The Sommerodde (Telychian, Silurian) positive carbon isotope excursion: why is its magnitude so variable?

A major surprise when integrating $δ$¹³C_org data with the graptolite biostratigraphy (Loydell et al. 2017) of the Sommerodde-1 core, Bornholm, Denmark was the discovery of a pronounced positive excursion (amplitude c. 4‰) in the Oktavites spiralis graptolite Biozone of the Telychian, with higher $δ$¹³C_org values (peaking at −25.4‰) than either the very well-known Hirnantian (HICE; peaking at −27.7‰) or early Sheinwoodian (ESCIE; peaking at −27.1‰) excursions (Hammarlund et al. 2019). This was clearly not the Valgu excursion (Munnecke and Männik 2009), which is stratigraphically older, nor the more recently identified Manitowoc excursion (McLaughlin et al. 2019), which is stratigraphically younger (see later discussion), but a previously largely unrecognized and unnamed perturbation in the carbon isotope record. Hammarlund et al. (2019) named it the Sommerodde positive organic carbon isotope excursion (SOCIE). Other high-magnitude Late Ordovician and Silurian positive $δ$¹³C excursions are associated with major global environmental (and biotic) changes and therefore the discovery of a new major excursion suggests that the Telychian might host another, previously unrecognized, interval of significant environmental change, potentially with similarly previously unrecognized impacts on biotic diversity.

This paper has three aims. First, to review further potential records of the SOCIE, mostly published after 2019, showing that the Sommerodde excursion varies dramatically in magnitude, ranging from a high-magnitude excursion (in sections representing deeper water environments), comparable with that seen in the Sommerodde-1 core, to either a low-magnitude excursion or unrecognizable in sections representing the shallowest environments. Second, to present new $δ$¹³C_org data from a section in central Spain in which the Sommerodde excursion is a very minor event and third, to discuss why the SOCIE shows such a wide variation in magnitude and is so pronounced in the Sommerodde-1 core $δ$¹³C_org record. We conclude that the most likely reason is a change in organic matter composition in deeper water areas. This is the first positive $δ$¹³C excursion in the Silurian to show a pattern of increasing magnitude with increasing depositional water depth. Our study demonstrates that the Silurian $δ$¹³C record and the causes of Silurian $δ$¹³C perturbations are more varied than previously recognized.

Figure 1 shows the Baltic localities discussed in this paper, including Sommerodde (Bornholm). The SOCIE was recorded in several samples extending over c. 10 m of the Sommerodde-1 core (between depths of c. 80 and 70 m; Fig. 2). Graptolites (Loydell et al. 2017) demonstrate that much of the excursion (77.02–70.59 m) lies within the middle part of the O. spiralis graptolite Biozone (Fig. 3). It began in strata lacking graptolites, but which are underlain by beds yielding lower spiralis Biozone graptolites (Fig. 3). The excursion must therefore begin within either the lower or middle spiralis Biozone and end within the middle spiralis Biozone.

Hammarlund et al. (2019) summarized previous carbon isotope studies through this interval and were able to recognize the SOCIE in two $δ$¹³C_carb curves from the Baltic states. In the Ventspils D-3 core from Latvia (Fig. 1), the positive $δ$¹³C_carb excursion reaches its peak in strata bracketed by lower and middle spiralis Biozone graptolite assemblages (Loydell and Nestor 2006; Hammarlund et al. 2019, their fig. 7). The SOCIE is the highest magnitude excursion in the Llandovery of this core, with the peak value exceeding that of the well-known mid-Homerian excursion, but of significantly lower magnitude than either the ESCIE or mid-Ludfordian excursions (Kaljo et al. 1998). In other sections, however, the SOCIE is either of limited magnitude (e.g. the Ruhnu (500) core, Estonia; Hammarlund et al. 2019, their fig. 7) or unrecognizable (the Viki core, Estonia; Kaljo et al. 2003; see following text for discussion of the $δ$¹³C_carb and $δ$¹³C_org data of Richardson et al. (2019) from Viki). It is also unrecognizable in the Knapp Creek core, Iowa, USA (McAdams et al. 2017).

It is possible that the SOCIE is represented in the $δ$¹³C_org record from the Zwierzyniec-1 core, Poland (see Fig. 1). Figure 4 summarizes the rather limited carbon isotope and graptolite biostratigraphic data from Sullivan et al. (2018). Stimulograptus vesiculosus (2952.95 m) is a middle–upper O. spiralis Biozone species, whereas Streptograptus wimani (2951.4 m) first appears at the base of the succeeding Cyrtograptus lapworthi graptolite Biozone. The presence of Monoclimacis vomerina sensu lato, identified by Zalasiewicz and Williams (in Sullivan et al. 2018, their appendix A) from 2956.50 to 2956.75 m, indicates an age no older than the Monoclimacis crenulata graptolite Biozone. The basal Sheinwoodian biozonal index Cyrtograptus murchisoni was present in the 2949–2950 m sample. The most positive late Telychian $δ$¹³C_org value is thus in a stratigraphic position consistent with that of the SOCIE.

Biostratigraphic data are similarly rather limited from the Altajme core, Gotland (see Fig. 1). The $δ$¹³C_carb record from here includes a small magnitude (c. 0.5‰) double-peaked positive excursion in strata at depths between 320 and 317 m (Hartke et al. 2021, their fig. 3); O. spiralis is recorded from a depth of 322.14 m and strata up to 316 m are assigned to the spiralis Biozone. The biostratigraphic data are thus consistent with this excursion representing the SOCIE. The magnitude of this excursion is far less than that of the ESCIE (c. 3‰, referred to by Hartke et al. 2021 as the Ireviken excursion), but is comparable with that in the Ruhnu (500) core, Estonia (Hammarlund et al. 2019, their fig. 7), which lies in the same facies belt (Fig. 1).

Richardson et al. (2019) provided both $δ$¹³C_carb and $δ$¹³C_org data from the Viki core, Estonia (previously analysed for $δ$¹³C_carb by Kaljo et al. 2003). The same Viki core data were also plotted by Yan et al. (2022). Based on the correlation of graptolite and conodont biozones (e.g. Männik 2007), the SOCIE should be looked for in the Pterospathodus amorphognathoides lennarti and lower Pterospathodus amorphognathoides lithuanicus conodont biozones. The supplementary data table of Richardson et al. (2019) enables the integration of their geochemical data with the conodont biozonation for the Viki core of Männik (2007). There are rather limited data through the key interval (four $δ$¹³C_carb analyses and three $δ$¹³C_org analyses between 160 and 150 m depth). The $δ$¹³C_carb value falls from 2.15‰ in the upper Pterospathodus amorphognathoides angulatus conodont Biozone (155.65 m) to 1.43‰ in the Pt. a. lennarti Biozone (153.6 m), rising to 1.73 and 1.78‰ in the lower Pt. a. lithuanicus Biozone (151.68 and 151.17 m, respectively). There is no $δ$¹³C_org analysis for the uppermost Pt. a. angulatus Biozone sample (155.65 m). The sample below (at 160.25 m) has a $δ$¹³C_org value of −26.78‰, the Pt. a. lennarti Biozone (153.6 m) sample is a little more positive (−26.46‰), rising further into the lower Pt. a. lithuanicus Biozone (−26.09‰ at 151.68 m) before falling back to −26.71‰ at 151.17 m. The limited data therefore support the presence of a minor positive $δ$¹³C_org excursion at a level correlative with the SOCIE in the Sommerodde core, with no obvious positive excursion revealed by the $δ$¹³C_carb data.

Young et al. (2020) presented $δ$¹³C_org data from the Aizpute-41 core, Latvia (Fig. 1; see Loydell et al. 2003 for the biostratigraphy of this core and Cramer et al. 2010 for the $δ$¹³C_carb data). A c. 1‰ positive shift is present in the core at a depth of 944.7 m (not shown on fig. 4 of Young et al. 2020, but given in table 2 of their supplementary data), rising from −27.40‰ at 945.7 m to −26.35‰ at 944.7 m and declining again to −27.27‰ at 942.7 m. Loydell et al. (2003) assigned strata from 944.62–941.88 m to the middle spiralis graptolite Biozone, so this positive shift is at the correct biostratigraphic level for the SOCIE (Fig. 5). This −26.35‰ value is the most positive in the entire dataset, which includes the sedgwickii excursion (referred to by Young et al. 2020 as the late Aeronian excursion), the Valgu excursion (discussed briefly in the following) and the ESCIE (with −26.56‰ recorded from the Monograptus riccartonensis graptolite Biozone at 910.3 m). The Aizpute-41 data therefore match those from the Sommerodde-1 core, with the SOCIE having the most positive recorded $δ$¹³C_org value. Like Sommerodde, Aizpute is one of the deepest water localities (Fig. 1) from which published Telychian carbon isotope data are available.

Young et al. (2020) showed the Valgu excursion extending over a much greater stratigraphic interval than other researchers, to a level (at 945.7 m depth in the Aizpute-41 core) high within the lower spiralis Biozone (and within the middle of the Pterospathodus celloni conodont Superzone). Munnecke and Männik (2009) showed the Valgu excursion to be confined to the Pterospathodus eopennatus conodont Superzone (see also Hammarlund et al. 2019, their fig. 6), the top of which correlates with a level in the lower to middle crenulata graptolite Biozone (Loydell et al. 2003, their fig. 17; Männik 2007, their fig. 5). The Telychian $δ$¹³C_orgdata of Young et al. (2020) are plotted in Figure 5 with the graptolite biostratigraphy corrected (from Loydell et al. 2003); the 944.7 m value, representing the SOCIE, is now included. Overall, the Telychian in the Aizpute-41 core exhibits significantly more positive $δ$¹³C_org values than the upper Rhuddanian, lower and middle Aeronian and lowermost Sheinwoodian (Young et al. 2020, their fig. 4) and, despite there being fluctuations in $δ$¹³C_org values, it is difficult to identify individual excursions, in part perhaps because of the condensation of much of the lower–middle Telychian, with the combined turriculatus, crispus, sartorius and griestoniensis graptolite biozones represented by only 6 m of strata, less than the individual thicknesses of either the crenulata (c. 7 m) or spiralis (c. 10 m) graptolite biozones.

Cichon-Pupienis et al. (2021) recorded a significant positive $δ$¹³C_org excursion, tentatively identified as the SOCIE, in the upper part of the Ju¯rmala Formation of the Vežaičiai-2 (V-2) core, western Lithuania (Fig. 1). This is of similar magnitude to the ESCIE, but of lesser magnitude than two excursions lower in the Telychian, tentatively identified by Cichon-Pupienis et al. (2021) as the Valgu and Kallholn excursions. The graptolite biostratigraphy of the V-2 core is yet to be established, with the biozones shown based on correlation with the succession in the Viduklė-61 core, 90 km to the ESE in south-central Lithuania (Fig. 1). A very interesting feature of the putative SOCIE in the V-2 core is that it is within the thickest development of greenish grey mudstones within the Telychian (Fig. 6). This parallels its development within a thick interval of predominantly paler strata within the Sommerodde-1 core (Fig. 2).

Strauss et al. (2020) integrated $δ$¹³C_carb and graptolite biostratigraphic data from the Mount Hare Formation, Road River Group of northern Yukon, Canada. A positive excursion (of c. 1.5‰) is seen within the spiralis Biozone (Fig. 7), beginning just over halfway through the strata assigned to the biozone (which is not divided biostratigraphically into lower, middle and upper and is overlain by 8.4 m of strata lacking diagnostic graptolites). Assuming that this is the SOCIE, its relative magnitude is similar to that in the Ventspils D-3 core, Latvia – that is, exceeding that of the mid-Homerian excursion and less than the ESCIE and Ludfordian excursions. The assumed SOCIE in the Road River Group is also of lesser magnitude than the Valgu excursion, which was not seen in the Ventspils core because no analyses were carried out through the relevant stratigraphic interval. In terms of environment, the analysed Road River Group samples were from towards the southern end of the Richardson Trough (Strauss et al. 2020), which is described as an intraplatformal deep-water basin formed by rifting along the Great American Carbonate Bank on the margin of Laurentia.

Yan et al. (2022, p. 9) refer to a positive $δ$¹³C_org excursion in the lower Pterospathodus amorphognathoides amorphognathoides Superzone of the Baizitian section, Sichuan, China ‘that is equivalent to the SOCIE of Denmark’. Their figure 7 shows a significant positive excursion (4‰) at this level (labelled Manitowoc OCIE), but this is younger than the age of the SOCIE on Bornholm; the base of the Pt. a. amorphognathoides Superzone correlates with that of the C. lapworthi graptolite Biozone (Loydell et al. 2010), which succeeds the O. spiralis Biozone, in the middle of which the SOCIE terminates. In addition, and very significantly, figure 5 of Yan et al. (2022) shows an unconformity in the Telychian of the Baizitian section, resulting in an absence of the conodont biozones (uppermost Pt. a. angulatus to P. a. lithuanicus) that are correlative with the O. spiralis graptolite Biozone and thus the SOCIE could not be represented in this section. The Baizitian data therefore support what is suggested in the following: that there are two separate $δ$¹³C excursions in the O. spiralis and C. lapworthi biozones.

Yan et al. (2022, p. 7) suggested that the Manitowoc excursion and the SOCIE are synonymous. We discuss the evidence for this in the following section and conclude that they are separate excursions.

McLaughlin et al. (2019) identified a new positive $δ$¹³C_carb excursion in the Telychian of the Michigan Basin, which they named the Manitowoc excursion. In those sections in Wisconsin where the excursion was identified, it occurred immediately above ‘the top of a cluster of hardgrounds’, a level which was used as a correlation datum. In the Sheboygan SH-541 core, Wisconsin (the core used for correlation with Baltica by McLaughlin et al. 2019 and for correlation with sections elsewhere in Laurentia, Baltica and South China by Yan et al. 2022), this correlation datum lies just above a depth of 250 m. In their summary correlation diagram, McLaughlin et al. (2019, their fig. 8) showed the Manitowoc excursion to be within the upper part of Silurian Stage Slice Te3, equating to the uppermost Pt. eopennatus conodont Biozone through to the top of the Pt. a. lithuanicus conodont Biozone (Cramer et al. 2011, their fig. 3). These conodont biozones correlate with the M. crenulata and O. spiralis graptolite biozones and, on this basis, the Manitowoc CIE and SOCIE would appear at first sight to be potentially synonymous. Note, however, that no conodont biostratigraphic data were presented from the Sheboygan SH-541 core.

McLaughlin et al. (2019) noted a ‘striking similarity’ of the $δ$¹³C_carb curve in the Sheboygan SH-541 core to that of ‘the biostratigraphically well-constrained Estonian Viki core’ of Kaljo et al. (2003) and correlated the Sheboygan and Viki $δ$¹³C_carb curves (McLaughlin et al. 2019, their fig. 2). The key correlation datum in Wisconsin, above which the Manitowoc excursion lies, is correlated with a level in the lower part of the Pt. a. amorphognathoides conodont Biozone. This, however, is within Stage Slice Te4, not Te3, of Cramer et al. (2011). On this basis, the Manitowoc excursion, as originally defined, is younger than McLaughlin et al. (2019, their fig. 8) indicated and would lie within the C. lapworthi graptolite Biozone, within which there was a well-documented eustatic fall in sea-level (Loydell 1998, 2007). McLaughlin et al. (2019) noted, as had Loydell (2007), that ‘the ascending limbs of the positive $δ$¹³C_carb excursions correspond with facies changes consistent with progressive sea-level fall’. If the correlation of McLaughlin et al. (2019, their fig. 2) is correct, then the Manitowoc excursion is likely to be associated with the mid-lapworthi Zone sea-level fall.

Yan et al. (2022) generated a rather different concept of the Manitowoc excursion from that originally presented by McLaughlin et al. (2019). They expanded the excursion to include strata below the correlation datum cluster of hardgrounds in the Sheboygan SH-541 core and divided the excursion into three parts. The conspicuous positive excursions shown in the Sheboygan SH-541 core within Interval 1 and the lower part of Interval 2 in Yan et al. (2022, their fig. 8) both lie below these hardgrounds and thus also below the level of the Manitowoc excursion as originally defined. The excursion within Interval 1 of Yan et al. (2022) is c. 50 m below the correlation datum and was shown to lie within the upper part of the Pt. eopennatus conodont Superzone. That within the lower part of Interval 2 is shown as in the lower part of the Pt. a. angulatus conodont Biozone. McLaughlin et al. (2013) is cited as the reference for the stratigraphy of the Sheboygan core, but no conodont data for the core is presented in this work to support the conodont biozonation presented. Indeed, McLaughlin et al. (2013, p. 15) stated that there is ‘poor biostratigraphic control’.

The Manitowoc CIE is also identified by Yan et al. (2022, their fig. 8) in the Viki core, Estonia using both the $δ$¹³C_carb curve of Richardson et al. (2019), which differs little from that of Kaljo et al. (2003) used for correlation by McLaughlin et al. (2019), and the $δ$¹³C_org curve of Richardson et al. (2019). The most pronounced $δ$¹³C_carb excursion (of c. 0.7‰) is within Interval 1, which lies in the upper half of the Pt. a. angulatus conodont Biozone (equivalent to the upper M. crenulata and lowermost O. spiralis graptolite biozones; Männik 2007). There are minor excursions within Interval 2 (of c. 0.4‰) and Interval 3 (a single analysis at 135.48 m of 1.49‰, with 15 stratigraphically lower and higher samples, between 141.27 and 122.07 m, showing little variation between 1.07 and 1.28‰), with only that in Interval 3 corresponding to the stratigraphic level of the Manitowoc excursion according to the correlation of McLaughlin et al. (2019).

Based on the available evidence, it seems best to continue to refer to the positive $δ$¹³C excursion in the spiralis graptolite Biozone (terminating in its middle) as the SOCIE and to use the term Manitowoc excursion for a later Telychian excursion in the lower Pt. a. amorphognathoides conodont Biozone (as shown by the integrated chemostratigraphic and biostratigraphic data of Yan et al. 2022 from Baizitian, China), which is likely to correlate with a sea-level fall in the middle C. lapworthi graptolite Biozone.

A richly fossiliferous Telychian–Sheinwoodian, largely black shale succession is very well exposed in the Jabalón River section, Corral de Calatrava, central Spain (Fig. 8). The graptolite and conodont biostratigraphy of the section were studied by Štorch et al. (1998) and Loydell et al. (2009). As lithological samples from all horizons were retained, this seemed an ideal opportunity to determine how well developed the SOCIE is in a graptolitically well-dated section in a facies and palaeogeographical setting different from that of Sommerodde (Bornholm).

This part of central Spain was part of the Central Iberian Zone (Fig. 8) of peri-Gondwanan Europe during the Silurian and was located at a much higher (southern) latitude setting than Bornholm and the Baltic states (which were part of Baltica) on the other side of the Rheic Ocean (Torsvik and Cocks 2013; Fig. 9). The tentative palaeogeographical reconstruction of the North Gondwanan regions during the Silurian Period by Robardet and Gutiérrez-Marco (2002) (see also Loydell et al. 2015, their fig. 3) shows the study area to be towards the outer part of what is described as an ‘inner shelf’.

The middle Telychian (upper M. crenulata graptolite Biozone) to lower Sheinwoodian (lower Wenlock M. riccartonensis graptolite Biozone) part of the section was studied in detail by Loydell et al. (2009) when it was hoped that the section would be a potential candidate replacement Global Boundary Stratotype Section and Point for the base of the Wenlock Series. Like so many apparently complete sections, however, the Jabalón River section contains unconformities (Fig. 8). Two significant unconformities were recognized, detectable biostratigraphically and marked in one case by a thin shell bed and in the other by an abrupt change in facies to an impure limestone overlain by brachiopod-rich shale. So, what at first sight looked to be a stratigraphically continuous Telychian section is in fact missing the uppermost Telychian (upper C. lapworthi graptolite Biozone and the entirety of the C. insectus and C. centrifugus graptolite biozones; Fig. 8). The presence of O. spiralis in the highest Telychian sample below the lower unconformity (overlain by Wenlock strata) is biostratigraphically very significant. In more complete graptolitic sections in the Czech Republic and Wales, O. spiralis makes its last appearance in the lower half of the lapworthi graptolite Biozone (Loydell and Cave 1996; Štorch 2023). Thus any biotic or environmental changes occurring during the latest Telychian (late lapworthi to centrifugus zones) can have left no record in the Jabalón River section in the fossil or isotope record. An absence of strata from this stratigraphic level has been widely reported (see Loydell et al. 2003, 2010 for examples from Baltica). The second unconformity in the Jabalón River section affects part of the lower Sheinwoodian (encompassing at least the entirety of the Monograptus firmus graptolite Biozone; Fig. 8).

Unconformities such as these can explain the sudden disappearance and appearance of taxa within a section and will also significantly influence the shape of isotope curves, either by partially or entirely removing excursions from the record (e.g. the absence of the HICE in the Road River Group of the Peel River section; Strauss et al. 2020) or by generating dramatically steepened limbs, such as seen in the Jabalón River section $δ$¹³C_org data for the rising limb of the ESCIE (Fig. 10).

The samples from the Jabalón River section were collected by the authors in June 1999. To retrieve the concentrations and isotopic composition of organic carbon in the samples, the powdered (<63 µm) samples were treated with 1 M HCl then washed and dried. The samples were analysed using elemental analyser isotope ratio mass spectrometry at Iso-Analytical in the UK in April 2020, calibrated with a set of International Atomic Energy Agency standards. Standards of known isotopic composition and with similar concentrations to the samples were analysed in parallel. The isotopic composition of carbon was reported relative to the Vienna Pee Dee Belemnite (V-PDB) standard. Uncertainty in the standardization gave an error for $δ$¹³C of 0.1‰.

The range of $δ$¹³C_org values in the Jabalón River section samples is fairly limited (Fig. 10), ranging from a low of −28.7‰ (recorded in the murchisoni graptolite Biozone) to a high of −27.5‰ (in the riccartonensis graptolite Biozone). It makes sense to consider the Telychian and Sheinwoodian isotope records individually as they are separated by a significant unconformity.

The overall trend through the upper crenulata graptolite Biozone through to the unconformity in the lower lapworthi graptolite Biozone is towards more negative values (Fig. 10). The penultimate crenulata Biozone sample records the most positive value of −27.7‰. In the Sommerodde-1 data (Hammarlund et al. 2019), values fluctuate above the Valgu excursion between strata yielding crenulata and lower spiralis Biozone graptolites, so it is not surprising to see something similar in the Jabalón River section $δ$¹³C_org data.

The negative trend is interrupted by a few slightly more positive values (labelled SOCIE on Fig. 10) within the lower to middle spiralis Biozone, but if the SOCIE was not being looked for, then these would be unlikely to be commented upon. Interestingly, four thin micaceous siltstones are present within this interval; Fig. 10). These are the only slightly coarser clastic horizons within the entire studied middle Telychian–lower Sheinwoodian part of the section, suggesting that some significant environmental change was occurring at this time, but, strangely, it was not one that has affected the $δ$¹³C_org record to any great extent. Biostratigraphically, these siltstones and the very minor positive excursion are at a comparable level to the SOCIE in the Sommerodde-1 core. In both sections, Streptograptus nodifer has its first appearance datum (FAD) within the SOCIE and Oktavites excentricus has its FAD just above it (Fig. 11). The FADs of both are consistently recorded elsewhere from the middle of the spiralis Biozone (e.g. Bjerreskov 1975; Loydell et al. 2003, 2010, 2017; Loydell and Nestor 2006; Štorch and Piras 2009).

The lower and middle murchisoni graptolite Biozone record the lowest $δ$¹³C_org values within the studied part of the section, with a positive shift of 0.4‰ occurring in the upper part of the biozone (Fig. 10). This is a typical level for the start of the ESCIE (Cramer et al. 2010; Loydell and Large 2019). The next sample, from the riccartonensis graptolite Biozone, is above an unconformity (the firmus graptolite Biozone is missing; Fig. 10). The $δ$¹³C_org value of −27.5‰ is the most positive in the section. In all sections studied worldwide the riccartonensis graptolite Biozone is within the protracted peak of $δ$¹³C values within the ESCIE (Cramer et al. 2010), so again the Jabalón River section is showing what is typically seen.

The results from the Jabalón River section in Spain revealed only a very small positive excursion at the same biostratigraphic level at which the SOCIE is recorded as a major positive excursion in the Sommerodde-1 core at Bornholm. The fact that the ESCIE in the Jabalón River section is at the same stratigraphic level (starting in the upper murchisoni graptolite Biozone with its peak value in the riccartonensis graptolite Biozone) from which it is well known worldwide (see Loydell and Large 2019 for a recent study and discussion) strongly suggests that the $δ$¹³C_org values from the Jabalón River section are a reliable record of environmental changes here. So, the question is: why is the SOCIE so weakly developed in the Jabalón River section? Or perhaps the question should be: why is the SOCIE so well developed in the Sommerodde-1 core data, extending over several samples through 10 m of strata with a magnitude exceeding that of the HICE and ESCIE? Possible answers to these questions are discussed in the following, together with the unusual features of the SOCIE.

The SOCIE exhibits a very unusual feature. The consistent pattern seen in other Silurian positive carbon isotope excursions is that their magnitude decreases into deeper water strata. This was highlighted by Loydell (2007) for $δ$¹³C_carb, with evidence presented for excursions from East Baltic and Arctic Canadian sections. Loydell (2007, p. 534) concluded: ‘Given this consistent pattern of offshore decline in the magnitude of isotope excursions it is to be expected that the $δ$¹³C record of deeper water sections will record clearly only the most significant events’.

The SOCIE, however, has its greatest magnitude in the deepest water setting studied thus far (the Sommerodde-1 core, $δ$¹³C_org). It is readily identifiable, but of lesser magnitude, in the $δ$¹³C_org and $δ$¹³C_carb records in what are interpreted to be slightly shallower settings (Aizpute, Ventspils and Vežaičiai, Fig. 1; probably also the Peel River, Yukon) and becomes a minor event in the $δ$¹³C_org and $δ$¹³C_carb records (Ruhnu and Altajme, Fig. 1; Jabalón River section, Fig. 10) or unrecognizable (e.g. in the $δ$¹³C_carb record of Viki, Fig. 1) with increased shallowing/proximity to palaeoshorelines.

Recent papers have demonstrated palaeobathymetry gradients in both $δ$¹³C_carb and $δ$¹³C_org in Lower Paleozoic sections. In South China, Li et al. (2018, their fig. DR2) recorded decreasing $δ$¹³C values (both carbonate and organic) with increasing depositional water depth in the upper Cambrian. A similar pattern is revealed in the data of Yang et al. (2020, fig. 9) from the Upper Ordovician: in the Linxiang Formation, shallow water $δ$¹³C_org values are c. −27 to −26‰ and deeper water values are c. −29‰; in the Wufeng, Daduhe and Tiezufeike formations shallow water $δ$¹³C_org values are c. −29 to −28‰ and deeper water values are c. −31 to −30‰. Unfortunately, $δ$¹³C_org data through the spiralis Biozone (and thus the SOCIE interval) are available only for the deeper water sections of the Baltic basin, preventing such a comparison. For the ESCIE (Sheinwoodian), however, it appears that there is no major variation in the lowest $δ$¹³C_org values in the Baltic region prior to the excursion, with c. −29‰ consistently recorded from the shallower water Gotland sections (e.g. Vandenbroucke et al. 2013; Hartke et al. 2021), the deeper water Vežaičiai (Cichon-Pupienis et al. 2021) and Ventspils (Young et al. 2020) sections, and the deepest water Sommerodde section (Hammarlund et al. 2019).

As a general rule, the larger Hirnantian and Silurian positive carbon isotope excursions are associated with facies changes indicating shallowing and the onset of the major excursions is coincident with graptolite extinctions (Loydell 2007). It was noted earlier that the strata hosting the SOCIE in the Jabalón River section are the only strata in the middle Telychian to lower Sheinwoodian part of the section to include clastic rocks of coarser grade than shale: the micaceous siltstones present are consistent with shallowing and shoreline progradation at this time.

The spiralis Biozone is not, however, associated with any obvious graptolite extinction event. The limited biotic impact of the environmental changes associated with the SOCIE may reflect the very high sea-levels at this time (Loydell 1998). It has also been noted (Loydell and Abouelresh 2021) that not all Silurian positive isotope excursions coincide with graptolite extinctions. Although the sedgwickii, ESCIE, mid-Homerian and mid-Ludfordian excursions all saw graptolite extinctions and species turnover (Loydell 2007), the smaller positive excursions (e.g. mid-Rhuddanian and early Aeronian), by contrast, occurred at times of graptolite species diversification. This suggests that the graptolites’ preferred habitat was not affected by the environmental changes occurring at these times. It is assumed that the same is true for the SOCIE, with this strongly suggesting that the SOCIE is not reflecting a major change in the global climate and that its amplitude has been somehow enhanced in the Sommerodde-1 $δ$¹³C_org record.

It could be argued (J. Frýda, pers. comm.) that the total organic carbon content (TOC) of Telychian samples from the Sommerodde-1 core is too low (0.48–0.87 wt% TOC) to enable reliable $δ$¹³C_org analyses. However, this cannot explain why there is a positive $δ$¹³C_carb excursion at the same biostratigraphic level in the Ventspils D-3 core, Latvia (Kaljo et al. 1998; Hammarlund et al. 2019) and a positive $δ$¹³C_org excursion at the same stratigraphic level in other Baltic sections. It is significant also that the well-known and widely recorded excursion straddling the Aeronian–Telychian boundary, the Rumba low (which appears in the Silurian carbon isotope compilations of Cramer et al. 2011 and Sullivan et al. 2018), is very well developed in the Sommerodde-1 core data (Hammarlund et al. 2019, their fig. 5), in samples with an even lower carbon content (0.31–0.51 wt% TOC) than within the SOCIE and that the Valgu excursion can be identified in the Sommerodde-1 core data at the same biostratigraphic level as in the Peel River, Yukon, Canada (Strauss et al. 2020). There is again a low carbon content (0.33–0.70 wt% TOC) in the Sommerodde-1 core Valgu excursion samples. It therefore seems reasonable to see the SOCIE as a genuine reflection of environmental change at this time.

Interestingly, Cichon-Pupienis et al. (2021) noted that the magnitude of carbon isotope excursions in the Silurian of Lithuania seems to be inversely proportional to the TOC content of the samples analysed. The carbon content of the Jabalón River spiralis graptolite Biozone samples varies between 5.15 and 7.86%, very much higher than the 0.48–0.87% in the Sommerodde core SOCIE samples, in agreement with this observation. Cichon-Pupienis et al. (2021) stated that this inverse relationship seems to be related to the quantity and composition of organic matter and its preservation.

Cichon-Pupienis et al. (2021) suggested that $δ$¹³C_org excursions could be the result of the early diagenetic partial oxidation of organic matter resulting in isotopically more ¹³C-enriched organic matter. The supporting reference cited (Hatch and Leventhal 1997) states that oxidation took place during subaerial exposure of the overlying carbonates. Subaerial exposure of the Sommerodde-1 sediments, deposited during a time of high global sea-levels (Loydell 1998) and in a deep-water environment, seems very unlikely (Fig. 1) because Bjerreskov and Jørgensen 1983 estimated a water depth of 1000 ± 300 m for Bornholm during the early Homerian. Hammarlund et al. (2019) also commented on the mixing of isotopic fingerprints associated with early diagenesis, noting that this only occurs in shallow water environments (Holmden et al. 1998; Fanton and Holmden 2007; Ahm et al. 2018; Higgins et al. 2018), so that an excursion in a deeper water setting (such as the SOCIE in the Sommerodde-1 core) should serve as a reliable record of a global perturbation of the carbon cycle.

The question arises, of course, as to whether the oxidation of organic matter on and within oxygenated sediments in a deep-marine environment could also similarly affect $δ$¹³C_org values and generate or enhance a positive $δ$¹³C_org excursion. Evidence from studies of more recent sediments (e.g. McArthur et al. 1992; Freudenthal et al. 2001; Lehmann et al. 2002), however, strongly suggests that this is unlikely.

In many areas, the Telychian as a whole has long been recognized as preserving sediments more characteristic of oxygenated environments, with more pale-coloured and more bioturbated strata and marine red beds (Ziegler and McKerrow 1975) and fewer graptolitic shales than the preceding Rhuddanian and Aeronian (Llandovery) or succeeding Sheinwoodian (Wenlock). The stratigraphic distribution of early Silurian marine anoxia was summarized diagrammatically by Page et al. (2007, their fig. 2), with a conspicuously largely oxygenated Telychian; recently, Hounslow et al. (2021) referred to a ‘Telychian oxygenation event’. There is an interval of strata encompassing the SOCIE in the Sommerodde-1 core that is noticeably paler grey on the optic televiewer image of the borehole wall (Fig. 2; Loydell et al. 2017, their fig. 20) and this can be seen in photographs of the core itself (e.g. Loydell et al. 2017, their fig. 5, a photograph of the 75.45–71.80 m interval). The change in colour is gradual and is associated with a decreased proportion of darker graptolitic horizons, although these are still present. The possibility that the oxidation of organic matter during an interval of very high bottom water oxygen levels has generated isotopically more ¹³C-enriched organic matter and thus a more pronounced excursion might therefore seem worth considering (but see the comments at the end of the previous paragraph). Questions then arise as to what caused the more widespread occurrence of oxygenated environments in the Telychian in general and why particularly in the middle spiralis graptolite Biozone in the Sommerodde-1 core.

Perhaps there was an influx of cooler, more oxygenated water into the Silurian Baltic basin at this time. No glacial deposit has been recorded from the Telychian of Gondwana (but precise dating of Silurian glacial deposits has only rarely been achieved; e.g. Caputo and dos Santos 2020). Both the sea-level curve of Loydell (1998; Fig. 8) and the oxygen isotope data from the Viki core, Estonia (Fig. 1) presented by Lehnert et al. (2010, their fig. 3) indicate that there were numerous glacial advances and retreats during the spiralis Zone (or correlative lennarti and lithuanicus conodont zones). Or perhaps there was a temporary change in bottom water circulation resulting from relatively nearby tectonic activity? Hurst et al. (1983) referred to substantial platform carbonate collapse over a large area following continental collision involving northern Baltica at the northern Iapetus Ocean margin; in Greenland, platform carbonates are replaced in the Telychian succession by basinal turbidites. Dating of the continental collision is enabled by the extraordinary thickness (500 m) of spiralis graptolite Biozone strata in Greenland (Bjerreskov 1986), reflecting the erosion of uplifted source areas. This collision therefore coincides at least in part with the SOCIE. Inevitably there is speculation here, but something exceptional and presumably short-lived is required to explain what appears to be an anomalous Sommerodde-1 core isotope record in the unusually pale strata of the spiralis Biozone and, in particular, the presence of the high-magnitude SOCIE.

There is a similar Late Ordovician example of an unusually high-magnitude (nearly 8‰) positive $δ$¹³C_org excursion. Pancost et al. (1999) suggested that the anomalously high $δ$¹³C_org values recorded by the Guttenberg excursion in the Katian of Iowa could reflect a change in organic matter composition: the colonial mat-forming organic-walled microfossil Gloeocapsomorpha prisca (sometimes referred to as an alga, more frequently as probably a cyanobacterium; Foster et al. 1989, 1990) became abundant in this excursion interval. Carbon isotope analyses of individual biomarkers of G. prisca by Pancost et al. (1999) indicated that these were enriched in ¹³C by 7‰ relative to compounds derived from other algal sources. G. prisca characterizes very shallow marine and intertidal environments in the Ordovician, so is unlikely to have been involved in generating the enhanced Sommerodde-1 $δ$¹³C_org values. However, this Ordovician example demonstrates the potential for a profound change in the nature and relative abundance of organic-walled organisms to enhance dramatically the magnitude of an isotope excursion. Unfortunately, many of these organic-walled organisms would have partially decayed (so as to be unidentifiable) as they descended through the water column to reach the seafloor as amorphous organic matter. Vandenbroucke et al. (2013, p. 95) emphasized the importance of (and need for) unidentifiable contributors within amorphous organic matter to explain the bulk $δ$¹³C_org values in the Telychian–Sheinwoodian of Gotland, Sweden.

Vandenbroucke et al. (2013) discussed various previous studies of the differing $δ$¹³C values of organic-walled fossils and analysed the $δ$¹³C of chitinozoans (traditionally viewed as metazoan egg cases, but recently re-interpreted as protists; Liang et al. 2020, but see Vodička et al. 2022) and scolecodonts (polychaete worm jaws), but both of these organic-walled microfossils had more negative $δ$¹³C values than the bulk values. Lecuyer and Paris (1997) recorded the following $δ$¹³C values from picked organic remains in a residue from the late Ludlow Kopanina Formation of the Czech Republic: graptolites −29.5‰; scolecodonts −28.0‰; and leiospheres −27.0‰. It is easy to see how the $δ$¹³C_org of a bulk organic sample dominated by graptolites could differ significantly from one dominated by leiospheres. Similarly, a change in the contribution to amorphous organic matter by organisms with very different $δ$¹³C values would inevitably influence the bulk-rock $δ$¹³C_org.

A change in the relative proportions of the organic matter preserved on the seafloor resulting from an organism enriched in ¹³C flourishing in waters overlying the deep-water environment at Sommerodde at the time of the SOCIE represents a potential explanation for both the excursion itself and, assuming that the organism was restricted to this deep-marine environment or the waters above it, also the atypical increasing magnitude of the SOCIE with increasing water depth.

The SOCIE extends through 10 m of pale, organic-poor, deep-marine strata in the O. spiralis graptolite Biozone of the Sommerodde-1 core, Bornholm. It has the highest magnitude of any Upper Ordovician–lower Silurian positive isotope excursion in this core. Elsewhere, however, with the exception of the Aizpute-41 core, Latvia (also a deeper water palaeoenvironment), the magnitude of the excursion is much less pronounced, becoming a minor event (or even unrecognizable) in shallower water settings, thus revealing a trend unique within the Silurian of increasing excursion magnitude with water depth.

Vandenbroucke et al. (2013) emphasized the challenges of interpreting changes in $δ$¹³C_org, stating:

… the interpretation of $δ$¹³C_org can be complex because of the various factors that influence the values, including varying sources of organic matter (e.g. bacteria, phytoplankton, zooplankton), varying production rates of the organic material, growth rates and geometry of phytoplankton cells, and the potential presence of carbon concentrating mechanisms, properties of the ambient sea water (temperature, pCO₂, light, nutrients), preservation and diagenesis, the inorganic C pool, and the analytical (e.g. acid digestion) method.

This host of factors must be borne in mind.

With due consideration of these listed factors, many of which are impossible to know or quantify in the geological record, we suggest that the unusually high magnitude of the SOCIE in the Sommerodde-1 core is most probably the result of a change in the relative contributions of different sources of the organic carbon preserved and analysed, with an organism with higher ¹³C flourishing at this time only in the seas above deeper water environments.

We thank Emma Hammarlund for her comments on a draft of this work, Jiří Frýda for discussions and Roy Smith for his work on the figures. We also appreciate the comments and suggestions of Sigitas Radzevičius and an anonymous referee.

DKL: writing – original draft (lead); JCG-M: writing – review and editing (supporting); PŠ: writing – review and editing (supporting).

We thank Emma Hammarlund for funding the carbon isotope analyses.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

All data generated or analysed during this study are included in this published article (and its Supplementary information files).