Linking the Progressive Expansion of Reducing Conditions to a Stepwise Mass Extinction Event in the Late Silurian Oceans Chelsie N

https://doi.org/10.1130/G46571.1

Manuscript received 28 May 2019 Revised manuscript received 19 July 2019 Manuscript accepted 6 August 2019

Linking the progressive expansion of reducing conditions to a stepwise mass extinction event in the late Silurian oceans Chelsie N. Bowman1, Seth A. Young1, Dimitri Kaljo2, Mats E. Eriksson3, Theodore R. Them II4, Olle Hints2, Tõnu Martma2 and Jeremy D. Owens1 1Department of Earth, Ocean and Atmospheric Science–National High Magnetic Field Laboratory, Florida State University, Tallahassee, Florida 32306, USA 2Department of Geology, Tallinn University of Technology, Ehitajate tee 5, 19086 Tallinn, Estonia 3Department of Geology, Lund University, Sölvegatan 12, SE-223 62 Lund, Sweden 4Department of Geology and Environmental Geosciences, College of Charleston, Charleston, South Carolina 29424, USA

ABSTRACT fauna and a resur gence of microbially mediated The late Ludlow Lau Event was a severe biotic crisis in the Silurian, characterized by sedimentary facies (e.g., Talent et al., 1993; resurgent microbial facies and faunal turnover rates otherwise only documented during the Jeppsson, 2005; Calner, 2005, 2008; Eriksson “big fve” mass extinctions. This asynchronous late Silurian marine extinction event preceded et al., 2009). an associated positive carbon isotope excursion (CIE), the Lau CIE, although a mechanism Despite the magnitude and complexity of for this temporal offset remains poorly constrained. Here, we report thallium isotope data the LKE, its mechanistic underpinnings are from locally reducing late Ludlow strata within the Baltic Basin to document the earliest not well constrained. An expansion of reduc- onset of global marine deoxygenation. The initial expansion of anoxia coincided with the on- ing conditions has been implicated as a po- set of the extinction and therefore preceded the Lau CIE. Additionally, sulfur isotope data tential driver of the observed stepwise extinc- record a large positive excursion parallel to the Lau CIE, interpreted to indicate an increase tion (e.g., Munnecke et al., 2003; Stricanne in pyrite burial associated with the widely documented CIE. This suggests a possible global et al., 2006). This hypothesis is also used to expansion of euxinia (anoxic and sulfdic water column) following deoxygenation. These data explain the possibility of extensive burial of are the most direct proxy evidence of paleoredox conditions linking the known extinction to organic carbon, resulting in the Lau positive the Lau CIE through the progressive expansion of anoxia, and most likely euxinia, across carbon isotope excursion (CIE; e.g., Saltzman, portions of the late Silurian oceans. 2005), but this cannot explain the temporal offset between the LKE and the Lau CIE. Fur- INTRODUCTION graptolite studies of deeper-water shale succes- ther, organic carbon burial can be affected by High rates of evolutionary turnover and sions (termed the Kozlowskii event; Koren’, other factors (e.g., Canfeld, 1994). Variations severe, punctuated extinctions of marine taxa 1993; Urbanek, 1993), herein referred to as the in eustatic sea level and carbonate weather- were hallmarks of the Silurian (e.g., Jeppsson, Lau/Kozlowskii extinction (LKE). The LKE ing rates have also been invoked as potential 1998; Crampton et al., 2016). These events is at least the tenth largest extinction event mechanisms for driving positive CIEs (e.g., occurred during the transition from the Late in Earth history, with ∼23% loss of genera Hirnantian CIE; Kump et al., 1999). A global Ordovician icehouse to Devonian greenhouse (e.g., Bond and Grasby, 2017, and references expansion of reducing conditions, however, worlds, when a dynamic ocean-atmosphere therein). In addition to conodonts and grapto- provides a kill mechanism and can be tested system oscillated between cool and warm lites, it affected a wide range of marine taxa, using combined traditional and novel paleo- conditions (e.g., Jeppsson, 1998). Recurrent including brachiopods (Talent et al., 1993), redox proxies. extinctions in graptolites and conodonts were fshes (Eriksson et al., 2009), and acritarchs This study investigated the relationship be- associated with the transitions between the (Stricanne et al., 2006). Extinctions of indi- tween the LKE and Lau CIE in the context of alternating climate states, the most notable vidual taxonomic groups were asynchronous, changing marine redox conditions in Upper Si- being the globally documented late Ludlow with documented extinctions in benthic and lurian (Ludfordian Stage) strata from the Baltic Lau extinction (Jeppsson, 1998; Calner, 2005; nektonic groups preceding the planktic organ- Basin. In order to reconstruct the evolution of Crampton et al., 2016). This extinction was frst isms (e.g., Munnecke et al., 2003; Stricanne global marine redox conditions, we measured recognized using conodonts from carbonate et al., 2006; Calner, 2008). The LKE shares thallium (Tl) isotopes, manganese (Mn) con- 34 platform successions (termed the Lau Event; similar characteristics to the “big fve” mass centrations, and pyrite sulfur isotopes (δ Spyr) e.g., Jeppsson and Aldridge, 2000) and then in extinctions, such as the survival of disaster from a distal shelf/slope setting (Latvia), and

CITATION: Bowman, C.N., et al., 2019, Linking the progressive expansion of reducing conditions to a stepwise mass extinction event in the late Silurian oceans: Geology, v. 47, p. 968–972, https://doi.org/10.1130/G46571.1

968 www.gsapubs.org | Volume 47 | Number 10 | GEOLOGY | Geological Society of America

Downloaded from https://pubs.geoscienceworld.org/gsa/geology/article-pdf/47/10/968/4830621/968.pdf?casa_token=bVkh5kLY-Z0AAAAA:359HkSvelJv13qxZCterkUhCPVlSWGIJq3TGaDdjftR2LMRqK1rxLmTc8uiLAhi7IzCyGQ by Florida State University user on 05 April 2020 GEOLOGIC SETTING isotopes (ε205Tl = [{R /R } – 1] × 104) Estonia sample reference Land In the late Silurian, the Baltic Basin was lo- have been used to investigate the earliest onset of Shallow Shelf cated in a tropical, epicratonic seaway on the global marine deoxygenation during Mesozoic Deep Shelf Latvia southern margin of the paleocontinent Bal- CIEs (Ostrander et al., 2017; Them et al., 2018). Ocean tica (e.g., Eriksson and Calner, 2008; Fig. 1; During Mn-oxide precipitation, Tl is adsorbed Fig. DR1 in the GSA Data Repository1). with a large positive isotope fractionation, thus Lithuania The northern and eastern edges of the basin leaving seawater isotopically lighter (as re- P-20 were delineated by rimmed carbonate shelves viewed in Nielsen et al., 2017). Precipitation and U-1 with parallel facies belts ranging from lagoonal burial of Mn-oxides require oxic bottom-water Sweden deposits in the north to deep shelf shales and conditions, the expanse of which represents the marls in the south, deepening toward the Rheic dominant seawater control of Tl isotope com- Poland Ocean (Eriksson and Calner, 2008). The Ud- position on time scales <∼5 m.y. (Nielsen et al., dvide-1 drill core and nearby outcrops on the 2017; Owens et al., 2017). The global seawater island of Gotland, Sweden, are predominantly Tl isotope signal is recorded in euxinic marine composed of carbonates from the shallow shelf settings and in anoxic waters with sulfde near area of the basin (for more details, see Eriks- the sediment-water interface from basins that Rheic Ocean son and Calner, 2008). The Priekule-20 drill are well connected to the open ocean (Owens core from southwestern Latvia predominantly et al., 2017). Consequently, reconstruction of the 0km 100 200 o consists of shales and marls from a correlative Tl isotope composition of late Silurian seawater N 30 S deep shelf setting (details in Kaljo et al., 1997). provides evidence for initial changes in global marine oxygenation by tracking the burial fux Figure 1. Paleogeographic reconstruction of METHODS AND RESULTS of Mn-oxides relative to the extinction, CIE, the late Silurian Baltic Basin region (modi- Two study localities were analyzed for δ13C and additional redox proxies (e.g., Ostrander fed from the Blakey Europe Series, Silurian records, organic or inorganic (micrite), to in- et al., 2017; Them et al., 2018). Importantly, a ca. 425 Ma; https://www2.nau.edu/rcb7/). vestigate carbon cycle dynamics. Pyrite sulfur large Tl isotope fractionation is not associated Locations of Gotland (Sweden) localities and the Latvian Priekule-20 drill core are was extracted from shale samples using a wide- with the burial of other Mn-bearing minerals marked by yellow stars. Detailed discussion ly accepted chromium reduction method, and (e.g., sulfdes, carbonates). Constraints on lo- of correlation and biostratigraphy of our two CAS was extracted from carbonates following cally reducing conditions using an independent localities can be found in the Data Repository standard methods. Full details of all analytical geochemical proxy are necessary to interpret Tl (see footnote 1). methods are given in the Data Repository. Sul- isotopes as a temporal seawater signature and to fur isotopes were analyzed to investigate global avoid contamination via local Mn-oxides (Ow- 34 carbonate-associated sulfate (CAS) sulfur iso- pyrite burial (δ SCAS) and the potential imprints ens et al., 2017). Low total Mn concentrations 34 34 topes (δ SCAS) from a shallow shelf setting (Got- on local signatures (δ Spyr). Sedimentary Tl [Mn] are indicative of locally reducing condi- land, Sweden; Fig. 1). This multiproxy, multili- tions (e.g., Boyer et al., 2011) and thus were thology approach aimed to establish frst-order utilized in this study. 1GSA Data Repository item 2019342, analytical 13 links among the fossil record of stepwise extinc- The Lau CIE is documented in the δ Ccarb methods, geochemical data, and cross plots, is 13 tion, carbon burial, and the progression from available online at http://www.geosociety.org/ and δ Corg records from the outer shelf setting more oxygenated to more reducing conditions datarepository/2019/, or on request from editing@ (Fig. 2A), with values increasing within the in the late Silurian seas. geosociety.org. upper part of the graptolite Bohemograptus

13 Mn-oxide burial Corg (‰ VPDB) Figure 2. Geochemi- graptolite Increase Decrease Age biostrat Fm. -29 -28 -27 -26 -25 -24 cal data from the Priekule-20 drill core, M. formosus ADB C Mituva 980 near Priekule, Latvia. Graptolite biozones are after Kaljo et al. (1997).

Lau/Kozlowskii extinc- erugnE 1000 tion interval is shaded in yellow. L.—Lobograp- tus; S.—Saetograptus; 1020 B.—Bohemograptus; N.— no graptolites present Neocucullograptus;

LUDFORDIAN M.—Monograptus. (A)

Nova Carbonate and organic N. kozlowskii 1040 Lau Kozlowskii carbon isotope data. Extinction Event B. bohemicus tenuis VPDB—Vienna Peedee

asybuD belemnite. (B) Pyrite S. lientward- inensis epušeŠ 1060 sulfur isotope data. VCDT—Vienna Canyon Diablo troilite. (C) Manga- L. scanicus xxxxxxxxxxxxx xxxxxxxxxxxxx 1080 nese concentration data, Average Crustal Value Average (m) with dashed line repre- GORSTIAN senting average crustal 012345 -30 -20-10 010 400500 600700 800 900 -4.5 -3.5 -2.5 shale orthocone brachiopod values. (D) Thallium iso- crinoid rugose coral 13C (‰ VPDB) 34S (‰ VCDT) [Mn] (ppm) 205 carb pyr Tl tope data plotted with 2σ marl burrows graptolite [Mn]carbcorr (ppm) xxxxx ash bed error bars.

Geological Society of America | GEOLOGY | Volume 47 | Number 10 | www.gsapubs.org 969

Downloaded from https://pubs.geoscienceworld.org/gsa/geology/article-pdf/47/10/968/4830621/968.pdf?casa_token=bVkh5kLY-Z0AAAAA:359HkSvelJv13qxZCterkUhCPVlSWGIJq3TGaDdjftR2LMRqK1rxLmTc8uiLAhi7IzCyGQ by Florida State University user on 05 April 2020 bohemicus tenuis–Neocucullograptus ko- overlying Burgsvik Oolite (conodont Ozarkodina reducing bottom-water oxidants such as oxy- zlowskii zones, reaching peak values of + 5.7‰ snajdri Zone). The overlying Hamra and Sundre gen and Mn-oxides, but not yet reducing them and −23.5‰, respectively, in the Nova Beds of Formations record the falling limb of the CIE, enough to increase widespread organic carbon the Dubysa Formation. The δ13C records return but not post-excursion baseline values. There is preservation and burial (e.g., Ostrander et al., 34 to baseline values in the overlying Engure For- a positive ∼30‰ excursion in the δ SCAS record 2017; Them et al., 2018). This early onset of mation. There is a positive ∼40‰ excursion in (Fig. 3B), with initial values ∼+11‰ in the När deoxygenation coincided with the initial phase 34 δ Spyr data (baseline values ∼–22‰ shift to peak Formation that rise through the Eke Formation to of extinction (e.g., brachiopods, fsh, and con- 34 values of up to ∼+15‰) that coincides with the values of ∼+22‰. The δ SCAS values then conodonts) that predated the Lau CIE (Fig. 4; e.g., Lau CIE within the upper Nova Beds and basal tinue rising through the Burgsvik Oolite, Hamra, Calner, 2008). Extinctions in these nektonic and Engure Formation (Fig. 2B). [Mn] values are and Sundre Formations to +40.6‰. benthic taxa coincide with the rising limb of low throughout the section, with average val- the positive Tl isotope excursion, which begins ues of 363 and 552 ppm in the lower to middle DISCUSSION ∼8 m before the Lau CIE. This suggests that Dubysa and Engure Formations, respectively The [Mn] values from the Latvia deep deoxygenation and the subsequent spread of (Fig. 2C). The ε205Tl record shows baseline shelf setting (Fig. 2C) are all below the aver- anoxia were responsible for the initial phases values in the lower to middle Dubysa section age crustal values and suggest depleted local of extinction in faunas living at or near the sed- of −4.1 to −4.6 (Fig. 2D). This is followed by Mn deposition, and therefore locally reduc- iment-water interface and within deeper waters a positive excursion in ε205Tl values that peaks ing conditions throughout the studied interval ∼175–270 k.y. prior to the Lau CIE (see the at −2.6 and averages −3.3 throughout the rest (Boyer et al., 2011). A cross-plot of [Mn] and Data Repository for calculations). For the frst of the drill core. Tl isotopes shows no signifcant correlation time in the Paleozoic, this stratigraphic relation- The Lau CIE is also documented in both the (Fig. DR2I). Thus, it is unlikely that local Mn- ship between extinction/turnover and carbon 13 13 δ Ccarb and δ Corg records from the inner shelf oxide burial infuenced the Tl isotope seawater isotopes and Tl isotopes is observed. A similar carbonates (Fig. 3A), with values beginning to signature. The observed positive shift in ε205Tl progression of events has been suggested for rise in the upper När Formation and peaking in from ∼−4.6 to − 2.6 begins within the Ludford- two Mesozoic oceanic anoxic events (OAEs; the overlying Eke Formation (conodont Icri- ian B. bohemicus tenuis graptolite biozone (Fig. Ostrander et al., 2017; Them et al., 2018), but odontid Zone) at +7.5‰ and −24.0‰, respec- 2D) and signifes a decline in the global buri- with varying magnitudes and durations. 13 tively. Within the Burgsvik Sandstone, δ Ccarb al of Mn-oxides. The reduction in Mn-oxide The positive carbon and sulfur isotope excur- 13 values decline to ∼+4.0‰, while δ Corg values burial was likely due to signifcant bottom- sions (Figs. 2 and 3) are consistent with transient increase to −22.9‰, which is followed by a water deoxygenation as anaerobic microbial increases in the amount of reduced carbon and 13 return to peak δ Ccarb values of +7.5‰ in the metabolisms kept pace with carbon export, sulfur buried globally as organic matter, pyrite, and possibly organic sulfur compounds (e.g., Gill et al., 2011; Owens et al., 2013; Raven

13 et al., 2019). Sea level may also have been a Corg (‰ VPDB) conodont contributing, but secondary, factor to the Lau Age biostrat Fm. -30-28 -26-24 0 CIE (see the Data Repository for further dis- Sundre A B cussion). This suggests that the Lau CIE began Zone

O. crispa Hamra Outcrop 10 as export and burial of organic carbon to the upper part seafoor outpaced consumption via remineral- Oolite Member 20 ization, which was likely dependent on a suf- fciently large portion of shelf and other marine environments being affected by deoxygenation 30 and expansion of anoxia. The excess organic O. snajdri Burgsvik matter available fueled microbial sulfate re- 40 duction (MSR) and ultimately increased pyrite Sandstone Member

lower part burial as MSR-produced H2S reacted with reac- 50 tive iron minerals in sulfdic environments. The Lau Kozlowskii Ludfordian Zone Extinction Event burial fractions of reduced carbon and sulfur

Eke 60 were preferentially enriched in 12C and 32S due Uddvide-1 Drill Core

Icriodontid to fractionation associated with biological pro- 70 cesses, and the remaining seawater was enriched 13 34

Zone in C and S. Euxinic conditions possibly ex- r panded into a greater portion of the oceans at Nä 80

siluricus the onset of the Lau CIE, as denoted by positive P. Upper/Main Part 34 34 (m) excursions in δ Spyr and δ SCAS (Figs. 2B and skeletal 1234567 10 20 30 40 sandstone 3B). This onset of euxinia temporally coincided pack/grainstones 13 34 Ccarb (‰ VPDB) SCAS (‰ VCDT) with the second wave of extinctions that affected oolitic siltstone oncoids grainstone planktic groups (i.e., ∼75% loss in biodiversity crinoids argillaceous of graptolites) and the rising limb of the CIE shale wackestone graptolites (Fig. 4A). Phytoplankton (e.g., acritarchs) actu- ally increased in abundance immediately prior Figure 3. Geochemical data from the Uddvide-1 drill core and nearby outcrops on Gotland, to and during the rising limb of the Lau CIE Sweden. Conodont biozones are after Jeppsson (2005). Lau/Kozlowskii extinction interval is shaded in yellow. P.—Polygnathoides; O.—Ozarkodina. (A) Carbonate and organic carbon (Stricanne et al., 2006) as reducing conditions isotope data. VPDB—Vienna Peedee belemnite. (B) Carbonate-associated sulfate (CAS) sulfur expanded, likely due to a lack of predation as isotope data. VCDT—Vienna Canyon Diablo troilite. zooplankton and larger ma rine taxa experienced

970 www.gsapubs.org | Volume 47 | Number 10 | GEOLOGY | Geological Society of America

in diversity conodonts 40 fish 40 work was performed at the National High Magnetic brachiopods 2006), graptolite (Manda 20 20 poorly constrained A et al., 2012), conodont Field Laboratory (Tallahassee, Florida), which is 6 (Calner, 2008), fsh (Eriks- supported by National Science Foundation Coop-

13 4 son et al., 2009), and erative Agreement No. DMR-1644779 and the State Ccarb Carbon 2 brachiopod (Talent et al., of Florida. (‰) B Burial 0 1993) extinctions. (B) Carbonate carbon iso- -3 tope record. (C) Thallium REFERENCES CITED 205Tl -4 Anoxia isotope record indicating Bond, D.P.G., and Grasby, S.E., 2017, On the causes C increased oceanic anoxia. of mass extinctions: Palaeogeography, Palaeocli- -5 (D) Carbonate-associated matology, Palaeoecology, v. 478, p. 3–29, https:// 40 sulfate (CAS) sulfur iso- doi.org/10.1016/j.palaeo.2016.11.005. 34S CAS Euxinia Boyer, D.L., Owens, J.D., Lyons, T.W., and Droser, (‰) 20 tope record suggesting ? D increased euxinia; dashed M.L., 2011, Joining forces: Combined biologi- 0 portions are expected, but cal and geochemical high-resolution palaeo- currently unconstrained oxygen history in Devonian epeiric seas: Palaeo- trends. (E) Depiction of geography, Palaeoclimatology, Palaeoecology, the biotic and marine v. 306, p. 134–146, https://doi .org/10.1016/ redox changes through- j.palaeo.2011.04.012. out the Gorstian and Calner, M., 2005, A Late Silurian extinction event and E Ludfordian. Stars mark anachronistic period: Geology, v. 33, p. 305–308, Go2 Lu1 Lu2 Lu3 positions of U-Pb zircon https://doi.org/10.1130/G21185.1. GORSTIAN LUDFORDIAN dates on K-bentonite Calner, M., 2008, Silurian global events—At the tip- SILURIAN ping point of climate change, in Elewa, A.M.T., 424.08±0.20 Ma 422.91±0.07 Ma ash beds from Podolia, Ukraine (Cramer et al., ed., Mass Extinctions: Berlin, Springer-Verlag, 2015). Constr uction of p. 21–57, https://doi.org/10.1007/978-3-540- this fgure is detailed in 75916-4_4. the Data Repository (see Canfeld, D.E., 1994, Factors infuencing organic car- footnote 1). bon preservation in marine sediments: Chemi- cal Geology, v. 114, p. 315–329, https://doi .org/10.1016/0009-2541(94)90061-2. earlier extinctions. Acritarchs fnally declined for the LKE that was associated with the pro- Cramer, B.D., Schmitz, M.D., Huff, W.D., and with an associated ∼95% drop in abundance dur- gressive expansion of reducing marine condi- Bergström, S.M., 2015, High-precision U-Pb ing the peak of the CIE just before the global tions. Increased anoxic and euxinic conditions zircon age constraints on the duration of rapid extent of euxinia reached a maximum, which is likely shoaled from deeper shelf/slope areas to biogeochemical events during the Ludlow Ep- inferred by the rising limb nearing the peak of shallow platform settings during the Lau CIE. och (Silurian Period): Journal of the Geological Society [London], v. 172, p. 157–160, https://doi 34 the δ SCAS values. Our Tl isotope data provide detailed evidence .org/10.1144/jgs2014-094. In both the Gotland and Latvia δ34S records, for expanding marine deoxygenation in the in- Crampton, J.S., Cooper, R.A., Sadler, P.M., and Foote, peak excursion values postdate the correspond- terval that preceded the Lau CIE, coinciding M., 2016, Greenhouse-icehouse transition in the ing peak δ13C values in the Lau CIE. The offset with the initial phase of extinction. This was Late Ordovician marks a step change in extinction regime in the marine plankton: Proceedings of in these records suggests that organic carbon followed by global carbon and sulfur isotope the National Academy of Sciences of the United burial fueled high MSR rates, but it may also perturbations that coincided with continued States of America, v. 113, p. 1498–1503, https:// be related to differences in oceanic residence marine extinction. This study highlights the doi.org/10.1073/pnas.1519092113. times and/or continued pyrite burial post-CIE role of oxygen depletion (i.e., nonsulfdic anox- Eriksson, M.E., Nilsson, E.K., and Jeppsson, L., 2009, (e.g., Owens et al., 2013). The falling limb of ia) near the onset of biotic change and provides Vertebrate extinctions and reorganizations during the Late Silurian Lau event: Geology, v. 37, 34 the δ Spyr record also lags the falling limb of a mechanism for the previously documented p. 739–742, https://doi.org/10.1130/G25709A.1 . the CIE, perhaps indicating the continued con- stepwise extinction event. The progressive ex- Eriksson, M.J., and Calner, M., 2008, A sequence strati- sumption of previously exported organic car- pansion of oceanic anoxia leading to euxinia graphical model for the late Ludfordian (Silurian) bon after the termination of the burial event, is a potential mechanism for extinction due to of Gotland, Sweden: Implications for timing between changes in sea level, palaeoecology, and which is corroborated by Tl isotopes not re- signifcant stress on marine ecosystems (Meyer the global carbon cycle: Facies, v. 54, p. 253–276, turning to baseline values (cf., Them et al., and Kump, 2008), which might be similar to https://doi.org/10.1007/s10347-007-0128-y. 2018). Regardless of the C-S offset and isotopic at least two Mesozoic OAEs. More broadly, Gill, B.C., Lyons, T.W., Young, S.A., Kump, L.R., magnitudes, these large-magnitude S-isotope this study indicates that global marine redox Knoll, A.H., and Saltzman, M.R., 2011, Geo- excursions that span 1 m.y. require a reduc- dynamics were a major driver in the evolution chemical evidence for widespread euxinia in the ∼ later Cambrian ocean: Nature, v. 469, p. 80–83, tion in the marine sulfate reservoir, which was of the late Silurian biosphere, and potentially https://doi.org/10.1038/nature09700. likely signifcantly lower than modern seawater other Paleozoic biotic crises. The multiproxy Jeppsson, L., 1998, Silurian oceanic events: Summa- (e.g., Gill et al., 2011). redox approach provides a more holistic global ry of general characteristics, in Landing, E., and view of redox changes and supports recent evi- Johnson, M.E., eds., Silurian Cycles: Linkages of Dynamic Stratigraphy with Atmospheric, Oce- CONCLUSIONS dence suggesting prevalent low-oxygen condi- anic, and Tectonic Changes: James Hall Centen- The integrated paleontological and geo- tions in the upper oceans of the Paleozoic (Lu nial Volume: Albany, New York, New York State chemical records suggest a stepwise extinction et al., 2018). Museum Bulletin 491, p. 239–257.

Geological Society of America | GEOLOGY | Volume 47 | Number 10 | www.gsapubs.org 971

atmospheric pCO2 during the Late Ordovician: Severmann, S., Kuypers, M.M.M., Woodfne, 0182(93)90126-4. Palaeogeography, Palaeoclimatology, Palaeoecol- R.G., and Lyons, T.W., 2013, Sulfur isotopes Them, T.R., II, Gill, B.C., Caruthers, A.H., Gerhardt, ogy, v. 152, p. 173–187, https://doi.org/10.1016/ track the global extent and dynamics of eux- A.M., Gröcke, D.R., Lyons, T.W., Marroquin, S0031-0182(99)00046-2. inia during Cretaceous oceanic anoxic event S.M., Nielsen, S.G., Trabucho Alexandre, J.P., Lu, W., et al., 2018, Late inception of a resiliently oxy- 2: Proceedings of the National Academy of and Owens, J.D., 2018, Thallium isotopes reveal genated upper ocean: Science, v. 361, p. 174–177, Sciences of the United States of America, v. protracted anoxia during the Toarcian (Early Ju- https://doi.org/10.1126/science.aar5372. 110, p. 18407–18412, https://doi.org/10.1073/ rassic) associated with volcanism, carbon burial, Manda, S., Štorch, P., Slavík, L., Fryda, J., Križ, J., and pnas.1305304110. and mass extinction: Proceedings of the Nation- Tasáryová, A., 2012, The graptolite, conodont and Owens, J.D., Nielsen, S.G., Horner, T.J., Ostrander, al Academy of Sciences of the United States of sedimentary record through the late Ludlow Ko- C.M., and Peterson, L.C., 2017, Thallium-isotope America, v. 115, p. 1–6, https://doi.org/10.1073/ zlowskii event (Silurian) in the shale-dominated compositions of euxinic sediments as a proxy for pnas.1803478115. succession of Bohemia: Geological Magazine, v. global manganese-oxide burial: Geochimica et Urbanek, A., 1993, Biotic crises in the history of the 149, no. 3, p. 507–531, https://doi.org/10.1017/ Cosmochimica Acta, v. 213, p. 291–307, https:// Upper Silurian graptoloids: A paleobiological S0016756811000847. doi.org/10.1016/j.gca.2017.06.041. model: Historical Biology, v. 7, p. 29–50, https:// Meyer, K.M., and Kump, L.R., 2008, Oceanic euxinia Raven, M.R., Fike, D.A., Bradley, A.S., Gomes, M.L., doi.org/10.1080/10292389309380442. in Earth history: Causes and consequences: An- Owens, J.D., and Webb, S.A., 2019, Paired organ- nual Reviews of Earth and Planetary Science Let- ic matter and pyrite δ34S records reveal mecha- Printed in USA

972 www.gsapubs.org | Volume 47 | Number 10 | GEOLOGY | Geological Society of America