Chemical Weathering of the Southern Tibetan Plateau: Study the Geochemical Weathering Indices in the Top Soils


 Continental chemical weathering has been suggested to affect the concentration of atmospheric carbon dioxide that influences global climate change at different time scales. Various indices for chemical weathering have been adopted to investigate past change in chemical weathering intensity and climate change on oceanic and lacustrine sediment archives. The reliability of the chemical weathering indices has been questioned as most sediments likely originate from multiple types of bedrock that may experience various degrees of chemical weathering and can thus be reliably robust indicators of climate and paleoclimate. Here we present Sr-type (e.g. Rb/Sr Sr/Ba) and Na-type (e.g. CIA CIW PIA CPA) chemical weathering indices for top soils across the southern Tibetan Plateau to discuss the chemical weathering characteristic in the Tibetan Plateau and to examine their response to regional climate variation. The results of chemical indices and the A-CN-K ternary plot show that the southern Tibetan Plateau is under the carbonate control of the primary chemical weathering stage with the cold-dry climate. Correlation analyses show shat Sr-type indices co-vary with mean annual temperature and annual precipitation while Na-type indices show little consistence with regional climate. The climate condition is the dominant control of Sr-type indices of top soils in the study area and the bedrock may be the dominant control for the Na-type indices. We also compared the corresponding indices at a Holocene lacustrine sediment profile in the Qaidam Basin in the northeast Tibetan Plateau with regional climatic records which strongly supports our observation in the top soils. The results of the study suggest that for the relative cold and dry climate in Tibetan Plateau the Sr-type indices are more sensitive to climate condition than Na-type indices. This suggests that the Sr-type indices are likely more suitable than Na-type indices to reflect the change of climate on the Tibetan Plateau. Caution should be taken for using the Na-type indices for reconstructing the past change in climate for the study area.


Introduction
Continental chemical weathering on global scale is suggested to impose a net drawdown of atmospheric CO 2 that in uences global climate changes (Walker et al. 1981;Berner et al. 1983; Berner and Berner 1997; Ruddiman et al. 1997). The Tibetan Plateau (TP) have been considered as a key region for chemical weathering that has lowered atmospheric CO 2 (Fedo et al. 1995) and CPA (Chemical Proxy of Alteration) (Buggle et al. 2011) and Sr-type indices (Buggle et al. 2011) including Rb/Sr and Sr/Ba ratios have been constructed to study changes in chemical weathering intensity climate change and geological events (Nesbitt et al. 1980;Chen et al. 1999; Muhs et al. 2001; Buggle et al. 2011;Velbel 1993; White and Blum 1995;Yang et al. 2004a; Deepthy and Balakrishnan 2005). Applications of the indices well reveal changes in weathering pro les dissolved matters in river water and aeolian sediments. For example Descourvieres et al. (2011) suggested that CIA re ected changes in chemical weathering associated with tectonic activity in Perth Basin in southwest Australia. Chen et al. (1999) showed that the Rb/Sr ratio could be used as an index for chemical weathering associated with past climatic changes at the Luochuan loess-paleosol pro le in the central Chinese Loess Plateau. However there is a lack of systematic studies to carefully examine the proxies against regional geochemical background and climate parameters in order to understand their suitability to sediments such as top soils and river sediments that are derived from different rock types and spanning different climatic conditions. The uplift of the TP and its environmental effects during the Cenozoic have been widely studied since tectonic uplift-weathering hypothesis was proposed by Raymo et al. (1988). Weathering from the river basins on the TP plays a key role in surface processes and geochemical cycles in earth supergene environments including global carbon cycle and chemical composition of the oceans (Berner et al. 1983; Kump et al. 2000). Previous studies have focused on main drainage basins around the TP such as the Yangze River catchment (Yang et al. 2006;Borges et al. 2008) and the Yellow river drainage basin (Li 2003;Yang et al. 2004a;Zhang et al. 1990) and so on. And the results indicate that there is a strong relationship between weathering and the climate change during glacial and interglacial periods (Liu et al. 2004;Yang et al. 2006;Borges et al. 2008). Unfortunately the weathering characteristic of the soils in the TP were discussed rarely. Theoretically the development of top soils by in situ weathering of earthy and rocky materials on direct exposure to atmospheric agents would result in a better preservation of chemical wreathing information (Qiu et al. 2014). Top soils on the TP mainly originate from in situ weathering of bedrocks or are transported from nearby high elevation mountains by surface waters (Lin and Feng 2015) which are ideal to investigate the response of various indices to climate change (Qiu et al. 2014). Here we present the geochemical analysis on top soils across the southern TP to 1) investigate the chemical weathering characteristic of top soils; 2) examine the relationship between various chemical weathering indices and regional climate condition (e.g. regional temperature and precipitation). We also compared various indices at a sediment core in Qaidam Basin with Holocene climate records to test our observation in the top soils.  (Sun et al. 2007). The Qiangtang block in the central is characterized by metamorphic rocks and Late Paleozoic (Carboniferous and Permian) shallow marine strata in the west and Triassic-Jurassic shallow marine carbonate rocks interbedded with terrestrial clastic and volcaniclastic strata in the east (Liu 1988). The Himalaya block in the north consists mainly of Precambian clastic sediments and metasedimentary rocks in the south (Brook eld 1993) late Proterozoic to early Cambrian metasedimentary rocks in the middle (Parrish and Hodges 1996) late Precambrian to early Paleozoic sedimentary and metasedimentary (Yin and Harrison 2000) and thick Permian to Cretaceous continental margin sequences in the north (Brook eld 1993). The sampling sites consists mainly of two east-west-trending structural blocks the Lhasa block to the south and the Qiangtang block to the north.

Geology And
The climate in the southern TP is generally cold and dry. Monsoon season from June to September is relatively warm and humid and the non-monsoon season from October to May is extremely cold and dry. During the non-monsoon the climate is under the control of westerly winds and dust storms happen frequently in the western TP. During the monsoon season moisture and heat are transported northward to the TP from the south Indian Ocean. The mean annual temperature (MAT) and the annual precipitation (MAP) are relatively high in the southeast area (0 o C and 300 mm/year) and lower in the northwest area (-3 o C and less than 50 mm /year) (Huang et al. 2008;Li et al. 2009). Thus chemical weathering processes are relatively weak and physical weathering is a dominating process for soil production. For instance glacier activities (e.g. glacial grinding) and frost weathering in high mountains play an important role (Sun et al. 2007).
The climate of the Qaidam Basin is the typical desert climate which is also cold and dry compared to the other area in China. The annual average temperature is ~ 5.33℃ with great seasonal difference and the annual precipitation is only 10mm ~ 30 mm. The evaporation could reach up to 3564.4 mm per year. And the climate of Dongtaiji Nuoer Lake belongs to the Qaidam Basin desert climate with the MAT of 2.0℃ and the MAP less than 25 mm.

Top soils on the southern TP
The 114 top soils developed on various rocks and with different climatic zone were collected across the southern TP ( Fig. 1 and supplementary Table 1). We collected 0-5 cm soil after removed the visible plants and roots. Each sample was ~ 500 g stored in plastic bags. All soil samples were easily get but far away from human activities and collected at relatively low elevations where the soils mainly originate from in situ weathering or are transported from nearby mountains (Chen et al. 1981;Lin and Feng 2015). Soil samples were collected from different parent rock (Table S1) but all located in Lhasa Block and Qiangtang Block. The geographic coordinate (longitude latitude and altitude) of each sampling sites was given by a handheld GPS device during sampling at the site (Table S1).
3.2 Lacustrine sediment samples from Lake Dongtaiji Nuoer Qaidam Basin Lake Dongtaiji Nuoer is located in the central Qaidam Basin northeastern TP (Fig. 1). An 11.3-m deep trench was excavated by a local lithium factory. We collected the lake sediment samples at a 10-cm interval and stored in plastic bags.

Methods
Soil samples were dried in a 50℃ oven for more than 24 hours before laboratory analysis. Visible stones and plant residues were picked out before grinding samples using an agate mortar. Then the dry soil samples were passed using a 200 mesh sieve (< 0.074 mm) for laboratory analysis. The lake sediment samples were air-dried at room temperature and grounded using an agate mortar which were passed through using a 200 mesh sieve prior to analysis.
All top soil samples were analysed for geochemical elements of Mo Zr Sr Rb Th Pb Se As Au Zn W Cu Ni Co Cr V Sc S Ba K Fe Mn Ca and Ti during the sampling process in the eld using a portable X-ray uorescence spectrometer (XRF) (Niton XL3t 950). San (2013) presented the complete dataset in her thesis of which only Rb and Sr were presented in this study. We selected 32 soil samples for laboratory analysis in this study in order to investigate the geochemistry characteristic and chemical weathering for the top soils (Table S2). The 32 samples were selected from various types of sampling soils which covers typical climate zones on the TP (Table S2). In the laboratory 32 soil samples were analysed for major elements (SiO 2 Al 2 O 3 Na 2 O K 2 O MgO CaO P 2 O 5 TFe 2 O 3 TiO 2 and MnO) using X-Ray Fluorescence Spectrometry (XRF) (Rigaku RIX-3000). Trace elements (Rb Sr and Ba) were analysed following nitrichydro uoric acid digestion (HNO 3 + HF) using Inductively Coupled Plasma-Mass Spectrometry (ICP-MS).
Replicate samples and reference materials (AGV-2 BHVO-2 BCR-2 and RGM-2) were used for quality control which shows that the standard deviation is less than 6%. Analysis of the soil samples was undertaken at the State Key Laboratory of Geological Processes and Mineral Resources in China University of Geosciences. The concentration of Sr and Rb were highly consistent between laboratory analysis and portable XRF measurements for the 32 samples.

Climate data
In order to examine the relationship between various chemical weathering indices from the top soils and climatic parameters we obtained the MAT and the MAP data for 1992 to 2012 for the 114 sampling sites using the China Meteorological Forcing Dataset (ITPCAS CMFD) (He 2010) ( Table S1). The MAP ranged from 69.91 mm to 1073.59 mm and the MAT ranged from − 11.14℃ to 18.13℃ for the study area. From the Fig. 1 and the Table S1 it can be found that there are big differences for the climate conditions for the sampling sites where the MAT is less than 0℃ and the MAP is less than 200 mm almost are located in the western Tibet.

Results
The concept of geochemical proxies of mineral alteration (e.g. weathering indices) relies on the selective removal of soluble and mobile elements from a weathering pro le compared to the relative enrichment for rather immobile and non-soluble elements (Yang et al. 2004b). To transport process the elements of Na Ca and Mg are highly mobile in the weathering environment using as the mobile element and Al is proposed as the immobile counterpart to minimize biases due to variable mineralogical composition of the parent material (Liu et al. 1984;Buggle et al. 2011). We calculated the following some Na-type indices for the study:

Sr-type indices in the top soils
Rb/Sr ratios varied signi cantly for the 114 top soils across the southern TP with the highest value observed in YF-3 in the southeastern Tibet and the lowest value in BH-9 in the northwestern Tibet ( Fig. 3A and Table S3). The mean Rb/Sr value of the 114 top soils was 0.55 with the range from 0.11 to 4.28. The soils signature YF-and YP-located in the eastern study area always had the relatively big Rb/Sr ratios compared to the other soils (Fig. 3A). And for the BH-samples sited in the western study area they signi cantly had the lowest Rb/Sr ratio to all the soils and the Rb/Sr ratios in the TPE-and TPW-soils around Lhasa were relatively moderate. It found that with the increasing longitude the tendency of the increasing Rb/Sr ratios was showed signi cantly. Meanwhile for the collected 32 soil samples measured in the laboratory the Rb/Sr ratios also showed similar distribution pattern (Fig. 3B). The maximum Rb/Sr was recorded in the YF-2 located in the eastern Tibet and the minimum Rb/Sr was recorded in the BH-22 located in the western Tibet. And the other soils had the relatively moderate Rb/Sr value.
For the Sr/Ba ratios we calculated for selected 32 soil samples with the laboratory measurement. The Sr/Ba ratios ranged from 0.32 to 8.34 with the mean value of 1.33 (Fig. 3C). The two highest Sr/Ba ratio were recorded in TPW-43 and BH-22 with the value of 8.35 and 8.00 respectively (Fig. 3C). These two sites were in close all located around Bangong Co in the western Tibet. The lowest Sr/Ba value of 0.32 (Fig. 3C) was recorded in the YF-7 located in the eastern Tibet. It can be found that for the YF-YP-and some TPE-signatured samples located in the western Tibet recorded relatively lower Sr/Ba ratios compared to the BH-and TPW-signatured samples (Fig. 3C). The Sr/Ba ratios were relatively moderate in the sampling sites almost around Lhasa. From the western to the eastern with the increasing longitude the value of Sr/Ba was increased apparently showing opposite to the Rb/Sr records.

Low degree of chemical weathering of the top soils
In rock-forming minerals alkali and alkaline-earth cations such as Na K and Ca form strongly ionic bonds with oxygen that are long and weak. They are thus more easily broken and cations consequently released than transition elements and metalloids forming tighter bonds with more covalent character (Taylor and McLennan 1995). While Ca Na and K are removed from feldspars by aggressive soil solutions the proportion of aluminum to alkali and alkaline-earth metals increases progressively in the products of weathering.

An Al 2 O 3 -(CaO*+Na 2 O)-K 2 O (A-CN-K) ternary diagram based on the mass balance principle feldspar
leaching experiments and the thermodynamic calculation of mineral stability has been used to predict the trend of continental chemical weathering and the alteration of mineralogical or geochemical components (Nesbitt and Young 1982). The extent of conversation of chemical proxies such as CIA which is 50 for unweathered feldspars and varies consequently from 50-65 for low chemical weathering to 65-75 for moderating chemical weathering and to 75-100 for strong chemical weathering (Nesbitt and Young 1982) (Fig. 4) The distribution of the top soils on a ternary A-CN-K diagram revealed a cluster of points together with a distribution along a line parallel to the A-CN join (Fig. 4) suggesting that the weathering of the soils was close to a non-steady state condition   Fig. 4 indicated that the products of weathering were smectite and plagioclase with no kaolinite. A simple weathering trend was de ned in A-CN-K diagram (red line in Fig. 4) which was subparalled to the A-CN boundary primary because removal rates of Na and Ca from plagioclase generally were greater than the removal rate of K from K-feldspar (Liu et al. 1984;Young 1984 1989;Nesbitt and Markovics 1997). In terms of the elemental geochemical signature the effect of decalci cation and the removal of Na (dissolution of plagioclase) on the top soil was accompanied by the evolution of an Al enrichment. The chemical weathering of the southern TP according to the A-CN-K ternary plot was characterized by the carbonate control of the primary weathering stage under cold-dry climatic conditions and a transition between Ca and/or Na removal. This nding is consistent with the result of Sun et al. (2007) that the chemical weathering processes are relatively weak for the TP. Should the degree of weathering continue the samples are predicted to continue along the A-CN join until it reaches the A apex and Ca Na and K are completely removed.

In uence of climate on the Sr-type weathering indices for the top soils
According to the ITPCAS CMFD we collected the data of the MAT and MAP for the 114 soil sampling sites (Table S1). The MAP varied with a huge range for the study area. The maximum MAP of 1073.59 mm was recorded in the soil of YF-1 located in the eastern Tibet with the longitude of 97.03°and the latitude of 28.47° (Table S1) while the minimum MAP of 69.91 mm was recorded in the soil of TPW-27 located the South of Bangong Co in the western Tibet (Table S1). Also the MAT ranged in a relatively large range from − 11.14℃ to 18.13℃. The maximum value of MAT was also recorded in the YF-1 and the minimum value was recorded in BH-15 around Bangong Co in the western Tibet (Table S1). From Fig. 1 Table S1 in general the value of the MAT and MAP were all greater in the eastern Tibet than in the western. Mostly the MAP and MAT increased with the increasing longitude.

and the
Sr is common in minerals such as plagioclase pyroxene amphibole and biotite which are susceptible to weathering (Liu et al. 1984;Nesbitt et al. 1980;Reeder et al. 2006). While regarding the choice of the immobile element ions of intermediate ionic potential i.e. ions that tend to from insoluble hydrolysated are generally employed. Rb and Ba can be immobilized by adsorption on clay minerals due to their large ionic radius are often used as immobile references. Therefore under intense weathering conditions the mobility of Sr is greater than Rb and Ba (Chen et al. 1996) and the Rb/Sr should be relatively higher and the Sr/Ba should be relatively lower.
From the distribution characteristics of the Sr-type indices in the top soils (Fig. 3) the relatively high Rb/Sr ratio and low Sr/Ba ratio found in the eastern Tibet suggested that the stronger chemical weathering occurred. And the relatively low Rb/Sr value and high Sr/Ba value in the western Tibet suggested that the relatively weak chemical weathering occurred (Fig. 3). For the soils around Lhasa the Sr-type indices value was relatively moderate indicated the central Tibet suffered the relatively moderate chemical weathering compared to the eastern and western sites. In general the soils in the eastern sites suffered the strongest chemical weathering. The chemical characteristic re ecting by the Sr-type indices looked like similar to the change of the MAT and MAP (Fig. 1) for the study area. Statistical analysis using SPSS shows that the MAT and MAP were signi cantly correlated with the Sr-type indices (r = 0.57 for Rb/Sr with n = 114 p < 0.05; r = 0.58 for Sr/Ba with n = 32 p < 0.05) (Fig. 5) suggesting an important in uence of climate parameters on the intensity of chemical weathering. The well correlation suggested that the Srtype indices could be used as the indicators of climate. Warm temperatures high precipitation as well as dense vegetation cover would result in a more acidi ed soil and a high content of organic ligands in the soil water (Stumm and Morgan 1996). This would promote removal of Sr from the soil. However Rb-rich minerals such as muscovite biotite and K-feldspar were more chemically stable than Sr-rich minerals such as carbonate plagioclase and amphibole (Goldich and Gast 1966). Therefore the mobility of Sr was greater than for Rb and Ba (Chen et al. 1996). This would explain the higher Rb/Sr and lower Sr/Ba ratios observed in the top soil in the eastern area.
The altitudes of the soil samples obtained in this study ranged from 1503 to 5262 m (Table S1) which showed slightly negative correlation with Rb/Sr ratios of top soils (r = 0.43) (Fig.S1) indicating a moderate in uence of altitude on the chemical weathering intensity. However the altitude co-varies with temperature and precipitation vegetation cover which would impose similar in uence on the chemical weathering. But the Sr-type indices were more dependent on climate condition than on altitude (the correlation coe cient of altitude with the Sr-type indices was smaller than that of climate condition with the Sr-type indices). The collected top soils were formed in situ or deposit in the low-lying places with the parent materials from surrounding source rocks (Lin and Feng 2015). According to the Table S3 and  Table 1 (Table S2) suggesting the parent rock may has little impact on the Sr-type indices. The value of Rb/Sr ratios were highly correlated between the 32 samples with the size < 200 mm measured by ICP-MS and the 114 bulk samples by portable XRF instrument in the eld (r = 1) indicating that for the study top soils the dependence for Sr-type indices on grain size was little.

3 Limited In uence Of Climate On The Na-type Indices
We calculated Na-type indices for 32 soils samples. The data of the MAP and the MAT with the 32 soils sites were also listed (Table S2) The maximum MAP was recorded in the YF-2 located in the eastern Tibet. Also in the eastern Tibet the MAT value was higher than that in the western Tibet. According to the climate data we discussed in 4.2 the eastern Tibet was much more moisture and warmer than the western Tibet.
However Na-type indices including CIA CIW PIA and CPA clustered apparently in the 32 top soils for the study area (Fig. 2). They all had similar value according to 4.1. In the TP under the more warm and moisture condition the Na-type indices did not show apparently much higher. There was no signi cant correlation between the Na-type indices and the MAT and MAP (at the 0.05 level) (Fig. 6) based on the correlation analyses suggesting that in uence of climate on the variation of Na-type indices was limited. And therefore the Na-type indices could not be used as the indicators of the climate change for the study area. The limited in uence of climate on the Na-type indices is consistent with observations on the bedload sediments of large rivers originating from the TP (Borges et al. 2008). But the most of the other researches about the chemical weathering with catchment (Li and Yang 2010; Shao and Yang 2012) and top soils (Qiu et al. 2014) in China showed that the Na-type indices usually were controlled by the climate dominantly.
As discussed in 5.1 the distribution of the top soil samples on a ternary A-CN-K diagram (Fig. 4) showed that most soil samples on the southern TP fallen in an early stage of chemical weathering and the carbonate control was the primary weathering stage under cold-dry climatic conditions. As pointed by the investigation chemical weathering like CIA actually re ected the integrated weathering history rather than present weathering intensity (Li and Yang 2010). The carbonate control of the primary weathering stage of the TP led to the Sr and Ca usually bound in carbonates more sensitive to present climate change than Na and K because Na and K generally bound in silicates. This might be caused the poor relationship between the climate condition and the Na-type indices in the 32 sampling sites. This might interpret why Rb/Sr and Sr/Ba may have more positive correlations with temperature and rainfall. It had been suggested that if chemical weathering was limited the weathering products (i.e. the top soils) would inherit the elemental composition of the parent silicate materials according to the CIA and the A-CN-K ternary plot (Tripathi et al. 2007). As we did not collect bedrock for top soils we compiled values of Na 2 O  (Table S4 and Fig.S2). For the results the Na-type indices showed assemble each other with the different rocks for the different sites. Like that in granite rock and basalt rock the Na-type indices varied at the similar range and had the similar value. The variation of the Na-type indices had no signi cantly different with different rock types for the collected data in the southern TP (Fig.S2). According to the results it meant that for the Na-type indices in the southern TP the in uence of the bedrock might be more dominant than the climate condition.

4 Sr-And Na-type Indices At Lake Dongtaiji Nuoer
In order to further investigate whether Sr-type and Na-type indices would re ect climate changes in TP we measured elemental concentration and calculated all indices on a lacustrine sediment pro le at Lake Dongtaiji Nuoer in the Qaidam Basin northeastern TP. The climate in the Qaidam Basin has the relatively similar characteristic of cold-dry with the Tibet and the geochemistry of the Qaidam Basin is in uenced by the Tibet signi cantly. Chronological controls on the pro le was established based on AMS 14 C ages. The 14 C age at 950 cm depth was 13830+/-50 year (16693 cal yr BP) indicating that most of the sediment in the pro le was accumulated during the Holocene. Since 14 C measurements in saline lakes are affected by the reservoir effect (Hou et al. 2012) our comparison of the results with regional climatic records was tentative. We compared the Sr-type and Na-type indices in the stratigraphic variation with a regional climate record ratio of Artemisia to Chenopodiaceae (A/C) at Hurleg Lake in Qaidam Basin that re ects changes in the effective precipitation ) and the salinity minerals variety re ects the climatic change apparently in Chaka Lake (Liu et al. 2008).
Both Sr-type and Na-type indices exhibited signi cant variations at Lake Dongtaiji Nuoer (Fig. 7). Higher values of Rb/Sr occurred below 900 cm and above 180 cm and relatively lower Rb/Sr values were observed between 180 cm and 900 cm ( 14 C age of 5000-13830 BP) (Fig. 7). The variation of Rb/Sr ratios is generally consistent with A/C records at Hurleg Lake and the salinity minerals records in Chaka Lake. A/C ratios suggest that the climate is relatively dry and variable from 9.5 to 5.5 ka in the Qaidam Basin ) which was supported by more saline minerals in Chaka Lake during the same period (Liu et al. 2008). Lower Rb/Sr ratios from 9.5 to 5.5 ka at Lake Dongtaiji Nuoer showed relatively weak chemical weathering.
All Na-type indices showed similar variation in the lake sediment pro le (Fig. 7). Higher values of the Natype indices occurred at 0-250 cm (0-5517calyr BP) and 450-650cm (6522-7527 cal yr BP). Lower Natype indices values were observed at 250-450 cm (5517-6522 cal yr BP) and 650-1135 cm (7527-23133 cal yr BP). Only in the upper pro le (above 200 cm) were the variations of the Na-type indices consistent with the A/C ratio and the salinity minerals variety which suggested that climate in uenced the Na-type indices. But below 200 cm no clear relationship was observed between Na-type indices and A/C records. For the whole sediment pro le the consistence of the Na-type indices stratigraphic variation with the climate records did not show apparently.
In general the comparison between the Sr-Na-type indices and climate records in the Qaidam Basin suggested that the records of Sr-type indices would be better re ect the in uence of climate change on chemical weathering which is consistent with the observation on the top soils. The Na-type indices probably could not be used as a reliable proxy of instantaneous chemical weathering with climate change (Li and Yang 2010; Shao and Yang 2012) for the sediment pro le in Qaidam Basin. Because the Na-type indices usually used for re ecting the integrated chemical weathering processes for sediment (Li and Yang 2010;Shao and Yang 2012). However the difference between the two different chemical weathering indices should be further studied. Clearly therefore caution should be needed when using the Na-type indices as proxies for climate change.

Conclusions
We have analysed both Sr-type (Rb/Sr and Sr/Ba) and Na-type (CIA CIW PIA and CPA) indices on the top soils on the southern TP. Based on the results of the chemical indices and the A-CN-K ternary plot the southern TP is under the carbonate control of the primary chemical weathering stage under the cold-dry climate. The statistical analysis results show that the Sr-type indices are more sensitively respond to mean annual temperature and annual precipitation. This suggests that Sr-type indices (Rb/Sr and Sr/Ba) would better re ect climate changes for the study area. Na-type indices may also reveal intensity of chemical weathering however which seems less sensitive to climate change. The comparison of Holocene records in the Qaidam Basin of Sr-and Na-type indices and climate changes strongly agree with the observation on the top soils. The different performance of the Na-type and Sr-type indices with the climate condition in the TP should also be further studied in the future.

Declarations
Availability of data and materials All data generated or analysed during this study are included in this published article and its supplementary information les.