Qinghai Lake (36°32′–37°15′N; 99°36′–100°47′E) is located in the northeastern part of the Qinghai–Tibet Plateau, with a length of about 109 km from east to west and a width of about 39.8 km from north to south, and an average water depth of 18.3 m. It is the largest inland saline lake in China. The geological and geographical environment and the current climate of this area have been examined in detail by Colman et al. (2007) and Jin et al. (2010). Qinghai Lake is developed within a basin surrounded by three mountain ranges, including Datong Mountain to the north, Riyue Mountain to the east, and the Qinghai Nanshan Mountains to the south (Bian et al., 2000). The lake is divided into two nearly equal sub-basins by an NNW-trending horst, from which an island (Mt. Haixin) emerges. Several minor fault scarps can also be found in the lake basin. The lake bed is generally flat, and the lake is hydrologically closed and evaporative (Zhang et al., 1994). The six largest rivers on the order of discharge inside the Lake Qinghai catchment are Buha, Shaliu, Hargai, Quanji, Daotang, and Heima rivers. These rivers are greatly impacted by the hydrological cycle linked to the area’s monsoonal climate, and they supply over 87% of the water (Li et al., 2007), ∼67% of the dissolved load (Jin et al., 2009a), and sediment discharge to the lake. Over half of these rivers start from the Buha River, whose water chemistry is affected by the underlain marine limestone and sandstones, which constrain carbonate-dominated compositions of the lake water (Jin et al., 2009b). Meltwater from surrounding mountain glaciers accounts for only 0.3% of the total runoff (Colman et al., 2007).

Qinghai Lake is a typically closed saline lake, and there is no output of surface water from the lake. The lake’s recharge water is mainly the recharging rivers, followed by natural precipitation and groundwater. There are more than 70 rivers of various sizes flowing into the lake, with an obvious asymmetrical distribution, the northwestern part being relatively large and having a high runoff volume, while the southeastern part is sparse and mostly seasonal. The runoff of the rivers is unevenly distributed within the year, with July to September having the highest runoff, accounting for 60%–80% of the total annual runoff, and January to February has the lowest runoff, accounting for only about 1% of the total annual runoff. Qinghai Lake receives its annual runoff recharge mainly from seven rivers, including the Buha, Shaliu, Hargai, Quanji, Ganzi, Daotang, and Heima rivers. The total annual runoff is about 2.5 billion m³, and the main rivers have a total annual runoff of about 2.2 billion m³, accounting for 90% of the total runoff in the basin.

The Buha River is the largest exogenous recharge river in the Qinghai Lake basin, with a total length of 286 km, a basin area of 14,387 km², an average multi-year runoff of 7.85 × 10⁸ m³, and a contribution of 46.9% of the runoff into the lake. The Shaliu River is the second largest recharge river in the Qinghai Lake basin, with a total length of about 106 km and a basin area of 1, 500 km², with an average annual runoff of 2.46 × 10⁸ m³, accounting for 14.5% of the total runoff into the lake. The third largest recharge river in the basin is the Hargai River. The Buha River, Shaliu River, and Hargai River account for more than 75% of the total runoff into the lake. Apart from these, the rest of the recharge rivers are relatively small in terms of basin area and average annual runoff and are mostly seasonal rivers that dry up seasonally (the Quanji River is 65 km long and has a basin area of 567 km², the Daotang River is 60 km long and has a basin area of 727 km², the Heima River is 17.2 km long and has a basin area of 107 km², and the Ganzi River does not directly drain into Qinghai Lake but is surface runoff, which is submerged underground and then partly fed into the lake as groundwater). It is worth noting that the Daotang River is the only river in the Qinghai Lake basin that flows in a southeast to northwest direction.

The water samples collected include Qinghai Lake and its exogenous recharge rivers. Among the recharge rivers, the Buha River, Shaliu River, Hargai River, and Daotang River were selected as the main study areas in view of the influence of various factors, such as the contribution rate of runoff from exogenous recharge rivers, the lithology control of the basins, and the seasonal differences of rivers. Due to the lack of seasonal water during the sampling period, no samples were collected from the Quanji River or the Heima River.

A sampling of the water samples required for the study continued from 2017 to 2020 and can be divided into four batches. In the first sampling batch, water samples were collected from Qinghai Lake in September 2017. The subsequent batches mainly sampled the duplicate samples from Qinghai Lake and the water samples from the exogenous recharge rivers of Qinghai Lake in June 2018, August 2019, and November 2020.

In the first batch, a total of 14 sampling sites were set up to sample the waters of Qinghai Lake, with a sampling range of 36°38′25′′∼37°07′52″N, 99°42′48′′∼100°41′23″E. Among them, three sampling sites took water samples from the surface layer (∼0.2 m) to the bottom layer of the lake at 3-m intervals, while the remaining nine sampling sites collected water samples from the surface layer of the lake only. One surface water sample was also taken during the second batch of sampling, and a total of 40 samples were collected from Qinghai Lake. The second, third, and fourth batches of water samples were collected from the exogenous recharge rivers of Qinghai Lake, including 13 samples from the upper, middle, and lower reaches of the Buha River and Shaliu River (6 from the Buha River and 7 from the Shaliu River), while six samples were collected from the Hargai River and Daotang River (3 from the Hargai River and 3 from the Daotang River), which are near the entrance of the rivers. A total of 19 samples were collected from the recharge rivers. The distribution of the sampling sites is shown in Figure 1.

Surface water samples from Qinghai Lake and its exogenous recharge rivers were collected directly in high-density polyethylene (HDPE) plastic bottles, while other water samples at different depths were collected in HDPE plastic bottles using a Plexiglas (PPMA) water collector. The first batch of 39 water samples collected from Qinghai Lake in September 2017 were transferred to a 50 ml centrifuge tube and centrifuged three times (7 min) at 3,500 rpm to remove insoluble impurities such as insoluble particles. After centrifugation, the supernatant of each sample was taken and stored in a new HDPE plastic bottle to analyze the follow-up chemistry procedure. Samples collected from the other batches were filtered within 24 h of collection using a Nalgene manual vacuum filtration system (Nalgene 300–4,100; 6,133–0,010) with a 0.8-μm pore size mixed cellulose membrane (Millipore AAWP04700). After pretreatment, the water samples were stored in HDPE plastic bottles and sent to the laboratory for testing. The details of sampling information are shown in Table 1.

In order to verify whether centrifugation and filtration methods to separate particles affect the experimental results, three samples were centrifuged and filtered through 41-, 0.8-, and 0.45-μm filters, respectively, to assess the proportion of particulate matter of different sizes in the total particulate matter. The results show that the particle sizes in the centrifuged samples were less than 41 μm. In addition, we compared ²³²Th concentrations filtered by 41-μm filters and 0.8-μm filters in four lakes from the Qaidam Basin (Table 2). The results show that the size of thorium-enriched particles is mostly between 0.8 and 41 μm, and the proportion of particles less than 0.8 μm is limited. Thus, the lake fluxes involved in this study are calculated from the ²³²Th concentration in a particulate matter less than 41 μm, and the ²³²Th concentration in river waters is in a particulate matter less than 0.8 μm. Therefore, we believe that Th concentration filtered by 0.8-μm filters and centrifugation three times could represent the content of dissolved Th in saline lake water to a certain extent.

The method of Th isotope analysis of water samples can be broadly divided into two stages: chemical separation and mass spectrometry, where the chemical separation stage consists of co-precipitation, low-speed centrifugation, ion-exchange resin separation, collection, and purification. The specific steps involved are as follows:

1) Co-precipitation separation: A certain amount of water sample was weighed in a pre-prepared HDPE plastic bottle, a certain amount of 14N HNO₃ (1 N = 1 M), and 1–2 drops of HClO₄ and tracer (²³³U–²³⁶U–²²⁹Th) were added to the water sample, heated and dried, and 2N HCl was used to dissolve the dried sample. Then, three to five drops of Fe solution was added to the sample, use NH₃·H₂O to reconcile the pH of the sample to 8 to 9, and mixed well and left for some time until a yellow precipitate appears at the bottom of the bottle.

The mass spectrometry procedure was relatively straightforward. The nuclide ratio of samples was determined by Thermo Fisher’s Neptune Plus MC-ICP-MS, and the concentrations of ²³⁰Th and ²³²Th in samples were calculated by isotope dilution. All measurements were carried out using the peak jump program in ion counting mode at the retarding potential quadruple. For Th, the measurement uncertainties involved propagated errors from ICP-MS isotope ratio measurements, spike concentrations, and blank corrections. The procedural blanks for chemical and mass spectrometric analyses at the Laboratory of Isotope Geochemistry in Xi’an Jiao Tong University are approximately 205 fg (5.3 e⁶ atoms) for ²³²Th, and 37 ag (1.0 e⁵ atoms) for ²³⁰Th. The methods used herein have been fully described by Cheng et al. (2000), Cheng et al. (2013) and Shen et al. (2002), Shen et al. (2012).

The measured ²³⁰Th concentrations were corrected for in-growth due to decay of ²³⁴U throughout the time the specimen was retained. To utilize ²³⁰Th-²³⁴U disequilibrium to obtain a Th residence time, ²³⁰Thxs concentrations must also be corrected for the proportion of ²³⁰Th liberated by the dissolution of lithogenic substances. This adjustment is made with concurrent measurements of ²³²Th, with a lithogenic ratio of ²³⁰Th/²³²Th = 4.0 × 10⁻⁶ mol/mol (Roy-Barman et al., 2009).

The particle-reactive isotopes in the U and Th decay series are helpful tracers of particulate flux in water. The geochemistry of ²³⁰Th in sediments is also important with respect to spatial and temporal variations in particulate flux (Ankney et al., 2017). Previous studies have shown that ²³⁰Th is produced by the decay of ²³⁴U at a constant rate. In natural waters, ²³⁰Th is adsorbed by particulate matter in the water and eliminated from the water as it sinks. This process takes much less time than its radioactive decay rate (half-life of ²³²Th = 14.1 × 10⁹ years; half-life of ²³⁰Th = 75.6 × 10³ years) (Edmonds et al., 2004). The scavenging rate of Th is equal to the reciprocal of its corresponding residence time, and the hydraulic residence time of Th (τ_Th) in the lake water column can be obtained by measuring the concentration of ²³⁰Thxs in water (Eq. 1) (Hayes et al., 2013). τ Th ( z ) = ∫ 0 z dissolve d 230 Thxsdz   ∫ 0 z activit y 234 U × λ 230 dz (1) where the multiplication of ²³⁴U activity and λ₂₃₀ is considered as the increment of ²³⁰Th, λ₂₃₀ = 9.1705 × 10⁻⁶ (Cheng et al., 2013).

In the following, the ²³²Th fluxes will be calculated, and their contributions will be assessed in relation to the ²³²Th distribution characteristics of the lake water column, respectively. The formula for calculating the ²³²Th fluxes in lake water is given in Equation 2 (Hayes et al., 2013). dissolve d 232 Thflux ( z ) = ∫ 0 z d 232 Thdz × ∫ 0 z ( λ 230 234 U ) dz ∫ 0 z d 230 Thxsdz = ∫ 0 z d 232 Thdz τT h (2)

The dissolved ²³²Th flux was then used to estimate a particulate lithogenic flux considering the content of ²³²Th in the basaltic lithogenic material around the QHH region (10.5 μg/g; Wang et al., 2016) and the fractional solubility of ²³²Th in lithogenic material (S_Th; from our preliminary Th result of particles from QHH and Qaidam Basin) as shown in Equation 3 (Hayes et al., 2013). Lithogenic flux ( z ) = Disslove d 232 Th flux  ( z )   [ Th ] QHH × S Th (3)

To obtain an estimation of the flux of lithogenic material using dissolved Th-isotopes, it is first necessary to convert ²³²Th concentrations into a dissolved flux using the residence time calculated. For this, dissolved ²³²Th was integrated from the surface to an integration depth and divided by the corresponding ²³⁰Th residence time (Eq. (1)).

Based on the surface water samples obtained from 15 different sampling locations in Qinghai Lake, the spatial distribution of Th isotope concentration data in the lake water is shown in Zhang et al. (2019). The details are shown in Supplementary Table S1. There are differences in the distribution of ²³²Th concentrations between different sampling sites at the same depth, ranging from 0.265 to 3.356 pmol/kg, with a mean value of 1.247 pmol/kg, with the maximum value occurring at 107 and the minimum value at 131. The details are shown in the supplementary material file as Supplementary Figure S1. The total dissolved ²³⁰Th concentration varied from 0.27 to 19.03 μBq/kg, with a mean value of 7.76 μBq/kg, with the maximum value occurring at site 127 and the minimum value at site 123, which fluctuated considerably. In the lake surface water, the ²³⁰Thxs content of the 14 sites ranges from 0.2 to 18.2 μBq/kg with an average value of 7.39 μBq/kg, except for site 123, where no excess ²³⁰Th occurred in the surface water. The maximum value appeared at site 127 and the minimum value appeared at site 201, with a large fluctuation and an average value of 7.39 μBq/kg.

In summary, there are significant differences in the concentrations of ²³²Th, ²³⁰Th, and ²³⁰Thxs in the surface water of Qinghai Lake. The concentration of ²³²Th at sites 107, 115, 116, 118, and 201, located at the edge of the northwest lake area, had significantly higher levels than other areas of the lake, with a mean value of 2.513 pmol/kg at the five sites, which was higher than the other samples. In contrast, the average content of ²³⁰Thxs in sites 131, 126, 127, and 132 in the south lake area was 11.98 μBq/kg, far higher than the average value of surface water.

Lake water was sampled at varying depths at three sites (107, 123, and 131) in Qinghai Lake, and the results are shown in Zhang et al. (2019). The details are shown in Supplementary Table S2, Supplementary Figure S2, and Supplementary Figure S3. Site 107 is located in the northern lake area. The concentration of ²³²Th varied from 3.356 to 14.995 pmol/kg, with an average value of 6.273 pmol/kg. From the surface to the bottom of the lake, the concentration of ²³²Th exhibited a gradual and limited downward increase before decreasing at the bottom of the profile. Total dissolved ²³⁰Th ranges from 7.08 to 21.59 μBq/kg, with an average value of 12.63 μBq/kg, with a maximum value at 21 m and a minimum value at 9 m. The concentration of ²³⁰Thxs ranges from 5.3 to 12.5 μBq/kg, with a mean value of 8.24 μBq/kg, maximum values at 15 m and 27 m, and a minimum value at 9 m. The total dissolved ²³⁰Th and ²³⁰Thxs increased with depth before decreasing at a depth of 24 m.

At site 123, the concentration of ²³²Th varied from 0.439 to 1.791 pmol/kg, with an average value of 1.005 pmol/kg. From the top to the bottom of the sampling profile, the concentration of ²³²Th increased slightly with depth before decreasing at 21 m. The total dissolved ²³⁰Th content ranges from 0.07 to 3.44 μBq/kg, with an average value of 1.35 μBq/kg, a maximum value at 21 m, and a minimum value at 3 m. The concentration of ²³⁰Thxs content ranges from 0.5 to 2.2 μBq/kg, with a mean value of 1.16 μBq/kg. The maximum value occurs at 21 m and the minimum at 26 m.

At site 131, which is located in the southern lake area, the concentration of ²³²Th ranged from 0.265 to 4.563 pmol/kg, which increased with the sampling depth, peaked at 21 m and then decreased, with an average value of 0.872 pmol/kg, lower than that of sites 123 and 107. The concentration of total dissolved ²³⁰Th ranges from 7.34 to 18.61 μBq/kg, with an average value of 12.52 μBq/kg. The concentration of ²³⁰Thxs ranges from 6.9 to 18.5 μBq/kg, with a mean value of 11.9 μBq/kg. The maximum and minimum values of total dissolved ²³⁰Th and ²³⁰Thxs occurred at 18m and 3m, respectively. Unlike sites 107 and 123, the concentration of total dissolved ²³⁰Th and ²³⁰Thxs in site 131 fluctuated widely and showed an overall decreasing trend with increasing depth.

Overall, 19 surface water samples were sampled from the Buha River, Shaliu River, Hargai River, and Daotang River. The characteristics of the distribution of Th isotope concentration in these recharged river waters are shown in Table 3.

In the Buha River, the concentration of ²³²Th ranges from 1.531 to 2.874 pmol/kg, with a mean value of 2.583 pmol/kg, showing a decreasing trend from the northwest (upstream) to southeast (downstream) regions. The concentration of total dissolved ²³⁰Th ranges from 0.72 to 3.60 μBq/kg, with an average value of 2.38 μBq/kg. The concentration of ²³⁰Thxs ranges from 0.8 to 1.7 μBq/kg, with a mean value of 1.37 μBq/kg, except for three sampling sites where no excess ²³⁰Thxs was observed.

In the Shaliu River, the concentration of ²³²Th ranges from 4.488 to 46.913 pmol/kg, with an average of 12.341 pmol/kg, which is higher than that in the Buha River and the Hargai River. The total dissolved ²³⁰Th concentration ranges from 3.15 to 39.84 μBq/kg, with an average value of 11.51 μBq/kg. The concentration of ²³⁰Thxs ranges from 0.5 to 7.1 μBq/kg with a mean value of 2.89 μBq/kg. The concentrations of ²³²Th and total dissolved ²³⁰Th and ²³⁰Thxs of SL-401 are much higher than those in other samples, which may be related to the sampling season. The sample SL-401 was sampled in winter, while other samples were sampled in summer.

In the Hargai River, the concentration of ²³²Th ranges from 0.433 to 3.132 pmol/kg, with an average value of 1.962 pmol/kg and little variation between sampling sites. The concentration of total dissolved ²³⁰Th ranges from 2.62 to 88.41 μBq/kg, with an average value of 31.30 μBq/kg. ²³⁰Thxs ranges from 0.4 to 88.1 μBq/kg, with a mean value of 29.93 μBq/kg. HEG-401, which was sampled in winter, had much lower ²³²Th concentrations than the other two samples. However, the concentration of total dissolved ²³⁰Th and ²³⁰Thxs of HEG-401 is approximately 40 times and 80 times more than that of the other two samples, respectively.

In the Daotang River, the ²³²Th concentration ranges from 2.350 to 61.941 pmol/kg, with a mean value of 24.532 pmol/kg. The concentration of total dissolved ²³⁰Th ranges from 10.03 to 101.34 μBq/kg, with an average value of 53.78 μBq/kg. The concentration of ²³⁰Thxs ranges from 3.5 to 99.7 μBq/kg, with a mean value of 36.6 μBq/kg. There were significant differences in the horizontal distribution of ²³²Th, total dissolved ²³⁰Th, and ²³⁰Thxs in the three samples from the Daotang River. The ²³²Th concentration of DT-401, which was sampled in winter 2020, was 7 times higher than the other samples.

Qinghai Lake is a typically closed saline lake, and the lake’s recharge water is mainly from exogenous recharge rivers, followed by natural precipitation and groundwater. Previous studies have suggested that the ²³²Th concentrations in surface water appear to be linked to the water column scavenging rate (Cochran et al., 1995; Edmonds et al., 1998; Edmonds et al., 2004). ²³²Th is non-radiogenic, so most of the ²³²Th in the lake water comes from the lake margins (bank erosion), rivers, or aeolian detrital particulate material. Th is insoluble in water and is rapidly eliminated from solution via scavenging on particulate material (Moore and Sackett, 1964). Therefore, if there were no contributions from other sources, the variation of ²³²Th concentrations in water should decrease with depth and show the highest values in surface water. Therefore, depth profiles can be used to assess the lithogenic fluxes in lake water. More specifically, the ²³²Th fluxes in deep water indicate a deep source of dissolved ²³²Th, possibly as a result of boundary scavenging. The sampling profiles indicate that the ²³²Th concentrations in Qinghai Lake increase slowly with depth, with higher bottom water concentrations in all three profiles. These trends suggest that an additional ²³²Th enters the water column at the bottom of the lake, possibly from the contribution of the bottom sediments. The ²³²Th concentrations at the lake surface (Supplementary Table S1 and Supplementary Figure S1 in the supplementary material file) suggest that the dissolved ²³²Th flux is primarily derived from the input of aerosol (dust) in shallow waters. Furthermore, the fluxes near the entrance of the recharging river appear to be derived from the ²³²Th contributed to the lake by the river. The ²³²Th in the lake is therefore thought to be due to the dissolution of dust aerosols in combination with inflow from the recharging rivers (Zhang et al., 2019). The characteristics of the horizontal and vertical distribution of ²³²Th in the lake water of Qinghai Lake suggest that there is an exogenous source contribution, of ²³²Th to its lake water, and that different areas of the lake do not receive exactly the same proportion of the exogenous source contribution which is from hydraulic sorting of grain size and types. Because several major recharge rivers flow through stratigraphic source rocks that are different, including felsic rocks, carbonate, metamorphic rocks, and clastic rocks. In view of this, it is reasonable to assume that the main sources of ²³²Th in the lake water column of Qinghai Lake include atmospheric deposition dust, exogenous recharge river, and the mix from the interface of water-rock with the water depth increase. Selecting an integration depth at which the lithogenic inputs can be effectively quantified has been a subject of discussion with varying particle types. Due to the sampling method and evaluation method of this study being different from Jin’s (Jin et al., 2009a), the lithogenic flux in this study is different from the flux in Jin’s. Jin et al. estimated the dust flux based on the suspended particulate matter supplied by the Buha River to Qinghai Lake, and the lithogenic flux of Qinghai Lake in our study includes two sources, atmospheric deposition and river recharge. The dust flux in this study refers to the particulate matter caused by atmospheric dry deposition.

The dissolved ²³²Th fluxes can be converted to lithogenic-material fluxes with knowledge about the composition of the material and its Th fractional solubility (Eq. (3)). In other words, this flux of lithogenic material is required to sustain the calculated dissolved ²³²Th and therefore the measured concentration of ²³²Th. The lithogenic material from the Qinghai region has a ²³²Th content varying from 10.5 to 11.7 μg/g (Wang et al., 2016). Therefore, we use this value for the calculations of the lithogenic flux in our area of study.

The largest source of uncertainty in the calculation of lithogenic fluxes is the solubility of ²³²Th from particles, which is a highly unconstrained parameter. From the few studies that report ²³²Th solubility in marine particles, highly variable values were found, ranging from 1% to 23%, depending on the particle size and depth (Arraes-Mescoff et al., 2001; Roy-Barman et al., 2002). There is no report about ²³²Th solubility in lake particles. According to our preliminary results from lake particles from the Qaidam Basin, ranging from 1 to 10%, depending on the particle size and types (unpublished data).

According to previous studies (Edmonds et al., 1998; Edmonds et al., 2004; Hayes et al., 2017), since most of the dissolved ²³²Th in seawater originates from atmospheric dust which dissolves into the ocean after settling to the sea surface, the ²³²Th concentration in marine waters decreases with depth and reaches its highest value at the sea surface. This is in contrast to the phenomenon we observed in the lake. All three depth profiles of Qinghai Lake showed a positive correlation between ²³²Th concentration and depth, and the high ²³²Th values in surface waters near site 107 suggested that this may be due to the input from exogenous recharge rivers. The advection of particulate matter from the nearshore shelf can increase regional scavenging rates and fluxes. The recharge rivers carried more exogenous detrital material into the lake and mixed it with the lake water, causing the particulate flux to increase with depth and reaching a maximum in the bottom water of the lake.

Based on Th isotope data and flux information of surface water in Qinghai Lake, the dust flux of lithologic material in September contributed by atmospheric dry deposition to the lake can be calculated, which is approximately 0.109 g/m²/yr. The lithogenic fluxes delivered to the lake by rivers can be quantified based on the data from stations located near the recharging rivers. The fluxes of lithologic material from the recharge rivers input into the lake are 120.55 g/m²/yr in the northwest basin and 12.42 g/m²/yr in the southeast basin, which are calculated by the mean value of lithogenic fluxes at depth profiles of 107 and 123 sites and the mean value at depth profile site 131, respectively. Sites 126, 127, 131, and 132 are located in the southeast of Qinghai Lake, far away from the estuary, and the ²³²Th of surface water is mainly derived from atmospheric dust. The mean lithogenic flux of the surface water at these four sites is 0.109 g/m²/yr. Sites 105, 107, 113, 115, 116, and 118 are located in the northwest and southwest of Qinghai Lake near the estuary. Their lithogenic fluxes mainly come from fluvial suspended particles and atmospheric dust. The mean lithogenic flux of the surface water at these six sites is 1.046 g/m²/yr. The contribution of atmospheric dust and river input to the lithogenic flux of Qinghai Lake can be assessed by these two mean values. For instance, the Buha River and Shaliu River contribute ∼90% of the lithogenic flux to the northwest basin of the lake, while the atmospheric dry dust deposition contributes only 10% of the flux. Therefore, unlike the northwest region, the source of particle material flux in the southeast basin of the lake is dominated by atmospheric deposition, and the contribution of the recharging river to the lake’s water flux is less than that of the atmospheric dissolution of dust at the lake surface. In jin et al. (2009a), the dust input represents 65% of the total input of dissolved and particulate forms, which is estimated based on the suspended particulate matter supplied by the Buha River to Qinghai Lake. Among the four recharge rivers of Qinghai Lake, from our result, the Buha and Shaliu Rivers have higher lithogenic fluxes, including dust input into the river and ²³²Th concentrations, contributing a large amount of particulate matter to the northwestern area of Qinghai Lake with a ratio of ∼90% of the detrital flux to the lake in September, which is actually not contradictory to Jin’s result with ∼65% of the dust input origin of the total input of dissolved and particulate forms in the Buha River and recharge into Qinghai Lake. This study supplies several rivers’ loads into the lake by a novel ²³⁰Th normalization method, which is further complementary to the detrital flux into Qinghai Lake with several major incharge rivers.

Hayes et al. (2013) suggest that there are two sources of Th in the ocean: one shallow due to dust dissolution and one deep associated with sediment dissolution/resuspension; and one removal mechanism, scavenging throughout the water column. Therefore, another reason for the higher concentration of ²³²Th in the deep water than in the surface water in the saline lakes may be due to the redissolution of ²³²Th in the sediment at the bottom of the lake. The other possibility is from accumulated sinking flux due to the slow deposition rate. In the future, we expect to use Th isotope information in lake sediments, combined with Th concentration data of depth profile particles’ composition and size, to further quantify the contribution of ²³²Th to lake sediments.

Overall, in September, the Buha and Shaliu rivers in the northwest basin contribute about 90% of the detrital ²³²Th flux to the lake. The lithogenic flux in the southeast lake is dominated by dust flux with a value of ∼0.109 g/m²/yr, while the higher lithogenic flux at the bottom of the lake was likely generated by accumulated sinking particulate matter and resuspension of bottom sediments in September.