Introduction

The Cretaceous period witnessed multiple episodes of intense volcanic activity, which are hypothesized to have acted as catalysts for global climate warming, shifts in ocean circulation patterns, and increased marine productivity. These oceanic anoxic events (OAEs)1,2 led to the widespread marine anoxia and the formation of organic-rich, black mudstone and shale deposits (total organic carbon (TOC) > 0.5 wt%). The Early Aptian OAE 1a is a notable example characterized by a significant increase in δ13C due to extensive accumulation of organic sediments in marine settings3,4. At the onset of the OAE 1a, an abrupt negative excursion in the δ13C curve (C3 segment sensu3) is commonly interpreted as a sign of isotopically light carbon entering the ocean-atmosphere system from marine volcanic emissions and the release of methane gas hydrates4,5. This negative spike is followed by a prolonged positive excursion in both organic and inorganic δ13C, which is recorded globally accross marine and terrestrial systems3,6.

During the OAE 1a, the Tethyan realm experienced various paleoenvironmental changes, including a warm climate interspersed with cooler periods7,8,9,10,11, enhanced chemical weathering6,12, carbonate platform drowning (northern Tethys13,14), the occurrence of Bacinella encrusters (central and southern Tethys15,16,17), and the development of restricted epicontinental seas (western Tethys18). These factors collectively contributed to to the development of OAE 1a, one of the most pronounced episodes of marine anoxia during the Cretaceous period19.

While the sedimentary, δ13C, and biogeochemical record of the OAE 1a is well documented globally, the connections between redox conditions and organic carbon burial remain poorly constrained20. Based on the analogy with modern oxygen-depleted marine environments, the redox-sensitive trace elements (RSTE) and rare earth elements plus yttrium (REY) are utilized as a proxy for the redox status of ambient waters21,22,23. They are commonly associated with organic matter, incorporated into sulfides, or bioconcentrated by planktonic fauna24,25,26,27,28. Since no single proxy can inarguably indicate a redox change and/or development of anoxia, it becomes imperative to adopt multiproxy strategies to unravel the paleoenvironmental evolution of OAE 1a. Furthermore, the majority of commonly utilized redox proxies are connected to redox conditions in bottom water and porewater.

To explore the climate sensitivity of ocean oxygen levels, both currently and across geological periods, a geochemical method to measure water oxygen content is essential. Iodine has emerged as a promising candidate due to its similarity to O2/H2O redox potential and its changing forms (iodide and iodate) in response to oxygen variations29. Iodate (IO3−) and iodide (I−) are the two thermodynamically favored species of inorganic iodine in seawater30,31. In the presence of oxygen, iodine is predominantly present as IO3−, and its conversion to I− occurs progressively with the reduction of dissolved oxygen29,32,33. Empirical investigations demonstrate the exclusive integration of IO3− into the crystalline structure of carbonate minerals, whereas I− is excluded32,34. The I/Ca ratio observed in contemporary planktonic foraminifera, approximately 5 µmol/mol, signifies adequately oxygenated water, while in oxygen-depleted regions, the I/Ca ratio significantly diminishes to 0.5 to 2.5 µmol/mol35,36. Consequently, the ratio of iodine to the combined concentrations of Ca and Mg in carbonates, denoted as I/(Ca + Mg) or I/Ca, serves as an indicator for assessing IO3− content in seawater, and thus dissolved oxygen concentration, during carbonate mineral formation. Elevated I/(Ca + Mg) values in carbonate rock samples correlate with higher oxygen concentrations in seawater32. However, this ratio can be diminished by diagenetic alterations, making I/(Ca + Mg) ratios in carbonate rocks conservative indicators, providing a lower limit estimation of IO3− presence (and free O2) during deposition37,38,39,40,41.

The Lower Dariyan Formation of the Kazhdumi Intrashelf Basin, Zagros Mountains, Iran offers a well-preserved archive of paleoenvironmental conditions during the Early Aptian OAE 1a along the NE margin of the Arabian Plate42. Throughout the Early Aptian, the basin, with maximum water depths of up to 135 m43, experienced continuous deposition of shallow-marine to hemipelagic facies, likely minimally impacted by meteoric diagenesis42,44 (Supplementary Information). The basin maintained a direct connection to the open ocean via a relatively narrow northern seaway43,45 (Fig. 1). Consequently, the Kazhdumi Basin is hypothesized to serve as a repository for a high-resolution record of regional paleoenvironmental changes, along with insights into climatic conditions42,44. This study investigates the environmental and diagenetic controls on I/(Ca + Mg) ratios in carbonate samples from the Lower Aptian Kazhdumi Basin. By integrating this proxy with other geochemical indicators (δ¹³C, Ce/Ce*, Mn/Ca, Fe/Ca), we aim to reconstruct paleoredox conditions during the OAE 1a. We hypothesis that, although diagenetic overprinting introduces variability, the I/(Ca + Mg) ratios—when interpreted alongside redox-sensitive trace elements and cerium anomalies—primarily record shifts in seawater oxygenation. These trends reveal a complex interplay among basin restriction, organic matter burial, and continental weathering as key drivers of paleoenvironmental change during the OAE 1a.

Fig. 1
figure 1

(A) Paleogeographic reconstruction of the western Tethys region during the Aptian (after88). (B) Aptian palaeogeography ot the Kazhdumi Intrashelf Basin with the location of the studied outcrop section modified from43. Figure 1 was originally generated in CorelDraw X7 by Abdus Saboor (co-author).

Full size image

Results and discussion

Environmental and diagenetic controls on iodine-to-calcium-magnesium ratios

The measured I/(Ca + Mg) ratios range from 0.005 to 1.30 µmol/mol (mean value 0.318 µmol/mol; Table 1). Most samples (> 80%) exhibit I/(Ca + Mg) values below 0.5 µmol/mol, whereas a small subset shows ratios above 0.6 µmol/mol. A positive correlation exists between I/(Ca + Mg) ratios and the manganese and iron contents (Figs. 2 and 3). The published δ13C values of the samples range from − 0.8 to 2.89‰ VPDB, with an average of 1.73‰ VPDB42,44. The negative δ13C values are not in phase with positive shifts in I/(Ca + Mg), nor do they correlate with Mn and Fe enrichment (Fig. 2). Mg/Ca and Mn/Sr values are mainly in the range of 0.01 to ~ 0.16 (mean 0.03 mol/mol), and ~ 0.02 to 1.5 (mean 0.16 ppm/ppm), respectively (Table 1). There is no correlation between I/(Ca + Mg) ratios and Mg/Ca, Mn/Sr, δ18O, Al, CaCO3, and TOC contents (Fig. 3).

Table 1 Results from the ICP-MS and ICP-OES analyses. Also shown are C-segments, and δ13C, δ18O, and total organic carbon (TOC) values from42,44.
Full size table
Fig. 2
figure 2

Lithologic log of the Lower Aptian succession, Kazhdumi Intrashelf Basin showing microfacies and depositional environments, depositional sequences42, C-isotopic segments (after3), carbon and oxygen isotopes42, TOC, I/(Ca + Mg) ratios, Fe/Ca and Mn/Ca (mmol/mol) values, and the evolution of key REY (rare earth plus yttrium) parameters. Red dashed line at 0.5 µmol/mol is considered the foundational baseline for I/(Ca + Mg) ratios in marine Proterozoic carbonates, based on numerous data points obtained from these carbonates38,39,48. MDST = lime mudstone, WKST = wackestone, PKST = packstone, FLST = floatstone, RDST = rudstone.

Full size image
Fig. 3
figure 3

Cross-plots of geochemical results from the studied succession. (A) δ13C vs. δ18O, (B) δ13C vs. microfacies, (C) I/(Ca + Mg) vs. Mg/Ca, (D) I/(Ca + Mg) vs. δ18O, (E) I/(Ca + Mg) vs. Mn/Sr, (F) I/(Ca + Mg) vs. Al, (G) I/(Ca + Mg) vs. CaCO3, (H) I/(Ca + Mg) vs. TOC, (I)/(Ca + Mg) vs. I, and (J) Mn/Ca vs. Fe/Ca.

Full size image

The pre-OAE 1a interval (upper C2 segment) exhibits a stepwise increasing trend of rather low I/(Ca + Mg) ratios that correspond to relatively high Ce/Ce* values (> 0.9), which suggests that the basinal waters were depleted in oxygen even before the onset of the global event. The latter is evident in the basin by a major negative δ13Ccarb excursion, and a decrease in I/(Ca + Mg) and Ce/Ce* ratios (segment C3, Fig. 2). Relatively uniform I/(Ca + Mg) ratios (mean 0.23 µmol/mol) characterize segments C4 and lower part of C5 + C6, suggestive of low oxygenation levels. A prominent stepwise increase in I/(Ca + Mg) ratios (from ~ 0.2 to ~ 1.3 µmol/mol) in the upper C5 + C6 segment is compatible with an increase in dissolved oxygen, which also matches an increase in Mn/Ca and Fe/Ca, and less pronounced increase in Ce/Ca ratios. The increase of these elements in basinal waters likely reflects an increase in continental weathering associated with relative sea-level fall at that stage (regressive systems tract of depositional sequence 2 sensu42). That continental weathering likely played an important role in increasing the amount of trace elements in basinal waters is also suggested by the relatively higher Mn/Ca, Fe/Ca, and Ce/Ca values in the upper C2 and C3 segments both of which form the regressive systems tract of the depositional sequence 1.

The range of I/(Ca + Mg) ratios from the Lower Aptian Kazhdumi Basin is similar to those recorded from the Proterozoic successions, albeit lower than the values from numerous Phanerozoic successions46, as well as from the modern well-oxygenated environments showing > 2.6 µmol/mol35,36. The values < 2.5 µmol/mol indicate the water-column iodate concentration of < 250 nM, which is commonly observed in marine environments, specifically within the oxycline situated directly above anoxic waters29,36. As a semi-quantitative guideline, an I/(Ca + Mg) value exceeding 0 µmol/mol may suggest the accumulation of iodate (IO3−) within seawater and an O2 level greater than 1–3 µM38,47. Conversely, the I/(Ca + Mg) value > 0.5 µmol/mol, representing a value higher than the baseline observed in the Precambrian time, signifies relatively well-oxygenated conditions39,48. But when the I/(Ca + Mg) ratio exceeds 2.6 µmol/mol, it indicates the absence of aquatic environments with O2 levels below 20–70 µM in contemporary shallow seawater36,38. Except for a small number of samples showing values near zero (Table 1), all the analyzed samples exhibit I/(Ca + Mg) ratios > 0 µmol/mol, suggesting porewater O2 conditions of at least 1–3 µM. Thus, the low but non-zero (mostly < 0.5 µmol/mol) I/(Ca + Mg) ratios of the studied samples are compatible with modern and ancient carbonates from suboxic and anoxic depositional environments.

The lack of field evidence for pervasive meteoric cementation, increased leaching, and/or dolomitization is supported by the δ13C values which do not suggest any pronounced impact of diagenesis in either the shallow or deeper-water facies of the Lower Dariyan Formation42 (Fig. 3A). Furthermore, there is no systematic variation of δ13C values (against microfacies), which mainly span from 0.6‰ to 2.80‰ VPDB (Table 1; Fig. 3B). This may point to the independency of δ13C values from the sedimentary (environmental) control. The weak correlation between Mg/Ca and I/(Ca + Mg) ratios (Fig. 3C) is compatible with the negligible impact of dolomitization on the I/(Ca + Mg) ratios. During the progression of meteoric diagenesis, a reduction in Sr and δ18O contents is typical, accompanied by concurrent increases in Mn and Fe concentrations49,50. The low δ18O values (Fig. 2) suggest that the original values underwent diagenetic alteration due to dissolution and the precipitation of micro-scale cement within a context of elevated temperatures in the deep burial environment42,50. The δ18O values generally vary between − 7.11‰ and − 4.44‰ VPDB and exhibit no distinct correlation with I/(Ca + Mg) ratios (Fig. 3D). This may suggest that, while a degree of (meteoric) diagenetic alteration is likely, the I/(Ca + Mg) values have not undergone substantial modification and consequently, these values can still serve as indicators of shifts in contemporary seawater redox conditions. The impact of meteoric diagenesis on I/(Ca + Mg) ratios becomes most evident when plotting I/(Ca + Mg) against Mn/Sr ratios (Fig. 3E). The majority of samples derived from the Lower Dariyan Formation exhibit indications of constrained meteoric diagenetic modification, characterized by Mn/Sr ratios < 1.5 (mean 0.16). The instances with the highest I/(Ca + Mg) ratios are observed in samples where Mn/Sr is < 1. Given the typical iodine concentration disparity between freshwater and seawater, where freshwater exhibits concentrations about one order of magnitude lower than seawater51,52, the admixture of freshwater and seawater would result in a reduction of iodine concentration53. This pattern implies that subsequent alterations likely contributed to variable decreases in I/(Ca + Mg) ratios. On a contrasting note, it is important to highlight that samples featuring notably high Mn/Sr content do not consistently exhibit correspondingly elevated I/(Ca + Mg) ratios (Table 1). The I/(Ca + Mg) values, albeit low, remain nonzero and are mainly distributed within the range of ≤ 0.5 µmol/mol (Table 1; Fig. 2; mainly from C4 to middle C5 + C6 segments sensu3), and suggest that specific instances of low I/(Ca + Mg) ratios might indeed represent original seawater signatures. In a broader context, our collected data support the fact that diagenetic processes may have fostered increased variability in I/(Ca + Mg) ratios, generally leading to lower values in some samples. However, the diagenetic alterations do not adequately explain occurrences of high I/(Ca + Mg) ratios38,39,40,53.

Finally, the presence of clay minerals and organic content is not likely to be a significant contributor to iodine in the studied carbonate samples. Our experimental approach involving the use of diluted acid (3% nitric acid32) helped to mitigate the liberation of iodine from organic matter and consequently, the I/(Ca + Mg) ratios of the samples are anticipated to be minimally influenced by potential organic matter contamination. This conclusion is supported by the weak correlations observed between I/(Ca + Mg) and aluminum content (R2 = 0.006; Fig. 3F), as well as between I/(Ca + Mg) and CaCO3 (R2 = 0.06; Fig. 3G) and TOC (R2 = 0.03; Fig. 3H). In contrast, a strong correlation exists between I/(Ca + Mg) and I content (R2 = 0.86; Fig. 3I), indicating that these values primarily capture the original seawater characteristics and/or the I/(Ca + Mg) ratio is regulated by the iodine content within the carbonates, rather than being dictated by the carbonate and TOC contents of the specimens39,54.

Cerium anomalies as possible redox proxies

Cerium (Ce) is distinguished from other rare earth elements (REEs) by its ability to adopt a tetravalent [Ce(IV)] state under oxidizing conditions, whereas other REEs remain trivalent55. This redox sensitivity allows Ce anomalies (Ce/Ce*) to serve as powerful paleoredox proxies. In modern oxygenated seawater, Ce3+ undergoes oxidation to Ce+4, mediated by manganese oxides and/or bacteria56. The insoluble Ce+4 is subsequently scavenged by Fe-Mn nodules and crusts, particularly in the deep ocean57. This process depletes seawater in cerium, creating a negative Ce anomaly (Ce/Ce* < 1), while concurrently enriching Fe-Mn oxides, resulting in a positive Ce anomaly. In anoxic waters, Ce remains primarily as soluble Ce³⁺, resulting in minimal fractionation, and potentially leading to weak or even positive Ce anomalies in associated sediments, especially when authigenic minerals preserve the seawater signal58,59. Consequently, lower Ce/Ce* values in our carbonates likely record incorporation of Ce-depleted seawater under oxic to suboxic conditions. Higher values (> 1.0), however, may signify either Ce retention under anoxic water column conditions or diagenetic modification within reducing porewaters. Overall, our interpretation of redox conditions relies on relative changes in Ce/Ce* values within the stratigraphic context, rather than absolute thresholds alone.

The basal part of the section (segment C2) shows uniform Ce/Ce* values (~ 0.85), followed by a stepwise decreasing trend into lower C4 where it reaches 0.67 (Fig. 2). This suggests an increase in seawater dissolved oxygen content before and at the onset of the OAE 1a, which is recorded in a predominantly carbonate interval with abundant benthic biota (ostreids, orbitolinids), extensive bioturbation, low productivity-sensitive trace elements (PSTE) and RSTE44. The Ce/Ce* values are highly variable in the middle and upper C4 and lower C5 + C6, reaching a maximum value of 1.17, i.e., the peak value of anoxia, in the upper Radiolarian Flood Zone (RFZ) (Fig. 2). The latter interval, represented by organic-rich deep open-marine and intrashelf basin pelagic dark gray shales and chert bands (TOC 4.5–9.0%60), is the intrashelf basin analog to the marine black shales. In the shallower Bab Basin and surrounding carbonate platform areas of the Arabian Peninsula, the coeval deposits record a significant expansion of the microencruster Bacinella that continued through C615,61. The remaining C5 + C6 segment overlying the RFZ exhibits a stepwise increase in Ce/Ce* values, suggesting a decrease in seawater-dissolved oxygen content up section. Coupled with relatively high δ13C values, lithology (dark to light gray deeper-water limestone and marl; Supplementary Information) with variable TOC, RSTE, and PSTE44, this is compatible with recurrent suboxic-oxic water conditions. Overall, the minimum Ce/Ce* value of the studied samples is 0.56, which is significantly higher than the modern-day Ce anomaly of 0.4 observed in the oxygenated upper ocean55 and close to the arbitrary cut-off value of 0.6 for poorly oxygenated Hauterivian-Aptian upper oceanic seawater62. Given the global distribution of Lower Aptian organic-rich shales2, the cerium anomaly observed in hemipelagic sediments of the Vocontian Trough62 likely reflects redox changes in water masses at a global scale. Although Ce has a short residence time in modern oceans (< 200 years63) and Ce anomalies are typically interpreted as indicators of local to regional redox conditions64, the anomaly recorded in Lower Aptian sediments of the Kazhdumi Basin was likely shaped by both regional and global deoxygenation during OAE 1a. The vast majority (97%) of the studied samples exhibit a Ce anomaly > 0.6 (mean: 0.84), indicating low oxygenation of the Aptian Kazhdumi Basin seawater (Fig. 2).

The Ce/Ce* vs. Pr/Pr* cross-plots are a useful tool when assessing whether Ce anomalies are genuine or were created artificially under the potential influence of lanthanum65,66,67 (Fig. 4). In the current study, only two samples, both from segment C5 + C6, exhibit positive Ce anomalies indicative of a genuine Ce enrichment (field IIIa in Fig. 4). Conversely, the majority of the samples exhibit negative Ce anomalies due to genuine Ce depletion (field IIIb in Fig. 4), which is similar to the coeval dataset from the Vocontian Trough of SE France62. Overall, the predominance of negative Ce anomalies suggests that the environmental conditions during the deposition were more conducive to Ce depletion rather than enrichment, assuming that the REY values incorporated into carbonate preserved an authigenic seawater signal59.

Fig. 4
figure 4

Lanthanum anomaly diagram (Ce/Ce* vs., Pr/Pr*) of the Lower Aptian succession, Kazhdumi Intrashelf Basin. La anomaly fields I-III after65.

Full size image

The alternating paleoredox conditions during the deposition of the Lower Dariyan Formation are also evident from the overall REY enrichment when compared to the modern oxic seawater68,69 and the overall positive Ce anomalies. In addition, the consistent and long-term decrease of the Ce/Ca ratio from an average value of 6.3 (before C3) to 3.5 µmol/mol (C3 and beyond) further suggests the shifting from the oxic-suboxic condition before OAE 1a into a more anoxic condition during OAE 1a, as indicated by other works on this proxy32.

The Y anomaly (Fig. 2) ranges from ~ 35 to 73.3 (mean 48.4), which suggests that the samples preserved an unaltered seawater REY signal, which is associated with Y/Ho > 3669. The Y anomaly is consistently low before and at the onset of the OAE 1a, and shows poor correlation with Ce anomaly, similar to modern seawater70. The Eu anomaly exhibits a broad range of values (Fig. 2), fluctuating between ~ 0.9 and 1.69 (mean 1.2), with high values (> 1.3) occurring in segments C4 and C5 + C6. These positive Eu anomalies possibly indicate periodically reducing conditions where Eu2+ was more soluble and mobile71. Shale-normalized Nd/Yb are consistently around 1, except for the three outliers in segment C3 that show values from 2 to > 3 (Fig. 2). This, along with a similar trend observed with shale-normalized Sm/Yb values, is suggestive of a significant periodic enrichment of LREEs compared to HREEs likely associated with scavenging of HREEs by organic matter72.

Redox sensitive elements

The Fe/Ca values range between 0.08 and 3.24 mmol/mol (mean 0.08 mmol/mol), while the Mn/Ca values vary from 0.02 to 0.4 mmol/mol (mean 0.02 mmol/mol), with both elements showing a notable positive correlation (R2 = 0.5; Fig. 3J). This correlation highlights their shared response to changes in redox states, as both manganese and iron are redox-sensitive elements.

Due to the redox-sensitive nature of both Mn and Fe elements, their covariation signifies a congruent reaction to changes in redox states73,74. In the initial stages of marine diagenesis, using oxygen and nitrate during the bacterial remineralization of organic matter coincides with the reduction of manganese. In dysoxic to anoxic environments, manganese depletion in sediments is attributed to the reductive dissolution of manganese oxyhydroxides, resulting in the mobilization of soluble manganese (Mn2+)75. Conversely, in oxygen-rich conditions, manganese predominantly forms less soluble compounds, adopting trivalent or tetravalent forms in the solid phase such as oxyhydroxides or oxides73. In extensively stratified redox settings, manganese is selectively deposited as Mn-oxyhydroxides in sediments located at or above the redoxcline. However, beneath this boundary, in sediments positioned lower, the element will experience reductive dissolution75,76. Consequently, fluctuations in Mn concentrations can transpire independently of sedimentary facies, sequence surfaces, and systems tracts, reflecting alterations in redox conditions. Hence, the enrichment of manganese within sediments serves as a potentially robust indicator for assessing paleoredox conditions77,78.

The relatively low Mn/Ca ratios in the studied succession suggest the absence of primary manganese-bearing minerals or the diffusion of soluble Mn²⁺ into the overlying water column28,79. The negative correlation between δ13C and Mn content (Fig. 2) supports the presence of authentic marine signatures, consistent with precipitation from primary marine fluids80. The simultaneous increase in Fe and Mn values, along with heightened Mn/Fe ratios, indicates precipitation within a reducing environment during burial diagenesis81,82 (Fig. 3J). This positive relationship also underscores a common source for Fe and Mn, reinforcing the idea of a shared origin for these elements.

Summary of redox evolution

In segment C2, geochemical signals indicate an initial phase of ocean deoxygenation (Table 2). The δ¹³C values begin to slightly increase, while I/(Ca + Mg) ratios remain low but show a stepwise increase, and Ce anomalies are relatively high (> 0.9). Periodically high Mo values suggest short periods of Mo removal into sediments under reducing conditions42. These trends are consistent with a transition from oxic to suboxic, likely driven by enhanced continental weathering and nutrient delivery to the basin. This phase sets the stage for the onset of Oceanic Anoxic Event 1a (OAE 1a).

Table 2 Summary of redox evolution in Kazhdumi intrashelf basin. Carbon-isotope segments C2–C6 after3.
Full size table

A sharp negative δ¹³C excursion defining segment C3 marks a major perturbation in the global carbon cycle, associated with large-scale volcanic CO₂ input and the release of methane gas hydrates4,5 (Table 2). During this interval, both I/(Ca + Mg) and Ce/Ce* ratios decline, indicating deterioration in preceding periodic suboxic conditions and return to well-oxigenated setting.

In segment C4 and the lower part of C5 + C6, δ¹³C values are characterized by a major positive excursion followed by relatively constant but high values, signifying increased burial of organic matter (Table 2). The low but non-zero (mostly < 0.5 µmol/mol) I/(Ca + Mg) ratios and periodically high Ce/Ce* ratios, coupled with high RSTE, PSTE, and TOC42 indicate suboxic to anoxic conditions, with little evidence for recovery.

Finally, in the upper portion of C5 + C6, the geochemical trends shift markedly (Table 2). δ¹³C values display little vriability, whereas I/(Ca + Mg) increases significantly—from around 0.2 to over 1.3 µmol/mol—in a stepwise manner and Ce/Ce* shows a slight rise. These changes suggest improving oxygenation, likely transitioning from anoxic or suboxic toward more dysoxic conditions. This redox recovery may reflect a combination of sea-level fall and enhanced weathering, promoting greater ventilation of the basin waters.

Methods

A 39.5-m-thick continuous hillside section of the Lower Dariyan Formation from the Kazhdumi Intrashelf Basin margin was studied near the village of Dareh Sefid, ~ 65 km north of Shiraz City, Fars Province. The major lithofacies types are summarized in Supplementary Information. A total of 62 bulk rock samples were collected every 0.6 to 1.3 m along the section (Fig. 2) and prepared for inductively coupled plasma mass spectrometry (ICP-MS) and inductively coupled optical emission spectrometry (ICP-OES). We also utilized recently published stable carbon and oxygen isotope data from this Sect42.

For iodine analysis (n = 62), ∼24 mg of MQ water-rinsed dry powders were utilized. 4 mL nitric acid (3%) was added for dissolution and then centrifuged to obtain supernatant. To stabilize iodine, 3% tertiary amine solution was added to the supernatant, and then diluted to 0.1% with MQ water. The iodine content was measured within 48 h to avoid any iodine loss, using an ICP-MS (Thermo iCAPTM TQ) at the National Research Center of Geoanalysis, Beijing. The sensitivity of iodine was tuned to ~ 50 kcps for a 1 ppb standard. Analytical uncertainties for iodine were monitored by a reference material JDo-1 (dolostone) and the detection limit of iodine is typically below 0.1 µmol/mol. For calcium (Ca), magnesium (Mg), manganese (Mn), and strontium (Sr) concentrations, supernatant was used and then diluted to 1:100.000 with 3% HNO3, were calibrated using JDo-1 and measured via ICP-MS (Thermo iCAPTM TQ) (Table 1). Rare earth elements and yttrium (REY) were analyzed by an Agilent 7700e inductively coupled plasma mass spectrometry (ICP-MS) on 55 samples at ALSGLOBAL Laboratory, China. Reproducibility was better than 5% for all REYs. Following the previous studies54,83, the powders of each (marl/shale) sample were rinsed with deionized water for two times to remove clay minerals and soluble materials, and then separated into aliquots for analysis. Their concentrations were standardized against the Post Archean Australian Shale (PAAS)84 (Table S1) to provide a comparative baseline. Ce anomalies were derived using the formula Ce/Ce* = [Ce/(Pr2/Nd)]SN85, while Pr anomalies were calculated as Pr/Pr* = [Pr/(0.5 Ce + 0.5 Nd)]SN86. Eu anomalies were determined by Eu/Eu* = [Eu/(0.67Sm + Nd)]SN65. The enrichment of middle rare earth elements (MREE) and light rare earth elements (LREE) was assessed by evaluating the ratios of (Sm/Yb)SN and (Nd/Yb)SN, respectively. Elemental analysis (no = 62; Mn/Ca and Fe/Ca) was carried out by inductively coupled optical emission spectrometry (ICP-OES; Table 1) using an Agilent 5110 VDV at University of Exeter, United Kingdom, following the methodology outlined in detail in Ullmann87. Bulk powders were weighed to 1 µg precision and 320 to 770 µg transferred into a 15 mL centrifuge tube. Samples were then digested in 2% v/v HNO3 with a target of achieving 25 µg/g nominal Ca concentration and dilution factors gravimetrically controlled. Ca concentrations and element/Ca ratios were calibrated against a set of a blank solution and three matrix-matched multi-element standards prepared from certified single-element plasma standards. Internal consistency of the measurements was checked with a quality control solution made of the same single-element plasma standards and the bias of the measurements was 1% or less for all element/Ca ratios. Repeat analyses of a stock solution of JLs-1 (n = 24; 2 s.d.) provided further reproducibility control yielding Mn/Ca of 0.032 ± 0.001 mmol/mol, and Fe/Ca of 0.17 ± 0.01 mmol/mol.