5.1 Evaluation of diagenetic influences
Assessing I/(Ca+Mg) and δ34SCAS values for potential diagenetic overprints is paramount to ensure the recorded signals represent changes in seawater. Importantly, a recently published study using δ44Ca and Sr/Ca ratios addressed the extent of diagenetic influence on both the Monitor Range and western Anticosti Island sections (Jones et al., 2020). This study suggests that deeper water settings like that of Monitor Range and western Anticosti Island, generally retain primary seawater geochemical signatures (i.e. sediment buffered), while strata deposited in shallower water settings are more likely to reflect geochemical signatures akin to sediment porewaters (Jones et al., 2020). Unfortunately, extensive diagenetic studies have not been performed on the Kardla drill core, Estonia, however this section is interpreted to be deposited in a deeper shelf setting (albeit shallower than the upper slope setting of the Monitor Range section), suggesting that this section likely preserves mostly primary geochemical signatures (Kaljo et al., 2011).
Meteoric diagenesis has been shown to decrease the concentrations of both iodine (Lu et al., 2010; Hardisty et al., 2017) and sulfate in carbonates, as freshwater typically contains lower concentrations of these ions. However, there are no known processes that can increase iodine in carbonates, and in the case of CAS, meteoric diagenesis itself imparts a negligible isotopic effect (Gill et al., 2008), however early diagenetic processes can still impart isotopic signatures. While there is an abundance of low I/(Ca+Mg) values recorded in our datasets, we interpret these as predominantly primary seawater signatures as significant diagenetic alteration cannot explain the very high I/(Ca+Mg) ratios, some of the highest in the early Paleozoic (Lu et al., 2018), recorded within the shallow marine patch reef facies of the western Anticosti Island section. These high I/(Ca+Mg) values were recorded during a lower stand of sea level in the Hirnantian, a stratigraphic interval, and carbonate facies that would have been most susceptible to extensive diagenesis (Fig. SI 3B). Studies of Cenozoic carbonates from the Great Bahamas Bank have shown that iodine concentrations may also be reduced during early diagenesis in carbonate settings (Hardisty et al., 2017). Intervals that were affected by meteoric diagenesis contained I/(Ca+Mg) values close to 0 μmol/mol (Hardisty et al., 2017), likely reflecting alteration by reducing fluids rather than primary seawater values. While it is possible that processes similar to these may have contributed to lowering general iodine concentrations in carbonates from this study, if fluid migration were to greatly affect primary geochemical signals it would be in the units that would have originally contained the highest porosity and lowest permeability (i.e. the shallow-water facies). In other words, the Lamframboise Member–Ellis Bay Formation, Anticosti Island and the Saldus Formation, Estonia, by this prediction would have low I/(Ca+Mg) values. However, these respective intervals within our carbonate successions contain the highest I/(Ca+Mg) values, while the lowest values are found in fine-grained carbonate and clay-rich facies where porosity and permeability would have likely inhibited fluid migration. Additionally, these Bahamian drill cores have shown other evidence for meteoric diagenesis in these intervals with near zero I/(Ca+Mg) values, including carbon isotopic signatures that are significantly depleted compared to the original aragonitic sediments that passively record primary seawater (Swart and Oehlert, 2018), whereas the Late Ordovician carbon isotopic data from our study sections do not show these types of signatures even surrounding intervals of glacioeustatic exposure (Brenchley et al., 2003; Desrochers et al., 2010; Young et al., 2010; Jones et al., 2016).
Geochemical crossplots are also a widespread tool used to assess the fidelity of geochemical signatures, where correlating trends with high R2 values can point to mixing of primary marine signal with those from diagenetic alteration. Recrystallization of carbonates during diagenesis can yield δ13C, δ18O, δ34S and I/(Ca+Mg) signatures that deviate from primary seawater values, reflecting a mixture of primary and secondary sources and producing linear or asymptotic relationships among the geochemical parameters (Ahm et al., 2018; Swart and Oehlert, 2018). Here we have cross-plotted δ18Ocarb, δ13Ccarb, [CAS], δ34SCAS, and I/(Ca+Mg) datasets, and show weak to no correlations, indicating that complete diagenetic overprint is absent from our datasets (supplemental material, Fig. S3). The only crossplots that show significant correlation are I/(Ca/Mg) vs δ18Ocarb and δ18Ocarb vs δ13Ccarb from Anticosti Island, however, these trends simply reflect two distinct data populations that internally do not correlate. Further, clear trends that continue across formational boundaries and major facies changes suggest that geochemical signatures found in these successions are largely primary and contain limited diagenetic alteration.
The likelihood of bulk CAS to faithfully record primary seawater sulfur isotope values has previously been called into question. Bulk δ34SCAS may incorporate sulfate from both primary and secondary carbonate phases potentially leading to more “noise” in isotopic datasets, while δ34SCAS from well-preserved brachiopod carbonate components in the same section show more invariant values (Present et al., 2015). Unfortunately, brachiopods within the sections presented here are relatively rare and thus not a viable option for performing high-resolution component-specific CAS measurements. Studies comparing early Cenozoic bulk CAS, planktonic foraminiferal CAS, and authigenic barite found that while species-specific foraminiferal data yielded vital effects up to ±1‰ versus barite, bulk CAS faithfully follow the recorded changes in secular δ34Ssulfate in both duration and magnitude (Toyama et al., 2020; Yao et al., 2020). Additionally, δ34SCAS records presented here across multiple paleobasins on separate paleocontinents from variable bathymetric depths show biostratigraphically, well correlated first-order trends (i.e. the fall δ34S in the late Katian and a return to heavier values within the late Hirnantian–Rhuddanian) also suggest preservation of primary seawater signatures (Fig. 5). While local diagenetic processes have may influenced the δ34SCAS data, it is very unlikely that each of the study sections would experience similar early and late diagenetic histories that resulted in similar first-order trends. These local diagenetic histories can more likely explain the smaller magnitude variations (~2-4‰) within and between δ34SCAS records from the study sites.