Shallow water redox conditions from the Permian–Triassic boundary microbialite: The rare earth element and iodine geochemistry of carbonates from Turkey and South China

Chemical Geology 351 (2013) 195–208

Contents lists available at SciVerse ScienceDirect

Chemical Geology journal homepage: www.elsevier.com/locate/chemgeo

Shallow water redox conditions from the Permian–Triassic boundary microbialite: The rare earth element and iodine geochemistry of carbonates from Turkey and South China Garrison R. Loope ⁎, Lee R. Kump, Michael A. Arthur Department of Geosciences, The Pennsylvania State University, University Park, PA 16802, USA

a r t i c l e

i n f o

Article history: Received 25 September 2012 Received in revised form 4 May 2013 Accepted 15 May 2013 Available online 24 May 2013 Editor: U. Brand Keywords: End-Permian extinction Permian–Triassic boundary microbialite Ce anomaly Anoxia South China Turkey

a b s t r a c t Redox sensitive elements serve as useful proxies of the oxygenation state of ancient environments, but their interpretation may not always be straightforward. To evaluate the inherent complexities, rare earths and yttrium (REY), iodine, and major element concentrations are determined in two carbonate sections spanning the Permian–Triassic (P–Tr) transition in Demirtas, Turkey and Cili, South China. We use major oxides to identify non-seawater REY sources such as siliciclastics, Fe-oxides, phosphates and diagenetic fluids. Additionally, we employ Y/Ho ratio, La anomaly, and light rare earth element depletion to identify which samples preserve a seawater-like REYSN distribution. In contrast to past interpretations, we find that the P–Tr boundary microbialites in both sections contain REY signatures indicative of deposition in an oxic environment. These boundary microbialites have their base at the extinction horizon and are widespread within the Tethyan region. In the Cili section, the underlying Permian limestone also preserves a seawater-like REY signature with a negative Ce anomaly. This indicates that the water column was oxygenated both before and after the extinction event. The Permian limestone in the Demirtas section does not preserve a seawater-like REYSN distribution, so the absence of a Ce anomaly cannot be used to distinguish prevailing redox conditions during deposition. However, in these samples, we find the presence of a diverse Permian benthic community sufficient to identify deposition in an oxic environment. The geochemical evidence for a continuously oxic environment during the deposition of the boundary microbialite presented in this study strongly supports work done using ostracods as redox indicators within the boundary microbialite in South China. The microbialite has been proposed as a disaster facies, in part resulting from the exclusion of grazers. Although anoxia is one of the suggested mechanisms for the exclusion of grazers, we find that it is not supported by geochemical and biotic evidence. © 2013 Elsevier B.V. All rights reserved.

1. Introduction The end-Permian mass extinction [~252 Ma] was the most severe biodiversity crisis of the Phanerozoic and is estimated to have caused the extinction of up to 96% of marine species (Raup, 1979). Determining the cause of this extinction has been the focus of much scientific investigation and has led to a wide range of hypotheses for the cause of the extinction (Hallam, 1989; Renne and Basu, 1991; Wignall and Hallam, 1992; Erwin, 1994; Isozaki, 1997; Krull and Retallack, 2000; Kump et al., 2005; Knoll et al., 2007; Brand et al., 2012; Joachimski et al., 2012). Many hypotheses (Wignall and Hallam, 1992; Kump et al., 2005; Meyer et al., 2008; Algeo et al., 2010) incorporate anoxia and euxinia (presence of aqueous hydrogen sulfide) because of the evidence supporting low O2 and high H2S in the Late Permian–Early Triassic oceans. The data that support anoxia and euxinia come from many ⁎ Corresponding author at: 1300 N 37th St. Lincoln, NE 68503, USA. Tel.: +1 814 769 0559. E-mail addresses: [email protected] (G.R. Loope), [email protected] (L.R. Kump), [email protected] (M.A. Arthur). 0009-2541/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.chemgeo.2013.05.014

different indicators from a wide geographic range, but there is still much disagreement on the timing and extent of anoxia and euxinia. Sedimentological and geochemical evidence including fine laminations, small pyrite framboids, dysaerobic benthos, high Th/U ratios, and a negative shift in δ238U suggest a nearly synchronous onset at the extinction event (Wignall and Twitchett, 2002; Wignall et al., 2005; Brennecka et al., 2011). In contrast, other studies have found evidence for anoxia and even biomarkers for photic zone euxinia starting well before the extinction (Isozaki, 1997; Cao et al., 2009; Hays et al., 2012). Across much of the Tethys, the diverse carbonate reef-deposits of the Late Permian were abruptly replaced by microbialites at the extinction horizon (Kershaw et al., 2011). An invasion of anoxic, CO2-rich waters onto the carbonate shelves has been proposed as a mechanism that would both stimulate carbonate precipitation by increasing microbial photosynthesis and eliminate the grazers that feed on microbial mats (Lehrmann et al., 2003; Pruss and Bottjer, 2004; Groves et al., 2005). Despite the interest in anoxia at the mass extinction, only two studies have investigated the paleo-redox conditions of the P–Tr boundary microbialite and these have found conflicting evidence. Liao et al. (2010) reported framboidal pyrite size distributions from

196

G.R. Loope et al. / Chemical Geology 351 (2013) 195–208

the Laolongdong microbialite and concluded that they likely represented the lower-dysoxic environment (Bond and Wignall, 2010). Although the size distribution of these framboids does suggest a low-oxygen environment, framboids of similar size are reported in sediments below fully oxic water bodies (Wilkin et al., 1996). Forel et al. (2009) used ostracod trophic groups to interpret redox condition of the boundary microbialite. They found that the deposit-feeding guild was dominant, which indicates a well-oxygenated environment. An oxic interpretation has also been recently suggested for Early Triassic (Spathian) microbialites in the Moenkopi Formation at Lost Cabin Spring, Nevada on the basis of low organic carbon and pyrite contents (Marenco et al., 2012). Although the Spathian microbialites were deposited after the boundary microbialite, they have also previously been suggested to represent episodic flooding of anoxic, high-alkalinity water onto carbonate shelves (Pruss and Bottjer, 2004; Mata and Bottjer, 2011). In this study we examine the distribution of rare earths and yttrium (REY) and iodine in two carbonate sections containing boundary the microbialite from Demirtas, Turkey, and Cili County, China (Fig. 1), in order to better constrain the redox conditions of shallow-water environments through the P–Tr transition. We find that a straightforward application of the Ce anomaly redox proxy for all samples leads to an apparent paradox with paleontological data as reported by Kakuwa and Matsumoto (2006). We then use techniques of REY analysis developed on modern and ancient microbialites (Webb and Kamber, 2000; Kamber and Webb, 2001; Van Kranendonk et al., 2003; Nothdurft et al., 2004) to integrate the geochemical redox indicators with paleontological data and provide a more parsimonious redox history. We also test the applicability of a recently proposed redox indicator, iodine, in our sections. By obtaining a better understanding of the redox conditions of shallow carbonate platforms during the Late Permian–Early Triassic, we can evaluate anoxia and euxinia as kill-mechanisms and as drivers of the associated abrupt transition to a microbially dominated carbonate ecosystem. 1.1. Cerium anomaly redox proxy The Ce anomaly is a widely used redox indicator that makes use of the high redox sensitivity of Ce and its relative depletion in oxic waters (German and Elderfield, 1990 and references therein). Microbial carbonates have been recently identified as an ideal REY archive because they both faithfully record the distribution of REY in the water column and have relatively high overall REY concentrations (Webb and Kamber,

2000). The Ce anomaly has been typically calculated as Ce/Ce* = 2 [CeSN]/([LaSN] + [PrSN]) (Bau and Dulski, 1996) where all concentrations are normalized to a shale composite (indicated as SN; usually PAAS; McClennan, 1995). Unfortunately, interpretation of Ce anomalies is complicated by the anomalous behavior of La in seawater. Although not widely considered until ICP-MS allowed the measurement of Pr, a positive La anomaly is characteristic of seawater (De Baar et al., 1991). To avoid the influence of La while calculating the Ce anomaly, Bau and Dulski (1996) proposed that a negative Ce anomaly be defined by Pr/Pr* > 1.0 where Pr/Pr* = 2[PrSN]/([CeSN] + [NdSN]). We prefer this notation because values of Pr/Pr* > 1.0 only occur in Ce depleted conditions (e.g. oxygenated waters). However, we report both Pr/Pr* and Ce/Ce* in this study in order to facilitate comparisons with studies that lack Pr data. 1.2. Iodine redox proxy Iodine is highly sensitive to changes in redox conditions and has been recently proposed as an indicator of shallow-water redox conditions in carbonates (Lu et al., 2010). The iodine redox proxy remains relatively unproven but may be of great use in shallow carbonate environments because of its potential to record more subtle changes in pO2 that are expected to occur in shallow-water environments. In this study we test the applicability of the iodine proxy in a period where it is thought that shallow waters became depleted in dissolved oxygen. 2. Geological background 2.1. Cili section The Cili section is located outside of Kangjiaping in Cili County, northwestern Hunan, China. During the Late Permian and Early Triassic Cili was part of the Jiangan carbonate platform (Wang et al., 2009) that stretched across much of central China in what was the equatorial, eastern Tethys (Fig. 1). The biostratigraphy of the section is described in Wang et al. (2009), while the isotopic geochemistry is published in Luo et al. (2010, 2011). The section spans the Upper Permian Changxing and the Upper Permian-Lower Triassic Daye Formations (Fig. 2), and shows no evidence of a disconformity at the boundary (Wang et al., 2009; Luo et al., 2010). The Changxing Formation is a matrix supported bioclastic limestone with abundant foraminifera, ostracods and echinoderms. Above this, at

30°N

Panthalassa Ocean

Cili

0°

Demirtas Neo-Tethys 30°S

Fig. 1. Paleogeographic map with location of the Cili and Demirtas sites presented in this study (modified from Muttoni et al., 2009).

G.R. Loope et al. / Chemical Geology 351 (2013) 195–208

Demirtas

-5

-10

-15

Bioclastic Limestone Micrite Microbialite

15

10

5

Oolitic Limestone

Sapadere Formation (Lower Triassic)

Daye Fm. (Lower Triassic)

20 (m)

0

-5

-10

Yüglük Tepe Formation (Upper Permian)

0

Changxing Fm. (Upper Permian)

5

H. Parvus Zone

10

Palaeofusilina-Colaniella Zone

15

I. staeschei Zone

Cili 20 (m)

197

-15

Fig. 2. A schematic framework showing the stratigraphic logs of the investigated sections and the stratigraphic levels of the extinction events and Permian–Triassic boundary. All measurements are reported as distance from the extinction horizon.

the base of the Daye Formation, sits the partially dolomitized P–Tr boundary microbialite. The microbialite has rare ostracodes, gastropods, fragments of bivalves and brachiopods and the first occurrence of H. parvus 3.5 m from its base (Wang et al., 2009). The placement of the P–Tr boundary at the first occurrence of H. parvus within the upper microbialite follows the precedent set in other sections in South China (Lehrmann et al., 2003; Wang, 2005). Above the microbialite there are thin beds of bioclastic limestone composed of recrystallized shell fragments in a radial-fibrous cement. The overlying micrite and oolitic limestone, at the top of the section, have horizons where recrystallization of calcite has resulted in a sparry texture. 2.2. Demirtas section The Demirtas section is located about 10 km north of Demirtas in the central Taurus Mountains of southern Turkey. The stratigraphy and biostratigraphy of this section have been described in Groves et al. (2005) and Baud et al. (2005), whereas the isotopic geochemistry

(δ13CCarb, δ34SCAS, δ18OCarb) was reported in Riccardi (2007). The P–T boundary has been assigned to the base of the stromatolite facies (Fig. 2) by Groves et al. (2005). This corresponds to the event horizon distinguished by Riccardi (2007) as the steepest shift in the δ13CCarb excursion. The base of the Demirtas section consists of the Permian Yüglük Tepe Formation composed of heavily recrystallized bioclastic wackestones and packstones with abundant calcareous algae, echinoderms, gastropods, bivalves, brachiopods and foraminifera in a micritic matrix. The age of these beds is constrained to the Changxingian by the presence of the foraminferan Paradagmarita mondi (Groves et al., 2005). The uppermost beds of the Yüglük Tepe Formation (53 cm) consist of oolitic grainstones with abundant fragments of Permian mollusks, ostracods and foraminifera (Fig. 2). The upper 8 cm of the oolitic limestone is deeply weathered and marks a proposed exposure surface (Groves et al., 2005). Approximately 5 m of finely laminated stromatolites at the base of the Lower Triassic Sapadere Formation rest on this disconformity. Above the stromatolite is ~10 m of highly recrystallized micrite with rare ooids and

198

G.R. Loope et al. / Chemical Geology 351 (2013) 195–208

shell fragments. Groves et al. (2005) documented the biodiversity of the lagenide foraminifers throughout the section and show that the lagenide community was reduced from mean of 6.8 species/sample in the Upper Permian Yüglük Tepe Formation to 0.3 species/sample in the Lower Triassic Sapadere Formation.

for Isotopes and Metals in the Environment (LIME). Measurements of geological standard JDo-1 were performed every 7–10 samples throughout each sample run for quality control. The percent relative standard deviation for the 11 JDo-1 analyses is 6% for each of the REEs and 15% for iodine. When the measurements of JDo-1 are normalized to PAAS and the Ce anomaly is calculated, the percent relative standard deviation for the Ce anomaly is 1.3%. Major elemental analysis was performed using lithium metaborate fusion on a Perkin-Elmer Optima 5300Dv ICP-AES at LIME using the procedure of Ingamells (1970). Percent relative standard deviation was calculated as less than 2% for each oxide using repeated measurements of USGS rock standard W-2. Organic carbon content for the Cili samples was determined by combustion of dissolved residues in an Elemental Analyzer (CE Instruments NC2500).

3. Methods A total of 40 samples from the Cili section (N 29° 24′ 05″, E 110° 54′ 28″) and 32 samples from the Demirtas section (N 36° 28′ 96″, E 32° 14′ 99″) were obtained and analyzed. Acetate peels were stained and examined petrographically for evidence of alteration (Fig. 3). Samples were cut, cleaned, powdered, dissolved and analyzed for REY and iodine by ICP-MS using a method modified from Lu et al. (2010). 10 mg powder was weighed and digested overnight in 7 ml of 2% HNO3 and 0.5% tertiary amine. Tertiary amine was used to prevent volatilization of iodine and reduce the memory effect (Muramatsu and Wedepohl, 1998; Lu et al., 2010). Prior to analysis, solutions were centrifuged for 5 min and the supernatant was decanted to remove the insoluble residue. Samples were run for iodine and all measureable REY using a 50 ppb indium internal standard on a Thermo X-Series II quadrupole ICP-MS at the Laboratory

4. Results 4.1. Cili section The bioclastic limestone of the Late Permian has a strong negative Ce anomaly (Ce/Ce* = 0.52 to 0.69; Pr/Pr* = 1.08 to 1.17) (Fig. 4, see supplement for data tables). This negative Ce anomaly persists through the

A F Cili

Demirtas 20 (m)

B 15

G

10

C

5

H D

0

-5

I E -10

-15 Fig. 3. Photomicrographs from acetate peels. Calcite has been dyed with Alizarin red-S and dolomite remains light gray.

G.R. Loope et al. / Chemical Geology 351 (2013) 195–208

Iodine

Ce anomaly

13

CCarb

(

199

Organic C

Carbonates

(ppm)

Ce/Ce*

(wt.%)

(wt.%) % Total Carbonate

Pr/Pr*

% Dolomite

10

10

10

10

10

0

0

0

0

0

-10

-10

-10

-10

-10

-20

-20

-20

-20

-20

-1

0

1

2

3

4

0

0.5

La-anomaly

0

1

0.2

0.4

0.6

0.8

0

20

40

60

0

REY

Major Oxides

Y/Ho

(La/La*)

80

0.025

0.05

NdSN/YbSN

(ppm)

(wt.%)

10

10

10

10

10

0

0

0

0

0

-10

-10

-10

-10

-10

-20

-20

-20

-20

-20

Fe2O3 Al 2O3 SiO2

1

1.2 1.4 1.6

30

40

50

60

0

1

2

3

4

0

10

20

30

40

0.6 0.7 0.8 0.9 1 1.1

Fig. 4. Geochemical profiles for the Cili section. All depths are measured from the extinction horizon at the base of the microbialite. δ13Ccarb data are from Luo et al. (2010). Total carbonate and percent dolomite in all figures and tables are calculated from CaO and MgO ICP-AES data. All Mg is assumed to be in dolomite and all dolomite assumed to have a 50:50 Ca:Mg ratio. For total carbonate we assume all Ca and Mg are part of either dolomite or calcite.

boundary microbialite but is gradually lost upward into the overlying Triassic carbonates. The pre-extinction samples also have relatively low REY concentrations and silicate contents compared to relatively high REY concentrations and silicate contents in the post-extinction samples (Table 1). Iodine concentrations are depleted in the microbialite and have a negative correlation with percent dolomite (r = −0.61, n = 20; Table 2). Dolomite approaches 80 wt.% in the microbialite and is abundant in the overlying carbonates. The REYSN distribution is shown for the pre-extinction and postextinction carbonates in Fig. 5A. Ce/Ce* has a negative correlation with iodine and total carbonate (r = −0.71, −0.66; n = 20; Table 2) and a positive correlation with organic carbon and Σ REY (r = 0.70, 0.52; n = 20), while Pr/Pr* has similar but inverted relationships with the same analytes (r = 0.65, 0.57, −0.57, −0.34; n = 20). The Σ REY is correlated with Al2O3, TiO2, SiO2, and total carbonate (r = 0.57, 0.57, 0.54, −0.50; n = 20; Table 1) indicating that some samples may have

Table 1 Summary of geochemical data across the extinction horizon for Cili and Demirtas.

Cili Demirtas

Post-extinction Pre-extinction Post-extinction Pre-extinction

Σ REY ppm

Al2O3%

Mean

Stdev

Mean

Stdev

Mean

Stdev

22.0 8.9 18.9 27.3

7.9 6.7 15.6 18.3

0.71 0.02 0.51 0.82

0.35 0.01 0.29 0.41

1.05 1.13 1.05 0.96

0.02 0.03 0.05 0.03

Pr/Pr*

n

13 7 18 14

REY derived from a siliciclastic source. XRF elemental maps show that Y, La, Ti and Al are evenly distributed across all mapped samples while Si and Fe had heterogeneous distributions particularly in the microbialite. 4.2. Demirtas section The upper-Permian bioclastic limestone beds lack a strong Ce anomaly (Ce/Ce* = 0.95 to 1.04; Pr/Pr* = 0.87 to 1.00) (Fig. 6). These strata have relatively high concentrations of silicate components and high concentrations of REY (Table 1). The shale-normalized distribution of REY is shown in Fig. 5B. The overlying microbialite beds display a negative Ce anomaly (Ce/Ce* = 0.61 to 0.76; Pr/Pr* = 1.03 to 1.12) that gradually approaches unity upward into the micrite (at top of section Ce/ Ce* = 0.86, Pr/Pr* = 1.00; Fig. 6). Silicate components and Σ REY both decrease upwards in the Triassic strata (Al2O3 falls from 1.11 to 0.23 wt.%, Σ REY falls from 60.8 to 1.9 ppm). The Σ REY is correlated with Fe2O3, Al2O3, SiO2, and TiO2 (r = 0.61, 0.59, 0.64; n = 32; Table 3) suggesting that a portion of the REY are associated with siliciclastic material. Organic carbon content is low (b 0.05 wt.% in all samples) and is not correlated to I (r = −0.10, n = 32; Table 3). Iodine concentrations fluctuate from below detection limits to 0.73 ppm and are not strongly correlated to any other analyte. Dolomite varies from 1.0% to 16.6% across the entire section (Fig. 6) and total carbonate content ranges from 91.5 to 99.5%. XRF element maps of all Permian and Triassic samples show that Y, La, Al and Ti are evenly

200

G.R. Loope et al. / Chemical Geology 351 (2013) 195–208

Table 2 Correlation matrix of r values for major and trace element data from the Cili section.

Pr/Pr* Ce/Ce* I Σ REY Y/Ho La/La* NdSN/YbSN Fe2O3 MnO SiO2 Al2O3 TiO2 SrO CaMg(CO3)2 Tot. Carb. Org. C.

Pr/Pr*

Ce/Ce*

I

Σ REY

Y/Ho

La/La*

NdSN/YbSN

Fe2O3

MnO

SiO2

Al2O3

TiO2

SrO

CaMg(CO3)2

Tot. carb.

Org. c.

1

−0.95 1.00

0.65 −0.71 1.00

−0.34 0.52 −0.71 1.00

0.50 −0.72 0.62 −0.81 1.00

0.81 −0.95 0.76 −0.66 0.85 1.00

−0.26 0.44 −0.50 0.73 −0.69 −0.57 1.00

−0.47 0.61 −0.71 0.53 −0.65 −0.74 0.45 1.00

−0.32 0.51 −0.70 0.62 −0.72 −0.70 0.57 0.90 1.00

−0.49 0.64 −0.61 0.54 −0.68 −0.75 0.51 0.94 0.88 1.00

−0.53 0.66 −0.62 0.57 −0.64 −0.75 0.51 0.88 0.82 0.97 1.00

−0.52 0.64 −0.60 0.57 −0.63 −0.73 0.48 0.87 0.80 0.96 1.00 1.00

−0.43 0.28 0.27 −0.22 0.06 −0.07 −0.08 −0.27 −0.45 −0.20 −0.13 −0.12 1.00

−0.24 0.42 −0.61 0.63 −0.65 −0.61 0.43 0.84 0.87 0.80 0.70 0.68 −0.42 1.00

0.57 −0.66 0.58 −0.50 0.57 0.71 −0.48 −0.88 −0.78 −0.96 −0.97 −0.97 0.01 −0.67 1.00

−0.57 0.70 −0.58 0.59 −0.74 −0.75 0.33 0.75 0.67 0.77 0.72 0.71 −0.11 0.74 −0.68 1

distributed throughout the samples while Si, and Fe had heterogeneous distributions. Less abundant REEs could not be mapped, but ICP-MS data suggest that all REEs correlate with Y and La. 5. Discussion The Ce anomaly pattern in the Demirtas section is similar to those of four other west Tethyan sites: Gartnerkofel, Austria (Attrep et al., 1991); Idrijca Valley, Slovenia (Dolenec et al., 2001); Julfa, Iran (Kakuwa and Matsumoto, 2006) and Velebit Mountain, Croatia (Fio et al., 2010). These sites, which all lack a negative Ce anomaly before the extinction event, have a negative Ce anomaly (expressed as a positive Pr/Pr*) immediately following the event, and return to an absence of a Ce anomaly in the Early Triassic (Fig. 7). On the contrary, the Cili section shows a completely different Ce anomaly pattern than the other five sites. The difference is most notable in the Permian section where the Cili samples show a strong negative Ce anomaly. A straightforward interpretation of these patterns with the assumption that all Ce anomalies record primary seawater chemistry would lead to different redox histories in the eastern and western Tethys. In this interpretation, Cili would have been well oxygenated during the deposition of the Permian bioclastic limestone and boundary microbialite and would have become anoxic during the deposition of the Early Triassic micrite and oolitic limestone. Under the same assumptions, the Demirtas data would be interpreted as evidence for anoxic conditions in the Late Permian and Early Triassic with an interruption by an oxygenation event at the extinction horizon. This interpreted redox history for Demirtas is much the same as has been proposed for the Idjrica Valley section (Dolenec et al., 2001). In both the Demirtas and

Idrijca Valley sections, this leads to an apparent paradox between the Ce anomaly redox proxy and paleontological observations. Although neither of these sites has undergone a rigorous diversity analysis, the Permian bioclastic limestone at each site has been reported to contain a rich benthic community including foraminifera, echinoderms, gastropods, bivalves, and brachiopods. These observations are in direct conflict with an anoxic water column as inferred from the Ce anomaly. This leads us to question the assumption that the Ce anomaly is derived from a primary seawater source in all samples. Recent work on the REYSN distribution of carbonates and other chemical precipitates indicates that the features characteristic of modern seawater, with the exception of a negative Ce anomaly, have existed since the Archean and can be used to identify a seawater REY fingerprint (Van Kranendonk et al., 2003; Bolhar et al., 2004; Kamber and Bolhar, 2004; Nothdurft et al., 2004). These characteristics, including a superchondritic Y/Ho ratio, light rare earth element (LREE) depletion, and positive La anomaly, are caused by differential complexation chemistry among the REY. The Y/Ho ratio of seawater, as recorded in primary marine carbonates, is characterized by values > 44 (Webb and Kamber, 2000), whereas typical shale is ~ 28 (PAAS; McClennan, 1995). LREE depletion is quantified by NdSN/YbSN and is typically b 0.30 in shallow environments (Zhang, 1996; Alibo and Nozaki, 1999; Webb and Kamber, 2000). The La anomaly is difficult to quantify because La has only one REE neighbor, Ce, which behaves anomalously in oxidized conditions. Therefore, the La anomaly must be calculated by extrapolation from Pr and Nd using La/La* = [LaSN]/(3[PrSN] − 2[NdSN]) (Bolhar et al., 2004). Webb and Kamber (2000) found that the shallow carbonate environment at Heron Reef was characterized by La/La* greater than ~ 1.2.

1

1

A

B

Post-extinction

Pre-Extinction

Sample/PAAS

Sample/PAAS

Pre-extinction

0.1

0.01

Post-extinction

0.1

0.01 La Ce Pr Nd Sm Eu Gd Tb Dy Y Ho Er Tm Yb Lu

La Ce Pr Nd Sm Eu Gd Tb Dy Y Ho Er Tm Yb Lu

Fig. 5. Averaged REYSN distributions for Cili (A) and Demirtas (B) with Y inserted into the REEs according to its ionic radius after Bau and Dulski (1996). All values are normalized to PAAS (McClennan, 1995).

G.R. Loope et al. / Chemical Geology 351 (2013) 195–208 13

Iodine

Ce anomaly

CCarb

Organic C

Carbonates

(ppm)

(

201

(wt.%)

(wt.%) % Dolomite

Pr/Pr* Ce/Ce*

% Total Carbonate

10

10

10

10

10

5

5

5

5

5

0

0

0

0

0

-5

-5

-5

-5

-5

-10

-10

-10

-10

-10

-15

-15

-15

-15

-15

1

2

3

4

0

0.5

La anomaly

1

0 0.2 0.4 0.6 0.8 1

Y/Ho

(La/La*)

0

20 40 60 80

0

REE

Major Oxides

0.05

0.1

NdSN/YbSN

(ppm)

(wt. %) SiO2 Fe 2O3 Al 2O3

10

10

5

5

5

0

0

0

0

-5

-5

-5

-5

-5

-10

-10

-10

-10

-10

-15

-15

-15

-15

-15

10

10

10

5

5

0

1

1.2

1.4

1.6

0

50

100

150

0

1

2

3

0 10 20 30 40 50

0

0.5

1

1.5

Fig. 6. Geochemical profiles for the Demirtas section in southern Turkey. All depths are measured from the lithologic P–T boundary at the base of the stromatolite. δ13Ccarb and organic carbon values are taken from Riccardi (2007).

There is a strong correlation between Σ REY and Y/Ho in both Cili (r = − 0.81; n = 20) and Demirtas (r = − 0.76; n = 32), which indicates that the inconsistencies between ancient sediments and modern seawater are likely due to a non-seawater source such as detrital material or diagenetic precipitates from pore waters. In order to interpret the geochemical history of the Tethys Ocean, first we need to distinguish which samples preserve primary seawater signatures. The three likely non-carbonate REY sources that require consideration are (1) terrestrially derived siliciclastic material (e.g. clays), which have a much higher concentration of REY than carbonates and have a flat REYSN

distribution (German and Elderfield, 1990; Webb and Kamber, 2000); (2) Fe and Mn oxides, which sorb high concentrations of REY but can be distinguished by their negative Y anomaly and positive Ce anomaly; and (3) phosphates, which also have high concentrations of REY and can be distinguished by their positive relationship between P and the REY (German and Elderfield, 1990). We attempted to eliminate the problem of non-carbonate REY sources by removing the insoluble residue with centrifugation prior to analysis by ICP-MS. Despite this, many samples have a siliciclastic REY pattern and relatively high REY concentrations. This is likely due to the liberation of REY from clays in acidic

Table 3 Correlation matrix of r values for major and trace element data from the Demirtas section.

Pr/Pr* Ce/Ce* I Σ REY Y/Ho La/La* NdSN/YbSN Fe2O3 MnO P2O5 SiO2 Al2O3 TiO2 SrO CaMg(CO3)2 Tot. Carb. Org C.

Pr/Pr*

Ce/Ce*

I

Σ REY

Y/Ho

La/La*

NdSN/YbSN

Fe2O3

MnO

P2O5

SiO2

Al2O3

TiO2

SrO

CaMg(CO3)2

Tot. carb.

Org c.

1

−0.92 1.00

−0.28 0.36 1.00

0.21 0.01 −0.26 1.00

−0.03 −0.11 −0.02 −0.76 1.00

0.37 −0.71 −0.34 −0.33 0.26 1.00

0.16 −0.03 0.08 0.35 −0.37 −0.22 1.00

−0.44 0.54 0.05 0.50 −0.51 −0.42 0.14 1.00

−0.53 0.66 −0.07 0.49 −0.50 −0.57 0.19 0.83 1.00

−0.10 0.27 −0.23 0.41 −0.21 −0.39 −0.12 0.58 0.56 1.00

−0.10 0.31 0.13 0.64 −0.71 −0.50 0.53 0.77 0.69 0.42 1.00

−0.04 0.21 0.03 0.61 −0.68 −0.37 0.41 0.68 0.67 0.35 0.89 1.00

0.02 0.15 0.08 0.59 −0.72 −0.34 0.50 0.71 0.70 0.35 0.89 0.91 1.00

−0.34 0.20 −0.12 −0.18 0.10 0.16 −0.29 −0.05 0.06 0.11 −0.20 −0.10 −0.17 1.00

0.15 −0.11 −0.23 0.06 0.08 0.01 −0.18 0.21 0.19 0.44 0.02 −0.05 0.00 −0.15 1.00

0.20 −0.38 0.01 −0.66 0.72 0.47 −0.42 −0.85 −0.80 −0.48 −0.93 −0.94 −0.89 0.11 −0.04 1.00

−0.63 0.70 −0.10 0.21 −0.23 −0.53 0.20 0.54 0.64 0.52 0.49 0.42 0.33 0.64 0.00 −0.54 1.00

202

G.R. Loope et al. / Chemical Geology 351 (2013) 195–208

Idrijca Valley

Velebit Mt.

15

Julfa

GK-1

430

10

220

5

0

0 -5

225

420

Depth (m)

Depth (m)

Depth (m)

425

5

-5 -10

Depth (m)

10

415

230

410 -10

-15

-15

-20

235 405

-25

-20 0

0.5

Ce/Ce*

1

0

0.5

1

400 -0.5

Ce/Ce*

-0.25

Ce/Ce*

0

0

0.5

1

Ce/Ce*

Fig. 7. Ce-anomaly from four sections in the western and central Tethys. The data are from the following sources, Idrijca Valley, Slovenia (Dolenec et al., 2001); Velebit Mt., Croatia (Fio et al., 2010); Julfa, Iran (Kakuwa and Matsumoto, 2006); GK-1, Austria (Attrep et al., 1991). Placement of the extinction horizon (solid horizontal line) and the P/T boundary (dashed horizontal line) in the Idrijca Valley, Julfa, and GK-1 sequences are taken from Korte and Kozur (2010). Placement of the extinction event and the P–Tr boundary in the Velebit Mountain section are from Fio et al. (2010). The Ce-anomaly from Julfa is not reported using the standard system and instead is reported on a scale where 0.0 is no anomaly.

conditions. Previous studies have attempted to address this issue by using acetic acid dissolutions, but none have found improvements over the nitric acid digestion (Nothdurft et al., 2004; Olivier and Boyet, 2006).

5.1. Diagenesis REY are often considered immobile elements, and it has been shown experimentally that the high affinity for the carbonate lattice causes the retention of Eu3+ (and likely all other REY) even during dissolution– reprecipitation (Stipp et al., 2003). Diagenetic pathways such as neomorphism often alter the concentrations of minor elements including Sr (Brand and Veizer, 1980), but recent work done on partially and fully neomorphosed corals from the Windley Key exposure of the Key Largo Limestone, Florida has suggested that REY behave conservatively during meteoric diagenesis (Webb et al., 2009). Azmy et al. (2011) argue that the Windley Key limestone underwent meteoric diagenesis in the vadose zone and therefore is not representative of most limestones, which are altered in a phreatic zone diagenetic setting. They show that limestones altered in the phreatic zone typically have REY contents enriched by an order of magnitude over unaltered coeval calcite. However, these REY enriched limestones have REYSN profiles that mimic the original REYSN profile with only a slight enrichment in the light rare earth elements (LREE) in comparison to the heavy rare earth elements (HREE). This retention of near-primary REYSN distribution is due to the close partition coefficients of REY and a seawater parent of the diagenetic fluid. The acetate peals from both the Cili and Demirtas sections evidence a large amount of textural recrystallization (Fig. 3), likely the result of the transition from aragonite and high-Mg calcite to the observed low-Mg calcite and dolomite. The samples from both Cili and Demirtas have elevated Σ REY contents (1–60 ppm) compared to most modern carbonates (0.05–3 ppm; Webb and Kamber, 2000) and display a slight enrichment of LREE over HREE compared to modern seawater (Elderfield and Greaves, 1982). This indicates that the Cili and Demirtas carbonates have undergone diagenetic enrichment in REY. Due to the near-micritic fabric and seawater-like REYSN profiles of our samples, we find it likely that this REY enrichment was caused by interaction with a diagenetic fluid of marine origin that mimicked the original REYSN profile as discussed in Azmy et al. (2011) and Azomani et al. (2013). At both Cili and Demirtas there are strong positive correlations

between Σ REY, MnO and Fe2O3 (>0.49, Tables 2 and 3) which is expected because REY, Mn and Fe all tend to accumulate during diagenesis. The Cili samples show a positive relationship between δ13C and δ18O when Permian and Triassic data are considered separately (Fig. 8A). This indicates that diagenetic fluids were likely a combination of meteoric and marine water. Fig. 8C and D show no correlation between Σ REY and Pr/Pr* at either Cili or Demirtas. This indicates that although some samples became more enriched in REY than others during diagenesis, the original Ce anomaly was preserved.

5.2. Cili 5.2.1. Contamination vs seawater The Permian bioclastic limestone samples from the Cili section have very little contamination from silicates (total carbonate >99.7%) and have low Σ REY concentrations compared to the post-extinction samples (Fig. 4). These samples also have high values of Y/Ho (mean = 47.5) and positive La anomalies (mean = 1.52) that point to a seawater REY source (Webb and Kamber, 2000; Bolhar et al., 2004). The low Y/Ho ratios in the microbialite and overlying Triassic samples are associated with a decrease in the percent total carbonate (Fig. 9B). This decrease in Y/Ho with percent carbonate follows the expected shale mixing line (solid line = PAAS; Nance and Taylor, 1976). Nearly all of the data fall within the upper and lower range of possible mixing lines as indicated by the two dashed lines representing different shale end-members. These two shale end-members are chosen from the 23 shales comprising PAAS (Nance and Taylor, 1976) in order to capture the largest possible range. Silicate contamination also explains the positive relationship between Y/Ho and Pr/Pr* in the Cili samples (Fig. 9C). The majority of the Triassic samples, including all of the microbialite, plot close to a siliciclastic mixing line with the purest carbonate sample and PAAS as end-members. This indicates that even though the samples lack an explicit oxic signature, they probably were deposited in oxic seawater and had their REYSN distributions masked by siliciclastics. In this figure only one sample, the uppermost Triassic oolitic limestone, plots in the anoxic seawater quadrant. The rest of the Triassic samples have Y/Ho ratios that have been heavily influenced by siliciclastic material. The uppermost oolitic limestone sample also has a more positive La anomaly (La/La* = 1.23), a lower NdSN/YbSN (0.71), and a higher total carbonate (99.24%) than the rest of the Triassic samples.

G.R. Loope et al. / Chemical Geology 351 (2013) 195–208

A

B

C

D

203

Fig. 8. (A) C and O isotope plot for the Cili section (data from Luo et al., 2010). (B) C and O isotope plot for the Demirtas section (data from Riccardi, 2007). (C) Plot of Pr/Pr* and REY abundance for the Cili section shows no relationship between Ce anomaly and diagenetic enrichment of REY. (D) Plot of Pr/Pr* and Σ REY for the Demirtas section shows no relationship between Ce anomaly and diagenetic enrichment of REY.

5.2.2. Paleo-redox history The Permian bioclastic limestone at the Cili section contains fossils including a diverse foraminifer assemblage, ostracodes, echinoderms and bryozoan fragments suspended in a fine-grained matrix, which implies that they were not transported prior to burial. Therefore, animals must have had sufficient O2 to respire while living on the seafloor. Also, the samples inspected lacked laminations, likely due to intense bioturbation. We find that this paleontological and sedimentological evidence for an oxic water column is supported by negative Ce anomaly data. These samples preserve a marine REY signature (Y/Ho, and La/La*) that supports the utility of Ce anomaly as a redox indicator. The microbialite contains foraminifers, ostracodes, gastropods, bivalves, and brachiopods fossils which indicates a oxygenated environment or transport of shelly material from a nearby oxygenated environment. The microbialite and overlying Triassic samples are heavily influenced by a large siliciclastic REY fraction, but most still record a negative Ce anomaly indicative of oxygenation (Fig. 9C). Of the samples that do not preserve an oxic signature (Pr/Pr* b 1.05), only one sample, in the uppermost oolitic limestone, preserves a seawater signature. The combination of a Ce anomaly at unity and a seawater signature do suggest water column anoxia, but we feel that more than one sample is needed to argue compellingly for anoxia in such a high-energy environment such as an ooid shoal. Whether or not anoxia occurs in the uppermost strata, the Cili section provides evidence for oxic conditions both before and immediately after the mass extinction. 5.3. Demirtas 5.3.1. Contamination vs seawater The Demirtas section shows seawater-like values of Y/Ho in the Triassic samples (15 out of 16 samples > 44), but the Permian limestone does not preserve a robust marine Y/Ho signature (only 6 of 12 samples >44). A clear rise in the La anomaly at the P–Tr boundary (from a mean of 1.15 in the Permian to a mean of 1.48 in the Triassic) also provides evidence for a strong marine signature in the Triassic

but not the Permian. LREE depletion is also absent in the Permian and basal 2 m of the Triassic (NdSN/YbSN = 1.18) but is present in the rest of the Triassic samples (NdSN/YbSN = 0.59). These data suggest that contamination from terrigenous clastics may dominate the REY signature in the Permian limestone, but have a somewhat reduced impact on the Triassic section. The strong correlation of Σ REY with Y/Ho, percent carbonate, Al2O3, SiO2, TiO2 as shown in Table 1 and Fig. 10A suggests that contamination by siliciclastic material has masked the REY pattern preserved in the carbonate fraction of the Permian samples. The strong negative relationship between Y/Ho and total carbonate shown in Fig. 10B further confirms that siliciclastic material is responsible for the deviations from a seawater-like REYSN distribution. This non-seawater-like REYSN distribution appears to be mostly limited to the Permian samples and is likely responsible for the absence of a Ce anomaly in these samples. A large REY contribution from Fe\Mn oxides is unlikely because Σ REY is more closely correlated with TiO2, SiO2, and Al2O3, than to Fe2O3 and MnO. Similarly, there is no correlation between Σ REY and P2O5, indicating that significant contributions from phosphates are unlikely. Biologically precipitated carbonates tend to have lower concentrations of REY than biologically induced carbonates such as microbialites (Webb and Kamber, 2000). We find that a lower concentration of carbonate-bound, seawater-derived REY in the Permian bioclastic limestone than in the microbialite is the best explanation for the reported siliciclastic and REY data. This would allow the similar silicate contributions observed in the Permian bioclastic limestone and the boundary microbialite to mask the REY pattern in the bioclastic limestone to a much greater extent than in the microbialite. Fig. 10C shows Ce anomaly (Pr/Pr*) as a function of contamination, as measured by Y/Ho*. A mixing line between PAAS (McClennan, 1995) and the bioclastic carbonate with the least siliciclastic input (from −26.1 m in the Cili section) shows a possible trajectory for most Permian samples. The six Permian samples that do plot in the anoxic seawater quadrant all lack a significant positive La anomaly and have NdSN/YbSN values close to that of shale. This calls question to a

204

G.R. Loope et al. / Chemical Geology 351 (2013) 195–208

Above microbialite Microbialite Pre-extinction

60

REY (ppm)

50

A

A

40 30 20 10 0 100

98

96

94

92

90

88

% Carbonate

B

80

B

70 60

Y/Ho

50 40 30 20 10 0 100

98

96

94

92

90

88

% Carbonate

C

1.2

C

Pr/Pr*

1.1 Oxic Anoxic 1

seawater 0.9

0

20

40

60

Y/Ho Fig. 9. Cross-plots of Cili data showing the effects of siliciclastic contamination on Ce-anomaly. (A) Permian and Triassic samples show an overall inverse correlation between percent carbonate and Σ REY. A siliciclastic mixing line is calculated using a relatively pure Permian carbonate (−17.6 m) as one end-member and PAAS (Nance and Taylor, 1976) as the other. The dashed mixing lines represent maximum and minimum mixing lines that use end-members from the Nance and Taylor (1976) data set (see section 6.2.1. for details). (B) When Y/Ho is plotted against total carbonate, samples follow the shale mixing lines (calculated in the same fashion as in A. (C) A Ce anomaly-Y anomaly plot is used to show which samples likely represent oxic or anoxic environments. An oxic– anoxic line has been plotted at Pr/Pr* = 1.05 after Bau and Dulski (1996) and the seawater Y/Ho lower threshold of 44 (Webb and Kamber, 2000) is marked by a second line.

REY seawater source as suggested by their high Y/Ho values. In Fig. 10C the Triassic samples show a wide range in Ce anomalies and do not plot along the mixing line. This may be due to variation within the carbonate fraction rather than a silicate source. 5.3.2. Paleo-redox history The Permian bioclastic limestone of the Demirtas section lacks laminations and preserves a diverse benthic community including 23 types of lagenide foraminifers, echinoderms, gastropods, ostracodes, bivalves and brachiopods all indicative of a well-oxygenated water column (Groves et al., 2005). The La anomaly and NdSN/YbSN values from these Permian samples suggest that the REY do not preserve a seawater signature and should not be used to interpret redox conditions. On the

Fig. 10. Cross-plots of data from the Demirtas section illustrating the effects of siliciclastic contamination. (A) Permian and Triassic samples show an overall inverse correlation between percent carbonate and Σ REY but have a steeper slope than the PAAS (solid) and upper and lower shale mixing lines (dashed lines, see text from Cili section for details). These mixing lines use the sample from +13.75 m as the carbonate end-member because it has both a strongly marine Y/Ho ratio and very little siliciclastic material. (B) Total carbonate and Y/Ho data follow shale mixing lines (same end-members as in A). (C) In a plot of Ce anomaly and Y/Ho neither the Triassic nor the Permian samples follow the plotted shale mixing line. None of the Demirtas samples have both oxidizing (>1.05) Pr/Pr* values and little siliciclastic contamination so the sample from −26.1 m in the Cili section was chosen as the carbonate end-member for the mixing line. The mixing line uses a siliciclastic end-member with a Pr/Pr* value of 0.95 because this is the approximate value in the Permian samples that have shale-like REYSN distributions (Y/Ho, La/La*, NdSN/YbSN).

other hand, high Y/Ho in half of the Permian samples suggest that they may have a marine signature. Due to these conflicting interpretations from the Permian REY data, we do not find sufficient reason to reject the strong paleontological and sedimentological evidence for Permian oxic conditions. The microbialite at Demirtas is extremely poor in macroinvertebrate fossils but this does not necessitate low oxygen conditions. The microbialite preserves a marine REYSN signature and a negative Ce anomaly indicating that the water column was well oxygenated. Although there is a substantial siliciclastic REY contribution to these samples, this source cannot be responsible for the negative Ce anomaly because mixing should drive the Ce anomaly towards unity. Approximately 4 m above the extinction horizon, the negative

G.R. Loope et al. / Chemical Geology 351 (2013) 195–208

Ce anomaly is lost while the REYSN distribution still closely resembles a seawater source. We find this strong evidence for the development of an anoxic water column as the microbialite transitioned into micrite. This change in redox conditions could represent a transgression in which anoxic waters flooded the site. This hypothesized redox history differs significantly from that proposed for the Demirtas site by Groves et al. (2005), who relied solely on paleontological and sedimentological observations. Groves et al. (2005) described the fossil-poor stromatolite as a disaster biota and attributed its presence to the exclusion of grazers by anoxic conditions. We find robust geochemical evidence that if anoxia did occur at the site, it did not occur until well after the development of the stromatolite. 5.4. Origin of the P–Tr boundary microbialite The prevailing redox conditions during the deposition of the boundary microbialite must be resolved before anoxia can be assessed as a trigger for the P–Tr boundary microbialite. Liao et al. (2010) reported framboidal pyrite size distributions from the Laolongdong microbialite and concluded that their average size (~8 μm) and standard deviation (1.8 to 3.5 μm) classified them in the lower-dysoxic environment (Bond and Wignall, 2010). Although the size distribution of these framboids does suggest a low-oxygen environment, framboids of similar size can also grow in sediments below fully oxic water bodies (Wilkin et al., 1996). Evidence supporting oxygenated conditions comes from an analysis of ostracod feeding guilds from across southern China (Forel et al., 2009) that found the deposit-feeding guild, a diagnostic indicator of a well oxygenated water column, remained dominant through the extinction event. Framboidal size analysis (Liao et al., 2010) has suggested low oxygen conditions, while ostracod feeding guild analysis (Forel et al., 2009) has suggested well-oxygenated conditions. Kershaw et al. (2011) tried to resolve the issue by proposing that periodic flooding of the platforms with anoxic water could account for both the ostracod and framboidal pyrite data. These transitions between oxic and anoxic conditions have not been observed in the framboid or the ostracod data sets or the new Ce anomaly data presented in this study. With the addition of the Ce anomaly from this study, the case for the deposition of the boundary microbialite in oxic conditions is greatly strengthened. This casts doubt on the hypothesis that exclusion of grazers by anoxia triggered the switch to microbial carbonate production. If anoxia did not coincide with the extinction event in these shallow-water sections, an additional kill-mechanism is required. If grazers were excluded from the microbialite by a shift in environmental conditions, hypercapnia, ocean acidification or warmer seawater may be better mechanisms than anoxia (Knoll et al., 2007; Shen et al., 2011; Brand et al., 2012; Joachimski et al., 2012). Anoxia and euxinia could still play important parts of the marine and terrestrial mass extinctions, but they require another partner that could decimate shallow carbonate platforms.

205

of these sites are re-examined (no chemical data beyond Ce/Ce* is available for Julfa, Iran), it becomes clear that the REY in many samples do not record original seawater conditions and that there is little robust Ce anomaly evidence for anoxia these sites. Both the Gartnerkofel-1 core and the Velebit Mountain section show a strong negative relationship between percent carbonate and Σ REY that closely matches the expected shale mixing line (Fig. 11A and C). This suggests that the primary REY source is from siliciclastics rather than the carbonate fraction. This siliciclastic source has caused the Ce anomaly in heavily contaminated samples to approach unity, which eliminates their utility as a paleo-redox proxy (Fig. 11B and D). In the Idrijca Valley section there is not a strong relationship between total carbonate and Σ REY (r = −0.34, n = 14), but Pr/Pr* is closely related to degree of dolomitization (r = 0.58, n = 14) and Σ REY (r = 0.80, n = 14; Fig. 11E). This suggests that REY accumulated during dolomitization of the samples and did not preserve the original REYSN profile. Regardless of the reason, none of the samples from the Idrijca Valley have Y/Ho distributions that match modern seawater (Fig. 11F). Kakuwa and Matsumoto (2006) propose that the negative Ce anomaly found at the extinction horizon at Julfa (Fig. 7) is evidence for the passage of the chemocline during a slow chemocline upward excursion. Their model hinges on the interpretation of the negative Ce anomaly at the extinction horizon as evidence for suboxic conditions. Both suboxic and oxic conditions are expected to produce negative Ce anomalies. This makes distinguishing the two environments difficult in ancient sediments. We see no reason to favor the interpretation of a negative Ce anomaly as evidence for suboxic conditions over the more standard oxic interpretation. Furthermore, in the Julfa section, all samples lack a Ce anomaly except those within ~2 m of the extinction horizon. We see no reason for the interpretation of the samples from above the extinction horizon as evidence for anoxia when the samples below the extinction horizon, with identical Ce anomalies, are interpreted as evidence for oxic conditions. The geochemical data from northern Italy (Brand et al., 2012) does not show a strong correlation between total measured REE concentrations and total carbonate content (r = 0.16, n = 11). This indicates that siliciclastic material is not an important source of REE. However, the REE data are from whole rock samples that are shown to have diagenetically altered Sr, Mg, Fe and Mn concentrations as well as δ18O, 87Sr/86Sr and δ13C. Given the relatively high REE contents of the bulk samples, diagenesis is likely to have significantly enriched the samples in REE (Azmy et al., 2011), but without a full compliment of REE it is difficult to assess alteration of the Ce anomaly or the REYSN profile. Nonetheless, our data from Cili and Demirtas support the conclusion of Brand et al. (2012) that anoxia was not pervasive across Tethys at the time of the extinction and could not have been the primary kill mechanism. 5.6. Iodine as a redox proxy

5.5. Application of Ce anomaly as a redox indicator in P–Tr carbonates In addition to the two P–T sections presented in this study, Ce anomaly data sets have been published for the Gartnerkofel-1 Core in Austria; Idrijca Valley, Slovenia; Julfa, Iran; Velebit Mountain, Croatia (Fig. 7); as well as Sass De Putia and Val Bruta in northern Italy (Brand et al., 2012). Except for the Italian sections, each of these published geochemical records includes samples from fossiliferous strata that lack the negative Ce anomaly typical of oxidized seawater. The interpretation of these four data sets has varied widely due to the apparent inconsistencies between a traditional view of the Ce anomaly and paleontological evidence. Although Kakuwa and Matsumoto (2006) and Fio et al. (2010) bring up the possibility of a siliciclastic influence on Ce anomaly, none of these studies have rigorously tested for a possible relationship between siliciclastic material and REY or documented a true REY seawater signature in their data. When the REY and major oxide data from three

If interpreted as a primary record using the Lu et al. (2010) model, the decrease in iodine within the microbialite at Cili could be interpreted as evidence for anoxia. We find it unlikely that iodine fluctuations in the Cili and Demirtas samples represent changes in the concentration of IO− 3 in seawater because the samples have been nearly entirely replaced or recrystallized. In the Cili samples, iodine has a strong negative relationship with the degree of dolomitization (r = −0.61, n = 20), which suggests that iodine was lost during diagenesis. The iodine trend in the Demirtas section fluctuates considerably but remains extremely low and does not correlate with the Ce anomaly data. Due to the highly recrystallized nature of these samples, we infer that their iodine distribution is more likely a result of loss in diagenesis than changes in seawater redox state. The new redox proxy presented in Lu et al. (2010) is enticing because the high redox sensitivity of I makes it ideal for discerning small

206

G.R. Loope et al. / Chemical Geology 351 (2013) 195–208

A

B

1.2

140

1

120 100

Ce/Ce*

REE (La+Ce+Sm+Lu+Eu+Tb +Dy+Yb)

Gartnerkofel

1.4 160

80

0.8 0.6

60 0.4

40

0.2

20 0 100

80

60

40

20

0 100

0

80

60

% Carbonate 100

20

0

1.7

C

90

Y/Ho<44 Y/Ho>44

80

1.5

70

1.4

60

1.3

50 40

Permian

Seawater

Triassic

1.2 1.1

30

1

20

0.9

10

0.8

0 100

D

1.6

Pr/Pr*

Velebit Mountain REY (ppm)

40

% Carbonate

Oxic Anoxic

0.7 95

90

85

0

20

% Carbonate

40

60

80

Y/Ho

1.3

1.7

E

1.2

F

1.6

1.1

1.4

1

1.3

Pr/Pr*

Idrijca Valley Pr/Pr*

1.5

0.9

1.2

Oxic

1.1

Anoxic

1

0.8

0.9

0.7

0.8

0.6

Seawater

0.7 0

50

100

150

200

250

0

REE

20

40

60

80

Y/Ho

Fig. 11. Previously published REY data from P–Tr carbonate sites. (A) Plot of total carbonate and the sum of the published REEs for the GK-1 core, Austria (Attrep et al., 1991). Solid line in all diagrams is a siliciclastic mixing line with the purest carbonate sample as the carbonate end-member and PAAS (Nance and Taylor, 1976) as the siliciclastic end member. Dashed lines in all diagrams represent the range of possible mixing lines by changing the siliciclastic end-member (see Figs. 7 and 9 for details). (B) Plot shows relationship between total carbonate and Ce/Ce* for the GK-1 core (Attrep et al., 1991). (C) Plot of total carbonate and net REY concentration from Velebit Mt., Croatia (Fio et al., 2010). (D) Redox diagram of Velebit Mountain data (Fio et al., 2010) using Pr/Pr* as a proxy for redox condition and Y/Ho as an indicator of REY source. An oxic–anoxic line has been plotted at Pr/Pr* = 1.05 after Bau and Dulski (1996) and the seawater Y/Ho threshold of 44 (Webb and Kamber, 2000) is marked by a second line. (E) Plot of Pr/Pr* and Σ REY for the Idrijca Valley section, Slovenia (Dolenec et al., 2001). (F) Redox diagram of Idrijca Valley data (Dolenec et al., 2001) with Pr/Pr* used to show Ce anomaly and Y/Ho to indicate REY source.

redox variations in shallow waters. The P–Tr results presented in this study do not provide a rigorous test of the utility of iodine as a redox indicator in carbonates because the diagenetic history of the samples is non-ideal for retention of a primary iodine record. Future work focusing on samples with unaltered carbonate mineralogy is required to test the model proposed by Lu et al. (2010). Given the susceptibility of carbonates to diagenesis and the mixed results of the proxy in ancient carbonates, an iodine content comparison of a carbonate sequence consisting of fresh, shallowly buried, and lithified carbonates is required to show the validity of this redox proxy. Although iodine is more sensitive to changes in redox conditions than Ce anomaly, the very strong affinity of REEs for the carbonate matrix makes Ce anomaly a more reliable indicator of redox conditions than iodine. The location of

cerium within the REEs also provides a large advantage for its use as a redox proxy because its neighboring REEs provide ideal redoxinsensitive controls for Ce. 6. Conclusions The REYSN patterns from both sections are consistent with deposition under a well-oxygenated water column lasting from the Late Permian up through the boundary microbialite. In the Cili section, oxic conditions prevail through the remaining Triassic section except the uppermost sample, which could possibly represent anoxia. In the Demirtas section, the transition from microbialite to micrite is accompanied by a shift from a negative Ce anomaly to the absence of a Ce

G.R. Loope et al. / Chemical Geology 351 (2013) 195–208

anomaly. This indicates a transition to an anoxic water column likely caused by a marine transgression. Reinterpretations of the Ce anomaly patterns from the Idrijca Valley, Velebit Mountain, and Julfa sites as well as the Gartnerkofel-1 core reveal that the paradox of conflicting Ce anomaly and paleontological data is due to contamination of samples with siliciclastics and, in the case of the Idrijca Valley, accumulation of REY during dolomitization. Although silicates mask the primary seawater REYSN distribution in many Upper Permian and Lower Triassic samples, the combined sections provide strong evidence that seawater at the P–Tr extinction was oxic in shallow carbonate environments. The boundary microbialite Ce anomaly data are interpreted as representing deposition in oxic waters. Together with evidence from ostracod feeding guilds, this data casts doubt onto the hypothesis that the boundary microbialite formed when anoxia caused the exclusion of grazers from carbonate platforms. This suggests that anoxia and euxinia, if factors, must have had a partner such as hypercapnia, climate change, or ocean acidification that caused the severe extinction in shallow-water carbonate platforms. Acknowledgments This project was supported by NSF (EAR 0807744) and the Penn State Department of Geosciences. We would also like to thank Genming Luo (China University of Geosciences) for providing the samples from the Cili section and two anonymous reviewers whose suggestions were invaluable. Appendix A. Supplementary data Supplementary data to this article can be found online at http:// dx.doi.org/10.1016/j.chemgeo.2013.05.014. References Algeo, T.J., et al., 2010. Changes in productivity and redox conditions in the Panthalassic Ocean during the latest Permian. Geology 38, 187–190. Alibo, D., Nozaki, Y., 1999. Rare earth elements in seawater: particle association, shalenormalization, and Ce oxidation. Geochimica et Cosmochimica Acta 63, 363–372. Attrep, M., Orth, C., Quintana, L., 1991. The Permian–Triassic of the Gartnerkofel-1 Core (Carnic Alps, Austria). Geochemistry of Common and Trace Elements II - INAA and RNAA. 1–15. Azmy, K., Brand, U., Sylvester, P., Gleeson, S.A., Logan, A., Bitner, M.A., 2011. Biogenic and abiogenic low-Mg calcite (bLMC and aLMC): Evaluation of seawater-REE composition, water masses and carbonate diagenesis. Chemical Geology 280, 180–190. Azomani, E., Azmy, K., Blamey, N., Brand, U., Al-Aasm, I., 2013. Origin of Lower Ordovician dolomites in eastern Laurentia: Controls on porosity and implications from geochemistry. Marine and Petroleum Geology 40, 99–114. Bau, M., Dulski, P., 1996. Distribution of yttrium and rare-earth elements in the Penge and Kuruman iron-formations, Transvaal Supergroup, South Africa. Precambrian Research 79, 37–55. Baud, A., Richoz, S., Marcoux, J., 2005. Calcimicrobial cap rocks from the basal Triassic units: western Taurus occurrences (SW Turkey). Comptes Rendus Palevol 4, 569–582. Bolhar, R., et al., 2004. Characterisation of early Archaean chemical sediments by trace element signatures. Earth and Planetary Science Letters 222, 43–60. Bond, D.P.G., Wignall, P.B., 2010. Pyrite framboid study of marine Permian–Triassic boundary sections: a complex anoxic event and its relationship to contemporaneous mass extinction. Geological Society of America Bulletin 122, 1265–1279. Brand, U., Veizer, J., 1980. Chemical diagenesis of a multicomponent carbonate system: 1, Trace elements. Journal of Sedimentary Petrology 50, 1219–1236. Brand, U., et al., 2012. The end-Permian mass extinction: a rapid volcanic CO2 and CH4climatic catastrophe. Chemical Geology 322–323, 121–144. Brennecka, G.A., Herrmann, A.D., Algeo, T.J., Anbar, A.D., 2011. Rapid expansion of oceanic anoxia immediately before the end-Permian mass extinction. PNAS 108, 17631–17634. Cao, C., et al., 2009. Biogeochemical evidence for euxinic oceans and ecological disturbance presaging the end-Permian mass extinction event. Earth and Planetary Science Letters 281, 188–201. De Baar, H., Schijf, J., Byrne, R., 1991. Solution chemistry of the rare-earth elements in seawater. European Journal of Solid State and Inorganic Chemistry 28, 357–373. Dolenec, T., Lojen, S., Ramovs, A., 2001. The Permian–Triassic boundary in Western Slovenia (Idrijca Valley section): magnetostratigraphy, stable isotopes, and elemental variations. Chemical Geology 175, 175–190. Elderfield, H., Greaves, M.J., 1982. The rare earth elements in seawater. Nature 296, 214–219.

207

Erwin, D.H., 1994. The Permo-Triassic extinction. Nature 367, 231–236. Fio, K., et al., 2010. Stable isotope and trace element stratigraphy across the Permian– Triassic transition: a redefinition of the boundary in the Velebit Mountain, Croatia. Chemical Geology 278, 38–57. Forel, M., Crasquin, S., Kershaw, S., 2009. Ostracods (Crustacea) and water oxygenation in the earliest Triassic of South China: implications for oceanic events at the endPermian mass extinction. Australian Journal of Earth Sciences 815–823. German, C.R., Elderfield, H., 1990. Application of the Ce anomaly as a paleoredox indicator: the ground rules. Paleoceanography 5, 823. Groves, J., Altiner, D., Rettori, R., 2005. Extinction, survival, and recovery of lagenide foraminifers in the Permian–Triassic boundary interval, central Taurides, Turkey. Journal of Paleontology 79, 1–39. Hallam, A., 1989. The case for sea-level change as a dominant causal factor in mass extinction of marine-invertebrates. Philosophical Transactions of the Royal Society of London, Series B 325, 437–455. Hays, L.E., Grice, K., Foster, C.B., Summons, R.E., 2012. Biomarker and isotopic trends in a Permian–Triassic sedimentary section at Kap Stosch, Greenland. Organic Geochemistry 43, 67–82. Ingamells, C., 1970. Lithium metaborate flux in silicate analysis. Analytica Chimica Acta 52, 323–334. Isozaki, Y., 1997. Permo-Triassic boundary superanoxia and stratified superocean: records from lost deep sea. Science 276, 235–238. Joachimski, M.M., et al., 2012. Climate warming in the latest Permian and the Permian– Triassic mass extinction. Geology 40, 195–198. Kakuwa, Y., Matsumoto, R., 2006. Cerium negative anomaly just before the Permian and Triassic boundary event: the upward expansion of anoxia in the water column. Palaeogeography, Palaeoclimatology, Palaeoecology 229, 335–344. Kamber, B., Bolhar, R., 2004. Geochemistry of late Archaean stromatolites from Zimbabwe: evidence for microbial life in restricted epicontinental seas. Precambrian Research 132, 379–399. Kamber, B.S., Webb, G.E., 2001. The geochemistry of late Archaean microbial carbonate: implications for ocean chemistry and continental erosion history. Geochimica et Cosmochimica Acta 65, 2509–2525. Kershaw, S., et al., 2011. Microbialites and global environmental change across the Permian–Triassic boundary: a synthesis. Geobiology 10, 25–47. Knoll, A.H., et al., 2007. Paleophysiology and end-Permian mass extinction. Earth and Planetary Science Letters 256, 295–313. Korte, C., Kozur, H.W., 2010. Carbon-isotope stratigraphy across the Permian–Triassic boundary: a review. Journal of Asian Earth Sciences 39 (4), 215–235. Krull, E., Retallack, G.J., 2000. δ13C depth profiles from paleosols across the Permian– Triassic boundary: evidence for methane release. Geological Society of America Bulletin 112, 1459–1472. Kump, L., Pavlov, A., Arthur, M., 2005. Massive release of hydrogen sulfide to the surface ocean and atmosphere during intervals of oceanic anoxia. Geology 33, 397–400. Lehrmann, D., et al., 2003. Permian–Triassic boundary sections from shallow-marine carbonate platforms of the Nanpanjiang Basin, south China: implications for oceanic conditions associated with the end-Permian extinction and its aftermath. Palaios 18, 138–152. Liao, W., et al., 2010. Shallow-marine dysoxia across the Permian–Triassic boundary: evidence from pyrite framboids in the microbialite in South China. Sedimentary Geology 232, 77–83. Lu, Z., Jenkyns, H.C., Rickaby, R.E.M., 2010. Iodine to calcium ratios in marine carbonate as a paleo-redox proxy during oceanic anoxic events. Geology 38, 1107–1110. Luo, G., et al., 2010. Isotopic evidence for an anomalously low oceanic sulfate concentration following end-Permian mass extinction. Earth and Planetary Science Letters 300, 101–111. Luo, G., et al., 2011. Stepwise and large-magnitude negative shift in delta C-13(carb) preceded the main marine mass extinction of the Permian–Triassic crisis interval. Palaeogeography, Palaeoclimatology, Palaeoecology 299, 70–82. Marenco, P.J., Griffin, J.M., Fraiser, M.L., Clapham, M.E., 2012. Paleoecology and geochemistry of Early Triassic (Spathian) microbial mounds and implications for anoxia following the end-Permian mass extinction. Geology 40, 715–718. Mata, S.A., Bottjer, D.J., 2011. Origin of Lower Triassic microbialites in mixed carbonatesiliciclastic successions: ichnology, applied stratigraphy, and the end-Permian mass extinction. Palaeogeography, Palaeoclimatology, Palaeoecology 300, 158–178. McClennan, S.M., 1995. Rare Earth Elements in Sedimentary Rocks: Influence of Provenance and Sedimentary Processes. In: Lipin, B.r, McKay, G.A. (Eds.), Geochemistry and Mineralogy of Rare Earth Elements. Mineralogical Society of America, Washington, DC. Meyer, K.M., Kump, L.R., Ridgwell, A., 2008. Biogeochemical controls on photic-zone euxinia during the end-Permian mass extinction. Geology 36, 747–750. Muramatsu, Y., Wedepohl, K., 1998. The distribution of iodine in the earth's crust. Chemical Geology 147, 201–216. Muttoni, G., et al., 2009. Opening of the Neo-Tethys Ocean and the Pangea B to Pangea A transformation during the Permian. GeoArabia 14, 17–48. Nance, W.B., Taylor, S.R., 1976. Rare earth element patterns and crustal evolution–I. Australian post-Archean sedimentary rocks. Geochimica et Cosmochimica Acta 40, 1539–1551. Nothdurft, L., Webb, G., Kamber, B., 2004. Rare earth element geochemistry of Late Devonian reefal carbonates, Canning Basin, Western Australia: confirmation of a seawater REE proxy in ancient limestones. Geochimica et Cosmochimica Acta 68, 263–283. Olivier, N., Boyet, M., 2006. Rare earth and trace elements of microbialites in Upper Jurassic coral- and sponge-microbialite reefs. Chemical Geology 230, 105–123. Pruss, S.B., Bottjer, D.J., 2004. Late Early Triassic microbial reefs of the western United States: a description and model for their deposition in the aftermath of the end-Permian mass extinction. Palaeogeography, Palaeoclimatology, Palaeoecology 211, 127–137.

208

G.R. Loope et al. / Chemical Geology 351 (2013) 195–208

Raup, D., 1979. Size of the Permo-Triassic bottleneck and its evolutionary implications. Science 206, 217–218. Renne, P., Basu, A., 1991. Rapid eruption of the Siberian traps flood basalts at the Permo-Triassic boundary. Science 253, 176–179. Riccardi, A., 2007. Carbonate-associated sulfate: assessment of and use as an isotopic proxy for global sulfur cycling during the end-Permian mass extinction. Pennsylvania State University. Department of Geosciences (PhD Thesis). Shen, S., et al., 2011. Calibrating the end Permian mass extinction. Science 334, 1367–1372. Stipp, S., Lakshtanov, L., Jensen, J., Baker, J., 2003. Eu3+ uptake by calcite: preliminary results from coprecipitation experiments and observations with surface-sensitive techniques. Journal of Contaminant Hydrology 61, 33–43. Van Kranendonk, M.J., Webb, G.E., Kamber, B.S., 2003. Geological and trace element evidence for a marine sedimentary environment of deposition and biogenicity of 3.45 Ga stromatolitic carbonates in the Pilbara Craton, and support for a reducing Archaean ocean. Geobiology 1, 91–108. Wang, Y., 2005. Calcimicrobialite after end-Permian mass extinction in South China and its palaeoenvironmental significance. Chinese Science Bulletin 50, 665–671. Wang, Q., Tong, J., Song, H., Yang, H., 2009. Ecological evolution across the Permian/Triassic boundary at the Kangjiaping Section in Cili County, Hunan Province, China. Science in China Series D: Earth Sciences 52, 797–806.

Webb, G., Kamber, B., 2000. Rare earth elements in Holocene reefal microbialites: a new shallow seawater proxy. Geochimica et Cosmochimica Acta 64, 1557–1565. Webb, G., Nothdurft, L., Kamber, B.S., Kloprogge, J., Zhao, J., 2009. Rare earth element geochemistry of scleractinian coral skeleton during meteoric diagenesis: a sequence through neomorphism of aragonite to calcite. Sedimentology 56, 1433–1463. Wignall, P., Hallam, A., 1992. Anoxia as a cause of the Permian Triassic mass extinction — facies evidence from Northern Italy and the Western United-States. Palaeogeography, Palaeoclimatology, Palaeoecology 93, 21–46. Wignall, P.B., Twitchett, R.J., 2002. Extent, duration, and nature of the Permian–Triassic superanoxic event. Geological Society of America Special Paper 356, 395–414. Wignall, P.B., Newton, R., Brookfield, M.E., 2005. Pyrite framboid evidence for oxygenpoor deposition during the Permian–Triassic crisis in Kashmir. Palaeogeography, Palaeoclimatology, Palaeoecology 216, 183–188. Wilkin, R., Barnes, H., Brantley, S., 1996. The size distribution of framboidal pyrite in modern sediments: an indicator of redox conditions. Geochimica et Cosmochimica Acta 60, 3897–3912. Zhang, J., 1996. Rare earth elements and yttrium in seawater: ICP-MS determinations in the East Caroline, Coral Sea, and South Fiji basins of the western South Pacific Ocean. Geochimica et Cosmochimica Acta 60, 4631–4644.