Fractionation Mechanism and Flux Estimation of Strontium Isotopes During Basalt Weathering

The fluxes of metal cations and isotopes released by weathering of silicate rocks are crucial and a prerequisite for constraining geochemical fluxes to rivers and oceans. This study presents mineral and elemental compositions along with 87Sr/86Sr and δ88Sr data from a basaltic weathering regolith on Hainan Island, South China to elucidate Sr isotope fractionation and weathering fluxes. The 87Sr/86Sr ratios vary from 0.703936 to 0.706338 as a result of differential weathering of the minerals. δ88Sr values in the weathering regolith range from −0.29 to 0.37‰, with the majority of the weathering regolith having lower δ88Sr values than the parent rock. Sr is leached into the soil solution during plagioclase decomposition, while 86Sr is preferentially adsorbed on the surface of secondary minerals. As weathering progresses, smectite decomposes and kaolinite desorbs under weakly acidic conditions, releasing the previously adsorbed 86Sr into the soil solution. The differential weathering of kaolinite and smectite controls the δ88Sr values of the weathering regolith, with pH being an important determinant of isotope fractionation. Furthermore, Sr elemental fluxes (SrFlux) and Sr isotopic fluxes (δ88SrFlux) of this weathering regolith were calculated using a mass balance model, yielding mean values of 0.20 (mg cm−3 Myr−1) and 0.052 (‰ (mg cm−3 Myr−1)), respectively. The δ88SrFlux exhibits a nonlinear positive correlation with the Chemical Index of Alteration (CIA), indicating that enhanced weathering leads to significant stable Sr isotope fractionation at CIA values below 95%. Our research promotes the understanding of Sr recycling and the fractionation behavior of stable Sr isotopes during consecutive weathering.


Introduction
Chemical weathering of silicate rocks is a process of decomposition and alteration of parent rocks in the Earth's epigenetic environment.It acts as a first-order control on elemental mobilization and redistribution in the critical zone and on input fluxes to river waters and the oceans (Blum & Erel, 1995;Das et al., 2022;Dessert et al., 2001;Gaillardet et al., 1999;Li & Elderfield, 2013;Palmer & Edmond, 1989;Pearce et al., 2015).Effective constraints on continental weathering fluxes are crucial for understanding the mass weathering output of the Earth's surface and the global mass cycle.Modeling weathering fluxes based on weathering intensity can provide a quantitative framework for determining the contribution of terrestrial crustal weathering to the hydrochemical composition and input fluxes of river water, and consequently the oceans.Nonetheless, the quantitative assessment of continental weathering fluxes, particularly those from weathering regolith, remains a significant challenge.
Our understanding of continental weathering fluxes is largely derived from studies of large global rivers.The geochemical compositions of dissolved elements and ions in river water serve as proxies for estimating continental weathering fluxes (Chao et al., 2015;Das et al., 2022;Harris et al., 1998;Krabbenhöft et al., 2010;Krishnaswami et al., 1992;Palmer & Edmond, 1989;Pearce et al., 2015;Su et al., 2021).However, these indirect estimates are based on catchment averages that integrate responses operating under variable hydrological and temperature conditions as well as diverse catchment lithologies.The weathering fluxes calculated from catchment dissolved load are controlled by factors such as runoff, groundwater discharge, catchment lithology, surface temperature, physical denudation, atmospheric inputs, and anthropogenic activities, which are difficult to quantify and distinguish.Additionally, the varied lithologies within a catchment complicate the accurate determination of the influence of silicate rocks on the hydrochemical composition of catchment waters (Andrews & Jacobson, 2017, 2018;Andrews & Taylor, 2019;Jacobson & Blum, 2000;Pearce et al., 2015;Stevenson et al., 2016Stevenson et al., , 2018;;Wei et al., 2013;Zhang et al., 2019).In this study, we aim to directly calculate the changes in weathering output fluxes from silicate weathering regolith.Our direct calculation method aims to eliminate the uncertainties associated with previous studies, which relied on the hydrochemical composition of river water.
We have employed a novel approach to directly calculate metal cations and isotopic fluxes during weathering at the scale of a weathering regolith.This method provides a clear insight into the role of cation discharge through the weathering regolith in the hydrochemical composition of river water and further permits the estimation of elemental fluxes to the ocean.Specifically, we have calculated Sr elemental fluxes and Sr isotopic fluxes from a basalt weathering regolith.The regolith derived from alkaline earth metal-bearing basalts is particularly suitable for assessing continental weathering fluxes.Basaltic rocks, which constitute approximately 3.5% of the continental surface, are predominantly situated near the ocean, such as in volcanic arcs or on oceanic islands.Regarding the global continental distribution, basalts represent the following proportions of the area of each continent: Asia 4.2%, Africa 3.3%, Antarctica 0.9%, Australia 2.2%, Europe 2.8%, North America 4.0%, and South America 4.2% (Hartmann & Moosdorf, 2012).Basaltic rocks undergo weathering at a faster rate than felsic rocks, and as a result, they contribute substantially to the transport of elements and materials to the ocean through rivers (Dessert et al., 2003;Eggleton et al., 1987;Gaillardet et al., 1999;Li & Elderfield, 2013;Ma et al., 2007;Nesbitt & Wilson, 1992;Poitrasson et al., 2008;Wille et al., 2018).Furthermore, highly reactive alkaline (Earth) metal elements such as Sr, Li, Mg, Ca, and K are significantly mobilized during basalt weathering, resulting in metal isotope fractionation that aids in tracing silicate weathering processes and constraining weathering fluxes (Andrews & Jacobson, 2017;Chen et al., 2020;Huang et al., 2012;X.-M. Liu et al., 2014;Ryu et al., 2014Ryu et al., , 2021)).Sr is a water-soluble alkaline earth metal element that is readily released and mobilized from continental rocks during chemical weathering (Nesbitt et al., 1980;Nesbitt & Wilson, 1992).This element has four naturally occurring stable isotopes: 84 Sr, 86 Sr, 87 Sr, and 88 Sr, with rnatural abundances of 0.56%, 9.86%, 6.98%, and 82.59%, respectively (Vizer, 1989).The isotope 87 Sr is radiogenic and produced by the decay of 87 Rb, whereas 84 Sr, 86 Sr, and 88 Sr are non-radiogenic stable isotopes.The stable Sr isotope system is considered a suitable proxy for determining the sources and mobilization of Sr during weathering as well as for assessing the fluvial output flux of Sr.The significant fractionation of δ 88 Sr values during weathering enables the use of both δ 88 Sr and 87 Sr/ 86 Sr ratios to delineate chemical weathering processes and quantify weathering output fluxes.Therefore, evaluating the weathering fluxes of Sr and its isotopes is crucial for comprehending global weathering fluxes.However, a thorough understanding of the mechanisms that influence δ 88 Sr values changes during weathering is imperative before these values can be effectively applied to estimate Sr output fluxes from weathering regolith.
The stable Sr isotope system provides insights into mass-dependent isotope fractionation through various processes that remove Sr and other cations from the weathering regolith.These processes include incongruent silicate weathering, carbonate precipitation, clay formation and adsorption, and plant uptake (Andrews et al., 2016;Andrews & Jacobson, 2017, 2018;Bullen & Chadwick, 2016;Chao et al., 2015;Halicz et al., 2008;Krabbenhöft et al., 2010;H.-C. Liu et al., 2016H.-C. Liu et al., , 2017;;Shalev et al., 2017;Stevenson et al., 2014Stevenson et al., , 2016Stevenson et al., , 2018;;Su et al., 2021Su et al., , 2022;;Vollstaedt et al., 2014;Wei et al., 2013).The δ 88 Sr values in silicate-dominated river water are significantly higher than the average composition of terrestrial silicates, suggesting that stable Sr isotopes are fractionated during continental weathering (Andrews & Jacobson, 2017, 2018;Andrews et al., 2016;Chao et al., 2015;Das et al., 2022;de Souza et al., 2010;Krabbenhöft et al., 2010;Pearce et al., 2015;Stevenson et al., 2018;Su et al., 2021).This may happen through the preferential release of isotopically heavier Sr during incongruent weathering or by the incorporation of isotopically light Sr isotopes into secondary minerals.Nonetheless, the fractionation behavior of stable Sr isotopes in these processes remains controversial, which complicates our understanding of continental weathering fluxes.For example, several studies have suggested that primary mineral weathering preferentially releases heavier Sr isotopes into the solution and that minerals with higher solid-phase δ 88 Sr values may be more soluble (Andrews & Jacobson, 2018;H.-C. Liu et al., 2017;Pearce et al., 2015;Su et al., 2022;Wei et al., 2013).In contrast, de Souza et al. (2010) suggest that incipient weathering of silicate rocks may not induce significant stable Sr isotope fractionation, implying that the δ 88 Sr values of the released Sr mirror those of the parent rocks.Similarly, acid leaching experiments on basalt have documented no or limited stable Sr isotope fractionation between the leachate and the residual solids (Chao et al., 2015).In addition, the mechanism of stable Sr isotope fractionation during secondary mineral formation is not well comprehended.The formation of secondary minerals during weathering promotes either incorporation of 86 Sr into their lattice or adsorption onto their surface (H.-C.Liu et al., 2022;Stevenson et al., 2016;Su et al., 2022).However, Bouchez and von Blanckenburg (2021) documented negligible fractionation when Sr was adsorbed onto the exchangeable sites of clay mineral surfaces.Investigating solute processes in subglacial discharge from the Russell Glacier on the Greenland Ice Sheet, Andrews and Jacobson (2018) found no Sr isotope fractionation associated with clay mineral formation.Similarly, Su et al. (2022) reported that their sequential extraction experiment revealed no isotope fractionation during the adsorption of Sr into the exchangeable pool.Furthermore, to date, no studies have been reported on stable Sr isotope fractionation resulting from clay mineral desorption in weathering regolith.Consequently, the current comprehension of stable Sr isotope fractionation in a weathering environment is rudimentary.Clarifying the mechanism behind stable Sr isotope fractionation during weathering is imperative.
The aim of this study was to determine the weathering fluxes of bulk Sr and Sr isotopes in the basalt weathering regolith as a function of weathering intensity and to assess the influence of metal cations released from the weathering regolith on the hydrochemical composition of river water.To this end, we analyzed radiogenic ( 87 Sr/ 86 Sr) and stable (δ 88 Sr) Sr isotope data as well as elemental and mineral compositions along a 49.5 m basalt weathering regolith on Hainan Island.We also elucidated the mechanism of stable Sr isotope fractionation at different weathering stages.Our findings will enhance the understanding of stable Sr isotope behavior during continuous chemical weathering and provide a more accurate framework for global Sr isotope budgets and cycling.

Study Area and Samples
The field site of this study is located in the northern part of Hainan Island, South China.During the Cenozoic, voluminous basaltic eruptions occurred in the northeastern part of Hainan Island, covering an area of approximately 4,160 km 2 (Ho et al., 2000;Wang et al., 2012).Cenozoic volcanism on the island began in the late Miocene (approximately 13 Ma) and terminated in the Holocene (Wang et al., 2012).Volcanism in Chengmai (the present study area) occurred in the Pleistocene epoch (∼1.64 Ma, K-Ar (Zhu & Wang, 1989)) and produced quartz tholeiites and olivine tholeiites (Ho et al., 2000;Zhu & Wang, 1989).Many lateritic weathering profiles have developed on these basalts under the tropical oceanic climate (Ma et al., 2007).
The Chengmai weathering regolith (19°46′34″N, 110°00′23″E) is located approximately 4 km north of the city of Chengmai (Figure 1a) (Z.Liu et al., 2017).The climate in this region is tropical and controlled by the East Asian monsoon.The mean annual temperature is approximately 25°C, and annual precipitation ranged from 800 to 2,500 mm during the past century, with an average of approximately 1,500 mm (Ma et al., 2007).Quaternary sediments cover lateritic weathered products in the north of Hainan Island.A small hill was chosen as the sampling site to exclude the influence of slope sediments and Quaternary sediments.After removing the vegetation and approximately 30 cm of overlying loose soil containing plant roots, samples were collected by using a technique involving a dual-tube core barrel.The outer steel tube was attached to a diamond drill bit to dig the hole, and an inner PVC tube with a 10 cm diameter × 1 m length was used to collect samples.Each sample was collected within a PVC tube with a 1-m step as the drilling advanced.The total depth of the entire weathering regolith drilled in this study was approximately 49.5 m.The weathering regolith from 0 to 33.4 m was removed.The elemental and isotopic data, as well as the mineralogical composition of the saprolites, show that the 0-33.4m of the weathering regolith was significantly influenced by exogenous inputs and is not representative of the initial chemical weathering process.The exogenous source is not a single source but a superposition of multiple sources, including rainwater, aeolian dust, and felsic detrital-derived compounds (as justified in the Supporting Information).The weathering regolith from 33.4 to 49.5 m was divided into the upper regolith (33.4-41.4m) and the lower regolith (41.4-49.5 m) based on the presence or absence of the primary mineral plagioclase (Figure 1b).At the bottom of the profile is the tholeiitic parent rock.Below this regolith, marine sediments were drilled, indicating that the basalt lava erupted in a marine environment.
In the laboratory, the outer layer of the core sample was removed to avoid contamination by residue in the slurry.Samples were collected by cutting a 1.5 cm wide groove into the PVC tube of the inner cores at an interval of 20 cm.Sample CM-R was recovered from the bottom of the weathering regolith, and although CM-R is slightly altered (chemical index of alteration (CIA) value of 45.3%), it is still the least weathered sample in the entire weathering regolith.We therefore treat it as the basaltic "parent rock" to quantify the elemental and isotopic variations in the overlying saprolites.A total of 150 samples were collected along the entire regolith.All samples were measured for pH value, mineral composition, and elemental composition.A subset of 17 samples, representing different weathering stages, were selected to measure Sr isotopes.

Analytical Methods
The bulk laterite samples and basaltic bedrock sample were first dried at 80°C and then ground into a powder with a particle size of less than 200 mesh.Before geochemical analysis, the powder samples were again baked at 105°C to remove moisture.Each sample powder aliquot was selected separately for the determination of mineralogical compositions, major elements, trace elements, pH values, and Sr isotopes.

Major and Trace Element Measurement
Major elements were measured using a Rigaku ZSX100e X-ray fluorescence (XRF) spectrometer.The powder samples were first baked at 900°C for 90 min to remove organic matter and carbonates.A total of 0.500 g baked powder was mixed with 4,000 g Li 2 B 4 O 7 and fused at 1,200°C to make glass discs for XRF measurement.The analytical precision for major elements was better than 1%.The concentrations of K 2 O, Na 2 O, CaO, and MgO in some of these saprolite samples were close to the XRF determination limit.These elements were double checked by ICP-OES using the solution prepared for trace element measurement.
Trace elements were analyzed using a Perkin Elmer Elan 6000 ICP-MS.The sample powder was baked at 700°C to eliminate the organic materials.Approximately 0.04 g of each sample powder was weighed into a 15 mL Teflon beaker, and a mixture of 2 mL HNO 3 -HF-HClO 4 (HNO 3 : HF: HClO 4 = 1.25 mL: 0.5 mL: 0.25 mL) was added and digested at 190°C for 48 hr.The evaporated and dried solution was dissolved with 2 mL of 1:1 HNO 3 , evaporated and dried again, and then diluted with 2% HNO 3 for trace element determination.Rh was added to each sample as an internal standard for calibrating the drift of the instrument during the measurement.The analytical precision for the trace elements was generally better than 5%.

Mineralogical Composition Analysis
The mineral compositions of the bulk samples were determined using a Bruker D8 Advance X-ray diffractometer (XRD).Approximately 2.0 g of the bulk sample was weighed and dried at 80°C for 4 hr.The dried sample powder was placed into the sample tank, flattened with a coverslip, and fixed on the sample table.X-ray diffraction patterns of the samples were recorded between 3°and 85°(2θ) at a scanning speed of 4/min with Cu/Kα radiation (30 mA and 40 kV).X-ray diffractograms of the analyzed sample powders were compared with the powder diffraction data to complete the mineral identification.We have processed the XRD data using DMI JADE 6 software, which allows phase searches to be performed using PDF standard cards.The intensity of the diffraction lines of the physical phases increases with the content.We have therefore analyzed the samples quantitatively using software corrections to obtain the relative proportions of minerals in the samples.

Sr Isotope Analysis
Sr isotope analyses were carried out at the State Key Laboratory of Isotope Geochemistry, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences (GIG-CAS).Sr purification and isotope measurements followed the detailed method of Ma et al. (2013).Approximately 80 mg of each sample powder was weighed in precleaned PFA beakers and digested by a mixture of HNO 3 + HF.The beakers were then placed on a hot plate at 140°C for 7 days and evaporated to dryness.After that, 1 mL of concentrated HNO 3 was added to the beaker and dried.This step was then repeated.Finally, the sample was dissolved in 4 M HNO 3 for Sr purification on a 50-100 mesh from Eichrom Company, USA Sr-spec resin column.The prepared sample solution was loaded onto precleaned Sr-spec resin and then leached with 15 mL of 4 M HNO 3 to remove the matrix.Finally, Sr was collected by rinsing with 5 mL of Milli-Q water.Sr isotopes were measured using a Thermo Fisher Scientific Neptune Plus Multi Collector-Inductively Coupled Plasma Mass Spectrometer (MC-ICP-MS).The 87 Sr/ 86 Sr ratios were obtained by internally calibrating the mass bias with a ratio of 88 Sr/ 86 Sr = 8.375209 (Nier, 1938).The stable Sr isotope (δ 88 Sr) values were obtained by the sample-standard bracketing calibration method and δ 88 Sr values were defined as δ 88 Sr (‰) = [( 88 Sr/ 86 Sr) sample /( 88 Sr/ 86 Sr) SRM987 1] × 1,000 (Fietzke & Eisenhauer, 2006).Standard materials, including the seawater standard IAPSO, the coral standard JCp-1, and the international standard BHVO-2, were repeatedly chemically treated and measured along with samples.The 87 Sr/ 86 Sr ratios and δ 88 Sr values are, respectively, 0.709176 ± 0.000012 and 0.390 ± 0.018 (2SD, n = 12) for the IAPSO standards; 0.709172 ± 0.000012 and 0.196 ± 0.018 (2SD, n = 12) for JCp-1; and 0.703475 ± 0.000013 and 0.220 ± 0.020 (2SD, n = 6) for BHVO-2.

Measurement of pH
For the pH determination of saprolites, 5.0 g of sample powder was put into a PTFE clean capped bottle, 10.0 g Milli-Q water was added and left for 8 hr, and then the supernatant was collected by filtration (Wang et al., 2018(Wang et al., , 2020)).The pH values of the supernatant were measured using a Thermo Orion Star A (520M-01A) with a precision better than 0.02.

Elemental Concentration and Mineralogical Composition
The results of the analysis are shown in Tables 1 and 2, with all data presented in Tables S1 and S2 in Supporting Information S1.The relative mobilization of an element during weathering is quantitatively evaluated by the τ value of elements in saprolites, defined as follows (Nesbitt, 1979;Nesbitt et al., 1980): Geochemistry, Geophysics, Geosystems 10.1029/2023GC011215 where C Sr and C Ti represent the concentrations of target elements Sr and Ti, respectively, and "s" and "p" refer to saprolite and parent rock, respectively.Ti is chosen as a conservative element to normalize other elements.The results of the calculation of the mobility (τ value) of each major element are shown in Table S3 in Supporting Information S1.The chemical index of alteration (CIA) is calculated as follows:  (Nesbitt & Young, 1982).The major elements used for calculation are listed in Table S1 in Supporting Information S1.
where CaO* is the CaO content in the silicate fraction of the sample (Nesbitt & Young, 1982).The basaltic "parent rock" (CM-R) is the least weathered sample with a CIA of 45.3% (Figure 3).
The "parent rock" consisted of plagioclase (55.9%) and smectite (44.1%) (Figure 2).Plagioclase is predominant in the lower section of the regolith, decreasing from 55.9% at 49.5 m to 13.4% at 41.4 m, and is absent above 41.4 m.Smectite is one of the main secondary minerals and can be observed along the entire regolith, with high contents below 41.The Sr concentration is 391 μg g 1 in the "parent rock."The Sr concentration decreases sharply in the portion above the "parent rock", ranging from 6.7 to 205 μg g 1 (Figure 3, Table 1).The τ Sr,Ti values show that most Sr has been leached out of the regolith.Specifically, from the bottom of the profile to 33.4 m, the τ Sr,Ti values gradually decreased from 58.9% to 99.1%.The 87 Sr/ 86 Sr ratios reported here were normalized to a constant ratio of 88 Sr/ 86 Sr of 8.375209, largely eliminating the possible mass-dependent Sr isotope fractionation, which occurs naturally.The 87 Sr/ 86 Sr ratios vary significantly across the regolith, ranging from 0.703936 to 0.706338 (Figure 3, Table 1).Samples from 49.0 to 44.4 m display 87 Sr/ 86 Sr ratios similar to that of the "parent rock" (0.703936), with an average value of 0.704054.Above this section, the 87 Sr/ 86 Sr ratios gradually increase from 0.704216 to 0.706338 between 44.4 and 33.4 m.The δ 88 Sr value in the "parent rock" is 0.23 ± 0.014‰, which is consistent with analytical error with the previously published value of 0.28 ± 0.08‰ in fresh basalt (Andrews & Jacobson, 2017;Charlier et al., 2012;H.-C. Liu, You, Huang, & Chung, 2012;Ma et al., 2013;Moynier et al., 2010;Ohno et al., 2008).The total variation of δ 88 Sr values in this weathering regolith amounts to 0.66‰, ranging from 0.29 to 0.37‰ (Figure 3, Table 1).

Variation in 87 Sr/ 86 Sr Ratios
Sr is generally hosted in plagioclase, pyroxene, biotite, and the glassy matrix in fresh basalt.If the minerals in fresh basalt have uniform initial 87 Sr/ 86 Sr ratios when the lava formed, then the 87 Sr/ 86 Sr ratios in saprolites change because different 87 Sr/ 86 Sr ratios have been accumulated as a result of long-term decay of 87 Rb to 87 Sr in minerals with different Rb/Sr ratios.However, we consider the contribution of radiogenic decay accumulation in minerals with different Rb/Sr ratios in the parent basalt to the 87 Sr/ 86 Sr ratios of saprolite to be negligible, since the age of the "parent rock" is only 1.64 Myr.Therefore, the weathering of different minerals may be the major reason for the variations in the 87 Sr/ 86 Sr ratios in the weathering regolith.Previous studies have documented that the variation of the 87 Sr/ 86 Sr ratios in the regolith is mainly controlled by the decomposition of plagioclase, hornblende, and biotite during weathering (Bain & Bacon, 1994;Su et al., 2022).Primary minerals such as biotite and plagioclase release large amounts of Sr during incipient dissolution.In this weathering regolith, the 87 Sr/ 86 Sr ratios gradually increase with increasing weathering intensity (Figure 3).This result is consistent with the lower loss of radiogenic Sr during weathering of plagioclase and hornblende (Hajj et al., 2017).In the lower regolith (41.4-49.5 m), the variation in 87 Sr/ 86 Sr ratios may be dominated by the decomposition of plagioclase and the formation of kaolinite.The experimental study by Brantley et al. (1998) found that different Sr isotopes can be bound differently in feldspars, resulting in different release rates during weathering.The formation of finegrained clay minerals may have also influenced the 87 Sr/ 86 Sr ratios of saprolites.In the upper regolith (33.4-41.4m), the variation of the 87 Sr/ 86 Sr ratios is mainly controlled by the secondary minerals smectite and kaolinite.Unfortunately, we were unable to determine the detailed mechanism of secondary mineral control of the 87 Sr/ 86 Sr ratio, especially for smectite.In the regolith above 33.4m, the 87 Sr/ 86 Sr ratios of saprolites change significantly (Table S7 in Supporting Information S1).At this depth, the primary minerals are completely decomposed.The significant variations in 87 Sr/ 86 Sr ratios in this interval may have been significantly influenced by exogenous materials (Supporting Information S1).Therefore, the mineralogical composition and the 87  ratios, such as plagioclase.It is also possible that the formation of fine-grained clay minerals may have influenced the 87 Sr/ 86 Sr ratios of the saprolites, but we are unable to assess this at present.

Variation in δ 88 Sr Values During Incongruent Weathering of Silicate Minerals
Previous studies on meteorite, lunar, Martian, and terrestrial samples have found significant variations in stable Sr isotopes, with glasses from evolved terrestrial rocks extending to δ 88 Sr values as low as 0.2‰ (Charlier et al., 2012;Moynier et al., 2010), markedly lower than the terrestrial average of 0.28‰ (Andrews & Jacobson, 2017;Charlier et al., 2012;H.-C. Liu, You, Huang, & Chung, 2012;Ma et al., 2013;Moynier et al., 2010;Ohno et al., 2008).These  fluctuations in stable Sr isotopic compositions are considered to be related to plagioclase presence, which tends to have higher δ 88 Sr values (Charlier et al., 2012;Su et al., 2022).These findings imply potential fractionation among different mineral phases in high-temperature rocks.Consequently, stable Sr isotopic compositions in saprolites may change with the weathering stage during the incongruent weathering of basalt minerals.Nevertheless, laboratory leaching experiments on basalt have shown that limited stable Sr isotope fractionation could result from incongruent silicate weathering (Chao et al., 2015;H.-C. Liu et al., 2017).In this regolith, the "parent rock" contains plagioclase and smectite, indicating that pyroxene and/or olivine, along with the glassy matrix, have completely decomposed into secondary minerals.The δ 88  In samples containing plagioclase within the lower regolith (41.4-49.5 m), Sr concentrations decrease rapidly with decreasing depth, the 87 Sr/ 86 Sr ratios increase gradually, and plagioclase contents are positively correlated with δ 88 Sr values (Figure 4).In general, the δ 88 Sr values in samples with high plagioclase contents are heavier than those with lower plagioclase contents.For instance, samples with plagioclase contents exceeding 30% (excluding CM-47-4, which contains 30.9% plagioclase) exhibit positive δ 88 Sr values, whereas those with less than 30% plagioclase display negative δ 88 Sr values.This pattern suggests that plagioclase decomposition releases Sr, leading to the accumulation of lighter Sr isotopes in saprolite and leching of heavy Sr isotopes into the soil solution.This conclusion is supported by the study of Su et al. (2022).Furthermore, this process could explain the heavier stable Sr isotope values observed in river water relative to the bedrock (Andrews & Jacobson, 2017, 2018;Chao et al., 2015;H.-C. Liu et al., 2017;Shalev et al., 2017;Stevenson et al., 2016Stevenson et al., , 2018;;Vollstaedt et al., 2014;Wei et al., 2013).In the upper regolith (33.4-41.4m), plagioclase is entirely decomposed and half of the samples exhibit negative δ 88 Sr values.In addition, the 87 Sr/ 86 Sr ratios gradually increase with decreasing depth and τ Sr,Ti values remain essentially constant, suggesting continuous Sr leaching.These observations indicate that plagioclase decomposition releases heavy Sr into the soil solution, leading to the accumulation of lighter Sr isotopes in saprolite.Free heavy Sr isotopes are leached from the weathering regolith.

Variation in δ 88 Sr Values During Adsorption by Secondary Minerals
The mobility of Sr in subsurface saprolites is largely influenced using the adsorption-desorption process of clay minerals.Clay minerals in soils at different locations can have differing sorption capacities and behave differently due to variability in chemical composition.Similar to the isotope fractionation of other elements such as Mg, Li, K, and Rb, stable Sr isotope fractionation can occur through adsorption by clay minerals (H.-C.Liu, You, & Tu, 2012;H.-C. Liu et al., 2022;Su et al., 2022).In this regolith, the clay minerals include smectite and kaolinite.Smectite is prevalent and constant in the regolith but decreases above 40.0m.Below 41.4 m, smectite originates from plagioclase decomposition.Smectite contents are negatively correlated with δ 88 Sr values throughout the regolith (Figures 5a and 5b).
Our study shows that isotopically heavy Sr is released into the soil solution, whereas isotopically light Sr is adsorbed onto smectite surfaces.Sr is predominantly adsorbed onto partially charged surfaces or incorporated into the crystal lattice during the neoformation of clay minerals.Smectite represents a group of expandable 2:1 phyllosilicate minerals with two silica-based tetrahedral sheets and water molecules present in the interlayer sites.
Sr can be adsorbed on the smectite surface or incorporated into its interlayer sites.In the absence of ionic competition, mineral-bound Sr is stable in smectite (H.-C.Liu et al., 2022).However, the regolith is an integrated system rich in multiple ions and the presence of competing cations (e.g., Ca 2+ and Mg 2+ ) can lead to the saturation of smectite interlayer sites, altering the adsorption of Sr (Ahmad, 1995;Sugiura et al., 2021).Therefore, stable Sr isotope fractionation is likely to be caused by the adsorption of isotopically light Sr onto the smectite surface.In this regolith, Sr is adsorbed onto the smectite surface to form outer-sphere complexes (Atun & Kaplan, 1996;Fuller et al., 2016;Wallace et al., 2012;Yamaguchi et al., 2018).The interatomic distance between Sr and O on the smectite surface is slightly longer than that between hydrated Sr and O (H.-C.Liu et al., 2022).Therefore, the isotopically light Sr tends to adsorb onto the smectite surface, forming slightly shorter Sr-O bonds.This is supported by qualitative rules of stable isotope fractionation, which indicate that heavy isotopes of an element tend to be concentrated in substances in which that element forms strong chemical bonds (e.g., substances with low coordination numbers or short chemical bond lengths) (Schauble, 2004).In summary, the formation of weakly bound outer-sphere surface complexes and changes in bond lengths during Sr adsorption to smectite surfaces lead to stable Sr isotope fractionation.Above 41.4 m in the regolith, changes in weathering conditions resulted in the decomposition of smectite.Isotopically, light Sr previously adsorbed is also released into saprolites during smectite decomposition.
Kaolinite is another secondary mineral and the final clay mineral in the weathering regolith.The correlation between kaolinite content and δ 88 Sr values is variable (Figure 5c) and similar to the relationship between pH and δ 88 Sr values (Figure 5d)  86 Sr, previously adsorbed on kaolinite, into the soil solution, resulting in relatively heavier δ 88 Sr values in saprolites (Figure 5c), as kaolinite remains stable in this regolith.Kaolinite is a 1:1 clay mineral where each crystal unit is composed of one octahedral with one tetrahedral and possesses a pH-dependent negative charge due to the dissociation of protons from terminal OH groups (Deepthi Rani & Sasidhar, 2012).The quantity of negative charges on the edges and exposed basal hydroxyls depends on pH and other ionic concentrations, with higher pH giving rise to more negative charges.The lower regolith has a higher pH (6.2-6.9) and more negative charges.Thus, Sr 2+ binds to the kaolinite surface by exchanging into the relatively abundant octahedral sites on the surface or the less common octahedral sites around the edges.Sr 2+ may also be adsorbed as a partially aqueous ion by clay minerals (Parkman et al., 1998) and its adsorption by kaolinite induces stable isotope fractionation in saprolites.Similar Sr isotope fractionation was observed in a granodiorite weathering regolith by Su et al. (2022).The Sr-O bond lengths of the kaolinite surface are significantly longer than those in solution and Sr 2+ is characterized by high coordination (Parkman et al., 1998).Therefore, we conclude that 86 Sr is adsorbed on the kaolinite surface, while 88 Sr is more likely to leach into the soil solutions, a conclusion supported by the qualitative rules of stable isotope fractionation (Schauble, 2004).In the upper regolith, the weakly acidic environment (pH 4.9-5.9)promotes the desorption of 86 Sr, corroborated by the positive correlation between kaolinite and δ 88 Sr values.At low pH, both the particle edges and surface of kaolinite are positively charged (Palomino & Santamarina, 2005), leading to the desorption of 86 Sr previously adsorbed.Similarly, other alkaline metal elements with properties similar to Sr, such as Li, Mg, and Rb, are subject to adsorption and desorption processes that are also kaolinitedominated and pH-dependent (Huang et al., 2012;Zhang et al., 2021;Zhu et al., 2023).In summary, secondary minerals control the δ 88 Sr values in the saprolites of this weathering regolith, with smectite and kaolinite adsorbing light Sr isotopes in the lower regolith.As the regolith transitions to a weakly acidic environment higher up, smectite decomposition and kaolinite desorption release the adsorbed 86 Sr into the soil solution.

Mechanism for Stable Sr Isotope Fractionation During Weathering
Chemical weathering of basaltic rocks releases significant amounts of Sr into rivers and oceans.Combined with 87 Sr/ 86 Sr ratios, stable Sr isotopic compositions can serve as a powerful new tool for tracing Sr cycling in river water (Andrews & Jacobson, 2017, 2018;Bain & Bacon, 1994;Chao et al., 2015;Das et al., 2022;Su et al., 2021;Wei et al., 2013;Zhang et al., 2019).The significant fractionation of stable Sr isotope during Sr transport from continental rocks to rivers underscores the importance of understanding stable Sr isotope behavior during continental weathering to shed light on riverine stable Sr isotope variations.We describe the Sr isotopic composition of river water using the Sr isotopic composition of basalt weathering regolith at different weathering stages.Basaltic lava has lower Rb/Sr ratios (<0.1) and 87 Sr/ 86 Sr ratios (<0.706), and thus the contribution of 87 Sr from the decay of 87 Rb in basaltic lava is quite small.The limited variation in 87 Sr/ 86 Sr ratios (0.703936-0.704788)compared to the larger variation in 87 Sr/ 86 Sr ratios in river water (0.703-0.734) indicates that the input of 87 Sr/ 86 Sr to rivers from the incongruent weathering of basalts at different weathering stages is constant (Andrews & Jacobson, 2017, 2018;Bain & Bacon, 1994;Das et al., 2022;Stevenson et al., 2018;Wei et al., 2013).In contrast, significant stable Sr isotope fractionation occurs throughout the weathering regolith, suggesting that basalt weathering contributes substantially to Sr in the river.The decomposition of primary minerals releases Sr, with light Sr isotopes being retained in the regolith while heavy Sr isotopes leaching into the soil solution.The extent of Sr elemental leaching and stable Sr isotope fractionation are controlled by weathering intensity.Here, we define ∆ 88 Sr saprolite parent = δ 88 Sr saprolite δ 88 Sr parent , where δ 88 Sr saprolite and 88 Sr parent represent the δ 88 Sr values in saprolites and parent rock, respectively.Samples from the lower regolith have lower CIA values (<75%) and lower Sr mobility (τ Sr,Ti < 80%), and show a smaller magnitude of stable Sr isotope fractionation (∆ 88 Sr saprolite parent = 0.33-0.08‰).In comparison, samples with CIA values above 75% and τ Sr, Ti values over 80% exhibit more significant stable Sr isotope fractionation ranging from 0.52 to 0.14‰ (Table S3 in Supporting Information S1).These isotopically light saprolites are consistent with the heavy stable Sr isotope compositions observed in river waters.In conclusion, while limited stable Sr isotope fractionation is found during incipient to moderate weathering stages, much greater fractionation is likely to occur with increasing weathering intensity and Sr mobility.This is similar to observations of saprolites from a mud volcano in Taiwan (Chao et al., 2015;Su et al., 2021).Our research elucidates the relationship between the magnitude of stable Sr isotope fractionation and weathering intensity, which contributes to a deeper understanding of the geochemical behavior of Sr during weathering.

Estimation of Sr Weathering Fluxes
Elements released from weathering regolith are a key source of dissolved elemental species in global rivers, especially the alkaline earth metal Sr.The isotopic composition of dissolved Sr in rivers often differs from that of the continental rocks.Recent studies indicate that the δ 88 Sr values of river water are heavier than those in the rocks, suggesting that chemical weathering of terrestrial rocks is responsible for the isotopic shift (Andrews & Jacobson, 2017, 2018;Chao et al., 2015;Krabbenhöft et al., 2010;Pearce et al., 2015;Stevenson et al., 2016;Su et al., 2021;Wei et al., 2013).Accurate quantification of cation release from continental weathering regolith is critical.The weathering regolith in this study, from 33.4 to 49.5 m, provides an optimal segment for calculating weathering fluxes, as it consists entirely of weathered basalt.We employ a mass balance calculation as follows: The calculation of weathering fluxes during basalt weathering has been of great interest.We postulate that the basaltic parent rock undergoes complete decomposition, and the Sr element flux is calculated as follows: where Sr parent rock represents the Sr concentration of the parent rock, ρ denotes the average density of the basalt including voids and infilled vesiculated tephra (1.25 g cm 3 ; (Chadwick et al., 2003)), S c is the collapse factor, and t is the age of the basalt (1.64 Myr).In a similar vein, the Sr elemental flux from the complete decomposition of saprolites is calculated using the Sr concentration of saprolites: where Sr saprolite is the Sr concentration in saprolites.The Sr in the loss component is derived from the release after the decomposition of the parent rock during weathering, so the Sr elemental fluxes in the loss component can be estimated as: Our calculations of Sr Flux values range from 0.08 to 0.29 (mg cm 3 Myr 1 ), with an arithmetic average value of 0.20 (mg cm 3 Myr 1 ) as shown in Figure 7 and Table 3.The Sr released by basaltic decomposition seeps into the river from the weathering regolith, suggesting that cations released from silicate rock weathering significantly influence the hydrochemical composition of rivers.However, the precise contribution of Sr from basalt The calculated Sr isotopic fluxes range from 0.016 to 0.066 (‰ (mg cm 3 Myr 1 )), with an arithmetic mean value of 0.052 (‰ (mg cm 3 Myr 1 )), as detailed in Figure 7 and Table 3. Sr Flux and δ 88 Sr Flux have a solid linear positive relationship, indicating significant fractionation of stable Sr isotopes during element mobilization (Figure 8a).This is further supported by the linear negative correlation observed between δ 88 Sr Flux and τ Sr,Ti values (Figure 8c).The relationship between Sr Flux and CIA as well as δ 88 Sr Flux and CIA are both nonlinear and positive correlations.Below a CIA of 95%, the relationship between Sr Flux and CIA increases exponentially, whereas above 95%, the trend becomes monotonic (Figure 8b).A similar pattern is observed for δ 88 Sr Flux and CIA (Figure 8d).Our findings demonstrate that isotope fractionation and element mobilization are significantly controlled by increased weathering intensity at CIA below 95%.At CIA values above 95%, the chemical weathering intensity in this regolith could be categorized as extreme, with pronounced Sr leaching.This study elucidates the connection between isotopic fluxes and weathering intensity, enhancing our comprehension of chemical weathering fluxes and processes.However, it is crucial to acknowledge that uncertainties in weathering fluxes calculations can arise from various factors, including climate, tectonics, and weathering rates.Climate significantly impacts weathering fluxes, with basalt weathering intensity and Sr mobilization rates varying under different climatic conditions.Accurate determination of the influence of climate on weathering fluxes is both challenging and essential.In addition, material transport and denudation rates vary widely across tectonic zones, and the influence of uplift and denudation rates on weathering fluxes is a notable concern.Further research is needed to collect additional data to confirm the reliability and relevance of the weathering fluxes.
Alkaline earth metal elements have similar chemical properties.In this weathering regolith, we observe a coupled elemental behavior between Sr and Ca.The positive correlation between Sr and Ca concentrations (Figure S2a in Supporting Information S1) suggests simultaneous leaching of Sr and Ca during weathering.However, Ca leaches before Sr due to its higher reactivity.The calculated weathering fluxes for both elements corroborate this finding (Figure S2b in Supporting Information S1), leading to the conclusion that Sr and Ca exhibit a strong coupling in their behavior during basalt weathering.Ca and Sr can provide additional insights into weathering processes.In contrast, the isotopic behavior of Rb and K during granite weathering is decoupled (Teng et al., 2020;Zhang et al., 2021) Liu et al., 2022;Sugiura et al., 2021).In summary, the coupled behavior of stable Sr isotopes and Ca isotopes during basalt weathering can provide valuable insights into the weathering process.Notably, Ca 2+ serves as a critical indicator for quantifying the contribution of silicate rock weathering to the global carbon cycle (Andrews & Taylor, 2019;Beerling et al., 2020;Berner et al., 1983;Gaillardet et al., 1999).However, Ca is susceptible to leaching, often approaching 100% in advanced weathering stages, which hampers accurate CO 2 consumption estimates.In contrast, Sr leaching is less pronounced than Ca, and stable Sr isotopes are sensitive to weathering.Therefore, we propose that Sr could be a promising new proxy for constraining CO 2 consumption during intense weathering, although further research is necessary to confirm this potential.

Conclusions
The basaltic weathering regolith in northern Hainan Island, South China, provides a natural opportunity for examining the dynamics of minerals, elements, and Sr isotopes during chemical weathering.Based on these data, the following conclusions can be drawn: 1. Isotopically heavy Sr is leached into the soil solution during the decomposition of plagioclase, while isotopically light Sr is preferentially absorbed onto secondary mineral surfaces such as smectite and kaolinite.Under acidic conditions, smectite decomposes releasing isotopically lighter Sr, and kaolinite releases the same as it desorbs.2. The stable Sr isotope values of the loss component calculated by mass balance are consistent with the stable Sr isotopic compositions of silicate-dominated river catchments.The release of metal cations during the weathering of silicate rocks significantly influences the chemical and isotopic composition of river water in the catchment.3. The average Sr elemental flux is calculated to be 0.20 (mg cm 3 Myr 1 ) and the average Sr isotopic flux is 0.052 (‰ (mg cm 3 Myr 1 )).The results indicate a nonlinear positive relationship between Sr elemental fluxes and CIA values as well as between Sr isotopic fluxes and CIA values, suggesting intense Sr leaching under extreme weathering.

Figure 1 .
Figure 1.(a) Simplified geological map of Hainan Island, South China, and location of sampling, with the red star representing the location of the weathering regolith (modified from Z. Liu et al. (2017)); (b) Schematic diagram of the basaltic weathering regolith from Chengmai.The weathering regolith was divided into two sections based on mineralogical composition, elemental contents, and isotope values: the upper weathering regolith (33.4-41.4m) and the lower weathering regolith (41.4-49.5 m).The weathering regolith from 0 to 33.4 m was removed (as justified in Supporting Information S1).

Figure 2 .
Figure 2. Variation in mineralogic composition in the weathering regolith as a function of depth.

Figure 3 .
Figure 3. Variation in the concentration of Sr, Rb, and TiO 2 , Sr isotopic composition, τ Sr,Ti value, CIA value, and pH value in the weathering regolith.
. The negative correlation between kaolinite and δ 88 Sr values in the lower regolith (41.4-49.5 m) suggests that the light Sr isotopes are adsorbed by kaolinite.Conversely, kaolinite contents show a positive correlation with δ 88 Sr values in the upper regolith (33.4-41.4m).This may be due to the desorption of

δ
88 Sr parent = (1 ƒ) δ88 Sr saprolite + ƒ δ 88 Sr loss , (3) where δ88 Sr parent , δ88 Sr saprolite , and δ 88 Sr loss represent the isotopic composition of the "parent rock," saprolites, and the loss component, respectively.The factor ƒ represents the mass fraction of Sr in the loss component normalized to the immobile element Ti in the parent rock.The positive ƒ value in soil solution samples indicates Sr removal from the weathering regolith.The hydrochemical composition of river water in the silicate rock catchment in the study area remains unknown.Fortunately, data for Sr concentrations and stable Sr isotopic compositions of river water in silicate rock catchments in Taiwan, which were published byChao et al. (2015) andSu et al. (2021), provide a useful comparison.The climate, rainfall, and weathering conditions in Taiwan are similar to those on Hainan Island.Our calculated Sr isotope values for the loss component (δ88 Sr loss ) agree with the stable Sr isotopic composition of the river (Figure6), demonstrating that heavy Sr preferentially leached into the hydrosphere during incipient basalt weathering, while light Sr remains in saprolites.The Sr isotopic composition of the silicate rock weathering into the river remains constant over time and space, reflecting the isotope signature from the original weathering process.

Figure 6 .
Figure 6.δ 88 Sr, δ 88 Sr loss , and δ 88 Sr* discriminant diagrams.δ 88 Sr (blue dots) and δ 88 Sr loss (red dots) represent the Sr isotopic compositions of saprolites and loss component (calculated from mass balance) of the Chengmai basaltic weathering regolith, respectively.δ 88 Sr* (green and black dots) represents the Sr isotopic compositions of river water in the silicate rock watershed in southwestern Taiwan (data from Chao et al. (2015) and Su et al. (2021)).The thick gray line represents the average Sr isotopic composition of rivers globally(Krabbenhöft et al., 2010;Stevenson et al., 2018).

Figure 7 .
Figure 7. Variation in weathering fluxes of Sr and its isotopes in the weathering regolith as a function of depth.δ 88 Sr loss represents the stable Sr isotopic composition in the loss component.

Figure 8 .
Figure 8. Plots of Sr elemental fluxes (Sr Flux ) versus (a) Sr isotopic fluxes (δ 88 Sr Flux ) and (b) CIA values.Plots of Sr isotopic fluxes (δ 88 Sr Flux ) versus (c) τ Sr,Ti values, and (d) CIA values.In panel (a), Sr Flux is positively correlated with δ 88 Sr Flux , indicating that elemental mobilization causes significant isotope fractionation.In panel (b) and (d), Sr Flux and δ 88 Sr Flux are nonlinearly and positively correlated with CIA values, respectively, indicating that increasing weathering intensity leads to significant elemental mobilization and stable Sr isotope fractionation at CIA values of 65%-95%.The trend of the curve flattens out at CIA values greater than 95%, indicating that the weathering regolith is almost depleted of Sr and that further increases in weathering intensity result in only minimal elemental mobilization and very limited stable Sr isotope fractionation.In panel (c), δ 88 Sr Flux is linearly and negatively correlated with τ Sr,Ti values, indicating that elemental leaching from the weathering regolith causes significant stable Sr isotope fractionation at τ Sr,Ti values of 55% to 90%.Stable Sr isotope fractionation is limited when τ Sr,Ti values are greater than 90%.

Table 2
Main Mineralogical Compositions in Samples From the Weathering Regolith

Sr Concentrations, 87 Sr/ 86 Sr Ratios, and δ 88 Sr Values of the Weathering Regolith
4 m, ranging from 40.2% to 66.2%.Above 41.4 m, the smectite content decreases sharply.Kaolinite is another secondary mineral in this regolith but rarely occurs below 44.4 m.The kaolinite content gradually increases in the saprolites above this section, with an abundance range of 13.5%-78.6%.Ilmenite was observed between 49.0 and 35.6 m with a content of 2.1%-8.8%.Approximately 3.0%-15.0%opal was measured in this regolith from 49.0 to 35.6 m.
Sr/86Sr ratios suggest that the variation in87Sr/86Sr ratios is determined by the weathering reactions of minerals with different 87 Sr/ 86 Sr LUO ET AL.

Table 3
Sr Elemental Fluxes, Sr Isotopic Fluxes, and Other RelevantParameters in Samples From the Weathering Regolith 88 Sr Flux = δ 88 Sr loss × Sr Flux , a Sr concentration in the saprolites.bStable Sr isotopic composition in the saprolites.cInitial stable Sr isotopic composition before weathering of the parent rock.dStable Sr isotopic composition in the loss component.δ (Deepthi Rani & Sasidhar, 2012;behavior of Ca and Sr observed in our study.Unfortunately, we do not have Ca isotope data for this basalt weathering regolith.The relationship between Ca concentration and stable Sr isotopes in the weathering regolith closely mirrors that between Sr concentration and stable Sr isotopes (FiguresS2c and S2din Supporting Information S1).Sr is capable of substituting Ca in the lattice of Ca-rich primary minerals.The adsorption-desorption behaviors of Ca and Sr in secondary minerals are highly consistent, with Ca 2+ being a major competitive cation for Sr 2+ during adsorption by these minerals(Deepthi Rani & Sasidhar, 2012; H.-C.