Changes in palaeoclimate and palaeoenvironment in the Upper Yangtze area (South China) during the Ordovician–Silurian transition

Mineralogy

The XRD analysis results are illustrated in Table 1 and Fig. 4. No significant qualitative difference was found in the mineralogical composition of samples from the Wufeng- Longmaxi formations. The samples are composed mainly of quartz (Fig. 2E, 22.1–71.3%, average of 39.1%), clay minerals (11.2–38.4%, average of 20.6%), calcite (4–28%, average of 17.7%) and dolomite (2.3–27.2%, average of 14.3%). Moreover, minor albite (1.6–6%, average of 3.6%), pyrite (0.5–5.9%, average of 3.3%) and microcline (0.4–3.4%, average of 1.3%) occur in the samples. The mixed layers of illite/smectite (43.0–65.0% average of 53.6%) and illite (22–40%, average of 30.8%) are the dominant constituents of the clay mineral contents of the Wufeng-Longmaxi formations. In addition, small numbers of minerals, including pyrite, molybdenite, and ilmenite, were detected simultaneously by XRD and scanning electron microscopy (SEM). Pyrite is mainly present as typical framboids and single euhedral pyrites in the studied samples (Fig. 2F).

Geochemistry

Provenance and tectonic setting

In general, some heavy minerals, such as zircon, are enriched gradually during sedimentary recycling⁶². According to Taylor et al.⁶⁴, the Th/Sc ratios reflect bulk source compositions, because Sc and Th are transferred quantitatively from the source to the sediment. McLennan⁶³ suggests that Zr/Sc ratios can be used as tracers for zircon or heavy mineral concentrations, as zirconium is mainly concentrated in zircons, in which less resistant minerals are preferentially destroyed.

Therefore, the Th/Sc and Zr/Sc ratios in this study and their bivariate plots were used to derive their sediment recycling degrees and mineral composition varieties^63,64. The Th/Sc and Zr/Sc ratios of these three formations fluctuated from 0.69 to 4.39 (average of 1.34) and from 5.64 to 43.51 (average of 9.76, Table 4). Both of these ratios indicate a lower fractionation, and the composition is close to PAAS. These data combined with the Th/Sc-Zr/Sc bivariate diagram (Fig. 8) indicate that during the deposition of the Wufeng-Longmaxi formations, sedimentary recycling was minor, and few clastic components came from older sediments. Furthermore, the sedimentary recycling of the studied samples is low, indicating that these geochemical data can be used to identify the provenance.

Geochemical data of detrital sediments have stable geochemical properties during weathering, transportation, and diagenesis; therefore, they can provide reliable information on provenance^65,66. To infer the provenance of sedimentary rocks, several authors have proposed major-based (e.g., Al₂O₃, TiO₂, and K₂O) discrimination diagrams in various studies of unknown basins^67,68. According to Hayashi et al.⁶⁹, the Al₂O₃/TiO₂ ratio change in igneous rocks is as follows: (a) the values of Al₂O₃/TiO₂ in mafic igneous rocks vary from 3 to 8, (b) the values of Al₂O₃/TiO₂ in intermediate igneous rocks fluctuate from 8 to 21, and (c) the values of Al₂O₃/TiO₂ in felsic rocks vary from 21 to 70. The Al₂O₃/TiO₂ ratio of the shales in the Wufeng Formation varies from 13.65 to 41.11 with an average of 22.11, the Al₂O₃/TiO₂ ratio of the carbonaceous marls in the Guanyinqiao Formation varies from 18.12 to 41.11 with a median of 19.22, and the Al₂O₃/TiO₂ ratio of the shales in the Longmaxi Formation varies from 20.375 to 60.44 (average of 23.59). These values suggest that the source rocks of the above formation are mostly between intermediate and felsic igneous rocks (Fig. 9A).

REE distributions in sedimentary rocks have played a vital role. The stable and unaffected properties during weathering, erosion, and early diagenesis make them valuable to provenance, leading to their special utility in tracing the source of sedimentary rocks^70,71. The Eu anomalies in sediments are generally considered to have been inherited from the source rocks^72–74. Generally, the LREE/HREE ratio of mafic rocks is low, and there is no Eu anomaly, while the LREE/HREE ratio of felsic rocks is usually high, and the Eu anomaly is significant^75,76. The normalized abundance and pattern of chondrites indicate that all samples are characterized by LREE enrichment, HREE deficits, and distinctly negative Eu anomalies (0.46–0.92; average of 0.6). All of these results suggest that felsic source rocks are the major source rocks for the Wufeng-Longmaxi formations.

In addition, several stable elements (e.g., La, Th, Hf, Yb, Zr, REEs) have been used to deduce the provenance of sedimentary rocks due to their immobility during sedimentation^76–78. On the bivariate diagram of La/Th-Hf⁶⁷, most samples plot in or near the felsic source field (Fig. 9B). In addition, the bivariate diagram of La/Yb vs. ∑REE reflects that the Wufeng-Longmaxi samples mainly plot in the granite source rocks (Fig. 9C). Overall, the provenance discrimination diagrams reveal that felsic (granitic) source rocks are the major source rocks for the studied Wufeng-Longmaxi formations.

Numerous studies have shown that the geochemical characteristics of detrital rocks are significantly controlled by the plate tectonic setting of the source area, therefore, the tectonic setting of the ancient terrains has been widely identified by using major-, trace- and rare earth element-based discrimination diagrams^79–81. Bhatia and Crook⁷⁰ proposed trace element-based discrimination diagrams to differentiate four tectonic settings: continental island arc, oceanic island arc, active continental margins, and passive margins. As shown in Fig. 10, most of the studied samples fall within or adjacent to the continental island arc and active continental margin domain in diagrams of La–Th–Sc, Th–Co–Zr/10, and Th–Sc–Zr/10.

To increase the success rate of identifying the tectonic setting, Verma and Armstrong-Altrin⁸² used common oxides (SiO₂, Al₂O₃, Fe₂O₃, MgO, CaO, Na₂O, K₂O, TiO₂, P₂O₅, and MnO) to develop two multidimensional diagrams based on the log_e-ratio transformation of major oxides to discriminate arc, continental rift, and collisional settings. All the major oxides must be adjusted to 100% after excluding the loss on ignition (LOI) and regarded as (X)_adj, where X represents the major oxides. These diagrams can be divided into two types based on the difference in (SiO2)_adj values, low-silica type (36%-63%), and high-silica type (63–95%). More details about the calculated equations for these two types of sediments are described in Verma and Armstrong-Altrin⁸². Recently, these diagrams have been successfully applied to discriminate the tectonic setting of older sedimentary basins^83–86. In the present study, the (SiO₂)_adj contents of the Wufeng-Longmaxi formations vary from 32.47 to 87.15, which can be classified as low-silica samples (n = 41) and high-silica samples (n = 13). For the high-silica diagram (Fig. 11A), eleven high-silica samples from the Wufeng and Longmaxi formations plot in the collision field, with two samples from the Guanyinqiao Formation (XD2P-B31 and XD2P-B29) plotting in the arc field. On the low-silica diagram (Fig. 11B), all forty-one low-silica samples plot in the collision field. All the discriminant-function diagrams above reflect that the sediments of the Wufeng and Longmaxi formations mainly originated from a collisional setting, while the Guanyinqiao Formation may be derived from rift and collisional settings.

Palaeoweathering indices and palaeoclimate implications

The intensity of chemical weathering is controlled mainly by the source rock composition, during weathering, climatic conditions, and rates of tectonic uplift of the source region (e.g., Refs.^87,88). During chemical weathering, alkaline metal elements such as Ca, Na, and K are largely removed from source rocks, while the concentrations of Al, Si, and Ba increase in the residue^89,90. The degree of weathering can be quantified by using mobile and immobile element oxides such as Na₂O, CaO, K₂O, and Al₂O₃. The ternary diagram of Al₂O₃ − (CaO^* + Na₂O) − K₂O (A–CN–K) (molecular proportions⁹¹) has been widely used to evaluate the differences in chemical composition related to chemical weathering, which can also be used to analyse the weathering history and palaeoclimate^92–96. As shown in Fig. 12, the samples of the three formations merged above the plagioclase-K-feldspar join, showed a narrow linear trend and were close to the muscovite and illite fields, which are consistent with the XRD results (Table 1). The weathering trend is parallel to the A-CN borderline and does not show any tilt to the K apex, indicating that the chemical weathering conditions are relatively stable, and excluding the effect of potassium salt metasomatism during diagenesis, which corrects CIA values for further analysis.

Among the different indices of weathering, the CIA is widely used in recent studies to quantify the degree of weathering^97–99. Young and Nesbitt¹⁰⁰ proposed that the degree of chemical weathering is related to the climatic conditions in the source area, and is considered a method to study palaeoclimate changes. This discovery also makes CIA indicators widely used to reconstruct palaeoclimate conditions^20,97–99.

The CIA values of the Wufeng-Longmaxi formation samples were calculated according to the method of McLennan⁶², and the results are listed in Table 2. Across the Ordovician–Silurian boundary, the CIA values of black shales from the Katian stage (from D. complanatus through P. pacificus biozones) are relatively consistent (Fig. 3), ranging from 67.48 to 73.37 (average of 69.72). In the lower M.extraordinarius zone (early Hirnantian), the CIA values, although they apparently decrease to a low, show a more fluctuating pattern in the M. extraordinarius-M. persculptus zone, varying between 58.30 and 68.74 (average of 62.62). At the top of the M. persculptus zone, they reached 64.66. Farther up, the CIA values return to persistently high values, from 63.47 to 72.85 (average of 67.77) in the Rhuddanian stage (from the A. ascensus through C.vesiculosus biozones).

In general, a higher CIA value (80–100) indicates tropical and humid climate conditions with strong chemical weathering, a medium CIA value (60–80) indicates moderate chemical weathering accompanied by warm and humid climate conditions, and a lower CIA value (50–60) indicates the palaeoclimate conditions of weak chemical weathering with a cold and arid climate^60,61,97. The lower CIA value (67.48–73.37) of the lower Wufeng Formation (D. complanatus—M. extraordinarius biozones) indicates a moderate degree of chemical weathering, which may reflect the warm and humid climatic conditions in the source area. Farther up, the CIA values during the Hirnantian stage (M. extraordinarius–M. persculptus biozones) first decreased from 69.23 to 61.51, then suddenly increased to 68.74, and then still showed a fluctuating pattern (58.30–64.66). These results indicate that the climatic conditions were mainly cold and dry before deposition, and the sediments experienced weak to moderate weathering, which is consistent with widespread Hirnantian glaciation^101–103. However, based on the data fluctuations, there may also be short-term pulses of climate warming during this interval. In summary, during the Hirnantian stage, the climate fluctuated rather than was persistently cold and arid.

A similar scenario has been reported in previous studies^97,104,105, including approximately five glacial-interglacial cycles during the glacial period with an estimated duration of ~ 500¹³ to ~ 1.0-Myr¹⁰⁶. As shown in Fig. 3, our fluctuating but relatively low CIA values correlate with a pronounced δ¹³C positive excursion, which represents a match with the global δ¹³C Hirnantian excursion (HICE)^107,108. Following this minimum, the CIA values increased above the Early Rhuddanian (at the bottom of the A. ascensus biozone), indicating that chemical weathering gradually increased, but compared to the pre-glaciation sediments (lower part of the Wufeng Formation) the CIA values remained relatively low, which means that the post-glacial climatic conditions remained relatively cool.

Palaeoredox conditions

Many trace elements, including U, V, Mo, Ni, and Co, show that their oxidation state and solubility change along with the variations in the redox state of the sedimentary environment; thus, they have been widely used to reconstruct palaeoredox conditions of sedimentary rocks and marine sediments^109–113. Among such elements, uranium (U) and molybdenum (Mo) have been extensively studied due to differences in their geochemical behaviour, and their covariation is considered to be related to specific redox conditions and processes in marine depositional systems¹¹⁴. The absorption of authigenic uranium by marine sediments starts under suboxic conditions, while the enrichment of authigenic molybdenum requires the presence of H₂S (i.e., euxinic conditions)^115–118. Moreover, aqueous U is completely unaffected through the particulate Mn/Fe-oxyhydroxide shuttle, but aqueous Mo may be enhanced during this process^119,120. Based on the above differences, the autogenous U–Mo covariance mode can be used to track the redox conditions, which are determined by its EF (see definition in “Major element”). In addition, some redox indices (U/Th, V/Cr, Ni/Co, and V/V + Ni) have been widely used to derive information on the palaeo-oxygen level of depositional environments^121–125. All four indices were calculated and reported as stratigraphic variations in Fig. 14.

All palaeoredox proxies indicate significant changes in redox conditions during the Late Ordovician to early Silurian, which can be identified as five distinct parts (Figs. 13, 14). For part I, during the accumulation of lower Wufeng formation (D. complantus, D. complexus, and lower part of P. Pacificus graptolite zones), all of the samples exhibit little or no enrichment for Mo and U, with (Mo/U)_auth ratios ranging from 0.05 to 1.33 with a median of 0.13, which are mostly between 0.1 and 0.3 times those of seawater (Fig. 13). The U/Th, V/Cr and Ni/Co values range from 0.16 to 0.83 (average of 0.35), 1.36–6.05 (average of 2.48) and 3.48–7.79 (average of 6.24), respectively, pointing to oxic to suboxic bottom water (Fig. 14). The above data suggest fluctuations between oxic and suboxic conditions.

For part II, during the upper Wufeng Formation (upper part of the P. Pacificus graptolite zone), most samples display elevated Mo enrichment, ranging from 5.36 to 18.1 ppm, with an average of 11.19 ppm. (Mo/U)_EF ratios mostly fluctuate between 0.3 and 1 times seawater, plotting closer to the anoxic end (Fig. 14). The U/Th, V/Cr, and Ni/Co values increase, varying between 0.56 and 2.30 (average of 1.00), 2.64–6.15 (average of 3.84), and 4.85–9.74 (average of 7.25), indicating fluctuations between suboxic and anoxic bottom waters. All of the above data reflect at least intermittent anoxic conditions below the Guanyinqiao Formation.

For part III, during the accumulation of Guanyinqiao Formation (M. extraordinarius and M. persculptus graptolite zones), most palaeo-redox proxies show first an increase and then a decrease, including Mo and U contents and U/Th, Ni/Co, V/Cr, V/V + Ni ratios. Most samples show low concentrations of Mo (< 10 ppm), but much higher Mo contents (32.60 and 71.80 ppm) occurred at the base of the Guanyinqiao Formation. (Mo/U)_EF ratios are generally less than that of seawater (0.3 × SW to 1 SW; Fig. 13), suggesting suboxic bottom water conditions. The U/Th, V/Cr, and V/V + Ni ratios indicate oxic to suboxic conditions (Fig. 14). Although most of the samples point to anoxic conditions, the Ni/Co ratio still shows a trend of increasing first and decreased afterwards. Thus, careful should be taken when applying these criteria to interpret the redox conditions.

For part IV, during the lower Rhuddanian Stage (A. ascensus-C. veiculosus graptolite zones), most palaeo-redox indices show similar variation patterns to that of part III, which increase first and then decrease. Most samples exhibit high average Mo concentrations of 49.67 ppm (and as large as 111 ppm), with (Mo/U)_EF ratios mostly approximately 1 time that of seawater (Fig. 13), suggesting a euxinic condition. Similar conclusions were found in the interpretations of U/Th, V/Cr, Ni/Co, and V/V + Ni ratios (Fig. 14), which indicate persistent strong euxinic bottom water during the deposition of the lower part of the Longmaxi Formation.

For part V, during the middle to upper Rhuddanian Stage (upper part of A. ascensus–C. veiculosus graptolite zone), most proxies remain at a steady level compared to those in part IV. The V/Cr, Ni/Co and U/Th ratios range from 3.2 to 7.31 (average of 4.75), 6.84–12.47 (average of 9.2), and 0.38–2.09 (average of 1.35), respectively, reflecting basic anoxic bottom waters. The Mo concentrations maintain a high level (average of 44.1 ppm), with (Mo/U)_EF ratios mostly approximately 1 times that of seawater (Fig. 13), revealing euxinic water conditions. The above data suggest that euxinia may have been maintained during the deposition of the lower Longmaxi Formation.

Relationship of δ¹³C_org variations to environmental changes and possible causes of carbon isotope excursions during the O–S interval

Our data that are derived from Xindi No. 2 well as described above clearly show water column redox variations pre-and-post Hirnantian glaciation, which can be summarized as two pulses of deepening degrees of anoxia. A parallel variation in δ¹³C_org values is also well documented in the Xindi No. 2 well. The positive δ¹³C_org excursion recognized as the HICE was initiated after the first pause of the Hirnantian glaciation from the M.extraordinarius zone to the lower M. persculptus zone (Figs. 12 and 14), followed by subsequent redox condition variations from increasing anoxia to sudden oxic-suboxic conditions. The δ¹³C_org values recorded above the HICE exhibit a − 3.2‰ negative excursion after the second pause of the Hirnantian glaciation from the upper M. persculptus zone to the lower A. ascensus zone, accompanied by rapid variations from oxic-suboxic to strong euxinic water conditions. These two abrupt shifts display strong covariations between δ¹³C_org and redox conditions, indicating a significant change in the relative carbon fluxes between ocean water, organic matter, and sediments.

In order to explore the spatial variation in the marine carbon isotopes during the Late Ordovician–early Silurian, we compared the lateral and vertical change in δ¹³C_org values across proximal to distal areas on the Yangtze Shelf Sea (Fig. 15). Three sections were selected for geochemical analyses^20,42, along with the Xindi No. 2 well in this study, and each section represents a lateral change from the inner- to mid-shelf. All sections are well established by graptolite biozonation, which can be used as stratigraphic correlation. As is shown in Fig. 15, the positive δ¹³C_org excursions characterize the Hirnantian interval across all four sections, but by comparing the positive shift, the records appear to be diachronous. In the shallow-water setting (NBZ) recorded by Yan et al.⁴⁰, the increase in δ¹³C_org occurs in the Metabolograptus extraordinarius biozone, while this excursion can be observed in the Metabolograptus persculptus biozone in a deeper water setting from the inner-shelf (SH section) to the mid-shelf (QL section), as well in the Xindi No. 2 well. The offset of δ¹³C_org values vary across all four sections, with a larger offset of ~ 3.7‰ in the Xindi No. 2 well, and of ~ 2.6‰ in the NBZ section, while relatively smaller offset values are present in the SH and QL sections (of ~ 1.2‰ and ~ 1.1‰). We suggest that this δ¹³C_org gradient might be explained by the different oceanic environments between shallow- and deep-water settings. SH and QL were under anoxic-euxinic conditions during Hirnantian glaciation according to Fe speciation and Mo concentrations²⁰, while the NBZ and Xindi No. 2 well have been established as oxic and oxic-suboxic during the glaciation by Zhou et al.¹² and by palaeoredox proxies in this study (Fig. 14). These spatial differences in water column redox conditions were probably due to differences in depositional water depths, with shallower waters associated with more oxic conditions.

Positive excursions (both δ¹³C_carb and δ¹³C_org) during the Hirnantian stage have been reported from sections of the Yangtze Platform (e.g. Refs.^{3,11,17,30,126,127}) as well as from different regions worldwide(e.g. Refs.^{13,14,108,128}). Proposed geological processes that may cause positive δ¹³C excursions include the following: (1) enhanced marine productivity and organic matter burial^{13,15,129,130}, (2) increased carbonate-platform weathering^128,131, (3) dissolved inorganic carbon (DIC) with high δ¹³C¹³², and (4) sea-level eustacy and shoaling of the marine chemocline^38,127,133. The positive δ¹³C excursion has been interpreted as the result of increased organic carbon burial, which lowered atmospheric pCO₂, leading to a consequent ¹³C enrichment^{108,128,134,135}. Due to the fact that photosynthetic carbon fixation of marine phytoplankton favours ¹²C relative to the carbon reservoir, the progressive increase in the burial of organic matter could have led to substantial enrichment of ¹³C in seawater, leading to a positive excursion of δ¹³C (e.g. Refs.^{15,128,129,136}. Based on these theories, the HICE is inferred to be indicative of a major carbon burial event during the Late Ordovician. However, this organic carbon burial causal model is yet to be reconciled with the marked lithofacies change, as the shelly limestone of the Guanyinqiao Formation is sandwiched between the pre-glaciation of the Wufeng Formation and the post-glaciation of the Longmaxi Formation black shales. Furthermore, data collected from recent studies have shown low TOC values during the Hinantian interval both on the Yangtze Platform (e.g. Refs.^30,43) and from global strata (e.g. Refs.^19,137–139). The carbonate weathering model proposed by Kump et al.¹²⁸ linked the δ¹³C excursion, especially that of δ¹³C_carb, to the enhanced carbonate platform weathering during glacioeustatic sea-level lowstand. The strong positive δ¹³C_org excursion from the Xindi No. 2 well coincident with the glacial Guanyinqiao Formation may reflect increased carbonate weathering during the Hirnantian glaciation. Although there may exhibit short-term pulses of climate warming during the Hirnantian glaciation (Fig. 12), which could accelerate carbonate platform weathering, high-resolution studies are still required to verify this conclusion. Apart from organic carbon burial and carbonate platform weathering, other factors, including dissolved inorganic carbon (δ¹³C_DIC) and glacio-eustatic sea-level change, cannot be excluded. Dissolved inorganic carbon (DIC) with high δ¹³C values, which were possibly generated by an authigenic carbonate precipitation mechanism, may play an important but poorly understood role in the acknowledgment of the Hirnantian excursion. According to Ahm et al.¹⁹, glacio-eustatic restriction of shallow epeiric seas may increase both lateral- and vertical gradients in δ¹³C_DIC from the shallow shelf to the deep basin. The lateral gradient exhibits higher seawater δ¹³C values in shelf-proximal than in shelf-distal environments due to increased photosynthesis and carbonate weathering, which can be observed in the Canada Arctic and Baltic basins^133,140 as well in the Yangtze shelf sea^127,141. This theory may explain such a high δ¹³C_org peak (~ − 25.3‰) in the Xindi No. 2 well, as its locality is closer to the shelf-proximal unit compared to previous studies (Fig. 15). Generally, organic matter is produced by photosynthetic organisms in surface water, as a result of the uptake of ¹²C, and is removed as sedimentary organic matter in deep water, causing even higher ¹³C values in proximal settings¹⁴². A surface- to deep-water gradient has been recorded in the modern Black Sea, which is a typical example of a bottom-sulfurized and stagnant ocean, where the δ¹³C_org of photosynthate in surface waters is ~ 4‰ heavier than the δ¹³C_org of sedimentary organic matter in sulfidic deep waters¹⁴³. It is worth noting that the Xindi No. 2 well was characterized by the highest δ¹³C_org value among all four sections (Fig. 15), which can be explained by two factors. On the one hand, the increasing circulation of organic carbon fluxes, either by photosynthetic activity or by dissolved O₂, contributed to the preservation of the heavier δ¹³C_org under more oxic water relative to those deposited under anoxia water conditions. On the other hand, increasing carbonate weathering exposed by the glacio-eustatic sea-level fall may have further accelerated carbon input to the proximal setting, contributing to the positive δ¹³C_org excursion. Thus, our high values of δ¹³C_org might be associated with increased photosynthetic activity and carbonate weathering when the sediments of the Xindi No. 2 well were deposited under oxic conditions in a shelf-proximal setting.

Environmental change in South China during the Hirnantian glaciation

The LOME was marked by two-phase extinction events¹⁴⁴. The first strike occurred in the M. extraordinarius–N. ojsuensis biozone (~ 445 Ma) coinciding with the onset of the Hirnantian glaciation and the positive shift in δ¹³C values of carbonates and organic carbon. The second wave of extinctions occurred in the early M. persculptus Zone (~ 444 Ma) coinciding with the subsequent abrupt rise in sea-level marked by a large negative shift in δ¹³C values^{1,6,11,20,129,145}. Our δ¹³C_org isotopic data provide a good constraint on the timing of the two-phase extinctions, combined with the aforementioned studies, which can also provide the relevant evolution of the temporal changes in climatic and ocean redox conditions (Figs. 8 and 15). From the late Katian (Late Ordovician), the Yangtze Block constantly collided with the Cathaysia Block due to Caledonian movement^146–150, resulting in northwest–southeast compression that caused the constantly expanding uplifts (i.e., Chengdu, Dianqian, and Jiangnan-Xuefeng uplifts) around the Sichuan basin and a deepening of the inner Yangtze Sea^151–155. During the deposition of the lower part of the Wufeng Formation, the Yangtze Sea evolved into a semi-isolated shelf sea, and gradually developed suboxic to anoxic bottom waters, which is consistent with all palaeoredox proxies (Mo-auth/U-auth, U/Th, V/Cr, and Ni/Co ratios). The climate of the Upper Yangtze basin during this time was generally warmer and more humid, consistent with CIA values from 67.48 to 73.37 (average of 69.72).

In the early Hirnantian, glaciation started on Gondwana, and its initial phase was related to the beginning of a cool climate. With the continuous formation of ice sheets, the global sea level began to fall, and the climate became colder and drier, which is consistent with the decline in CIA values (from 69.23 to 61.51) at the bottom of the Normalograptus. extraordinarius biozone (upper part of the Wufeng Formation). As temperatures decreased and eustatic sea-level dropped, redox conditions varied from anoxic to oxic-suboxic, which is consistent with the geochemical evidence discussed in our previous chapter. The overlying Guanyinqiao Formation thus deposited carbonaceous marls with abundant and diverse shelly fauna. During the early Hirnantian period, the oceanic redox evolution of the entire Yangtze Sea was mainly caused by sea level decline, and the sea level decline itself was the result of rapid global cooling. This result highlights the close relationship between the redox evolution of the Late Ordovician water column and climate change at that time^26,40,156. These conclusions are supported by the δ¹³C_org evidence in the Upper Hirnantian samples (Fig. 15).

In the late Hirnantian stage (M. persculptus biozone), glaciation ended, marked by an abrupt eustatic sea level rise. The climate started getting warmer and moister, which is consistent with the increase in the CIA values in the upper part of the Guanyinqiao Formation (from 58.30 to 62.60). As temperatures increased and sea level rose, the water-column redox rapidly returned to euxinia, coincident with the rising ratios (Mo-auth/U-auth, U/Th, V/Cr, Ni/Co, and V/V + Ni ratios) and Mo concentrations. In the early Rudanian stage, sea level rise and widespread euxinia jointly promoted black shale deposition in the Longmaxi Formation.

These data, together with previous research from other areas in the Yangtze Sea, indicate significant climatic fluctuations, from warming to cooling at the end of the late Katian stage, and then back to a warm climate at the end of the late Hirnantian, possibly fluctuate during these two-time intervals. Correspondingly, the marine redox system changed from anoxic to oxygenated, and then suddenly evolved into euxinic bottom waters, consistent with two pulses of the LOME. Climatic fluctuations, coupled with oceanic anoxia, were likely responsible for the gradual end of the Late Ordovician biotic crisis.