with K⁺, B₂O₃ correlates well with Na⁺, and B₂O₃ correlates well with Cl⁻, indicating that the B₂O₃ in the Ca-Cl brine is partially derived from recharge after redissolution with existing evaporites in the basin. The Li⁺ in brines also shows similar characteristics with the boron. In fact, the moderate correlation between Cl⁻ and Ca²⁺, Ca²⁺ and Na⁺ also supports the above inference. Together, these modes comprise the sources of B3⁺ and Li⁺ elements in the Ca-Cl brine.

The δ⁷Li values of samples in the Qaidam Basin display significant variation, ranging from 10.48‰ to 15.97‰ (Table 3); this variation implies that the source of Li in Ca-Cl brine at the sampling site is relatively consistent. The corresponding Li content in the sample was relatively high.

The B isotopes of Ca-Cl brines in the western Qaidam Basin were also analyzed and the results are presented in Table 4. The variation of the B isotope composition in this region ranged from −5.68‰ to 26.69‰, with an average of 6.10‰, which is much lower than the value of seawater (39.5‰).

According to the results of ion correlation analysis, the major elements (especially for Na and Cl) in Ca-Cl brines clearly originate from the dissolution of NaCl. If this conclusion is true, the process inevitably also dissolved other relatively soluble minerals during the Ca-Cl brine formation. The correlation between Mg²⁺, Ca²⁺ SO₄ ²⁻ and HCO₃ ⁻ indicate incomplete dissolution of dolomite during Ca-Cl brine migration. When dolomite is completely dissolved, both Ca²⁺ and Mg²⁺ are completely released; however, when dissolution is incomplete, part of the Mg²⁺ is released earlier, and the remaining carbonate is precipitated as CaCO₃. This preliminary dissolution process tends to result in a certain correlation between Mg²⁺ and CO₃ ²⁻ and a lack of correlation between Ca²⁺ and CO₃ ²⁻ (Eqs 2, 3). In addition, considering the correlations between K⁺, Na⁺, Ca²⁺, Cl⁻, B₂O₃ and Li⁺ as well as the correlations between Na⁺, Cl⁻ and TDS, it is reasonable to speculated that the K⁺ in Ca-Cl brines may originate from both the dissolution of existing evaporites in the basin and other sources. Such dissolution can occur because of weathering or hydrothermal water-rock interaction (Li, 2013; Miao et al., 2022). Although similar geochemical mobility of ions also results in correlation between data, the supply of these ions by the source region is an indispensable and important factor. This conclusion is also supported by the observed correlation between Ca²⁺, Li⁺ and Sr²⁺. This correlation shows that the sources of Li⁺ and Sr²⁺ in Ca-Cl brines most likely originate from the same source area related to minerals or rocks with high Ca²⁺ levels. Furthermore, the correlation between B₂O₃ and Li⁺ in Ca-Cl brines is moderate. B³⁺ and Li⁺ elements have many similar geochemical properties and both tend to be enriched in bedrock in acidic granite bodies and in weathered sedimentary rocks. It appears that some of the B³⁺ and Li⁺ elements in Ca-Cl brines may originate from bedrock weathering or water-rock interaction under hydrothermal conditions (He et al., 2020; Miao et al., 2022). Also, B₂O₃ and K, B₂O₃ and Na as well as B₂O₃ and Cl⁻ correlates much better (0.53, 0.73 and 0.75, respectively). This result indicates that the B₂O₃ in the Ca-Cl brine is partially derived from recharge after redissolution with existing evaporites in the basin. The moderate correlation, such as between Cl⁻ and Ca²⁺, Ca²⁺ and Na⁺, also supports the above conclusion. CaMg CO 3   2 + 2 H 2 CO 3   → Ca 2+ + Mg 2+ + 2 HCO 3 ‐ (2) CaMg CO 3   2 + H 2 CO 3   → CaCO 3 + Mg 2+ + 2 HCO 3 ‐ (3)

In brief, the findings present a typical “binary structure” in terms of sources. The main elements in Ca-Cl brines are mainly recharged by re-soluble evaporites, and they are possibly recharged by weathering or hydrothermal water-rock reactions, too. These recharge modes may represent the main supply pattern of elements within the Ca-Cl brines of this basin.

The main ion concentration combinations are Cl⁻, Na⁺, K⁺, Ca²⁺, SO₄ ²⁻, Mg²⁺, and HCO₃ ⁻. Anions were dominated by Cl-with a content of up to 189.35 g/L, and cations were mainly Na⁺, with a content of up to 118.51 g/L; moreover, the linear relationship between the Na⁺ content and TDS was excellent, with a correlation coefficient of 0.92. This result suggests that the high TDS characteristic of Ca-Cl type brines is closely related to the evaporative concentration or dissolution of salt minerals. The correlation coefficient between Na⁺ and Cl⁻ is 0.93, indicating that Na⁺ and Cl⁻ in the Ca-Cl brine are mainly derived from the dissolution of halite within the basin (Figure 2).

The genesis of Ca-Cl brines in inland basins has been studied before. The hydrochemical composition of this water body is rich in Ca²⁺ and poor in Mg²⁺; moreover, it contains small amounts of SO₄ ²⁻ and HCO₃ ⁻, and almost has no CO3- (Zhang et al., 1987; Lowenstein et al., 1989; Hardie, 1996; Lowenstein, 2003; Li et al., 2015). Ca-Cl brine is a special type of water body. Its hydrochemical composition is difficult to explain only by weathering between surrounding rock and atmospheric precipitation (river water), evaporation, and concentration on the Earth’s surface. However, it plays a crucial role in the formation of large ancient marine, rift, and terrestrial basin (Ca-Cl brine) potash deposits (Lowenstein et al., 1989; Hardie, 1990; Lowenstein et al., 2001).

The Qaidam Basin in China is a typical terrestrial salt-bearing basin; its western Tertiary strata are rich in Ca-Cl brines. How did this type of brine form? For the genesis of Ca-Cl brines, different scholars provided different explanations. Examples are the diagenetic transformation effect (albitization, dolomitization, and reduction), evaporation and concentration of ancient residual seawater, and hydrothermal genesis. Essentially, Ca-Cl brine is a water-salt system with high Ca²⁺ and low Mg²⁺ contents, suggesting that in nature, Ca-Cl brine is relatively enriched in Ca²⁺. According to the chemical composition as well as H, O, Sr, S, and B isotope data, scholars proposed that the water bodies in the Ca-Cl brine in the Qaidam Basin show characteristics of atmospheric rainfall and are mixed with deep fluids in the inner Earth (Tan et al., 2011; Li, 2013; Li et al., 2021). The hydrogeological data and B-Li isotope analysis presented above show that the B and Li elements in Ca-Cl brine have clearly originated from water-rock interaction at the basin margin. However, the main share of Na⁺ and Cl⁻ in the Ca-Cl brine originates from the dissolution of existing evaporites in the basin during the migration of the water body. These findings clearly indicate that there are two origins of the relatively abundant Ca²⁺ in Ca-Cl brines in the western part of the Qaidam Basin: It can originate from the contribution of dissolution of the existing evaporites; It can also originate from the magmatic hydrothermal fluids contributed by the water-rock reaction under suitable temperature-pressure conditions in the deep part of the basin margin and the magmatic hydrothermal fluids in the orogenic belt of the basin margin during the orogenic period. In fact, the Ca²⁺ in Ca-Cl brine that originates exclusively from the dissolution of existing evaporites in the basin is often in balance with CO₃ ⁻ and SO₄ ²⁻ in the evaporites. Consequently, the formation of Ca-Cl brines is prevented, and additional non-equilibrium Ca²⁺ input is required to allow the Ca-Cl brine to evolve into Ca-Cl type water bodies. Therefore, Ca²⁺ present in magmatic hydrothermal fluids contributed by water-rock interaction under appropriate temperature-pressure conditions at the deep basin margin and by magmatism at the basin margin orogenic belt during the orogenic period may be an important mechanism for the formation of this particular type of water. The results of ion correlation analysis further support this conclusion as correlation coefficients are 0.90 and 0.77 between Ca²⁺ and Li⁺, as well as between Ca²⁺ and Sr²⁺, respectively.

In the history of Ca-Cl brine research, Carpenter (1978) was the first to suggest that Ca-Cl brines in the basin were originally derived from marine evaporite. Carpenter further suggested that these Ca-Cl brines underwent a series of water-rock interactions with the surrounding rocks at normal temperature-pressure conditions during their burial, thus, resulting in Ca-Cl brines. However, the results of computer simulations indicate that the chemical composition of Ca-Cl brines that formed under such conditions differs significantly from that of Ca-Cl brines (Hanor, 1994). Therefore, Lowenstein. (2003) and Bottomley et al. (2005) re-evaluated the genesis of Silurian-Devonian Ca-Cl brines in the Illinois Basin in the United States and the western basin of Canada. Their results suggest that the Ca-Cl brines in these areas mainly inherited the chemical composition of Silurian-Devonian paleo seawater from a Ca-Cl type sea (Lowenstein et al., 2001; Holt et al., 2014), rather than being led by later water-rock interactions and transformation of hydrochemical characteristics. These studies highlight the importance of the chemical composition of paleo seawater with high Ca²⁺ and low Mg²⁺ contents in the Ca-Cl sea for the formation of Ca-Cl brines in the basin (Hardie, 1996; Lowenstein et al., 2001; Lowenstein, 2003; Bottomley et al., 2005). However, the Qaidam Basin in China is a typical terrestrial basin without residual or recharge of paleo seawater during its formation (Zhang et al., 1987). This rules out the possibility that Ca-Cl brines in the Qaidam Basin were formed by the diagenetic transformation effect (i.e., albitization, dolomitization, and reduction) and the evaporation and concentration of ancient residual seawater.

The concept that Ca-Cl brines are the cause of hydrothermal fluids has been proposed by many scholars (Hardie, 1990; Lowenstein and Risacher, 2008; Tan et al., 2011; Lowenstein et al., 2016). Hardie (1990) showed that Ca-Cl hydrothermal fluids that are rich in K⁺, B³⁺, and Li⁺ are distributed at mid-ocean ridges or terrestrial tectonic activity zones. Such hydrothermal fluids are driven by thermal anomalies or topographic dynamics to rise along fracture zones and emerge at the surface to form thermal springs or salt Ca-Cl brines. Example are areas such as the Reykjanes Ridge in Iceland (Bjornsson et al., 1972), the Salton Trough in the United States (Thompson, 1983; Hardie, 1990), and the Red Sea (Brewer and Spencer, 1969). In these areas, thermal Ca-Cl brine reacts under high temperature (100°C–400°C) conditions between seawater and ocean floor basalts and forms serpentinized peridotites and Ca-Cl hydrothermal Ca-Cl brines (Hardie, 1990). In terrestrial tectonic activity zones, deep-source hydrothermal fluids participate in and may be accompanied by albitization, chlorination, and epidotization, thus forming Ca-Cl brines (Hardie, 1990). Elemental composition and isotopic analyses by Tan et al. (2011) showed that both Ordovician thermal Ca-Cl brines in the central Tarim Basin and Tertiary oilfield Ca-Cl brines in the western Qaidam Basin have a deep source hydrothermal mixing genesis. Petrographic thin section studies and the components analysis of fluid inclusions in halite suggest that mantle-derived Ca-Cl hydrothermal fluids exert an important influence on the evolution of Ca-Cl brines and mineral assemblage therein (Lowenstein et al., 2016). The basement of the Qaidam basin is a relatively stable terrene (Tang et al., 2000), and decades of research have not found any clear records of hydrothermal activity in the interior of the basin. Therefore, the water-rock interaction in the deep part of the basin margin and the recharge and mixing of magmatic hydrothermal fluids during the orogenic period are important mechanisms for the formation of Ca-Cl brines in this basin. The conclusion of this study is important for a better understanding of the genesis of Ca-Cl brines in the western part of the Qaidam Basin.