Diagenetic effects on strontium isotope ( 87 Sr/ 86 Sr) and elemental (Sr, Mn, and Fe) signatures of Late Ordovician carbonates

: Understanding the effect and extent of diagenesis on the isotopic compositions of Sr in marine carbonates is a critical prerequisite for their use to unravel past environments. Here, we explore the dominant controls on carbonate 87 Sr/ 86 Sr of a Late Ordovician section from the Monitor Range, USA. Our results reveal a distinct increase in 87 Sr/ 86 Sr from 0.70794 to 0.70830 in the mid-upper D. ornatus zone, which is markedly higher than the published datasets of contempor-aneous samples with a relatively lower and stable 87 Sr/ 86 Sr ratio of ~0.7079. These elevated 87 Sr/ 86 Sr ratios suggest a local and post-depositional overprint and cannot be interpreted to reflect the 87 Sr/ 86 Sr of the coeval seawater. Furthermore, 87 Sr/ 86 Sr exhibits statistically significant positive correlations with geochemical indicators for diagenesis ([Mn], [Fe], Mn/Sr, Fe/Sr), indicating that diagenetic alteration is the principal control on the observed radiogenic 87 Sr/ 86 Sr values. Using a numerical model of marine diagenetic fluid-rock interaction, we demonstrate that the observed Sr isotopic and elemental data can be best explained by the chemical variations in bulk carbonates associated with diagenetic alteration. Our results highlight that diagenesis may significantly alter the pristine 87 Sr/ 86 Sr ratios of carbonates than previously thought, although the samples satisfy the stricter geochemical criteria of Sr isotope preservation ([Sr] > 300 ppm, [Mn] < 300 ppm, [Fe] < 1000 ppm, Mn/Sr < 0.2, Fe/Sr < 1.6), pointing to the need for more caution when using bulk carbonate 87 Sr/ 86 Sr as a tracer of paleoenvironmental changes.


Introduction
The ubiquitous deposition of sedimentary carbonates along passive continental margins over the last 3.8 billion years provides the potential for long-term, continuous, and highresolution records of past environmental conditions [1] .These records can be examined by geochemical analyses of carbonates based on their stable isotope compositions (e.g., δ 13 C, δ 18 O, δ 238 U) and/or trace element contents (e.g., I/Ca, Ce/ Ce * ) [2−5] .In particular, carbonate strontium isotope ( 87 Sr/ 86 Sr) records could offer important constraints on the magnitude and timing of continental weathering, paleoclimate changes, and tectonic activities in the geological past [6−14] .In addition, 87 Sr/ 86 Sr stratigraphy has long been used as a chemostratigraphic technique to correlate strata over the last billion years [15−19] , especially when biostratigraphic data are deficient or unavailable.However, multiple factors can influence the extent to which the 87 Sr/ 86 Sr of bulk carbonates traces the seawater Sr isotope compositions at the time of deposition; for instance, post-depositional diagenesis is the most prevalent and significant mechanism that could alter the primary 87 Sr/ 86 Sr signatures of carbonate [15, 20−22] .As a result, diagenetic alteration must be carefully assessed prior to interpreting carbonate 87 Sr/ 86 Sr as records of secular changes in seawater 87 Sr/ 86 Sr.
The mid-late Katian immediately before the Hirnantian glaciation (latest Ordovician) is an excellent time interval during which to investigate the major controls on 87 Sr/ 86 Sr of bulk carbonates because large published datasets have shown a relatively stable and lower seawater 87 Sr/ 86 Sr value of ~0.7079 [6,15,16] .This low 87 Sr/ 86 Sr ratio has commonly been attributed to enhanced weathering of fresh volcanic rocks, leading to a more juvenile seawater 87 Sr/ 86 Sr value, as well as lowering of atmospheric CO 2 levels, global cooling, and the first significant ice age of the Phanerozoic (Hirnantian glaciation, ca.443 Ma) [6,11,23,24] .Thus, if the observed coeval 87 Sr/ 86 Sr records do not coincide with this value (i.e., ~0.7079), then other factors must have been involved to change the primary 87 Sr/ 86 Sr of the studied samples.
In this study, we examined the 87 Sr/ 86 Sr and elemental concentrations of Sr, Mn and Fe ([Sr], [Mn], [Fe]) in bulk carbonate rocks from the Monitor Range section situated in central Nevada, USA.We focus on this locality because it provides continuous and undisturbed Late Ordovician sedimentary records with well-established biochronological and geochemical data [25−28] , offering a promising opportunity to explore the potential factors that could have modified the original 87 Sr/ 86 Sr signals of carbonate.We develop a fluid-rock interaction modeling to further evaluate the resetting of 87 Sr/ 86 Sr and elemental concentrations of Sr, Mn, and Fe during postdepositional diagenesis and compare the model predictions with observed records to demonstrate how incorporating geochemical data alongside diagenetic model results can support the identification of pristine seawater 87 Sr/ 86 Sr ratios.

Geological settings
The Monitor Range section (39°12.660'N,116°24.257'W), in north-central Nevada, USA, is exceptionally exposed and allows high-resolution collection of samples spanning the Katian-Hirnantian interval.Previous studies of this section have established a robust stratigraphic framework, including biostratigraphy, sedimentology, and chemostratigraphy [25−32] .The biostratigraphy, based mainly on graptolites, is well constrained and composed of the D. ornatus and P. pacificus zones in ascending order for the Katian stage (Fig. 1).These biozones can be correlated regionally and globally with coeval successions [25,26] .
Sedimentologically, the D. ornatus zone consists mainly of clayey lime mudstone to brown-gray to dark brown calcareous mudstone interbedded with dark gray lime mudstone (Fig. 1).The overlying lower-middle P. pacificus zone is dominated by the lithology of thin-to medium-bedded dark gray lime mudstone and interbeds of clayey lime mudstone to calcareous mudstone.Stratigraphically higher, the upper part of the P. pacificus zone is composed predominantly of medium-to thick-bedded lime mudstone interbedded with thin calcareous mudstone and chert nodules, blebs, bulbous, and stringers.

Sample preparation
A total of 36 bulk carbonate samples spanning the mid-late Katian from the Monitor Range section are washed thoroughly with deionized water to remove surface contaminants and sliced into small pieces or chips to remove any visually weathered parts and altered materials.The dried sample pieces were then pulverized into powder (< 200 mesh) using a vibratory disc mill for geochemical analyses.

Analyses of elemental concentrations and 87 Sr/ 86 Sr
Approximately 0.12 g of each sample was dissolved in 30% acetic acid using a 10-mL precleaned centrifuge tube at room temperature to avoid dissolution of noncarbonate phases.The samples were then centrifuged to separate the supernatant from insoluble residues.Aliquots of clear supernatant were evaporated to dryness and brought up in 2% HNO 3 for elemental concentration analyses on an ICP AES instrument.Typical precision was better than 10% (2σ) based on  [26] .Solid curves are the locally weighted scatterplot smoothing (LOWESS) fits for the geochemical data.duplicate analyses of in-run check standards.
For the 87 Sr/ 86 Sr measurements, the sample supernatants were dried completely and taken up in 2.5 mol/L HCl for Sr purification.Purification of Sr was conducted following standard cation-exchange procedures [33] .The 87 Sr/ 86 Sr of purified samples was measured on a Thermo Scientific Triton thermal ionization mass spectrometer.The reported 87 Sr/ 86 Sr ratios were corrected for instrumental mass discrimination using a normal Sr ratio of 86 Sr/ 88 Sr = 0.1194, with typical precision better than 0.00001 (2σ).

Results
We observed distinct stratigraphic variations in 87 Sr/ 86 Sr and elemental concentrations (i.e., [Sr], [Mn], [Fe]) in the Monitor Range section (Fig. 1, Table 1).The 87 Sr/ 86 Sr in the lower D. ornatus zone declines from 0.70814 to 0.70794 over the interval of 0-34.7 m.Synchronously, [Sr] increases from 257.0 ppm to 700.6 ppm, accompanied by relatively low concentrations of Mn and Fe, thus resulting in low ratios of Mn/Sr and Fe/Sr in this interval. 87Sr/ 86 Sr then rises [Fe], Mn/Sr, and Fe/Sr ratios display congruent increasing trends in the corresponding interval.The 87 Sr/ 86 Sr then returned to a lower ratio of 0.70792 at 101.6 m at the top of the D. ornatus zone.Again, [Sr] displays an increasing trend coincident with prominent decreases in the [Mn], [Fe], Mn/Sr, and Fe/Sr ratios in this interval.Higher section, 87 Sr/ 86 Sr shows limited variations between 0.70792 and 0.70806 at the P. pacificus zone over the interval of 103.7-135.1 m, coinciding with lower and stable values of [Mn], [Fe], Mn/Sr, and Fe/Sr.We note that the Mn/Sr ratios are consistently very low, falling below 0.4 in almost all of the data and averaging approximately 0.19 (Table 1, Fig. 1).

Comparison with the published Katian 87 Sr/ 86 Sr datasets
To evaluate the validity of our newly obtained Sr isotope record from the Monitor Range section, we compared them with the published coeval 87 Sr/ 86 Sr datasets after calibration to the age framework based on Geologic Time Scale 2020 [34] .Sixteen out of 36 87 Sr/ 86 Sr values from the Monitor Range section vary between 0.70808 and 0.70830, showing a remarkable offset (with an average of 0.00029) with the coeval seawater 87 Sr/ 86 Sr ratio of ~0.7079 [16,19] (Fig. 2).These anomalously radiogenic values suggest that local effects have shifted the primary Sr isotopic signals and thus cannot represent the secular changes in 87 Sr/ 86 Sr of global seawater.

Controls on 87 Sr/ 86 Sr records of the Monitor Range section
Carbonates from geological records have suffered various degrees of diagenetic alteration since their deposition, which is the most common factor that can modulate the 87 Sr/ 86 Sr ratios and the degree to which they track the variations in coeval seawater 87 Sr/ 86 Sr [10,[20][21][22]35] .During diagenesis, the trace element concentrations of carbonates would change systematically with the importation of foreign Sr, which could significantly alter the original 87 Sr/ 86 Sr. To evalate the importance of diagenesis that may influence the 87 Sr/ 86 Sr records at the Monitor Range section, we examined the [Sr], [Mn], [Fe], Mn/Sr, and Fe/Sr ratios (Fig. 3).
[Mn], [Fe], Mn/Sr, and Fe/Sr ratios are important for evaluating the preservation of carbonate 87 Sr/ 86 Sr values, as Mn and Fe can act as tracers of nonmarine diagenetic fluids, such as continental, meteoric, and metamorphic fluids, which are commonly characterized by higher Mn and Fe concentrations and radiogenic 87 Sr/ 86 Sr ratios [21,22,35,36] .Post-depositional diagenetic processes commonly result in high 87 Sr/ 86 Sr ratios along with an enrichment of Mn and Fe and depletion of Sr in the final carbonates [10, 20-22, 35, 37] , as Mn and Fe have larger partition coefficients, leading to preferential incorporation into carbonate from diagenetic fluids [22,37,38] .The diagenetically altered carbonates, therefore, may be theoretically expected to have higher [Mn] and [Fe] with lower [Sr] and thus higher Mn/Sr and Fe/Sr ratios than the primary phase [10,[20][21][22] .Hence, the comprehensive study of 87 [43] , Shields et al. [44] , Saltzman et al. [16] , and Edwards et al. [15] , representing the best fit for the published Ordovician 87 Sr/ 86 Sr data [19] .The time scale of the Ordovician is based on Geologic Time Scale 2020 [34] , and the stage slices are from Bergström et al. [45] .Hi.: Hirnantian.
Comprehensive identification of primary carbonate geochemical signals Hu et al.
promising to assess the intensity of post-sedimentation alteration on the original 87 Sr/ 86 Sr of carbonates [38,39] .Previous studies have attempted to use [Sr], [Mn], [Fe], Mn/Sr, and Fe/Sr values as geochemical criteria for identifying the least altered carbonates.For instance, Burke et al. [18] constructed the Phanerozoic 87 Sr/ 86 Sr curve by using bulk carbonates with [Sr] > 200 ppm, and Denison et al. [35] used the values of [Sr] > 900 ppm, [Mn] < 300 ppm, and Mn/Sr < 0.5 for late Paleozoic shelf carbonates.Edwards et al. [15] analyzed the paired 87 Sr/ 86 Sr of well-preserved conodont apatite and bulk carbonates from the Ordovician and suggested that the original seawater 87 Sr/ 86 Sr can be faithfully recorded in bulk carbonate with [Sr] > 300 ppm, and Wang et al. [40] proposed that samples with [Mn] < 300 ppm, [Fe] < 1000 ppm, and Mn/Sr < 1 are capable of preserving signals of Permian seawater 87 Sr/ 86 Sr.When including the Fe/Sr ratios, Kuznetsov et al. [41] suggested that carbonates are likely to preserve the primary Sr isotopic composition if Mn/Sr < 0.2 and Fe/Sr < 5, and Rud'ko et al. [42] evaluated the 87 Sr/ 86 Sr of Upper Jurassic carbonates under stricter values of Mn/Sr < 0.2 and Fe/Sr < 1.6.Collectively, [Sr] > 300 ppm, [Mn] < 300 ppm, [Fe] < 1000 ppm, Mn/Sr < 0.2, and Fe/Sr < 1.6 can be considered as more severe and stricter geochemical criteria for Sr isotope preservation in Ordovician carbonate rocks.
In the Monitor Range section, all the samples have [Sr] > 300 ppm and vary between 332.4 and 1251.7 ppm with an average of 709.7 ppm (n = 35), except for a value of 248.1 ppm obtained in one sample at the base of the studied section (Fig. 1, Table 1).[Mn] is lower than 300 ppm and ranges from 27.6 to 293.0 ppm, with a mean of 125.2 ppm (n = 36).Twenty-four out of 34 samples have [Fe] < 1000 ppm and range from 19.5 to 2548.4 ppm with an average of 629.9 ppm (n = 34) (Figs. 1 and 3, Table 1).Correspondingly, the Mn/Sr of the Monitor Range section exhibits low values ranging from 0.04 to 0.46 with a mean of 0.19, while Fe/Sr ranges from 0.04 to 4.02 with a mean of 1.00 (Table 1).On the basis of the abovementioned geochemical criteria of preservation of the Sr isotope system in Ordovician limestones (i.e., [Sr] > 300 ppm, [Mn] < 300 ppm, [Fe] < 1000 ppm, Mn/Sr < 0.2, Fe/Sr < 1.6), the 87 Sr/ 86 Sr of the samples from the Monitor Range section with [Fe] > 1000 ppm, Fe/Sr > 1.6, and Mn/Sr > 0.2 may be considered to have been diagenetically altered.
Although some samples meet the geochemical criteria, their 87 Sr/ 86 Sr values are still higher than that of coeval seawater (~0.7079,Fig. 2), suggesting that diagenesis or other postdepositional processes remain the dominant controls on the observed 87 Sr/ 86 Sr.To further check whether these elevated 87 Sr/ 86 Sr ratios could be interpreted entirely by diagenetic alteration, we dissect the correlations of 87 Sr/ 86 Sr with [Mn], [Fe], Mn/Sr, and Fe/Sr (Fig. 3).The scatter plots of 87 Sr/ 86 Sr vs.
6 Developing a theoretical model for 87 Sr/ 86 Sr variation during carbonate diagenesis To more systematically and quantitatively constrain the influence of diagenetic alteration on the 87 Sr/ 86 Sr records at the Monitor Range, we develop a numerical open system fluidrock interaction model that involves the evolution of [Sr], [Mn], [Fe], Mn/Sr, Fe/Sr, and 87 Sr/ 86 Sr ratios during carbonate diagenesis to interpret the observed trends [20,21] .In this model, the concentration of element i (e.g., Mn, Sr, Fe) in the carbonate ( ) is expressed using the following equation: where and are the initial concentrations of element i in the carbonates and fluid, respectively.is the effective fluidrock distribution coefficient and is defined by the ratio of (i.e., = / ).N is the weight ratio of fluid to carbonate.), the effective fluid-rock distribution coefficient ( ), and the weight ratio of fluid to carbonate (N), which is described by the following formula:

Sr
The modeling outputs are illustrated in Figs. 4 and  5, providing a theoretical basis for the evolution of 87 Sr/ 86 Sr with [Sr], [Mn], [Fe], Mn/Sr, and Fe/Sr.The element concentrations in fluid ( ) used in this model are variable to capture a broader significance for the natural burial environments.For fluids with relatively high Sr concentrations, such as Sr-rich brine (e.g., < 10), the 87 Sr/ 86 Sr of carbonates could be easily altered via fluid-rock interactions with low fluid-to-rock ratios (N < 1, Fig. 4a).Our model predicts that 87 Sr/ 86 Sr would increase simultaneously with increasing [Mn], [Fe], Mn/Sr, and Fe/Sr ratios during diagenesis of carbonates (Fig. 5).When applying the modeling results to the studied samples, the observed covariations between 87 Sr/ 86 Sr and diagenetic indicators ([Mn], [Fe], Mn/Sr, Fe/Sr) can be largely reproduced through fluid-rock interactions (Fig. 6), further supporting that the anomalously radiogenic 87 Sr/ 86 Sr values in the Monitor Range section can be best explained by diagenetic alteration.

Implications of carbonate diagenesis for the Sr isotope paleoenvironmental proxy
Our results have two substantial implications for the explanation of Sr isotopic records in geological carbonates.First, our results demonstrate that post-depositional diagenesis can result in significant stratigraphic variations in 87 Sr/ 86 Sr of carbonate rocks that do not represent primary changes in isotopic   ).The and of the primary carbonates are assigned to be 0.7079 and 1300 ppm, respectively, comparable with those of the least-altered samples from the Monitor Range section.The effective fluid-rock distribution coefficient of Sr ( ) and 87 Sr/ 86 Sr of fluid ( ) are set to 1 and 0.711 [21,22] , respectively.and chemical compositions of the coeval seawater, although the samples satisfy the stricter geochemical criteria of preservation of Sr isotope systems (i.e., [Sr] > 300 ppm, [Mn] < 300 ppm, [Fe] < 1000 ppm, Mn/Sr < 0.2, Fe/Sr < 1.6).Second, the comprehensive examination of correlations between 87 Sr/ 86 Sr and diagenetic indicators ([Mn], [Fe], Mn/Sr, Fe/Sr) and modeling results reveal that the covariations between different elemental proxies provide a geochemical characterization for identifying the effects of diagenesis on carbonate 87 Sr/ 86 Sr records, highlighting that the combination of geochemical data and numerical modeling has the potential to improve our understanding of the fidelity of carbonate Sr isotope paleoenvironmental proxy.

Conclusions
The carbonates from the Late Ordovician Monitor Range section provide important constraints on the importance of diagenetic alteration controlling 87 Sr/ 86 Sr in bulk carbonate sediments.Analyses of these carbonates, where the 87 Sr/ 86 Sr of contemporaneous seawater has been well constrained, provide critical insights into our ability to reconstruct the 87 Sr/ 86 Sr of paleo-seawater from carbonates.The main conclusions of this study are summarized below: The bulk carbonates from the Monitor Range section record distinctly higher 87 Sr/ 86 Sr ratios than that of the coeval seawater, suggesting that post-depositional processes have shifted the primary Sr isotopic compositions significantly and cannot be interpreted to represent the variations in seawater 87 Sr/ 86 Sr values.
The samples exhibit statistically significant positive correlations of 87 Sr/ 86 Sr with [Mn], [Fe], Mn/Sr, and Fe/Sr, as well as [Mn] with [Fe] and Mn/Sr with Fe/Sr, which is consistent with our modeling results that predict 87 Sr/ 86 Sr trends to higher values with enrichment of Mn and Fe and depletion of Sr during diagenetic alteration.These observations and modeling results strongly suggest that the radiogenic 87 Sr/ 86 Sr values of bulk carbonate from the Monitor Range section could be fully attributed to diagenetic resetting.
Our results highlight that although the carbonates meet the stricter geochemical criteria of retention of Sr isotope systems (i.e., [Sr] > 300 ppm, [Mn] < 300 ppm, [Fe] < 1000 ppm, Mn/Sr < 0.2, Fe/Sr < 1.6), their original 87 Sr/ 86 Sr values may still can be diagenetically altered, urging caution in identifying Sr isotope variations as global until potential diagenetic-related changes can be excluded.We propose that multiple examinations, including correlations between 87 Sr/ 86 Sr and diagenetic indicators ([Mn], [Fe], Mn/Sr, Fe/Sr) combined with numerical modeling, can be utilized as a useful strategy to demonstrate the preservation of primary seawater 87 Sr/ 86 Sr values.150 [21,39] .Other constant parameters are same as in Fig. 4.

Fig. 4 .
Fig. 4. Modeling results illustrating the evolution of carbonate 87 Sr/ 86 Sr versus the weight ratio of fluid to rock (N) and [Sr] during fluid-rock interactions in an open system condition.(a) 87 Sr/ 86 Sr versus N, (b) 87 Sr/ 86 Sr versus [Sr].An increasing N value representing a greater volume of fluid has reacted with carbonate.The final 87 Sr/ 86 Sr of carbonate is simulated with different initial Sr concentrations of fluid (i.e., = 3, 13, 26, and 65).The and of the primary carbonates are assigned to be 0.7079 and 1300 ppm, respectively, comparable with those of the least-altered samples from the Monitor Range section.The effective fluid-rock distribution coefficient of Sr ( ) and 87 Sr/ 86 Sr of fluid ( ) are set to 1 and 0.711[21,22] , respectively.

Table 1 .
87Sr/ 86 Sr and elemental data from the Monitor Range section, Nevada, USA.
continuously in the middle D. ornatus zone and reaches a maximum of 0.70830 at 78.5 m.This increase in 87 Sr/ 86 Sr coincides with an overall decline in[Sr], showing a roughly reverse stratigraphic trend of the two.Additionally, the [Mn], Comparison of carbonate 87 Sr/ 86 Sr records in the present study with published datasets derived from conodonts.The gray solid line is the LOWESS curve of 87 Sr/ 86 Sr datasets from Qing et al.