Universidade de São Paulo

2014-08-23

Enhanced primary productivity and

magnetotactic bacterial production

in response to middle Eocene warming in the

Neo-Tethys Ocean Palaeogeography, Palaeoclimatology, Palaeoecology, v. 414, p. 31-45, 2014http://www.producao.usp.br/handle/BDPI/48818

Downloaded from: Biblioteca Digital da Produção Intelectual - BDPI, Universidade de São Paulo

Biblioteca Digital da Produção Intelectual - BDPI

Departamento de Oceanografia Física, Química e Geológica -

IO/IOF

Artigos e Materiais de Revistas Científicas - IO/IOF

http://www.producao.usp.brhttp://www.producao.usp.br/handle/BDPI/48818

Enhanced primary productivity and magnetotactic bacterial productionin response to middle Eocene warming in the Neo-Tethys Ocean

Jairo F. Savian a,⁎, Luigi Jovane b, Fabrizio Frontalini c, Ricardo I.F. Trindade d, Rodolfo Coccioni c,Steven M. Bohaty e, Paul A. Wilson e, Fabio Florindo f, Andrew P. Roberts g,Rita Catanzariti h, Francesco Iacoviello b

a Departamento de Geologia, Instituto de Geociências, Universidade Federal do Rio Grande do Sul, Av. Bento Gonçalves, 9500, 91501-970 Porto Alegre, Brazilb Departamento de Oceanografia Física, Instituto Oceanográfico, Universidade de São Paulo, Praça do Oceanográfico, 191, 05508-120 São Paulo, Brazilc Dipartimento di Scienze della Terra, della Vita e dell'Ambiente, Università degli Studi di Urbino “Carlo Bo”, Campus Scientifico, Località Crocicchia, 61029 Urbino, Italyd Departamento de Geofísica, Instituto de Astronomia, Geofísica e Ciências Atmosféricas, Universidade de São Paulo, Rua do Matão, 1226, 05508-090 São Paulo, Brazile Ocean and Earth Science, University of Southampton, National Oceanography Centre, Southampton SO14 3ZH, UKf Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, 00143 Rome, Italyg Research School of Earth Sciences, The Australian National University, Canberra ACT 0200, Australiah Istituto di Geoscienze e Georisorse CNR, 56124 Pisa, Italy

a b s t r a c ta r t i c l e i n f o

Article history:Received 23 August 2013Received in revised form 5 August 2014Accepted 9 August 2014Available online 23 August 2014

Keywords:PaleoproductivityMECOPutative magnetofossilsMonte CagneroItaly

Earth's climate experienced awarming event known as theMiddle Eocene Climatic Optimum (MECO) at ~40Ma,which was an abrupt reversal of a long-term Eocene cooling trend. This event is characterized in the deepSouthern, Atlantic, Pacific and IndianOceans by a distinct negative δ18O excursion over 500 kyr.We report resultsof high-resolution paleontological, geochemical, and rock magnetic investigations of the Neo-Tethyan MonteCagnero (MCA) section (northeastern Apennines, Italy), which can be correlated on the basis of magneto- andbiostratigraphic results to the MECO event recorded in deep-sea sections. In the MCA section, an interval witha relative increase in eutrophic nannofossil taxa (and decreased abundances of oligotrophic taxa) spans theculmination of the MECO warming and its aftermath and coincides with a positive carbon isotope excursion,and a peak in magnetite and hematite/goethite concentration. The magnetite peak reflects the appearance ofputative magnetofossils, while the hematite/goethite apex is attributed to an enhanced detrital mineralcontribution, likely as aeolian dust transported from the continent adjacent to the Neo-Tethys Ocean during adrier, more seasonal climate during the peak MECO warming. Based on our new geochemical, paleontologicaland magnetic records, the MECO warming peak and its immediate aftermath are interpreted as a period ofhigh primary productivity. Sea-surface iron fertilization is inferred to have stimulated high phytoplanktonproductivity, increasing organic carbon export to the seafloor and promoting enhanced biomineralization ofmagnetotactic bacteria, which are preserved as putative magnetofossils during the warmest periods of theMECO event in the MCA section. Together with previous studies, our work reinforces the connection betweenhyperthermal climatic events and the occurrence (or increased abundance) of putative magnetofossils in thesedimentary record.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

The early part of the Cenozoic Era was characterized by greenhouseconditions through the early Eocene, followed by a ~17 Myr-longcooling trend (e.g. Zachos et al., 2008). This long-term cooling trendwas interrupted by the Middle Eocene Climatic Optimum (MECO) — awarming event that peaked at ~40 Ma (base of Chron C18n.2n) (e.g.Bohaty and Zachos, 2003; Jovane et al., 2007; Bohaty et al., 2009). TheMECO event has been recognized from multiple sites in the Southern,

Atlantic, Pacific, Indian, and Tethyan Oceans (Fig. 1a). It was first identi-fied in foraminiferal stable isotope records from the Atlantic and Indiansectors of the Southern Ocean (Barrera and Huber, 1993; Bohaty andZachos, 2003) and the hallmark of the event is a distinct negative δ18Oexcursion that spanned 500 kyr (Bohaty et al., 2009). The end of theevent wasmarked by a prominent negative shift in benthic foraminifer-al δ13C of up to ~1.0‰ (e.g. Bohaty et al., 2009; Edgar et al., 2010). Thelong-lasting δ18O excursion, with a b100 kyr warming peak (MECOwarming), has been interpreted to indicate a ~4–6 °C temperature in-crease of both surface and intermediate deep waters (Bohaty et al.,2009; Edgar et al., 2010). Organic molecular paleothermometry in thesouthwest Pacific revealed absolute sea surface temperatures of 24 °C

Palaeogeography, Palaeoclimatology, Palaeoecology 414 (2014) 32–45

⁎ Corresponding author. Tel.: +55 51 33086364; fax: +55 51 33086337.E-mail addresses: jairo.savian@ufrgs.br, jairosavian@gmail.com (J.F. Savian).

http://dx.doi.org/10.1016/j.palaeo.2014.08.0090031-0182/© 2014 Elsevier B.V. All rights reserved.

Contents lists available at ScienceDirect

Palaeogeography, Palaeoclimatology, Palaeoecology

j ourna l homepage: www.e lsev ie r .com/ locate /pa laeo

http://crossmark.crossref.org/dialog/?doi=10.1016/j.palaeo.2014.08.009&domain=pdfhttp://dx.doi.org/10.1016/j.palaeo.2014.08.009mailto:jairo.savian@ufrgs.brmailto:jairosavian@gmail.comhttp://dx.doi.org/10.1016/j.palaeo.2014.08.009http://www.sciencedirect.com/science/journal/00310182

to 26 °C just below the onset of MECO, and MECO peak temperaturesexceeding 28 °C (Bijl et al., 2010). The temperature increasecorresponded to a concomitant pCO2 increase by a factor of 2 to 3 (Bijlet al., 2010). An atmospheric pCO2 rise has also been inferred at othersites by changes in deep ocean chemistry as revealed by the net declinein carbonate accumulation that reflects widespread calcite compensa-tion depth (CCD) shoaling (Bohaty et al., 2009). The abrupt pCO2 in-crease during the MECO event has been tentatively ascribed tomassive decarbonation during subduction of TethyanOceanpelagic car-bonates under Asia as India drifted northward (Bohaty and Zachos,2003; Bijl et al., 2010).

The Contessa Highway (CHW) section in Central Italy represents thefirst outcrop section of marine sediments in which MECO was

documented (Fig. 1a, b; Jovane et al., 2007). More recently, the MECOevent has also been recognized at the Alano di Piave section, northeastItaly (Fig. 1a; Luciani et al., 2010). In the Alano section the MECOevent is followed by deposition of two organic-rich intervals (ORG1and ORG2; Spofforth et al., 2010), which are thought to representrapid organic carbon burial contemporaneous with the global pCO2drawdown during the post-MECO recovery interval. Significantpaleoredox, foraminifera and calcareous nannofossil assemblagechanges have been documented in the same section that point to ashift towardmore eutrophic waters and a lowering of oxygen availabilityat the end of MECO during deposition of organic-rich beds (Luciani et al.,2010; Spofforth et al., 2010; Toffanin et al., 2011). Deposition of organic-rich sediments after the transient warming event has been associated

a

b c

Fig. 1. (a) Paleogeographic reconstruction at 40Ma and global distribution of sites fromwhich theMECO event has been studied. The sites include: Kerguelen Plateau (ODP Leg 119, Sites738, 744; ODP Leg 120, Site 748),Weddell Sea (ODP Leg 113, Sites 689, 690), Angola Basin (DSDP Leg73, Site 523), Islas Orcadas Rise (ODP Leg114, Site 702), BlakeNose (ODP Leg 171, Site1051), ShatskyRise (ODPLeg198, Site 1209), Equatorial Pacific (ODPLeg199, Sites 1218, 1219; IODPLegs 320–321, SitesU1331–U1333),MascarenePlateau (ODPLeg115, Sites 709, 711),Walvis Ridge (ODP Leg 208, Site 1263), Demerara Rise (ODP Leg 207, Sites 1258, 1260), East Tasmanian Plateau (ODP Leg 189, Site 1172), Labrador Sea (ODP Leg 105, Site 647) and theonshore Monte Cagnero (MCA), Contessa Highway (CHW), Alano, Kršteňany, and Seymour Island sections. The reconstruction was made with web-based software available at http://www.odsn.de/odsn/services/paleomap/paleomap.html. (b) Location of the MCA and CHW sections. (c) Simplified geological map of the MCA section (Lat. 43°38′50″N, Long. 12°28′05″E, 727 m above sea level) with formation names and boundaries indicated (Coccioni et al., 2013).

33J.F. Savian et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 414 (2014) 32–45

http://www.odsn.de/odsn/services/paleomap/paleomap.htmlhttp://www.odsn.de/odsn/services/paleomap/paleomap.html

with enhanced delivery of terrestrial material that would have bothincreased nutrient availability to the sea surface and stimulated primaryproductivity (Spofforth et al., 2010).

Inmarine environments, changes in primary productivity are directlyrelated to the distribution and availability of nutrients and can be trackedby several proxies. One proxy is the relative abundance of magnetofossilsin pelagic sediments (e.g. Hesse, 1994; Tarduno, 1994; Tarduno andWilkison, 1996; Lean and McCave, 1998; Yamazaki and Kawahata,1998; Kopp and Kirschvink, 2008; Yamazaki, 2008, 2009; Roberts et al.,2011; Chang et al., 2012; Larrasoaña et al., 2012; Yamazaki, 2012).Magnetofossils are the inorganic remains of magnetotactic bacteria,which intracellularly biomineralize magnetically non-interacting singledomain crystals, composed of magnetite, griegite, or both, which arearranged in chains within the cell (Bazylinski and Frankel, 2003; Faivreand Schüler, 2008; Kopp and Kirschvink, 2008; Moskowitz et al., 2008).These chains of nanocrystals (40–300 nm) are used by themagnetotacticbacteria to orient themselves relative to Earth's magnetic field(Blakemore, 1975) to find optimal living conditions in strongly chemi-cally stratified aquatic environments (Bazylinski and Frankel, 2004).Magnetosomes can be preserved in sediments and in some cases mayaccount for 20–60% of their bulk magnetization (Egli, 2004; Housenand Moskowitz, 2006; Roberts et al., 2011). Although magnetotacticbacteria are found in various marine environments (e.g. Petersenet al., 1986; Vali et al., 1987; McNeill, 1990; Petermann and Bleil,1993; Hesse, 1994; Housen and Moskowitz, 2006; Jovane et al., 2012),detection of fossil magnetosomes can be complicated by their mixturewith other magnetic minerals, or withmagnetic particles with differentmagnetic domain structures. Nevertheless, magnetofossils have beenreported from ancient pelagic marine environments of different ages,such as the Cretaceous–Paleogene boundary (e.g. Abrajevitch andKodama, 2009), the Paleocene–Eocene boundary (e.g. Chang et al.,2012; Larrasoaña et al., 2012), the Eocene–Oligocene (e.g. Robertset al., 2011, 2012; Yamazaki et al., 2013), the Oligocene–Miocene(Channell et al., 2013; Florindo et al., 2013; Ohneiser et al., 2013), andthe Pliocene (Yamazaki, 2009; Yamazaki and Ikehara, 2012).Magnetosome abundance in sediments is strongly controlled by theavailability of particulate iron and organic carbon flux to the seafloor(Roberts et al., 2011), which can be related to climate, includinghyperthermal conditions (e.g. Schumann et al., 2008; Chang et al.,2012; Larrasoaña et al., 2012).

Here we present a high-resolution environmental magnetic,micropaleontological and stable isotopic investigation of a westernTethyan pelagic marine section at Monte Cagnero (MCA), Italy, whichwas deposited toward the end of the middle Eocene. Environmentalmagnetic data are used to detect magnetofossil variations across theMECO event and just after the warming event. These results, combinedwith paleoecological data from nannofossils and foraminifera, are usedto assess paleoenvironmental scenarios associated with the MECOevent in the Neo-Tethys realm.

2. Geological setting andmagneto- andbio-stratigraphic framework

A continuous Paleogene sedimentary record is preserved in theScaglia limestone, which includes pelagic limestones and marlylimestones of the Umbria–Marche succession, central Italy. The MCAsection is exposed on the southeastern slope of Monte Cagnero (Lat.43°38′50″N, Long. 12°28′05″E, 727 m above sea level) near the townof Urbania, northeastern Apennines, Italy (Fig. 1a, b). Because of itscompleteness and stratigraphic continuity, the MCA section is animportant pelagic sedimentary succession for studying Eocene andOligocene climatic events (Coccioni et al., 2008; Hyland et al., 2009;Jovane et al., 2013; Fig. 1c). Following Coccioni et al. (2008), the metersystem in the MCA section was established with meter level 100 asthe stratigraphic equivalent to meter level 0 of the Global BoundaryStratotype Section and Point (GSSP) for the Eocene/Oligocene boundaryatMassignano (Premoli Silva and Jenkins, 1993).We focus on the lower

part of the section, from 58 to 72meters stratigraphic level (msl) (Fig. 2).This 14-m-thick interval of the MCA section belongs to the ScagliaVariegata Formation and consists of bundles of limestone-marl couplets.From foraminifera assemblages, a lower bathyal (1000–2000 m) waterdepth was inferred for the depositional environment of the MCA sectionduring the middle Eocene (Guerrera et al., 1988; Parisi et al., 1988),which gradually shoaled to mid-bathyal (800–1000 m) and upperbathyal (400–600 m) water depths in the late Eocene and earlyOligocene, respectively.

High-resolution magneto- and bio-stratigraphic calibration of theMCA section was carried out by Coccioni et al. (2008), Hyland et al.(2009), and Jovane et al. (2013). The studied interval spans from themiddle–upper part of Chron C18r to the lowermost part of C18r.1rwith an average sedimentation rate of ~8.57 m/Myr following thegeomagnetic polarity timescale (GPTS) of Cande and Kent (1995)(confirmed by Ogg, 2012). The primary nannofossil and foraminiferalbiostratigraphic events identified are (from bottom to top) the:(l) lowest occurrence (LO) of Orbulinoides beckmanni at 63.2 msl;(2) highest occurrence (HO) of O. beckmanni at 65.5 msl; and (3) HOof Chiasmolithus solitus at 70.0 msl (Jovane et al., 2013). It is worthnoting, however, that recognition of the LO of O. beckmanni can besubjective because this taxon originates from Globigerinatheka euganea,and some transitional forms of problematic taxonomic assignmentoccur from 62.8 to 63.2 msl. Additionally, identification of the HO of O.beckmanni is hampered by its scarce abundance, moderate preservationof planktic (P) and benthic (B) foraminifera and some problematicspecimen identifications up to 66.60 msl. Nevertheless, the MCAbiostratigraphic record in the 58–72 msl interval is likely continuousand is interpreted to span planktonic foraminiferal Zones P12 to P14of Berggren et al. (1995) and Zones E11 to E13 of Wade et al. (2011),and calcareous nannofossil Zones NP16 to NP17 of Martini (1971) andCP14a to CP14b of Okada and Bukry (1980).

3. Materials and methods

Two-hundred-sixty-five bulk rock samples were collected at ~5 cmstratigraphic intervals from the MCA section, corresponding to atemporal spacing of ~6 kyr between samples. High-resolution calciumcarbonate, stable isotopic, rock magnetic and micropaleontologicalanalyses were performed. Carbonate contents were measured at theNational Oceanography Centre, Southampton (NOCS). One-hundred-twenty-five bulk rock samples were reduced to fine powder in anagate mortar and their CaCO3 content was obtained using a Dietrich-Frühling calcimeter thatmeasures the CO2 volumeproduced by completedissolution of pre-weighed samples (300 ± 1 mg each) in 10% vol. HCl.Total carbonate contents (wt.% CaCO3) were computed with a precisionof 1% taking into account pressure and temperature of the laboratoryenvironment, the amount of bulk sample used, and the volume of CO2in the calcimeter. Standards of pure calcium carbonate (e.g. CarraraMarble) were measured every ten samples to ensure proper calibration(Appendix A). For coarse-fraction analyses, weighed freeze-driedsamples were soaked in deionized distilled water, and were washedthrough 63 μm sieves. The N63 μm fraction residue was collected, dried,and weighed. The coarse fraction is defined as the weight percent ratioof the N63 μm size fraction to the weight of the bulk sample (~100 g foreach sample) (Broecker and Clark, 1999, 2001). This index has beenlargely used as an indicator of carbonate dissolution (e.g. Hancock andDickens, 2005; Colosimo et al., 2006; Leon-Rodriguez and Dickens,2010; Luciani et al., 2010).

Semi-quantitative mineralogy of the bulk fraction was determinedon 6 representative samples distributed along the section (58.15,63.25, 63.50, 64.90, 65.10, and 69.05 msl) using powder X-raydiffraction (PXRD) analysis. For bulk analyses, an aliquot of about 1 gwas crushed and pulverized in a mortar. About 15 mg was used forPXRD analysis made on an automated Olympus® BTX diffractometer,using Co-K radiation, operated at 30 kV and 0.326 mA, over the range

34 J.F. Savian et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 414 (2014) 32–45

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

40 50 60 70 80 90 100

Str

atig

raph

ic th

ickn

ess

(msl

)

CaCO3

(%)

0 2 4 6 8 10 12

Coarse Fraction (%)

1.4 1.6 1.8 2 2.213C

(%o, VPDB)

-2 -1.5 -1

18O (%o, VPDB)

0 1 2 3 4 5 6 7

Susceptibility(x10 m3/kg)

0 2.4 4.8 7.2 9.6

IRM900 mT(x10 -5 Am2/kg)

0 0.2 0.4 0.6 0.8 1

S-ratio

(1) (2) (3) (4) (5) (6) (7)

40.1

3040

.145

P12

E11

P13

P14

E12

E13

NP

16C

P14

aN

P17

CP

14b

1

2

3

4

5

a

b

c

d

e

f

g

h

imarlcalcareous marl

lithology:

marly limestonelimestone

C18

rC

18n

-8

0 10 20 30 40

Susceptibility CFB (x10-8 m3/kg)

0 1 2 3 4 5

ARM (x10

-6 Am2/kg)

0 6 12 18 24 30

ARM CFB (x10-6 Am2/kg)

0 2 4 6 8

IRM900mT CFB (x10-5 Am2/kg)

0 0.5 1 1.5 2

HIRM (x10 -5 Am2/kg)

0 6 12 18(x10-6 Am2/kg)

HIRM CFBδ

δ

Fig. 2. Changes in CaCO3 content, coarse fraction, δ13C and δ18O in bulk sediments, low-fieldmagnetic susceptibility (χ), and rockmagnetic properties (anhysteretic remanentmagnetization, ARM; isothermal remanentmagnetization, IRM at 900mT;hard isothermal remanentmagnetization, HIRM; S-ratio300) across the 14-m-thick studiedMCA section. χ, ARM, IRM900mT, andHIRM300mTwere also calculated on a carbonate-free basis (CFB, dashed lines). Themagnetobiostratigraphy is from Jovaneet al. (2013). Numerical ages (1) are fromCande andKent (1995) (star) andOgg (2012) (diamond). Biostratigraphy is based on the planktonic foraminiferal Zones of (2) Berggren et al. (1995) and (3)Wade et al. (2011) and calcareous nannoplanktonZones of (4) Martini (1971) and (5) Okada and Bukry (1980). The areas shaded in yellow, dark yellow and green highlight the intervals (2–4) that represent the MECO, MECO warming peak and the post-MECO periods.

35J.F.Savian

etal./Palaeogeography,Palaeoclimatology,Palaeoecology

414(2014)

32–45

5–55° of 2θ, with a step size of 0.05° and 100 exposures, for an analysistime of 22min.Mineral identification and analysis was carried out usingthe XPowder software (Version 2010.01.15 PRO), which uses the PDF-2International Centre for Diffraction Data (ICDD) database. In order tominimize subjective influences, the baseline was determined automat-ically with XPowder defaults. X-ray diffraction identification criteriawere based on the indications of Biscaye (1965), Brown and Brindley(1980), and Moore and Reynolds (1997). For quantitative analysis, theReference Intensity Ratio (RIR) method, which is based upon scalingall diffraction data to the diffraction of standard reference materials,was used (Chung, 1974). The Xpowder software automatically com-putes an amorphous phase that is characterized by all the phases notconsidered in the quantitative analysis. Minerals were identified bytheir basal reflection at 3.85 and 3.03 Å (calcite), 4.26 and 3.34 Å(quartz), 15–16 Å (smectite), and 14.2, 7, and 4.72 Å (chlorite).

Oxygen and carbon isotope analyses were conducted using VGOptima and VG Prism dual-inlet isotope ratio mass spectrometersat NOCS. One-hundred-twenty-five samples were reacted in acommon acid bath at 90 °C using an automated carbonatepreparation system with a carousel device. NBS-19, Atlantis II, and anin-house Carrara Marble standard were included in all sample runs.All values are reported in standard delta notation (δ) in parts per mil(‰) relative to VPDB (Vienna Pee Dee Belemnite), and analyticalprecision is estimated at 0.06‰ (1σ) for δ13C and 0.08‰ (1σ) for δ18O(Appendix A).

Paleomagnetic and environmental magnetic measurements werecarried out at NOCS and at the University of São Paulo (USP). Themagnetic remanence of 253 unoriented block samples (5-cmresolution) was measured using a three-axis 2-G Enterprises cryogenicmagnetometer (model 755R), housed in a magnetically shielded roomat NOCS.

Low-field magnetic susceptibility (χ) was measured with aKappabridge KLY-3 (AGICO) magnetic susceptibility meter. All datawere normalized by mass, due to the irregular sample volumes. Ananhysteretic remanent magnetization (ARM) was imparted in a100 mT alternating field (AF) with a direct current bias field of0.05 mT. An isothermal remanent magnetization (IRM) was impartedin a direct field of 900 mT (IRM900mT) and was demagnetized in back-fields of 100 mT (−IRM100mT) and 300 mT (− IRM300mT). From thesemeasurements, we calculated the S-ratio (S300mT = [− IRM300mT /IRM900mT]) and “hard” isothermal remanent magnetization(HIRM300mT = [IRM900mT + IRM300mT] / 2), in order to investigate thecoercivity of magnetic minerals. Hysteresis loops and first-orderreversal curve (FORC) diagrams were analyzed for eight samples fromthe MCA section spanning the MECO interval and from intervalsimmediately before and after thewarming event. FORCsweremeasuredwith an averaging time of 200 ms, a smoothing factor (SF) of 4, andusing the input parameters of Egli et al. (2010) (Hc1 = 0 mT, Hc2 =110 mT; Hu1 = −15 mT, Hu2 = +15 mT; δH= 0.63 mT). For sampleswith natural remanent magnetization above 10−6 Am2/kg we per-formed only one FORC run using a vibrating sample magnetometer(Princeton Measurements Corp., PMC) at NOCS, whereas for weaksamples we performed multiple runs with a PMC alternating gradientmagnetometer to increase the signal-noise ratio of the FORC distribu-tion (Heslop and Roberts, 2012). FORC diagrams derived from singleruns were calculated using the MATLAB routine of Egli et al. (2010).FORC diagrams obtained with multiple runs were stacked andcalculated using the MATLAB routine of Heslop and Roberts (2012).Magnetic mineralogy was further investigated at USP for selectedsamples through the acquisition of an IRMandmeasurement of thermo-magnetic curves. IRM acquisition curves were obtained for six sampleswith a 2-G Enterprises pulsemagnetizer and a cryogenicmagnetometer(model 755UC). IRM acquisition curves were analyzed with cumulativelog-Gaussian (CLG) functions using the software of Kruiver et al. (2001).The CLG function is described by three parameters (SIRM, B1/2 and dis-persion parameter, DP) that characterize magnetic minerals

(Robertson and France, 1994; Kruiver et al., 2001). Thermomagneticcurves up to 700 °C were obtained using a KLY-4S AGICOmagnetic susceptibility meter with high-temperature attachmentat USP to determine the Curie or Néel temperatures of magneticminerals.

For calcareous nannofossil analyses, samples were prepared fromunprocessedmaterial as simple smear slides using standard preparationmethods (Bown and Young, 1998) at the Istituto di Geoscienze eGeorisorse CNR di Pisa (Italy). Smear slides were studied under a LeitzLaborlux 12 Pol light microscope both under crossed nicols and trans-mitted light at a magnification of 1250×. Most nannoplankton specieswere identified according to the taxonomy of Perch − Nielsen (1985)except for sphenoliths and Dictyococcites that were classified followingFornaciari et al. (2010) and Reticulofenestra umbilicus that was definedfollowing the taxonomic criteria adopted by Backman and Hermelin(1986), ascribing to this species all specimens N14 μm and groupingas Reticulofenestra spp. all specimens b14 μm. Assemblages were stud-ied following quantitative countingmethods based on at least 300 spec-imens (Appendix B). Following Toffanin et al. (2011), the relativeabundances of species belonging to the genus Sphenolithuswere deter-mined by counting 100 specimens. Rare taxa, such as the genusHelicosphaera and species belonging to the genus Chiasmolithus werecounted in a prefixed area of 10 mm2 (three–four transects). The stan-dard calcareous nannofossil zonations of Martini (1971) and Okadaand Bukry (1980) are widely used for low- and middle-latitude Eo-cene–Oligocene (E–O) biostratigraphic studies, and were used in thispaper. To infer probable temperature and trophic variations of surfacewaters, most calcareous nannofossils were, when possible, allocatedinto groups of environmental affinities, largely following Haq andLohmann (1976), Aubry (1992), Gardin and Monechi (1998),Bralower (2002a,b), Tremolada and Bralower (2004), Persico and Villa(2004), Gibbs et al. (2006), Villa et al. (2008), Raffi et al. (2009)and Agnini et al. (2011). Based on this literature, the following environ-mental groups were used: eutrophic taxa (Dictyococcites bisectus,Dictyococcites scrippsae, Reticulofenestra daviesii), and oligotrophictaxa (Cribrocentrum reticulatum, Ericsonia spp., Sphenolithus spp.,Zygrhablithus bijugatus) (Appendix B).

For foraminiferal analyses, samples were treated at theUniversità degli Studi di Urbino, following the cold acetolysetechnique of Lirer (2000), by sieving through a 63 μm mesh anddrying at 50 °C. The cold acetolysis method enabled extractionof generally easily identifiable foraminifera even from induratedlimestones. This technique offered the possibility of accurate tax-onomic determination and detailed foraminiferal assemblageanalysis. For planktonic foraminifera, all samples were studiedfor biostratigraphy and quantitative analysis was performed ona subset of 101 samples. The residues were studied with a binoc-ular stereomicroscope to characterize assemblages and to identi-fy biostratigraphic marker species. A representative split of atleast 300 specimens was picked from the N63 μm fraction,mounted on micro-slides for permanent record and identificationpurposes, and classified following the taxonomic criteria ofBerggren and Pearson (2005). The planktonic foraminiferal zona-tions of Berggren et al. (1995) and Wade et al. (2011) werefollowed. Following Hancock and Dickens (2005), the fragmenta-tion index (FI) was calculated using at least 300 specimens and in-cluding the whole test, fragments and dissolved tests to estimatecarbonate dissolution effects (Appendix C). For benthic foraminifera,a quantitative study of sixty-four selected samples was performed. Arepresentative split of the N90 μm fraction was used to pick approx-imately 300 specimens. The sample–split weight used to pick ben-thic foraminifera was determined so that the foraminiferal density(FD), expressed as the number of foraminifera per gram of dry sedi-ment, could be calculated. The planktonic to planktonic and benthic(P/P + B) ratio, expressed as a percentage, and the percentages ofagglutinants were also calculated (Appendices C and D).

36 J.F. Savian et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 414 (2014) 32–45

4. Results

4.1. Micropaleontology

Benthic foraminiferal assemblages are generally diverse andwell-preserved throughout the studied MCA section, except at63.2–64.0 msl where evidence of partial dissolution is observed(Fig. 3). The assemblages are dominated by calcareous-hyaline formswith variable percentages of agglutinants that are more abundant atthe base of the section and at 63.2–64 msl (Fig. 3). The P/(P + B) ratiofluctuates throughout the sequence, with the lowest values at63.2–64 msl where both the highest FD value and percentage fragmen-tation are documented (Fig. 3). The 63.2–65.5 msl interval is thereforecharacterized by pronounced paleoecological and paleoenvironmentalchanges. The increase of benthic to planktonic foraminifera and ofagglutinated forms that are less prone to dissolution might reflectshallowing of the lysocline. In addition, increased FD values mightsuggest greater nutrient availability at the seafloor probably related toincreased detrital mineral influx associated with intensified hydrologicaland weathering cycles. In two intervals eutrophic taxa increase inabundance relative to oligotrophic taxa: 63.5–65.5 msl and 67.0–70.2 msl(Appendix B). Eutrophic and oligotrophic percentages vary throughoutthe section, but within the background level of variation.

4.2. Stable isotopes

Bulk δ13C and δ18O values average 1.8‰ and −1.4‰, respectively,through the studied MCA section (Figs. 2c, d). δ13C data havebackground values of ~1.8‰ and are characterized by an increasefrom ~1.6 to 2.1‰ at 63.2–65.5 msl (~40.2–40.1 Ma), with a peakvalue of 2.1‰ at 63.35 msl (~40.1 Ma) (Fig. 2c). The peak value isfollowed by a smoothly decreasing trend up to 66.34 msl (~39.8 Ma).Bulk δ18O varies from−0.2‰ to−3.3‰ and is noisy (Fig. 2d). Similarlynoisy δ18O in the correlative CHW section was interpreted as resultingfrom burial diagenesis and/or meteoric water diagenesis (Jovane et al.,2007). It is well known that bulk carbonate δ13C is more robust to

diagenetic alteration than δ18O because the carbon content of fluidsis often too low compared to that of carbonate rocks to modifysignificantly the carbonate carbon isotopic composition (e.g. Veizerand Hoefs, 1976).

4.3. Environmental magnetism

Low-field magnetic susceptibility (χ) along the studied MCA sectionvaries between 0.57 × 10−8 and 6.42 × 10−8 m3/kg (Fig. 2e), whereasCaCO3 contents vary between 47% and 93% (Fig. 2a). Two narrowintervals (63.30–63.90 msl and 64.15–65.30 msl) have lower CaCO3contents and higher χ values (Fig. 2a, e). CaCO3 and χ have a significantnegative correlation (r=−0.72), i.e. lower CaCO3 is related to higher χvalues, which correspond to peak abundances of paramagnetic (e.g.clays) and ferrimagnetic minerals. Six representative samples fromdepths of 58.15, 63.25, 63.50, 64.90, 65.10, and 69.05 msl have similarXRD patterns, with slight differences between samples. The mainphases present, in order of abundance, are: calcite (Cal), quartz (Qtz),smectite (Sm) and chlorite (Chl) (Fig. 4). Calcite contents vary betweena minimum of 61.3% (63.50msl; Fig. 4c) and 77.2% (58.15 msl; Fig. 4a);samples from 58.15 to 69.05 msl (Fig. 4a, f) have relatively higheramounts of calcite (77.2% and 71.8, respectively) compared to samplesfrom 63.25, 63.50, 64.90 and 65.10 msl (61.5%, 61.3%, 65.9% and 66.1%,respectively, Fig. 4b, c, d, e). These results are compatible with thesmaller amounts of CaCO3 in the same interval (Fig. 2a). Quartz valueshave an opposite trend, with higher values (4.6%) in sample63.50 msl, and lower amounts (2.7%) in sample 58.15 msl. Smectiteand chlorite have similar trends, with lower values (8.3%, 10.3% and5.7%, 7.9%, respectively) in samples 58.15 msl and 69.05 msl. Incontrast, higher percentages of smectite and chlorite are found insamples 63.25, 63.50, 64.90 and 65.10 msl. The amorphous phase alsohas a comparable trend to that of quartz, smectite and chlorite, withhigher contents in samples from 63.25, 63.50, 64.90 and 65.10 msl.

To account for dilution effects by the carbonatematrix,we calculatedmagnetic parameters (χ, ARM, IRM900mT and HIRM300mT) on acarbonate-free basis (CFB, dashed lines in Fig. 2e–h). To achieve this,

limestone

NP

16N

P17

CP

14b

CP

14a

P12

P14

E13

E11C

18r

C18

n

58

60

62

66

68

70

72

71

69

67

65

61

59

P13

E12Bar

toni

an

Lithology:calcareous marl

marly limestonemarl

(6)(4)(5)(2) (3)(1) (7) 0 10 20 30

Faunal Density(x100)

0 10 20 30 40

Agglutinants(%)

100 98 96 94 92

P/(P+B)(%)

64

63

10 20 40 60 80

Fragmentation Index(%)

1

2

3

4

5

Str

atig

raph

ic th

ickn

ess

(msl

)

40.1

3040

.145

Fig. 3. Changes in selected benthic and planktonic foraminiferal parameters [P / (P + B) ratio], foraminiferal density (FD), agglutinants, and fragmentation index across the studied MCAsection. Magnetobiochronostratigraphy and shaded areas are in Fig. 2.

37J.F. Savian et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 414 (2014) 32–45

we normalized the magnetic parameters by (100 wt.% CaCO3). Allmagnetic parameters that depend on the concentration of ferrimagneticminerals are well correlated before and after this normalization(Fig. 2e–h). For example, the ARM and ARM CFB, IRM900mT andIRM900mT CFB, and HIRM300mT and HIRM300mT CFB have coherentpeaks at 63.30–63.90 msl and 64.15–65.30 msl, as mentioned above,but also a less pronounced peak at 66.6–68.8 msl.

The relative concentration of different magnetic minerals across thestudied interval can be inferred from the S-ratio300mT and HIRM300mT(Bloemendal et al., 1992; Liu et al., 2007). The large IRM300mT andS-ratio300mT fluctuations indicate variable proportions of low andhigh-coercivity magnetic minerals, i.e. magnetite (predominant whenS-ratio is near 1) and hematite/goethite (present in significantconcentrations when HIRM is high and S-ratio is close to 0). Thepresence of mixtures of magnetic minerals with contrasting coercivitiesis corroborated by stepwise IRM acquisition curves. IRM acquisitioncurves were measured at fields up to 1 T for six representative samplesalong theMCA section (Fig. 5a; Table 1) and components due to differentmagnetic minerals were fitted using CLG functions (Fig. 5b–d) (Kruiver

et al., 2001; Heslop et al., 2002). Three magnetic components wereidentified (Table 1). They correspond to a low-coercivity component(B1/2 = 16 mT, DP = 0.40–0.45), a medium-coercivity component(B1/2 = 50–70 mT, DP= 0.30–0.40) and a high-coercivity component(B1/2 = 250–479 mT, DP = 0.24–0.30). Following Roberts et al. (2011,2012), component 1 (lowest coercivity component in Table 1) isinterpreted as related to relatively coarse-grained magnetite of detritalorigin. However, we cannot discount that component 1 represents thecoarse end of a distribution of fine-grained, largely superparamagneticparticles, perhaps produced in situ by dissimilatory iron-reducingbacteria (e.g. Egli, 2004). The magnetofossil component is correlatedto themedium-coercivity component 2 (Table 1) with coercivity valuesaround 40 mT and a small DP (~0.3), similar to the IRM signal ofmagnetotactic bacteria found in other studies (e.g. Egli, 2004; Chenet al., 2007; Jovane et al., 2012; Roberts et al., 2012). Component 3 isinterpreted to represent a high coercivity hematite fraction, but canalso include some fraction of goethite. Inside the magnetic mineralconcentration peaks, the medium-coercivity component usuallydominates the IRM signal, with contributions varying between 31%

Sm ChlQtzCal AmorphousLEGEND

Rel

ativ

e In

tens

ity

Cal

Cal

CalCal

CalSm Chl Qtz CalChl Qtz

MCA-58.158.3

5.72.7

77.2

6.1

Rel

ativ

e In

tens

ity

Cal

Cal

Cal Cal CalSm Chl QtzCalChl

Qtz

MCA-63.2514.3

11.2

3.8

61.6

9.1

Rel

ativ

e In

tens

ity

Cal

Cal

CalCal

CalSm Chl Qtz

CalChl Qtz

MCA-63.5014.4

11.0

4.6

61.3

8.7

Rel

ativ

e In

tens

ityCal

Cal

CalCal Cal

SmChl Qtz CalChl Qtz

MCA-64.9014.1

8.5

3.7

65.9

7.8

0 10 20 30 40 50

Rel

ativ

e In

tens

ity

Cal

Cal

CalCal

CalSmChl

QtzCalChl Qtz

MCA-65.1012.9

9.4

3.7

66.1

7.9

Rel

ativ

e In

tens

ity

Cal

Cal

CalCal

CalSmChl Qtz CalChl Qtz

MCA-69.0510.3

7.9

3.2

71.8

6.8

a b

c d

e f

Degrees 2θ 0 10 20 30 40 50

Degrees 2θ

0 10 20 30 40 50

Degrees 2θ 0 10 20 30 40 50

Degrees 2θ

0 10 20 30 40 50

Degrees 2θ 0 10 20 30 40 50

Degrees 2θ

Fig. 4. Representative XRD diffractograms for six representative samples (a) before theMECO, (b, c) during theMECOwarming peak, (d, e) during the post-MECO period, and (f) after theMECO event. Pie charts contain the semi-quantitative summaries for each sample.

38 J.F. Savian et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 414 (2014) 32–45

and 52%. The low-coercivity component varies from 5% to 17%. It occursinside the magnetic concentration peaks but is absent (or below thedetection level) outside these stratigraphic levels. In these samples(MCA-67.55, MCA-68.60), the high-coercivity component can reachup to 82% of the total IRM.

Hysteresis data (Fig. 6; Table 2), including the ratio of saturationremanence to saturation magnetization (Mrs/Ms) and the coercivity ofremanence to coercive force (Bcr/Bc), from MCA samples lie within thepseudo-single domain (PSD) field of Day et al. (1977). The presence ofhematite mixed with fine-grained magnetite is indicated by thewasp-waisted shape of the loops (Roberts et al., 1995). The mixture of

low- and high-coercivity phases does not affect significantly the Dayplot because SD hematite has similar hysteresis ratios to those of SDmagnetite (Roberts et al., 1995). The significant departure of bulkhysteresis parameters from values expected for uniaxial SD magnetitefor both theMCA and CHW samples (Fig. 6) can be related to significantmixtures of non-SD detrital magnetite grains (superparamagnetic, PSDand multi-domain grains) in the studied samples (Dunlop, 2002;Roberts et al., 2012) as suggested by the IRM acquisition curves.

FORC diagrams provide detailed information about magneticinteractions and microcoercivity distributions (Roberts et al., 2000).High-resolution FORC measurements following the specifications of

0

0.2

0.4

0.6

0.8

1

0 200 400 600 800 1000

63.25

M/M

r

Field(mT)

63.5064.5565.1067.5568.60

a

1

2

3

4

5

6MCA - 63.25 msl

Raw dataComponent 1Component 2Component 3Sum of components

Syn

thet

ic IR

M (

x10

-4A

m2 /

kg)

1

2

3

4

5

6

1 1.5 2 2.5 3

MCA - 65.10 msl

1 1.5 2 2.5 3

0.5

1

1.5

2

2.5MCA - 68.60 msl

0.5

1

1.5

2

2.5

3

1 1.5 2 2.5 3

1

2

3

4

5

6

1

2

3

4

5

6

b c dRaw dataComponent 1Component 2Component 3Sum of components

Raw dataComponent 2Component 3Sum of components

Gra

dien

t (x1

0-4

)

10Log Applied Field (mT) 10Log Applied Field (mT) 10Log Applied Field (mT)

Fig. 5. (a) IRM acquisition curves for six representative samples from the MCA section. (b–d) IRM unmixing analyses (Kruiver et al., 2001; Heslop et al., 2002) for three representativesamples; (b) and (c) represent samples collected inside themagneticmineral concentration peaks and (d) fromoutside the peaks. Rawdata (circles) and calculated IRMacquisition curvesare shown for two and three fitted components after fitting of a spline function.

39J.F. Savian et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 414 (2014) 32–45

Egli et al. (2010) can be used to infer the presence of biogenicmagnetite(e.g. Egli et al., 2010; Roberts et al., 2011, 2012; Jovane et al., 2012;Larrasoaña et al., 2012; Yamazaki, 2012; Yamazaki and Ikehara, 2012).FORC distributions are nearly identical in the high-magnetizationintervals between 63.2 and 65.5 msl (Figs. 2, 7). These FORC diagramshave a sharp horizontal ridge (Fig. 7) at Hb = 0 that indicate negligiblemagnetic interactions and a dominance of non-interacting SD particles(Roberts et al., 2000) that is characteristic of intactmagnetosome chains(e.g. Egli et al., 2010; Roberts et al., 2011; Jovane et al., 2012; Larrasoañaet al., 2012; Roberts et al., 2012; Yamazaki, 2012; Yamazaki and Ikehara,2012). The coercivity distribution has a broad peak between 3 and25 mT, with a maximum at 15–20 mT (Fig. 7) that falls within therange expected for magnetite magnetosomes (Egli, 2004; Kopp andKirschvink, 2008; Egli et al., 2010). Outside the 63.2–65.5 msl interval,the magnetic signal of the samples is much weaker and no meaningfulFORC distribution could be obtained from these samples (Fig. 7a, h).Multiple runs were then performed for intervals with low magnetiza-tion and also within the 66.6–68.8 msl peak. Intervals with lowmagne-tization have no meaningful FORC distributions even after stacking ofnine runs (Fig. 7j). On the other hand, statistically significant FORC

results are estimated for the 66.6–68.8 msl interval after stacking ofnine measurement runs; resulting distributions are similar to thoseobserved within the major magnetic peaks (Fig. 7k).

To compare our results with those from a nearby section, we alsoperformed FORC analyses on samples from the MECO interval at theCHW section, between 135 and 139 msl, as defined by Jovane et al.(2007). This interval is characterized by a peak in magneticconcentration-dependent parameters (e.g. χ, ARM, IRM). A horizontalridge due to non-interacting SD magnetite with coercivities between10 and 50 mT is also evident in these samples (Fig. 7i).

5. Discussion

5.1. The MECO event in the Neo-Tethys

Benthic and planktonic foraminiferal and calcareous nannofossilassemblages in the MCA section across the Chron C18r/C18n boundaryare affected by important faunal turnovers. Together with rockmagnetic,geochemical and stable isotope data, they enable subdivision of thestudied 14-m-thick section into five discrete intervals (Figs. 2, 3, 8):

Table 1Parameters associated with components identified from IRM analysis of six representative samples from the Monte Cagnero section.

Sample Component 1 Component 2 Component 3 Component 1 Component 2 Component 3

SIRM (Am2/kg) SIRM (Am2/kg) SIRM (Am2/kg) Contribution (%) Contribution (%) Contribution (%)

MCA-63.25 6.00E−05 3.30E−04 2.00E−04 10 56 34MCA-63.50 3.00E−05 2.00E−04 4.20E−04 5 31 65MCA-64.55 1.20E−04 3.20E−04 2.50E−04 17 46 36MCA-65.10 5.00E−05 3.20E−04 2.50E−04 8 52 40MCA-67.55 NA 1.40E−04 2.40E−04 NA 37 63MCA-68.60 NA 4.00E−05 1.80E−04 NA 18 82

B1/2 (mT) B1/2 (mT) B1/2 (mT) Contribution (%) Contribution (%) Contribution (%)

MCA-63.25 15.8 63.1 478.6 10 56 34MCA-63.50 15.8 50.1 251.2 5 31 65MCA-64.55 15.8 63.1 446.7 17 46 36MCA-65.10 15.8 70.8 446.7 8 52 40MCA-67.55 NA 70.8 239.9 NA 37 63MCA-68.60 NA 70.8 251.12 NA 18 82

DP (log10 mT) DP (log10 mT) DP (log10 mT) Contribution (%) Contribution (%) Contribution (%)

MCA-63.25 0.40 0.33 0.24 10 56 34MCA-63.50 0.40 0.30 0.30 5 31 65MCA-64.55 0.45 0.34 0.24 17 46 36MCA-65.10 0.45 0.31 0.24 8 52 40MCA-67.55 NA 0.40 0.30 NA 37 63MCA-68.60 NA 0.40 0.30 NA 18 82

0

0.1

0.2

0.3

0.4

0.5

0.6

0 1 2 3 4 5 6

Mrs

/Ms

Bcr/Bc

SD

PSD

MD-3 10-6

-2 10-6

-1 10-6

0

1 10-6

2 10-6

3 10-6

-1 -0.5 0 0.5 1

MCA-63.85 msl

Mom

ent (

Am

2 )

Field (T)

a b

Fig. 6. (a) Hysteresis loop for one representative sample fromMCA pelagic carbonate samples. (b) Day plot (Day et al., 1977) for nine representative samples from the MCA section. Thedata fields represented in the Mrs/Ms versus Bcr/Bc diagram are for single domain (SD), pseudo-single domain (PSD) and multi-domain (MD) titanomagnetite particles.

40 J.F. Savian et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 414 (2014) 32–45

Interval 1 (58–61 m, ~370 kyr), Interval 2 (61–63.2 m, ~250 kyr),Interval 3 (63.2–64 m, ~70 kyr), Interval 4 (64–65.5 m, ~200 kyr) andInterval 5 (65.5–72 m, ~780 kyr). Following Bohaty et al. (2009), thetotal duration of the MECO event is estimated to be ~500 kyr with awarming maxima at the end of the event. During the warming peak, ashoaling of the CCDbyup to 500–1500mhas been reported in the Pacific,Atlantic and Indian oceans where a total loss of carbonate is reported atsites with paleodepths below ~3000 m (Bohaty et al., 2009; Pälikeet al., 2012). Interval 3 of the MCA section is characterized by the lowestCaCO3 contents in the studied section and the highest bulk carbonate δ13Cvalues. It also has the lowest concentrations of the coarsest fraction,which is used as a proxy for carbonate dissolution (e.g. Hancock andDickens, 2005; Colosimo et al., 2006; Leon-Rodriguez and Dickens,2010; Luciani et al., 2010). Dissolution during this interval is alsosupported by increased values of the planktonic foraminiferal FI, greaterrelative abundances of agglutinated foraminifera that are less prone todissolution, and a clear decrease of the P/P + B ratio (Fig. 3). Theincrease in dissolution likely contributes to the higher concentrationsof magnetic particles in this interval indicated bymagnetic susceptibility,ARM (andARMCFB), IRM900mT (and IRM900mT CFB), andHIRM300mT (andHIRM300mT CFB) (Fig. 2). In contrast, there is no evidence for carbonatedissolution upsection in Intervals 4 and 5. Interval 3 falls within the low-ermost part of Chron C18n, therefore, it can be magnetostratigraphicallycorrelated with peak MECO warming in marine sediment cores (Bohatyet al., 2009), and in the Tethyan CHW (Jovane et al., 2007) and Alano sec-tions (Luciani et al., 2010; Spofforth et al., 2010; Toffanin et al., 2011).Thus, observed changes in paleoenvironmental proxies provide cluesthat point to seafloor CaCO3 dissolution and lysocline shallowing at theMCA site during the MECO warming. Interval 3 is immediately followedby an intervalwith lower CaCO3 contents associatedwith highermagnet-ic susceptibility values. On the basis of available magneto- and bio-stratigraphic results, Interval 4 correlates well with the organic-richORG1 unit identified at the Alano section (Luciani et al., 2010; Spofforthet al., 2010). These organic-rich layers are thought to represent rapidorganic carbon burial events at the end of the MECO event, probablyinduced by enhanced delivery of terrestrial material to the ocean(Spofforth et al., 2010).

5.2. Ocean iron fertilization and magnetotactic bacterial abundance duringthe MECO event

Detrital magnetic minerals can be delivered to the deep sea throughdifferent transportation pathways, including ice-rafting, mass flows,hemipelagic sediment plumes or wind (e.g. Evans and Heller, 2003;Liu et al., 2012). Fine-grained terrigenous inputs that reached thedeep, carbonate-dominated MCA site were most likely transported bysuspended plumes or wind. Both processes result in similar grain-sizedistributions and mineralogy (Rea, 1994). Hemipelagic sediment hasbeen reported in modern environments to distances of 500 km fromthe coast and possibly as far as 900 km (Rea, 1994 and references there-in). Such distances are within the expected range of the MCA sectionfrom the paleoshoreline of the Neo-Tethys. Nevertheless, processes

that control hemipelagic and aeolian inputs to the deep sea tend to beout of phase;wet conditions (and enhanced runoff) in the source regionwould favour hemipelagic fluxes, whereas drier source region climateswould enhance aeolian dust fluxes. Enhanced aeolian supply relativeto detrital fluxes into theNeo-Tethys has been reported on the southernTethyanmargin (central Egypt) as a result of drier, likelymore seasonal,climatic conditions during the Paleocene–Eocene Thermal Maximum(PETM) (Schulte et al., 2011). Similar observations have been madefor the PETM in the Bighorn Basin, Wyoming (Wing et al., 2005; Krausand Riggins, 2007; Smith et al., 2009), East Africa (Handley et al.,2012) and the southernKerguelenPlateau, IndianOcean, on theAntarcticmargin (Larrasoaña et al., 2012).

Environmental magnetic parameters can serve as sensitive aeoliandust proxies, given that hematite forms in oxidizing, dehydrating desertenvironments (Larrasoaña et al., 2003; Liu et al., 2012 and referencestherein). Hematite concentrations can be easily tracked using the“hard” IRM on a carbonate free basis (HIRM900mT CFB). In the MCAsection, HIRM900mT CFB has two pronounced peaks in Intervals 3 and4 (Fig. 8c), which coincide with the MECO peak warming and thepost-MECO recovery. We interpret the increased hematite concentra-tions in Intervals 3 and 4 to be related to the presence of an enhancedaeolian dust component similar to scenarios proposed for other Eocenesites (Roberts et al., 2011; Larrasoaña et al., 2012). For example, rapidcontinental aridification has been recorded by lithofacies and pollenrecords from the Xining Basin dated at 40.0 Ma (Bosboom et al.,2014). The onset of continental aridification and obliquity-dominatedclimate cyclicity coincides with the MECO-peak and continuesafterward into the post-MECO cooling.

Aeolian dust is a potential source for iron fertilization of the ocean(e.g. Maher et al., 2010; Roberts et al., 2011; Larrasoaña et al., 2012;Liu et al., 2012). Roberts et al. (2011) reported enhanced iron supplyby aeolian dust in Eocene sediments (ODP Hole 738B) of the southernKerguelen Plateau (IndianOcean),where amarked increase in hematiteconcentrations coincided with a switch from oligotrophic to eutrophicconditions. The interval with higher hematite concentrations is alsocoincident with increased magnetofossil abundances. Roberts et al.(2011) argued that iron input and increased organic carbon deliveryto the seafloor were the main factors that controlled the abundance ofmagnetotactic bacterial populations. Likewise, in the MCA section themagnetic proxies classically used to trace low-coercivity magnetite(ARM CFB) and high-coercivity hematite/goethite (HIRM900mT CFB)co-vary, which indicates a concomitant increase in concentration ofboth magnetic minerals in Intervals 3 and 4 (Figs. 2 and 8b, c). Asignificant fraction of the magnetite in these intervals is due to non-interacting SD particles, as indicated by a FORC central ridge signature(Fig. 7), which we attribute to putative magnetofossils. The sameintervals are characterized by a relative increase in eutrophic calcareousnannofossil taxa (Fig. 8d). Eutrophication could have been stimulatedby sea surface iron fertilization produced by addition of iron-richaeolian dust to surface waters. This promoted enhanced organic carbonexport to deeper waters and burial on the seafloor, which stimulatedincreased magnetotactic biomineralization (e.g., Villa et al., 2014). Theiron needed for biomineralization by magnetotactic bacteria is likelyto have been provided by diagenetic iron reduction that released Fe3+

from the most reactive iron-bearing minerals, including hydrous ferricoxide and lepidocrocite (Poulton et al., 2004). Magnetite and hematiteare more resistant to dissolution (Yamazaki et al., 2003; Poulton et al.,2004; Garming et al., 2005; Roberts et al., 2011) and, therefore, survivedthis mild iron reduction that occurred under iron-reducing but notanoxic conditions. Thus, simultaneous delivery of enhanced organiccarbon, reactive iron-bearing aeolian dust particles and non-reactiveaeolian hematite particles to the seafloor would have released theexisting limitation on key nutrients (carbon and iron) for an existing,but small, population ofmagnetotactic bacteria to produce the observedmagnetic signatures. An increase in ARMCFB and HIRM900mT CFB is alsoobserved outside the MECO interval, but with lower intensity, at the

Table 2Measured hysteresis parameters for studied sediments from the MCA and CHW sections.

Sample ID Depth(msl)

Mr (Am2/kg) Ms (Am2/kg) Mr/Ms Bc(mT)

Bcr(mT)

Bcr/Bc

CAG-58.15 58.15 18.61E−09 44.89E−09 0.41 28.65 287.27 10.00CAG-64.55 64.55 394.32E−09 1.03E−06 0.37 27.47 72.34 2.63CAG-63.25 63.25 428.86E−09 1.83E−06 0.23 20.02 57.05 2.85CAG-64.90 64.90 497.47E−09 1.26E−06 0.39 30.42 73.96 2.43CAG-63.50 63.50 741.31E−09 2.01E−06 0.36 29.24 84.64 2.89CAG-65.10 65.10 545.76E−09 1.70E−06 0.31 28.79 79.39 2.76CAG-63.85 63.85 768.56E−09 2.70E−06 0.28 22.17 60.15 2.71CAG-66.30 66.30 46.93E−09 997.2E−09 0.04 5.04 9.45 1.87CAG-64.50 64.50 429.7E−09 1.08E−06 0.39 33.79 91.15 2.70

41J.F. Savian et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 414 (2014) 32–45

c) Monte Cagnero 63.25 msl

10 20 30 40 50 60 70 80 90 100

10505

10

- 0.

4

- 0.

20

0.2

0.4

0.6

0.81

e) Monte Cagnero 63.50 msl

20 40 60 80 100

10505

10

- 0.

4

- 0.

20

0.2

0.4

0.6

0.81

g) Monte Cagnero 64.50 msl

20 40 60 80 100

10505

10

- 0.

8

- 0.

6

- 0.

4

- 0.

20

0.2

0.4

0.6

0.81

a) Monte Cagnero 58.15 msl

10 20 30 40 50 60 70 80 90 100

10505

10

- 0.

8

- 0.

6

- 0.

4

- 0.

20

0.2

0.4

0.6

0.81

b) Monte Cagnero 64.55 msl

20 40 60 80 100

10505

10

- 0.

4

- 0.

20

0.2

0.4

0.6

0.81

d) Monte Cagnero 64.90 msl

20 40 60 80 100

10505

10

- 0.

4

- 0.

20

0.2

0.4

0.6

0.81

f) Monte Cagnero 65.10 msl

10 20 30 40 50 60 70 80 90 100

10505

10

- 0.

8

- 0.

6

- 0.

4

- 0.

20

0.2

0.4

0.6

0.81

h) Monte Cagnero 66.30 msl

10 20 30 40 50 60 70 80 90 100

10505

10-

1

- 0.

8

- 0.

6

- 0.

4

- 0.

20

0.2

0.4

0.6

0.81

i) Contessa Highway 137.15 msl

10 20 30 40 50 60 70 80 90 100

10505

10

- 0.

6

- 0.

4

- 0.

20

0.2

0.4

0.6

0.81

Hc [mT]

Hc [mT] Hc [mT]

Hb

[mT

]

Hb

[mT

]

Hb

[mT

]

Hb

[mT

]

Hb

[mT

]

Hb

[mT

]

Hb

[mT

]

Hb

[mT

]

Hb

[mT

]

Hb

[mT

]

Hb

[mT

]

Hc [mT] Hc [mT]

Hc [mT] Hc [mT]

Hc [mT] Hc [mT]

Hc [mT] Hc [mT]

20 40 60 80 100

10505

10

00.20.40.60.8

20 40 60 80 100

10505

10

0.60.40.20.81

j) Monte Cagnero 65.70 msl

k) Monte Cagnero 67.55 msl

42 J.F. Savian et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 414 (2014) 32–45

66.6–68.8 msl peak. FORC distributions for this interval are similar tothose observed across the MECO (Fig. 7) and also coincide with anincrease in eutrophic taxa (Fig. 8d). In this case, the same mechanismcan be advocated for the coeval increase in magnetofossils andhigh-coercivity phases (Fig. 8b, c) and suggests the presence of amagnetotactic bacteria assemblage throughout the whole studiedinterval that bloomedwith increased supply of limiting nutrients duringepisodic warming events.

6. Conclusions

An integrated high-resolution stable isotope, geochemical, micropa-leontological and environmentalmagnetic analysis has been carried outover a 14-m-thick interval of the Monte Cagnero section (Umbria-Marche Basin), Italy, which corresponds to the 40.8–39.1 Ma periodaround the Middle Eocene Climatic Optimum (MECO). Magneticparameters indicate a concomitant increase of aeolian iron supply in theform of hematite, and a higher abundance of magnetite magnetofossilsproduced by magnetotactic bacteria as indicated by FORC diagrams thatare typical of non-interacting SD magnetite between 63.2 and 65.5 msl.This interval corresponds to the peak MECO warming and its aftermath.Intervals with enhanced putative magnetofossil concentrationscorrespond to those for which other proxies systematically point to anincrease in primary productivity, which was probably stimulated by

increased aeolian supply of iron to surface ocean waters. Such a scenariohas been recently envisaged for the PETM event (e.g. Chang et al., 2012;Larrasoaña et al., 2012), and we now confirm a similar connectionbetween putative magnetofossil abundance and paleoproductivitythrough the MECO event. It reinforces the connection betweenhyperthermal climatic events and the occurrence (or increasedabundance) of putative magnetofossils. Further work is needed to assesswhether the preserved inorganic remains of magnetotactic bacteria canprovide a useful paleoproductivity proxy in ancient carbonate sediments.

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.palaeo.2014.08.009.

Acknowledgements

We acknowledge financial support from the Marie Curie Actions(FP7-PEOPLE-IEF-2008, proposal n. 236311). Jairo F. Savian acknowl-edges the Brazilian CNPq (Process 201508/2009 5) for funding a VisitingFellowship at the National Oceanography Centre Southampton (NOCS),where the magnetic and isotopic measurements were completed.Andrew Roberts acknowledges the Australian Research Council forproviding funding through grant DP140104544. Luigi Jovane andFrancesco Iacoviello acknowledges the Fundação de Amparo à Pesquisado Estado de São Paulo (FAPESP) for financial support grant numbers2011/22018-3 and 2012/18304-3, respectively.

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

40 50 60 70 80 90 100

Str

atig

raph

ic th

ickn

ess

(msl

)

CaCO3

(%)

(1) (2) (3) (4) (5) (6) (7)

40.1

3040

.145

P12

E11

P13

P14

E12

E13

NP

16C

P14

aN

P17

CP

14b

1

2

3

4

5

a b c

marlscalcareous marl

lithology:

marly limestonelimestone

C18

rC

18n

0 1 2 3 4 5

ARM (x10

-6 Am

2/kg)

0 6 12 18 24 30

magnetofossil production- +ARM CFB

(x10-6 Am2/kg)

0 0.5 1 1.5 2

HIRM (x10

-5 Am

2/kg)

0 6 12 18

eolian dust- +

(x10-6 Am2 /kg) HIRM CFB

0 10 20 30 40 50

Eutrophic species (%)

20 30 40 50 60

Oligotrophic species(%)

d

Fig. 8. Productivity proxies based on selected calcareous nannofossil genera and magnetic parameters from theMCA section. (a) CaCO3, (b) ARM and ARM CFB, (c) HIRM and HIRM CFB,and (d) eutrophic and oligotrophic nannofossil taxa. The increase in eutrophic taxa coincides with an increase in ARM and HIRM during the MECO interval. These results suggest anincrease in iron fertilization and high primary productivity during MECO. Magnetobiochronostratigraphy and shaded areas are in Fig. 2.

Fig. 7. FORCdiagrams for ten samples from theMCA section (within and outside theMECOevent) andone sample from theContessaHighway section (during theMECOevent).Within theMECO event, FORC diagrams have sharp ridges that are indicative of non-interacting SD particles (Roberts et al., 2000; Egli et al., 2010). Samples 58.15 and 66.30 do not have the samebehaviour. FORC diagrams in (a) to (i)weremeasuredwith a single run in a VSM,whereas those in (j) and (k)were obtained after stacking ninemeasurement runs using amore sensitiveAGM instrument. FORC diagrams obtained with multiple runs were stacked and calculated using theMATLAB routine of Heslop and Roberts (2012). The smoothing factor (Roberts et al.,2000) is 4 in all cases.

43J.F. Savian et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 414 (2014) 32–45

http://dx.doi.org/10.1016/j.palaeo.2014.08.009http://dx.doi.org/10.1016/j.palaeo.2014.08.009

References

Abrajevitch, A., Kodama, K., 2009. Biochemical vs. detrital mechanism of remanence ac-quisition inmarine carbonates: a lesson from the K–T boundary interval. Earth Planet.Sci. Lett. 286, 269–277.

Agnini, C., Fornaciari, E., Giusberti, L., Grandesso, P., Lanci, L., Luciani, V., Muttoni, G., Pälike,H., Rio, D., Spofforth, D.J.A., Stefani, C., 2011. Integrated biomagnetostratigraphy of theAlano section (NE Italy): a proposal for defining the middle-late Eocene boundary.Geol. Soc. Am. Bull. 123, 841–872.

Aubry, M.-P., 1992. Late Paleogene calcareous nannoplankton evolution: a tale of climaticdeterioration. In: Prothero, D.R., Berggren, W.A. (Eds.), Eocene–Oligocene Climaticand Biotic Evolution. Princeton University Press, Princeton, New Jersey, pp. 272–309.

Backman, J., Hermelin, J.O.R., 1986. Morphometry of the Eocene nannofossilReticulofenestra umbilicus lineage and its biochronological consequences. Palaeogeogr.Palaeoclimatol. Palaeoecol. 57, 103–116.

Barrera, E., Huber, B.T., 1993. Eocene and Oligocene oceanography and temperatures inthe Antarctic Indian Ocean. In: Kennett, J.P., Warnke, D.A. (Eds.), The AntarcticPaleoenvironment: A Perspective on Global Change. AGU Ant. Res. Series. 60, pp.49–65.

Bazylinski, D.A., Frankel, R.B., 2003. Biologically controlled mineralization in prokaryotes.In: Weiner, S., Dove, P.M. (Eds.), BiomineralizationRev. Mineral. 54. MineralogicalSociety of America, Washington, DC, pp. 95–114.

Bazylinski, D.A., Frankel, R.B., 2004. Magnetosome formation in prokaryotes. Nat. Rev.Microbiol. 2, 217–230.

Berggren, W.A., Pearson, P.N., 2005. A revised tropical to subtropical Paleogene planktonicforaminiferal zonation. J. Foraminifer. Res. 35, 279–298.

Berggren, W.A., Kent, D.V., Swisher, C.C., Aubry, M.P., 1995. A revised Cenozoic geochro-nology and chronostratigraphy. In: Berggren, W.A., Kent, D.V., Aubry, M.P.,Hardenbol, J. (Eds.), Geochronology, Time Scales and Global Stratigraphic Correlation.SEPM Special Publication. 54, pp. 129–212.

Bijl, P.K., Houben, A.J.P., Schouten, S., Bohaty, S.M., Sluijs, A., Reichart, G.-J., Damsté, J.S.S.,Brinkhuis, H., 2010. Transient middle Eocene atmospheric CO2 and temperature var-iations. Science 330, 819–821.

Biscaye, P.E., 1965. Mineralogy and sedimentation of recent deep-sea clay in the AtlanticOcean and adjacent seas and oceans. Geol. Soc. Am. Bull. 76, 803–832.

Blakemore, R., 1975. Magnetotactic bacteria. Science 190, 377–379.Bloemendal, J., King, J.W., Hall, F.R., Doh, S.-J., 1992. Rock magnetism of late Neogene and

Pleistocene deep-sea sediments: relationship to sediment source, diagenetic process-es, and sediment lithology. J. Geophys. Res. 97, 4361–4375. http://dx.doi.org/10.1029/91JB03068.

Bohaty, S.M., Zachos, J.C., 2003. Significant Southern Ocean warming event in the latemiddle Eocene. Geology 31, 1017–1020.

Bohaty, S.M., Zachos, J.C., Florindo, F., Delaney, M.L., 2009. Coupled greenhouse warmingand deep-sea acidification in the middle Eocene. Paleoceanography 24, PA2207.http://dx.doi.org/10.1029/2008PA001676.

Bosboom, R.E., Abels, H.A., Hoorn, C., van den Berg, B.C.J., Guo, Z.J., Dupont-Nivet, G., 2014.Aridification in continental Asia after the Middle Eocene Climatic Optimum (MECO).Earth Planet. Sci. Lett. 389, 34–42.

Bown, P.R., Young, J.R., 1998. Techniques. In: Bown, P.R. (Ed.), Calcareous NannofossilBiostratigraphy. Kluwer Academic Publications, Dordrecht, The Netherlands, pp.16–28.

Bralower, T.J., 2002a. Evidence of surface water oligotrophy during the Paleocene–Eocenethermal maximum: nannofossil assemblage data from Ocean Drilling Program Site690, Maud Rise: Weddell Sea. Paleoceanography 17. http://dx.doi.org/10.1029/2001PA000662.

Bralower, T.J., 2002b. Correction to “Evidence of surface water oligotrophy during thePaleocene–Eocene thermal maximum: nannofossil assemblage data from OceanDrilling Program Site 690, Maud Rise, Weddell Sea” by Timothy J. Bralower.Paleoceanography 17, 1060. http://dx.doi.org/10.1029/2002PA000832.

Broecker, W.S., Clark, E., 1999. CaCO3 size distribution: a paleocarbonate ion proxy?Paleoceanography 14, 596–604. http://dx.doi.org/10.1029/1999PA900016.

Broecker, W.S., Clark, E., 2001. Re-evaluation of the CaCO3 size index paleocarbonate ionproxy. Paleoceanography 16, 669–671. http://dx.doi.org/10.1029/2001PA000660.

Brown, G., Brindley, G.W., 1980. X-ray diffraction procedures for clay mineral identifica-tion. In: Brindley, G.W., Brown, G. (Eds.), Crystal Structures of Clay Minerals andTheir X-ray Identification. Mineral. Soc. Monogr. 5, pp. 305–359

Cande, S.C., Kent, D.V., 1995. Revised calibration of the geomagnetic polarity timescale forthe Late Cretaceous and Cenozoic. J. Geophys. Res. 100, 6093–6095. http://dx.doi.org/10.1029/94JB03098.

Chang, L., Roberts, A.P., Williams, W., Fitz Gerald, J.D., Larrasoaña, J.C., Jovane, L.,Muxworthy, A.R., 2012. Giant magnetofossils and hyperthermal events. Earth Planet.Sci. Lett. 351–352, 258–269.

Channell, J.E.T., Ohneiser, C., Yamamoto, Y., Kesler, M.S., 2013. Oligocene–Miocenemagnetic stratigraphy carried by biogenic magnetite at sites U1334 and U1335(equatorial Pacific Ocean). Geochem. Geophys. Geosyst. 14, 265–282. http://dx.doi.org/10.1029/2012GC004429.

Chen, A.P., Egli, R., Moskowitz, B.M., 2007. First-order reversal curve (FORC) diagrams ofnatural and cultured biogenic magnetic particles. J. Geophys. Res. 112, B08S90.http://dx.doi.org/10.1029/2006JB004575.

Chung, F.H., 1974. Quantitative interpretation of X-ray diffraction patterns. I. Matrix flush-ing method of quantitativemulticomponent analyses. J. Appl. Crystallogr. 7, 519–525.

Coccioni, R., Marsili, A., Montanari, A., Bellanca, A., Neri, R., Bice, D.M., Brinkhuis, H.,Church, N., Macalady, A., McDaniel, A., Deino, A., Lirer, F., Sprovieri, M., Maiorano, P.,Monechi, P., Nini, C., Nocchi, M., Pross, J., Rochette, P., Sagnotti, L., Tateo, F.,Touchard, Y., Van Simaeys, S., Williams, G.L., 2008. Integrated stratigraphy of the Ol-igocene pelagic sequence in the Umbria-Marche basin (northeastern Apennines,

Italy): a potential Global Stratotype Section and Point (GSSP) for the Rupelian/Chattian boundary. Geol. Soc. Am. Bull. 120, 487–511.

Coccioni, R., Sideri, M., Bancalà, G., Catanzariti, R., Frontalini, F., Jovane, L., Montanari, A.,Savian, J., 2013. Integrated stratigraphy (magneto-, bio-and chronostratigraphy)and geochronology of the Palaeogene pelagic succession of the Umbria-MarcheBasin (central Italy). In: Jovane, L., Herrero-Bervera, E., Hinnov, L.A., Housen, B.A.(Eds.), Magnetic Methods and the Timing of Geological Processes. Geol. Soc. Lond.Spec. Publ. 373, pp. 111–131. http://dx.doi.org/10.1144/SP373.4

Colosimo, A.B., Bralower, T.J., Zachos, J.C., 2006. Evidence of lysocline shoaling at thePaleocene/Eocene Thermal Maximum on Shatsky Rise, northwest Pacific. In:Bralower, T.J., et al. (Eds.), Proc. Ocean Drilling Program, Sci. Res. 198. College Station,Texas, pp. 1–36

Day, R., Fuller, M., Schmidt, V.A., 1977. Hysteresis properties of titanomagnetites: grainsize and composition dependence. Phys. Earth Planet. Inter. 13, 260–266.

Dunlop, D.J., 2002. Theory and application of the Day plot (Mrs/Ms versus Hcr/Hc) 1. The-oretical curves and tests using titanomagnetite data. J. Geophys. Res. 107 (B3), 2056.http://dx.doi.org/10.1029/2001JB000486.

Edgar, K.M., Wilson, P.A., Sexton, P.F., Gibbs, S.J., Roberts, A.P., Norris, R.D., 2010. New bio-stratigraphy, magnetostratigraphy and isotopic insights into the Middle Eocene Cli-matic Optimum in low latitudes. Palaeogeogr. Palaeoclimatol. Palaeoecol. 297,670–682.

Egli, R., 2004. Characterization of individual rock magnetic components by analysis ofremanence curves, 1. Unmixing natural sediments. Stud. Geophys. Geod. 48,391–446.

Egli, R., Chen, A.P., Winklhofer, M., Kodama, K.P., Horng, C.S., 2010. Detection of noninter-acting single domain particles using first-order reversal curve diagrams. Geochem.Geophys. Geosyst. 11, Q01Z11. http://dx.doi.org/10.1029/2009GC002916.

Evans, M.E., Heller, F., 2003. Environmental magnetism. Principles and Applications ofEnviromagneticsAcademic Press, (299 pp.).

Faivre, D., Schüler, D., 2008. Magnetotactic bacteria and magnetosomes. Chem. Rev. 108,4875–4898.

Florindo, F., Farmer, R.K., Harwood, D.M., Cody, R.D., Levy, R., Bohaty, S.M., Carter, L.,Winkler, A., 2013. Palaeomagnetism of sediments from Southern Ocean ODP Site744 (southern Kerguelen Plateau): implications for early-to-middle Miocene climatein Antarctica. Global Planet. Chang. 110, 434–454. http://dx.doi.org/10.1016/j.gloplacha.2013.05.004.

Fornaciari, E., Agnini, C., Catanzariti, R., Rio, D., Bolla, E.M., Valvasoni, E., 2010. Mid-Latitude calcareous nannofossil biostratigraphy and biochronology across the middleto late Eocene transition. Stratigraphy 7, 229–264.

Gardin, S., Monechi, S., 1998. Palaeoecological change in middle to low latitude cal-careous nannoplankton at the Cretaceous/Tertiary boundary. Bull. Soc. Geol. Fr.169, 709–723.

Garming, J.F.L., Bleil, U., Riedinger, N., 2005. Alteration of magnetic mineralogy at the sul-fate–methane transition: analysis of sediments from the Argentine continental slope.Phys. Earth Planet. Inter. 151, 290–308.

Gibbs, S.J., Bralower, T.J., Bown, P.R., Zachos, J.C., Bybell, L., 2006. Shelf-open ocean calcar-eous phytoplankton assemblages across the Paleocene–Eocene Thermal Maximum:implications for global productivity gradients. Geology 34, 233–236.

Guerrera, F., Monaco, P., Nocchi, M., Parisi, G., Franchi, R., Vannucci, S., Giovannini, G.,1988. La Scaglia Variegata Eocenica nella sezione di Monte Cagnero (bacinoMarchigiano interno): studio litostratigrafico, petrografico e biostratigrafico. Bull.Soc. Geol. Italy 107, 81–99.

Hancock, H.J.L., Dickens, G.R., 2005. Carbonate dissolution episodes in Paleocene andEocene sediment, Shatsky Rise, West-Central Pacific. Proc. Ocean Drilling Program,Sci. Res. 198. College Station, Texas, pp. 1–58

Handley, L., O'Halloran, A., Pearson, P.N., Hawkins, E., Nicholas, C.J., Schouten, S., McMillan,I.K., Pancost, R.D., 2012. Changes in the hydrological cycle in tropical East Africaduring the Paleocene–Eocene thermal maximum. Palaeogeogr. Palaeoclimatol.Palaeoecol. 329–330, 10–21.

Haq, B.U., Lohmann, G.P., 1976. Early Cenozoic calcareous nannoplankton biogeography ofthe Atlantic Ocean. Mar. Micropaleontol. 1, 119–194.

Heslop, D., Roberts, A.P., 2012. Estimation of significance levels and confidence intervalsfor first-order reversal curve distributions. Geochem. Geophys. Geosyst. 13, Q12Z40.http://dx.doi.org/10.1029/2012GC004115.

Heslop, D., Dekkers, M.J., Kruiver, P.P., van Oorschot, I.H.M., 2002. Analysis of isothermalremanent magnetisation acquisition curves using the expectation–maximizationalgorithm. Geophys. J. Int. 148, 58–64.

Hesse, P.P., 1994. Evidence for bacterial palaeoecological origin of mineral magnetic cyclesin oxic and sub-oxic Tasman Sea sediments. Mar. Geol. 117, 1–17.

Housen, B.A., Moskowitz, B.M., 2006. Depth distribution of magnetofossils in near surfacesediments from the Blake/Bahama Outer Ridge, western North Atlantic Ocean, deter-mined by low-temperature magnetism. J. Geophys. Res. 111, G01005. http://dx.doi.org/10.1029/2005JG000068.

Hyland, E., Murphy, B., Varela, P., Marks, K., Colwell, L., Tori, F., Monechi, S., Cleveland, L.,Brinkhuis, H., vanMourik, C.A., Coccioni, R., Bice, D.,Montanari, A., 2009. Integrated strat-igraphic and astrochronologic calibration of the Eocene–Oligocene transition in theMonte Cagnero section (northeastern Apennines, Italy): a potential parastratotype forthe Massignano global stratotype section and point (GSSP). In: Koeberl, C., Montanari,A. (Eds.), The Late Eocene Earth-Hothouse, Icehouse, and Impacts. Geol. Soc. Am. Spec.Paper. 452, pp. 303–322.

Jovane, L., Florindo, F., Coccioni, R., Dinarès-Turell, J., Marsili, A., Monechi, S., Roberts, A.P.,Sprovieri, M., 2007. Themiddle Eocene climatic optimum (MECO) event in the ContessaHighway section, Umbrian Apennines, Italy. Geol. Soc. Am. Bull. 119, 413–427.

Jovane, L., Florindo, F., Bazylinski, D.A., Lins, U., 2012. Prismatic magnetite magnetosomesfrom cultivatedMagnetovibrio blakemorei strainMV-1: amagnetic fingerprint inmarinesediments? Environ. Microbiol. Rep. 4, 664–668. http://dx.doi.org/10.1111/1758-2229.

44 J.F. Savian et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 414 (2014) 32–45

http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0495http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0495http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0495http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0005http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0005http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0005http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0010http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0010http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0010http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0015http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0015http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0015http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0500http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0500http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0500http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0500http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0505http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0505http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0505http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0025http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0025http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0030http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0030http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0510http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0510http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0510http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0510http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0035http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0035http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0035http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0095http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0095http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0040http://dx.doi.org/10.1029/91JB03068http://dx.doi.org/10.1029/91JB03068http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0050http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0050http://dx.doi.org/10.1029/2008PA001676http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0060http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0060http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0070http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0070http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0070http://dx.doi.org/10.1029/2001PA000662http://dx.doi.org/10.1029/2001PA000662http://dx.doi.org/10.1029/2002PA000832http://dx.doi.org/10.1029/1999PA900016http://dx.doi.org/10.1029/2001PA000660http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0520http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0520http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0520http://dx.doi.org/10.1029/94JB03098http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0110http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0110http://dx.doi.org/10.1029/2012GC004429http://dx.doi.org/10.1029/2006JB004575http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0115http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0115http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0120http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0120http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0120http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0120http://dx.doi.org/10.1144/SP373.4http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0535http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0535http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0535http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0535http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0130http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0130http://dx.doi.org/10.1029/2001JB000486http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0135http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0135http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0135http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0135http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0140http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0140http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0140http://dx.doi.org/10.1029/2009GC002916http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0550http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0550http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0150http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0150http://dx.doi.org/10.1016/j.gloplacha.2013.05.004http://dx.doi.org/10.1016/j.gloplacha.2013.05.004http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0160http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0160http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0160http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0165http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0165http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0165http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0170http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0170http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0170http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0175http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0175http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0175http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0180http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0180http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0180http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0555http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0555http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0555http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0185http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0185http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0185http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0190http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0190http://dx.doi.org/10.1029/2012GC004115http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0195http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0195http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0195http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0200http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0200http://dx.doi.org/10.1029/2005JG000068http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0565http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0565http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0565http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0565http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0565http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0565http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0220http://refhub.elsevier.com/S0031-0182(14)00406-4/rf0220http://dx.doi.org/10.1111/1758-2229

Jovane, L., Savian, J.F., Coccioni, R., Frontalini, F., Bancalà, G., Catanzariti, R., Luciani, V., Bohaty, S.M.,Wilson, P.A., Florindo, F., 2013. Integratedmagnetobiostratigraphy of themiddle Eocene–Ol-igocene interval from the Monte Cagnero section, central Italy. In: Jovane, L., Herrero-Bervera, E., Hinnov, L.A., Housen, B.A. (Eds.),MagneticMethods and the Timing of GeologicalProcesses. Geol. Soc. Lond. Spec. Publ. 373, pp. 79–95. http://dx.doi.org/10.1144/SP373.13

Kopp, R.E., Kirschvink, J.L., 2008. The identification and biogeochemical interpretation offossil magnetotactic bacteria. Earth-Sci. Rev. 86, 42–61.

Kraus, M.J., Riggins, S., 2007. Transient drying during the Paleocene–Eocene ThermalMaximum (PETM): analysis of paleosols in the Bighorn Basin,Wyoming. Palaeogeogr.Palaeoclimatol. Palaeoecol. 245, 444–461.

Kruiver, P.P., Dekkers, M.J., Heslop, D., 2001. Quantification of magnetic coercivity compo-nents by the analysis of acquisition curves of isothermal remanent magnetization.Earth Planet. Sci. Lett. 189, 269–276.

Larrasoaña, J.C., Roberts, A.P., Rohling, E.J., Winklhofer, M.,Wehausen, R., 2003. Threemillionyears of monsoon variability over the northern Sahara. Climate Dynamics 21, 689–698.

Larrasoaña, J.C., Roberts, A.P., Chang, L., Schellenberg, S.A., Fitz Gerald, J.D., Norris, R.D.,Zachos, J.C., 2012. Magnetotactic bacterial response to Antarctic dust supply duringthe Paleocene–Eocene thermal maximum. Earth Planet. Sci. Lett. 333–334, 122–133.

Lean, C.M.B., McCave, I.N., 1998. Glacial to interglacial mineral magnetic andpalaeoceanographic changes at Chatham Rise, SW Pacific Ocean. Earth Planet. Sci.Lett. 163, 247–260.

Leon-Rodriguez, L., Dickens, G.R., 2010. Constraints on ocean acidification associatedwith rapidand massive carbon injections: the early Paleogene record at Ocean Drilling Program Site1215, equatorial Pacific Ocean. Palaeogeogr. Palaeoclimatol. Palaeoecol. 298, 409–420.

Lirer, F., 2000. A new technique for retrieving calcareous microfossils from lithified limedeposits. Micropaleontology 46, 365–369.

Liu, Q.S., Roberts, A.P., Torrent, J., Horng, C.S., Larrasoaña, J.C., 2007. What do the HIRM andS-ratio really measure in environmental magnetism? Geochem. Geophys. Geosyst. 8,Q09011. http://dx.doi.org/10.1029/2007GC001717.

Liu, Q., Roberts, A.P., Larrasoaña, J.C., Banerjee, S.K., Guyodo, Y., Tauxe, L., Oldfield, F., 2012.Environmental magnetism: principles and applications. Rev. Geophys. 50, RG4002.http://dx.doi.org/10.1029/2012RG000393.

Luciani, V., Giusbert, L., Agnini, C., Fornaciari, E., Rio, D., Spofforth, D.J.A., Pälike, H., 2010.Ecological and evolutionary response of Tethyan planktonic foraminifera to the mid-dle Eocene climatic optimum (MECO) from the Alano section (NE Italy). Palaeogeogr.Palaeoclimatol. Palaeoecol. 292, 82–95.

Maher, B.A., Prospero, J.M., Mackie, D., Gaiero, D., Hesse, P.P., Balkanski, Y., 2010. Globalconnections between aeolian dust, climate and ocean biogeochemistry at the presentday and at the Last Glacial Maximum. Earth-Sci. Rev. 99, 61–97.

Martini, E., 1971. Standard tertiary and quaternary calcareous nannoplankton zonation.In: Farinacci, A. (Ed.), Proc. Second Planktonic Conf., Rome 1970. Tecnoscienza. 2,pp. 739–785.

McNeill, D.F., 1990. Biogenic magnetite from surface Holocene carbonate sediments,Great Bahama Bank. J. Geophys. Res. 95, 4363–4372. http://dx.doi.org/10.1029/JB095iB04p04363.

Moore, D.M., Reynolds Jr., R.C., 1997. X-Ray Diffraction and the Identification and Analysisof Clay Minerals. Oxford University Press, Oxford, (378 pp.).

Moskowitz, B.M., Bazylinski, D.A., Egli, R., Frankel, R.B., Edwards, K.J., 2008.Magnetic prop-erties of marine magnetotactic bacteria in a seasonally stratified coastal pond (SaltPond, MA, USA). Geophys. J. Int. 174, 75–92.

Ogg, J.G., 2012. Geomagnetic polarity time scale. In: Gradstein, F.M., Ogg, J.G., Smith, A.G.(Eds.), A Geologic Time Scale. Cambridge University Press, Cambridge, pp. 85–113.

Ohneiser, C., Acton, G., Channell, J.E.T., Wilson, G.S., Yamamoto, Y., Yamazaki, T., 2013. AMiocene relative paleointensity record from the Equatorial Pacific. Earth Planet. Sci.Lett. 374, 227–238.

Okada, H., Bukry, D., 1980. Supplementary modification and introduction of code num-bers to the low-latitude coccolith biostratigraphic zonation (Bukry, 1973, 1975).Mar. Micropaleontol. 5, 321–325.

Pälike, H., et al., 2012. A Cenozoic record of the equatorial Pacific carbonate compensationdepth. Nature 488, 609–615.

Parisi, G., Guerrera, F., Madile, M., Magnoni, G., Monaco, P., Monechi, S., Nocchi, M., 1988.Middle Eocene to early Oligocene calcareous nannofossil and foraminiferal biostratig-raphy at Monte Cagnero section, Piobbico (Italy). In: Premoli Silva, I., Coccioni, R.,Montanari, A. (Eds.), The Eocene/Oligocene boundary in the Marche-Umbria basin(Italy). International Subcommission on Paleogene Stratigraphy, Ancona, pp.119–135.

Perch–Nielsen, K., 1985. Cenozoic calcareous nannofossils. In: Bolli, H.M., Saunders, J.B., Perch–Nielsen, K. (Eds.), Plankton Stratigraphy. Cambridge University Press, UK, pp. 427–554.

Persico, D., Villa, G., 2004. Eocene–Oligocene calcareous nannofossils fromMaud Rise andKerguelen Plateau (Antarctica): paleoecological and paleoceanographic implications.Mar. Micropaleontol. 52, 153–179.

Petermann, H., Bleil, U., 1993. Detection of live magnetotactic bacteria in South Atlanticdeep-sea sediments. Earth Planet. Sci. Lett. 117, 223–228.

Petersen, N., von Dobeneck, T., Vali, H., 1986. Fossil bacterial magnetite in deep-sea sedi-ments from the South Atlantic Ocean. Nature 320, 611–614.

Poulton, S.W., Krom, M.D., Raiswell, R., 2004. A revised scheme for the reactivity of iron(oxyhydr)oxide minerals towards dissolved sulfide. Geochim. Cosmochim. Acta 68,3703–3715.

Premoli Silva, I., Jenkins, D.G., 1993. Decision on the Eocene–Oligocene boundarystratotype. Episodes 16, 379–382.

Raffi, I., Backman, J., Zachos, J.C., Sluijs, A., 2009. The response of calcareous nannofossil as-semblages to the Paleocene Eocene Thermal Maximum at the Walvis Ridge in theSouth Atlantic. Mar. Micropaleontol. 70, 201–212.

Rea, D.K., 1994. The paleoclimatic record provided by eolian deposition in the deep sea:the geologic history of wind. Rev. Geophys. 32, 159–195. http://dx.doi.org/10.1029/93RG03257.

Roberts, A.P., Cui, Y.L., Verosub, K.L., 1995. Wasp-waisted hysteresis loops: mineral mag-netic characteristics and discrimination of components in mixed magnetic system.J. Geophys. Res. 100, 17909–17924.

Roberts, A.P., Pike, C.R., Verosub, K.L., 2000. First-order reversal curve diagrams: a newtool for characterizing the magnetic properties of natural samples. J. Geophys. Res.105, 28461–28475. http://dx.doi.org/10.1029/2000JB900326.

Roberts, A.P., Florindo, F., Villa, G., Chang, L., Jovane, L., Bohaty, S.M., Larrasoaña, J.C.,Heslop, D., Fitz Gerald, J.D., 2011. Magnetotactic bacterial abundance in pelagic ma-rine environments is limited by organic carbon flux and availability of dissolvediron. Earth Planet. Sci. Lett. 310, 441–452.

Roberts, A.P., Chang, L., Heslop, D., Florindo, F., Larrasoaña, J.C., 2012. Searching for singledomain magnetite in the “pseudo-single-domain” sedimentary haystack: implica-tions of biogenic magnetite preservation for sediment magnetism and relativepaleointensity determinations. J. Geophys. Res. 117, B08104. http://dx.doi.org/10.1029/2012JB009412.

Robertson, D.J., France, D.E., 1994. Discrimination of remanence-carrying minerals inmixtures, using isothermal remanent magnetisation acquisition curves. Phys. EarthPlanet. Inter. 84, 223–234