Earth-Science Reviews 127 (2013) 111–139

Contents lists available at ScienceDirect

Earth-Science Reviews

j ourna l homepage: www.e lsev ie r .com/ locate /earsc i rev

Magnetic properties of pelagic marine carbonates

Andrew P. Roberts a,⁎, Fabio Florindo b, Liao Chang a, David Heslop a,Luigi Jovane c, Juan C. Larrasoaña d

a Research School of Earth Sciences, The Australian National University, Canberra, ACT 0200, Australiab Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata, 605, I-00143 Rome, Italyc Departamento de Oceanografia Física, Instituto Oceanográfico, Universidade de São Paulo, Praça do Oceanográfico, 191, 05508-120 São Paulo, Brazild Instituto Geológico y Minero de España, Unidad de Zaragoza, C/Manuel Lasala 44, 9B, Zaragoza 50006, Spain

⁎ Corresponding author. Tel.: +61 2 61253887; fax: +E-mail address: andrew.roberts@anu.edu.au (A.P. Rob

0012-8252/$ – see front matter © 2013 Elsevier B.V. All rihttp://dx.doi.org/10.1016/j.earscirev.2013.09.009

a b s t r a c t

a r t i c l e i n f o

Article history:Received 12 June 2013Accepted 28 September 2013Available online 10 October 2013

Keywords:Pelagic carbonateLimestoneMagnetic mineralsBiogenic magnetiteMagnetofossilsDiagenesisRemagnetisation

Pelagic carbonates are deposited far from continents, usually at water depths of 3000–6000 m, at rates below10cm/kyr, and are a globally important sediment type. Recent advances, with recognition of widespread preser-vation of biogenicmagnetite (the inorganic remains ofmagnetotactic bacteria), have fundamentally changed ourunderstanding of the magnetic properties of pelagic carbonates. We review evidence for the magnetic mineralstypically preserved in pelagic carbonates, the effects of magnetic mineral diagenesis on paleomagnetic and envi-ronmental magnetic records of pelagic carbonates, and what magnetic properties can tell us about the open-ocean environments in which pelagic carbonates are deposited. We also discuss briefly late diageneticremagnetisations recorded by some carbonates. Despite recent advances in our knowledge of these phenomena,much remains undiscovered. We are only at early stages of understanding how biogenic magnetite gives rise topaleomagnetic signals in sediments and whether it carries a poorly understood biogeochemical remanentmagnetisation. Recently developed techniques have potential for testing how different magnetotactic bacterialspecies, which produce different magnetite morphologies, respond to changing nutrient and oxygenation condi-tions. Future work needs to test whether it is possible to develop proxies for ancient nutrient conditions fromwell-calibrated modern magnetotactic bacterial occurrences. A tantalizing link between giant magnetofossilsand Paleogene hyperthermal events needs to be tested; much remains to be learned about the relationship be-tween climate and the organisms that biomineralised these large and novel magnetite morphologies. Ratherthanbeing awell-worn subject that has been studied for over 60years, themagnetic properties of pelagic carbon-ates hold many secrets that await discovery.

© 2013 Elsevier B.V. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1122. Formation, preservation and distribution of pelagic marine carbonates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1123. Paleomagnetic characteristics of pelagic marine carbonates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1144. Magnetic mineral diagenesis in pelagic marine environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1145. Magnetic properties of pelagic marine carbonates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119

5.1. Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1195.2. High-temperature magnetic measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1205.3. IRM acquisition, hysteresis and unmixing analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1215.4. FORC diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1225.5. Ferromagnetic resonance (FMR) spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1255.6. Low-temperature magnetic measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1255.7. Transmission electron microscope observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126

6. Origin of magnetic minerals in pelagic marine carbonates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1286.1. Biogenic magnetite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1286.2. Detrital magnetic minerals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1306.3. Authigenic magnetic minerals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1316.4. Exotic magnetic particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

61 23 61255105.erts).

ghts reserved.

112 A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

7. Paleomagnetic recording in pelagic marine carbonates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1328. Remagnetisations in pelagic marine carbonates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1339. Outstanding questions concerning the magnetisation of pelagic marine carbonates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

9.1. Are magnetotactic bacteria always gradient-organisms? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1339.2. Are biogeochemical remanent magnetisations globally important? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1349.3. At what depths do magnetotactic bacteria live? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1349.4. Can magnetofossils provide useful paleoproductivity or paleoenvironmental information? . . . . . . . . . . . . . . . . . . . . . . . 134

10. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

1. Introduction

Carbonate sediments are deposited on the seafloor of ~50% of theworld ocean (Fig. 1a). In addition, marine carbonates crop out on 10%of the global land surface (Blatt et al., 1980; Fig. 1a). The total world vol-ume of hydrocarbons hosted by carbonates has been estimated at 50%,and the ease with which fluids can flow and react with carbonates hasmade them a common host of ore deposits (Blatt et al., 1980). Muchof the carbon on Earth is stored in carbonate sediments, which makesthese sediments a crucial part of the global carbon cycle (e.g., Archeret al., 2000). Carbonate sediments are, therefore, a globally importantsediment type. In this paper, we focus on pelagic carbonates, whichare defined as those deposited far from the continents, usually atwater depths of 3000–6000 m, and at rates slower than 10 cm/kyr(Berner, 1980)with carbonate contents greater than 30%. Paleomagnet-ic studies have frequently targeted pelagic carbonate sediments, espe-cially since development of superconducting rock magnetometers(Goree and Fuller, 1976), because they often contain high-fidelityrecords of the ancient geomagnetic field. Classic studies of marine car-bonates demonstrated that the latitudinal distribution of ancient car-bonate rocks mirrors that of modern carbonates once corrected forpaleolatitude (e.g., Briden and Irving, 1964; Irving, 1964). They havealso been used to demonstrate the geocentric axial dipole (GAD) fieldhypothesis (e.g., Opdyke and Henry, 1969), which is a cornerstone ofpaleomagnetism. Extensive studies of Mesozoic and Cenozoic Tethyancarbonates have made fundamental contributions to tectonic recon-structions (e.g., Channell and Tarling, 1975; Lowrie and Alvarez, 1975;VandenBerg et al., 1978), magnetobio-chronology (e.g., Alvarez et al.,1977; Lowrie and Alvarez, 1977a; Roggenthen and Napoleone, 1977;Lowrie and Alvarez, 1981; Lowrie et al., 1982; Napoleone et al., 1983;Speranza et al., 2005; Jovane et al., 2007; Coccioni et al., 2008; Jovaneet al., 2009, 2013), and have provided evidence that led to the Creta-ceous–Tertiary boundary impact hypothesis (e.g., Lowrie and Alvarez,1977a; Alvarez et al., 1980). High-quality magnetic polarity records,whichmirror the expected polarity pattern of the geomagnetic polaritytimescale (GPTS), have also been reported from Cenozoic pelagic car-bonates in the Atlantic (e.g., Tauxe et al., 1983; Mead et al., 1986;Tauxe and Hartl, 1997; Channell et al., 2003; Edgar et al. 2010), Indian(e.g., Touchard et al., 2003; Savian et al., 2013), Pacific (e.g., Valet andMeynadier, 1993; Schneider et al., 1997; Lanci et al., 2004, 2005;Channell et al., 2013; Guidry et al., 2013; Yamazaki et al., 2013) andSouthern Oceans (e.g., Channell and Stoner, 2002; Roberts et al., 2003;Florindo and Roberts, 2005; Fuller et al., 2006; Florindo et al., in press).

Lowrie and Heller (1982) and Freeman (1986) reviewed the mag-netic properties of marine carbonates and outlined many detailsconcerning their magnetic stability, magnetic properties, and magneticmineralogy. Our understanding of the magnetic properties of pelagiccarbonate sediments has undergone a recent revolution (e.g., Robertset al., 2011, 2012; Larrasoaña et al., 2012; Yamazaki, 2012; Yamazakiand Ikehara, 2012; Channell et al., 2013) through use of techniquesthat enable better discrimination of the sources of fine-grainedmagnet-ic particles in sediments. In this paper, we summarise these develop-ments and document the magnetic properties observed in globally

distributed pelagic carbonates. Despite the substantial recent advancesin our understanding of the magnetic properties of marine carbonates,many issues remain unresolved.We outline some of these important is-sues in the hope that concerted research will be undertaken to resolvethese outstanding problems.

2. Formation, preservation and distribution of pelagicmarine carbonates

Carbonate formation in marine environments is controlled by watertemperature and concentration of dissolved CO2. Carbonates will onlyprecipitate in waters that are low in CO2; because CO2 is more solublein cooler waters, carbonate is more likely to form in tropical seas (±30° latitude). Likewise, carbonate precipitation is favoured thermody-namically at high temperatures and lowpressures,whichmakes low lat-itude surface waters an ideal environment for carbonate precipitation.However, even if carbonate precipitates in surface waters, it might notsurvive export from the photic zone to the seafloor. There is a well-known relationship between water depth and carbonate deposition;the so-called carbonate compensation depth (CCD) results from the ef-fect of pressure on calcite solubility (Archer, 1996), and is the depth atwhich the amount of CaCO3 delivered from above is equal to the amountremoved by dissolution. The lysocline occurs at shallower depths thanthe CCD and is the depth at which carbonate dissolution rapidly in-creases; it separates the upper waters in which planktonic calcareousskeletons are well preserved from the lower waters where they aremore poorly preserved. The position of the CCD is variable in spaceand time because of deep-sea acidification that results from re-mineralisation of organic matter as it settles from the photic zone. Thiseffect means that the Atlantic Ocean, which has recently ventilateddeep waters with high pH, supports carbonate preservation at greaterwater depths (average of ~5500m) than the Pacific Ocean (average of~4500m) where deep water is acidified by organic matter respiration(Archer, 1996). The global distribution of carbonate in seafloor sedi-ments (Fig. 1a) is, therefore, a function of water temperature, waterdepth (pressure), pH, and relative dilution by terrigenous sedimentcomponents (Archer et al., 2000).

Many carbonates form on shallow water platforms at low latitudes,wherewater depths are typically b20m.Modern shallowwater carbon-ates account for only 5% of the global distribution of carbonate sedimen-tation, while 95% occurs on oceanic slopes and on the deep seafloor(Blatt et al., 1980).Whilemany of themagnetic mineral types discussedin this paper also occur in shallow water carbonates (carbonate plat-forms, atolls or reefs), we focus explicitly on carbonate sediments de-posited in open-ocean pelagic environments because of their greatercomparative geographic and stratigraphic importance. Pelagic carbon-ate sediments are primarily oozes that consist of the remains of plank-tonic foraminifera and coccolithophores, although aragonite (frompteropods and heteropods),which ismore soluble than calcite, can con-tribute to the carbonate fraction of sediments deposited at interme-diate depths such as those near mid-ocean ridge crests. Calcareousnannofossils (produced by coccolithophores) have contributed to

90°N

60°N

30°N

0°

60°S

30°S

180°W 180°E0°90°W 90°E

100

80

60

40

20

0

(a)

(b)

(c)

90°N

60°N

30°N

0°

60°S

30°S

180°W 180°E0°90°W 90°E

Chlorophyll a concentration (mg/m )3

0.01 0.03 0.1 0.3 1 3 10

Wor

ld o

cean

CaC

O

(wei

ght %

)3

ODP 748

ODP 744ODP 738

ODP 758

ODP 711

ODP 1263

ODP 689ODP 690

Italian sections

180°W 180°E0°90°W 90°E

90°N

60°N

30°N

0°

60°S

30°S

Argolis, Greece

Fig. 1.Maps of global oceanfloor carbonate sedimentation and carbonate rock outcrops,marineproductivity and carbonate sequences fromwhich results are presented in this paper. (a)Worldocean CaCO3 distribution (contouredweight-% values) from surface sediments (data are from http://orfois.wdc-mare.org/Results/CaCO3.html). Carbonate rock outcrops on the continents areindicated in red (from theWorld Map of Carbonate Rock Outcrops v3.0 (Williams and Fong; http://web.env.auckland.ac.nz/our_research/karst/) and are based on sources acknowledged byWilliams and Ford (2006)). (b) Oceanmap of primary productivity, as measured by themean annual concentration of chlorophyll-a in phytoplankton (mg of chlorophyll per m3 of seawater)detected using SeaWiFS (Sea-viewingWide Field-of-view Sensor; courtesy of NASA) for the period from Sept. 4, 1998 to Dec. 11, 2010 (http://oceancolor.gsfc.nasa.gov/). (c) Locations of car-bonate sedimentary sequences for which results are presented in this paper.

113A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

114 A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

pelagic sedimentation since the Late Jurassic and planktonic forami-nifera have contributed since the Late Cretaceous.

Pelagic marine sedimentation is controlled by threemain factors: (i)distance from land (and, therefore, supply of terrigenous particles thatdilute the biogenic sediment fraction), (ii) water depth, as discussedabove, which affects preservation of not only biogenic carbonate butalso biogenic silica, and (iii) ocean fertility, which controls productionof biogenic particles in surface waters. The effects, and relative impor-tance, of these factors can be visually assessed in Fig. 1. Comparison ofmodern pelagic carbonate sedimentation patterns (Fig. 1a) with mod-ern productivity, indicated using an average (1998–2010) of globalocean fertility (Fig. 1b), reveals a correspondence between nutrient “de-serts” where primary productivity is low (darkest areas in Fig. 1b andlow CaCO3 contents in Fig. 1a). This pattern is disrupted, however,along the trace of the global mid-ocean ridge system, which representsthe shallowest parts of the deep ocean that lie well above the CCD andthat extend from the North Atlantic Ocean through the South Atlantic,into the Southern Ocean, where it bifurcates into two strands, one intothe Indian Ocean that terminates in the Red Sea, and the otherextending south of Australia, and into the Eastern Pacific Ocean whereit terminates at the Costa Rica Trench. Fig. 1a is a contour plot of theCaCO3 contents of surface sediment samples. These data represent afar greater time average than the 12years shown in Fig. 1b. Thus, eventhough the equatorial North and South Atlantic gyres are generallylow productivity regions (the so-called high-nitrate low-chlorophyll(HNLC) regions), nutrient limitations on phytoplankton productivitycan be overcome, for example, through iron fertilisation of the surfaceocean (Martin et al., 1991) by Saharan and Sahelian dust inputs (e.g.,Sarthou et al., 2003; Moore et al., 2009), which can give rise to pulsesof higher productivity. Carbonate produced during these higher produc-tivity episodes can then be exported to the deep-sea and preserved,particularly at relatively shallower water depths on either side of theMid-Atlantic Ridge (Fig. 1a). Much coring of deep-sea carbonate sedi-ments is aimed at understanding details of the interplay betweenclimate, terrigenous sediment supply, nutrient availability, primaryproductivity, the biogeochemical relationships between these factors,and their temporal variations. In contrast, the nutrient deserts in theNorth and South Pacific Ocean mainly lie below the CCD and onlysmall concentrations of carbonate are preserved (Fig. 1a), exceptaround areas with elevated bathymetry (islands, seamounts, and theEast Pacific Rise). Pelagic sedimentation in these deep Pacific regionsis dominated by slowly deposited red clays, which result from lack ofcarbonate preservation and long-distance transportation and deposi-tion of eolian dust. At higher latitudes, red clays give way to depositionof siliceous oozes.

3. Paleomagnetic characteristics of pelagic marine carbonates

Pelagic marine carbonates have provided outstanding records ofgeomagnetic field variability that have been used widely in tectonicsand geochronological studies (e.g., Fig. 2). Lowrie et al. (1980) reviewedthe history of how Tethyan Cretaceous pelagic carbonates were used tocalibrate the GPTS, to date biostratigraphic events, to determine an ap-parent polar wander path for Umbria, and to assess local vertical-axistectonic rotations. Since that time, pelagic carbonates have continuedto provide high-fidelity records of geomagnetic polarity history thathave been used for a range of purposes, but largely for stratigraphic dat-ing (e.g., Tauxe et al., 1983; Mead et al., 1986; Channell et al., 2003;Roberts et al., 2003; Jovane et al., 2004; Florindo and Roberts, 2005;Jovane et al., 2007; Coccioni et al., 2008; Channell et al., 2013; Jovaneet al., 2013; Florindo et al., in press). In general, pelagic carbonates areweakly magnetised, which is why they did not become a routine targetfor paleomagnetic investigations until superconducting rock magne-tometers became available (cf. Goree and Fuller, 1976). Despite theirgenerally weak magnetisations, pelagic carbonates often have high pa-leomagnetic stability, which is responsible for the superb magnetic

polarity records obtained from such sediments (e.g., Fig. 2). Thesehigh-quality magnetostratigraphic records usually have a square-waverecord of polarity changes, with inclinations that are consistent withthose expected for a GAD field at the site latitude (Figs. 2, 3). The clearlybimodal distributions of paleomagnetic inclinations (corresponding tonormal and reversed polarity; Fig. 3) from continuously deposited sed-iments assists with unambiguous correlation to the GPTS, which en-ables calibration of biostratigraphic datums (first and last occurrencesfor key taxa; Fig. 2) and underpins biostratigraphic dating.

The generally exceptional paleomagnetic stability of pelagic carbon-ates is evident in vector demagnetisation diagrams (Fig. 4). These dia-grams are used to determine the characteristic remanent magnetisation(ChRM) that defines the most stable paleomagnetic component isolatedduring stepwise demagnetisation (Zijderveld, 1967). The ChRM direc-tion is usually isolated after removal of a small secondary componentduring progressive alternating field (AF; Fig. 4a-e) or thermal (Fig. 4f,g) demagnetisation. The paleomagnetic direction associated with theChRM is usually determined using principal component analysis(Kirschvink, 1980), where a low value of the maximum angular devia-tion associated with the fit to this component (b2°) is indicative of anexceptionally well-defined ChRM vector. In thermal demagnetisationresults, the ChRM generally persists to temperatures between 550 and600 °C (Fig. 4f), which indicates the dominant presence of magnetite.In other cases, a magnetisation persists to 600–700 °C (Fig. 4g), whichis indicative of the presence of hematite. Likewise, nearly complete AFdemagnetisation at peakAFs of 70–100mT indicates thatmagneticmin-eral assemblages are dominated by a low coercivity magnetic mineralsuch asmagnetite (Fig. 4a–e), although some samples have a high coer-civity component that remains undemagnetised at 100 mT (Fig. 4h).Persistence of a magnetisation above 580 °C in a sister sample(Fig. 4g) suggests that this high coercivity component is due to hema-tite. These results provide useful insights into the magnetic mineralogyof pelagic carbonates, which is discussed in greater detail in Sections 5and 6.

In addition to providing information about magnetic mineralogy, AFdemagnetisation results have been used to provide information aboutthe domain state (i.e., particle size) of magnetic minerals. Lowrie andHeller (1982) presented results of the modified Lowrie–Fuller test(Lowrie and Fuller, 1971; cf. Johnson et al., 1975) for a range of tecton-ically upliftedMesozoic Europeanand southern Tethyanpelagic carbon-ates and argued that the magnetisations of these sediments aredominated by single domain (SD) magnetic behaviour for samplesthat containmagnetite only and, in some cases,magnetite andhematite.We present results below (Section 5) from a range of more modernmethods that provide definitive evidence concerning the domain stateof magnetic minerals in pelagic carbonates. Nevertheless, the resultsof Lowrie and Heller (1982) provided an important justification forthe long-term geological stability of magnetisations carried by pelagiccarbonates.

While many pelagic carbonates are paleomagnetically stable, this iscertainly not universally the case. It is possible that no stable paleomag-netic direction is preserved (Fig. 4i) if the concentration of magneticminerals is low or if the magnetic mineral assemblage has undergonediagenetic dissolution (see Section 4). Pelagic carbonates can also beremagnetised (see Section 8). We first discuss how organic matter dia-genesis affects paleomagnetic recording in pelagic marine environ-ments, and then document the magnetic minerals that are typicallypreserved in pelagic carbonates that have not undergone extensive dia-genetic modification.

4. Magnetic mineral diagenesis in pelagic marine environments

Diagenesis associated with post-depositional degradation of organicmatter is a fundamentally important process for determining the long-term fate of magnetic minerals in sedimentary environments. Deep-sea sedimentary organic matter accumulates largely through export,

PolarityLithology VGP latitudeC

ism

on

90°S 90°N0

Gub

bio

marl

limestone

100

m

Paleontological zonesPlanktonicForaminifera

Calcareousnannofossils

Globigerinaeugubina

Abathomphalusmayaroensis

Glt.gansseri

Glt.confusa

Glt.tricarinata

Glt.calcarata

Glt.elevata

Glt.concavatacarinata

Glt.concavataconcavata

Glt.schneegansi

Glt.helveticaH.lehmanniRtl.cushmani

not zoned

Planomalinabuxtorti

T. breggiensisG. ferreolensis

Schk. cabriG. blowiH. similis

Paleocene

Maa

stric

htia

nC

ampa

nian

San

toni

anC

onia

cian

Tur

onia

nC

enom

ania

n

Albian

Aptian

Barremian

Hauterivian

EiffelithusturriseiffeliP. cretacea

Parhabdolithusangustus

Chiastozyguslitterarius

Micrantholithusobtusus

Lithraphiditesbollii

Stage

-90 -60 -30 0 30 60 90

120

125

130

135

140

145

150

Inclination (°) (@20 mT)

Dep

th (

mbs

f)

Lithology Polarity

14H

15H

16H

siliceous ooze

nannofossil ooze

C15n

C16n.1n

C16n.2n

C17n.1n

C18n.1n

C17n.2n

C17n.3n

LO Reticulofenestraoamaruensis

FO Reticulofenestraoamaruensis

FO Isthmolithus recurvus

LO Reticulofenestrareticulata

Fig. 2. Illustration of the exceptional magnetostratigraphic records (with biostratigraphy) often recorded by pelagic carbonate sediments. Results are illustrated for (a) the tectonically uplifted Tethyan sections at Gubbio (central Apennines; solidsymbols) and Cismon (southern Alps; open symbols), Italy (redrawn from Lowrie et al., 1980), and (b) ODP Hole 689B, Maud Rise, Weddell Sea. Depths are in metres below seafloor (mbsf; redrawn from Florindo and Roberts, 2005).

115A.P.Roberts

etal./Earth-ScienceReview

s127

(2013)111

–139

0

100

200

300

400

500

600

-90 -60 -30 0 30 60 90

Inclination (°)

0

50

100

150

200

250

300

ODP Hole 744An = 6336

ODP Hole 748Bn = 4859

Fre

quen

cy

-90 -60 -30 0 30 60 90

Inclination (°)

0

100

200

300

400

500

-90 -60 -30 0 30 60 90

ODP Hole 689Dn = 6030

Inclination (°)

Fre

quen

cy

0

50

100

150

200

250

300

ODP Hole 690Cn = 2707

-90 -60 -30 0 30 60 90

Inclination (°)

Fig. 3. Histograms of paleomagnetic inclinations determined from u-channel measurements of Southern Ocean Eocene–Oligocene sediments from ODP Holes 744A and 748B, KerguelenPlateau (Roberts et al., 2003), and 689D and 690C,MaudRise (Florindo and Roberts, 2005). The bimodal inclination distributions are distributed around the expected values for a GAD fieldat the respective site latitudes (vertical lines). This demonstrates that secondarymagnetic overprintswere effectively removedbyprogressive alternating field demagnetisation (see Fig. 4)and that the recorded square-wave inclination records (Fig. 2) provide high-fidelity records of the ancient geomagnetic field. All studied sites are from the southern hemisphere Fig. 1c;negative inclinations correspond to normal polarity fields and positive inclinations correspond to reversed polarity fields.

116 A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

or rain-out, of planktonic remains fromoceanic surfacewaters.Microbi-al respiration of this organic matter is ubiquitous in both the water col-umn and marine sediments, where microbes derive energy throughuptake of oxygen and release of CO2 from the organic matter, whichleads to oxidative degradation of reactive organic compounds in the or-ganic matter. Microbial metabolism of sedimentary organic matteroccurs in a sequence in which the free energy yielded by different oxi-dants (per mole of organic carbon oxidised) progressively decreases(Froelich et al., 1979). The order of electron acceptor use from oxidantsis as follows: oxygen (oxic diagenesis), nitrate, manganese oxides, iron(oxyhydr)oxides (suboxic diagenesis), and sulphate (anoxic sulphidicdiagenesis). When one oxidant is depleted, the next most efficient(i.e., most energy producing) oxidant is used, etc., until either all oxi-dants are consumed or all reactive organic matter is consumed(Froelich et al., 1979). Progressive consumption of these reactants is il-lustrated in the idealised sedimentary pore water profile in Fig. 5a.Also shown are the products of these reactions (Mn2+, Fe2+, HS−,CH4). That is, as manganese oxides, iron (oxyhydr)oxides, and sulphateare reduced, the concentration of Mn2+, Fe2+, and HS−, respectively, inthe sedimentary pore waters increases. Once pore water sulphate is en-tirely reduced, anoxic diagenesis changes from being sulphidic tomethanic (Fig. 5a). Diagenetic increases in the concentrations ofMn2+, Fe2+, andHS− in sedimentary porewaters give rise to the forma-tion of authigenic minerals, the presence of which in the geological

record provides an indication of former early diagenetic conditions(Berner, 1981). The most relevant reactions for sedimentary magneticproperties involve authigenic formation of iron-bearing minerals(right-hand side of Fig. 5a).

In oxic diagenetic environments, which are common in pelagic car-bonates, hematite and goethite can form (e.g., Channell et al., 1982). Asource of dissolved iron would be required for such reactions. Pelagicbottom waters normally have negligible dissolved iron contents (Boydand Ellwood, 2010), and are not expected to provide a significant sourceof iron for such reactions. Themost likely source of dissolved iron, there-fore, is upward diffusion from underlying suboxic sediments that sup-port active iron reduction. The most reactive iron (oxyhydr)oxides thatare readily reduced in suboxic environments are ferric hydrous oxide,ferrihydrite and lepidicrocite (Poulton et al., 2004). Fe2+ released bythis process is oxidised to Fe3+ as it diffuses upward into sedimentswith oxic pore waters and can precipitate as authigenic hematite or goe-thite depending on Eh/pH conditions (e.g., Chan et al., 2000; Beitler et al.,2005).

In anoxic sulphidic sediments, dissolution of less reactive iron ox-ides, such as magnetite and hematite (Poulton et al., 2004), becomesubiquitous (Canfield and Berner, 1987). Dissolved Fe2+ released fromdetrital iron-bearing minerals then reacts with dissolved HS−, a by-product of sulphate reduction, to form sedimentary iron sulphides, par-ticularly pyrite (Berner, 1984; Kao et al., 2004). Dissolution of detrital

ARC132NRM = 2.6E-4 A/m

E

up/N

500°C

500°C

300°C

300°C

700 °C

NRM

NRM

CH110.90 mNRM = 1.55E-7 Am2/kg

N

up/W

up/N

E

ODP 744B-8H4-51-64.51NRM = 5.21E-3 A/m

NRM

NRM

NRM

300°C

300°C

500°C

500°C580 °C

NRM

100 mT

10 mT

50 mT

up,N

E

E

ODP 689B-14H-6-128.55 NRM = 4.69E-3 A/m

up/NNRM

10 mT

50 mT

ODP 690C-8H-4-69.22 NRM = 5.19E-3 A/m

90 mT

50 mT

NRM

10 mT

30 mT

NRM

30 mT

50 mT

NRM

up/W

N

BOS31NRM = 1.6E-4 A/m

10 mT

30 mT

45 mT70 mT

NRM

30 mT30 mT

(a) (b) (c)

(f)(e)

(g)

ODP1263B-5H4-127-89.60NRM = 1.17E-3 A/m

up/NE

NRM

NRM

10 mT

30 mT30 mT

50 mT(d)

up/N

CH110.90NRM = 1.49E-7 Am2/kg

NRM

NRM

E

10 mT

30 mT

100 mT

GOR40NRM = 3.8E-5 A/m

(h) (i)

up/N

E

NRM

NRM

200 °C300 °C450 °C

520 °C

Fig. 4. Typical AF and thermal demagnetisation results for pelagic carbonates from: (a–d) four SouthernOcean, and (e–i) four land-based sections. Demagnetisation steps are expressed inmilliTesla (mT) and °C. “NRM” represents the natural remanent magnetisation prior to demagnetisation. Open and solid symbols represent projections onto the vertical and horizontalplanes, respectively (Zijderveld, 1967). In detail: (a) is from late Eocene (Priabonian) and (b) late Oligocene (Chattian) nannofossil oozes from Maud Rise (Florindo and Roberts,2005); (c) is from a middle Miocene (Serravallian) nannofossil ooze from southern Kerguelen Plateau (Florindo et al., 2013); (d) is from a Late Eocene (Priabonian) nannofossil oozefrom Walvis Ridge (unpublished); (e–f) are from Speranza et al. (2005), where (e) is from the Late Jurassic (Kimmeridgian–Tithonian) Calcari ad Aptici limestone, Arcevia section,Italy; (f) is from the Early Cretaceous (Berriasian) Maiolica limestone, Bosso, Italy; (g, h) are sister samples from themiddle Eocene (Lutetian) Scaglia Variegata limestone, Contessa High-way section, Italy (Jovane et al., 2007); and (i) is from the early Cretaceous (Barremian) Maiolica limestone, Gorgo a Cerbara, Italy (Satolli et al., 2007).

117A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

magnetite and hematite in sulphidic diagenetic environments, and re-placement by paramagnetic pyrite, which does not carry a permanentmagnetisation, can destroy the primary paleomagnetic record (Karlinand Levi, 1983; Channell and Hawthorne, 1990; Karlin, 1990; Rowanet al., 2009). This process is ubiquitous in continental margin sedimentswith high organic carbon contents. If the rate of Fe2+ supply exceeds thatof HS− production, intermediate iron sulphides that form as precursorsto pyrite, such as mackinawite and ferrimagnetic greigite (Berner,1984; Roberts and Turner, 1993), can be preserved (Kao et al., 2004).Greigite can form early (e.g., Reynolds et al., 1999; Blanchet et al.,2009) or it can grow later and remagnetise the host sediment (Roberts

andWeaver, 2005) depending on the timing of availability of the neces-sary reactants. In general, sulphate-reducing diagenetic environmentsare destructive to paleomagnetic recording, and oxic to suboxic diage-netic conditions are more likely to preserve primary paleomagneticsignals.

The extent to which pore waters evolve to reveal the entire set of re-actions described in Fig. 5a depends on the rate of organic carbon supplyand how quickly it can be oxidised before each oxidant is progressivelyconsumed. The spectrum of diagenetic conditions expected in pelagicenvironments is illustrated in Fig. 5b–d, and is discussed below. Organiccarbon contents of carbonate sediments can vary between about 0.1%

Oxic

Suboxic

Anoxic

Processes

Aerobic respirationDenitrification

Manganesereduction

Ironreduction

Sulphatereduction

Methanogenesis

Authigenic iron-bearingmineral products

Hematite, goethite

Siderite

Pyrite (±mackinawite, greigite)

FeOOH

Concentration of dissolvedpore water species

Dep

th (

arbi

trar

y)

O2

NO3-

Mn2+

Fe2+

HS-

SO42-

CH4

Sulphidic

Non-sulphidic

Fe2+

Dep

th (

mbs

f)

O2

SO42-

(e.g., South Pacific Gyre)(e.g., Eirik Drift,North Atlantic Ocean)

(e.g., Ontong-Java Plateau,Equatorial Pacific Ocean)

Basalt(oceanic

lithosphere)

0

~70

Concentration of dissolvedpore water species

Mn2+ NO3-

Dep

th (

mbs

f)

SO42-

0

200

Concentration of dissolvedpore water species

Mn2+

100

CH4

Concentration of dissolvedpore water species

O2NO3

-

Mn2+

Fe2+

HS-

SO42-

Dep

th (

mbs

f)

0

12

Decreasing organic matter diagenesis

(a) (b) (c) (d)

Fig. 5. Schematic illustration of sedimentary porewater profiles associatedwith organicmatter diagenesis in pelagic settings. (a) Idealised porewater profiles for the full range of early diagenetic conditions, with authigenic iron-bearingminerals thatform in the respective early diagenetic zones (modified from Froelich et al., 1979; Berner, 1981; Kasten et al., 2003; Roberts andWeaver, 2005). Depths of concentration profiles for dissolved pore water species depend largely on sedimentation rateand reactive organic matter supply. Profiles are shown for reactants (O2, NO−

3, SO2−4) and products (Mn2+, Fe2+, HS−, CH4) in the redox zonation. (b–d) Schematic pore water profiles for three pelagic settings, with progressively decreasing in-

fluence of organic matter diagenesis (depths are reported in metres below seafloor (mbsf)), to illustrate the spectrum of diagenetic environments in pelagic settings. (b) Pelagic carbonates from ODP Hole 806A, Ontong–Java Plateau (after Tarduno,1994); the profile is shownonly for the uppermost 12mof the sediment record. Thepaleomagnetic signal is largely destroyed below this depth. (c)Hemipelagic sediments from IODP SiteU1305, Eirik Drift, North Atlantic Ocean (after Kawamura et al.,2012). Bottomwaters are oxic, butmoderate rates of organic carbon burial result in sulphate reduction at depth.Highmagnetite contentsmean that the paleomagnetic signal is preserved throughout the sediments even to down-core depths of 180m.(d) Pelagic red clay sediments from the ultra-oligotrophic South Pacific gyre (after Fischer et al., 2009), where sedimentation rates are only ~1mm/kyr (D'Hondt et al., 2009). Dissolved oxygenmeasurements at station 1 (23°51′S; 165°39′W;waterdepth=5697m) have been extrapolated using different model simulations through the full sediment thickness of 71m (Fischer et al., 2009). We illustrate the case with the greatest decrease in O2 concentrations (to microaerobic levels) of all thestations studied. In these oxic sediments, any organic matter that is delivered to the seafloor is rapidly oxidised. The sediments therefore only experienced oxic diagenesis, with oxygen diffusing through the entire sediment column to oxidise theunderlying basalt of the oceanic lithosphere.

118A.P.Roberts

etal./Earth-ScienceReview

s127

(2013)111

–139

119A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

and 10% (Blatt et al., 1980). In pelagic environments, organic mattercontents are usually b0.2% (Morse, 2005) in areas where primary pro-ductivity is low (Fig. 1b). As a result, slowly deposited pelagic sediments(≤1 cm/kyr) tend to undergo oxic diagenesis because organic carboncontents are low and dissolved oxygen from the overlying bottom wa-ters readily diffuses into sediments so that any organic matter respira-tion occurs under oxic conditions. Magnetic mineral diagenesis,therefore, is not as important as in continental margin sediments,where organic carbon fluxes are high (Berner, 1980), and oxic pelagicsediments often preserve high-quality paleomagnetic recordswith littleobvious indication of diagenetic alteration. Pelagic red clays from theultra-oligotrophic South Pacific gyre, where sedimentation rates areonly ~1mm/kyr (D'Hondt et al., 2009), provide an extreme illustrationof diagenesis that does not proceed beyond oxic conditions for an entiresedimentary succession (Fig. 5d). In this setting, dissolved oxygen hasdiffused through the entire sediment column to oxidise the underlyingocean crust basalt (Fischer et al., 2009). Hemipelagic sediments fromEirik Drift, North Atlantic Ocean, provide an intermediate examplewhere oxic and suboxic diagenesis persist over sediment thicknessesof tens of metres (Fig. 5c; Kawamura et al., 2012). Although bottomwa-ters are oxic, moderate rates of organic carbon burial result in sulphatereduction at depths of ~70m below seafloor (mbsf). However, the pa-leomagnetic signal is preserved to down-core depths of 180 m(Kawamura et al., 2012) because of high initial magnetite contents. Fi-nally, in more productive waters, organic carbon fluxes can be moder-ately high so that suboxic diagenesis occurs at shallow depths (e.g.,deoxygenation, nitrification, denitrification, and iron and manganeseoxide reduction; Berner, 1980). In the most productive waters, diagen-esis can have a controlling influence on sediment magnetic properties,although these changes will occur at greater depths than those in conti-nental margin sediments (Tarduno, 1994; Fig. 5b).

The discussion above assumes that diagenesis has occurred understeady-state conditions where pore waters are in equilibrium and thepore water zonation does not change significantly through time. Thiscondition is most likely to be satisfied in pelagic environments becausegenerally low organic carbon fluxes will not significantly perturb thesteady-state scenario. However, productivity can vary significantlythrough time as a result of intermittent supply of key limitingmicronutrients required by plankton (e.g., iron fertilisation of oligotro-phic waters through supply of eolian dust). On orbital timescales,larger-scale shifts in nutrient supply can be controlled by latitudinalmi-gration of oceanic fronts, or by changes in upwelling of nutrient-enriched deep waters. Time varying pulses of enhanced organic carbondelivery to the seafloor can give rise to marked changes in magneticmineral diagenesis. Such non-steady-state diagenesis can give rise tomagnetite dissolution cycles that were driven by, for example, orbitaleccentricity-scale productivity variations (e.g., Tarduno, 1992),precession-scale African monsoon driven deposition of organic-richsapropels (e.g., Larrasoaña et al., 2003), or instantaneous delivery of en-hanced organic carbon concentrations to the seafloor by turbidites (e.g.,Robinson et al., 2000). Non-steady-state magnetic mineral dissolutionassociated with relatively short-lived enhanced organic carbon burialevents are possible and could explain cyclicity between relativelystrong and weak magnetisations. Alternatively, however, enhanced or-ganic carbon supply can result in suboxic conditions that liberate dis-solved iron into sedimentary pore waters through iron reduction anddissolution of labile iron-bearing minerals (ferric hydrous oxide, ferri-hydrite, lepidocrocite, but not magnetite or hematite; Poulton et al.,2004), thereby providing the iron needed to enhance magnetite pro-duction by magnetotactic bacteria (e.g., Roberts et al., 2011;Larrasoaña et al., 2012). In this scenario, enhanced organic carbon sup-ply can give rise to increased magnetisations. Discrimination betweensuch scenarios requires use of other (e.g., geochemical, magnetic) prox-ies for diagenesis, of which there are many.

Finally, diagenetic microenvironments can bemore important in pe-lagic sediments than in continental margin sediments where reducing

diagenetic conditions aremore pervasive. For example, reducingmicro-environments can develop within foraminiferal tests (e.g., Stumm andMorgan, 1996) or around large pieces of organic matter, where sedi-mentary pyrite formation is only localised. Such situations are evidentby the preservation of iron oxides in unaffected parts of the sedimentand by pyrite formation, with lack of detrital iron oxides, in locally re-duced sediments.

In summary, diagenesis has variable effects on themagnetic proper-ties of pelagic carbonates, ranging from minimal to controlling influ-ences. Hematite or goethite formation in oxic diagenetic environmentscan be important in some settings (e.g., Henshaw and Merrill, 1980;Channell et al., 1982), while diagenetic dissolution of magnetite and de-struction of the paleomagnetic record can be important below produc-tive surface waters (e.g., Tarduno, 1994).

5. Magnetic properties of pelagic marine carbonates

It has long been recognised that pelagic marine carbonates usuallycontain amixture ofmagneticminerals, with potentially different origins(e.g., Lowrie and Heller, 1982; Freeman, 1986). Sophisticated interpreta-tions have often been made about the relative timing of magnetisationscarried by different magnetic minerals based on detailed paleomagneticand rock magnetic analyses (e.g., Heller, 1977; Lowrie and Alvarez,1977b; Channell et al., 1982; Lowrie and Heller, 1982). Recent develop-ment of measurement techniques that allow identification of multiplemagnetic mineral components and analysis techniques that enabletheir quantitative separation and analysis (see summary by Liu et al.,2012) assists with routine assessment of stratigraphic variations inmag-neticmineralogy and their potential paleomagnetic consequences. In thefollowing discussion, we present results from a range of rock magneticand other measurements to provide a representative overview of themagnetic properties of pelagic carbonates from globally distributed sed-imentary sequences (Fig. 1c). We then, in the succeeding sections, dis-cuss the origins of these magnetic minerals and their implications forpaleomagnetic studies.

5.1. Methods

Results are presented below from a range of methods, which are de-scribed briefly here. Paleomagnetic measurements were made with 2-GEnterprises superconducting rock magnetometers, with progressivestepwise demagnetisation performed using in-line alternating field(AF) demagnetisation coils. In some cases, results are presented fromu-channel samples that were measured at high spatial resolutions(Weeks et al., 1993). Stepwise thermal demagnetisation was carriedout with a range of ovens. The magnetic mineralogy of pelagic carbon-ates was assessed with temperature-dependent measurements of thelow-field magnetic susceptibility (χ). These measurements were madeusing a Kappabridge KLY-2 susceptibility meter with a CS-2 high-temperature attachment (Hrouda, 1994), with χ− T results presentedafter subtraction of the diamagnetic susceptibility of the empty sampleholder. IRM acquisition and backfield demagnetisation of the IRM, hys-teresis and first-order reversal curve (FORC) measurements (Pike et al.,1999; Roberts et al., 2000) were performed with Princeton Measure-ments Corporation Micromag instruments (either an alternating gradi-ent magnetometer or a vibrating sample magnetometer) to maximumapplied fields of 1T. IRM acquisition curves were subjected to unmixinganalysis (Robertson and France, 1994; Kruiver et al., 2001) to detect dif-ferentmagneticminerals using the endmember approach of Heslop andDillon (2007). The generally weakmagnetisations mean that IRM acqui-sition curves are often noisy. Individual curves were smoothed with aconstrained least squares spline (de Boor, 1994) prior to unmixing. Asoften as possible, FORC measurements were made at high-resolution,with up to 247 FORCs measured, to enable optimal recognition of thecentral ridge feature for identifying magnetically non-interacting uniax-ial SD particles, which in sediments are usually due to the presence of

120 A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

magnetofossils (Egli et al., 2010). FORC diagrams were processed usingthe approach of Heslop and Roberts (2012a) to identify regions of theFORC distribution that are statistically significant at the 0.05 level.

Ferromagnetic resonance (FMR) measurements were made with X-band Bruker EMXmicrospectrometers on 100–200mg of air-dried sed-iment. Samples were loaded in glass tubes and were microwaved at~9.4 GHz and power of ~0.632mW. The spectra were integrated overthree field sweeps from 0 to 700mT. Low-temperature magnetic mea-surements were made with Quantum Designs Magnetic PropertiesMeasurement Systems. Samples were cooled from room temperatureto low temperatures (10 or 20 K) either in zero-field cooled (ZFC) orfield-cooled (FC) conditions in a 2.5 T field. A saturation IRM (SIRM)was then imparted in a 2.5T field at 20K andmeasured in approximate-ly zero field (the residual field after a magnet reset from 2.5 T is ~200–300 μT) at 5 K intervals during warming to room temperature. For anSIRM imparted at room temperature (RTSIRM), low-temperature cy-cling (LTC) measurements were made from room temperature to 10K

0

10

20

30

40

50

60

70

0 100 200 300 400 500 600 700

Sus

cept

ibili

ty (

10-6

SI)

Temperature (°C)

0 100 200 300 400 500 600 700

Temperature (°C)

0 100 200 300 400 500 600 700

Temperature (°C)

744A-11H-5-4996.19 mbsf

heating

cooling

0

1

2

3

4

5

6

7

8

97.75 m

0

2

4

6

8

121.95 m(marl)

Sus

cept

ibili

ty (

10-6

SI)

Sus

cept

ibili

ty (

10-6

SI)

Contessa Highway

Contessa Highway

(marly limestone)

(a)

(c)

(e)

(nannofossil ooze)

Fig. 6. Temperature-dependent low-field magnetic susceptibility measurements as indicatorsstone andmarl samples from the Contessa Highway section, Italy (Jovane et al., 2007). (d) Oligoand Roberts, 2005). (e–f) Middle Miocene nannofossil ooze from ODP Hole 744, Kerguelen Plashown in (f). The results indicate the presence of magnetite (Curie temperature at 580 °C), heto 350 °C in (c) and (d)), and rare possible goethite (dehydration up to 100 °C in (a)).

and back to room temperature in zero field. Transmission electron mi-croscope (TEM) observations were made on magnetic particles thatwere extracted from bulk sediments by adapting the methods ofStoltz et al. (1986) and Hesse (1994). Magnetic extracts were viewedand analyzed using a CM300 FEI TEM operated at 300 kV, which isequipped with an EDAX Phoenix retractable X-ray detector (ultra-thinwindow) and a Gatan model 694 slow-scan digital camera.

5.2. High-temperature magnetic measurements

High-temperature analyses are among themost diagnostic measure-ments for assessingmagnetic mineralogy because themagnetisation (orsusceptibility) decreases sharply at the Curie or Néel temperature of fer-rimagnetic and imperfect antiferromagneticmaterials, respectively. Rep-resentative χ − T measurements for a range of pelagic carbonatesprovide evidence for the typically mixed nature of magnetic mineral as-semblages in such sediments (Fig. 6). ForMiddle Eocenemarly limestone

0

2

4

6

8

10

12

14

0 100 200 300 400 500 600 700

Temperature (°C)

0 100 200 300 400 500 600 700

Temperature (°C)

0 100 200 300 400 500 600 700

Temperature (°C)

0

2

4

6

8

10

110.70 m(marly limestone)

Sus

cept

ibili

ty (

10-6

SI)

Contessa HighwayS

usce

ptib

ility

(10

-6 S

I)

(b)

(d)

(f)

0

10

20

30

40

50

Sus

cept

ibili

ty (

10-6

SI)

cooling

heating

690B-10H-788.9 mbsf

(nannofossil ooze)

of magnetic mineralogy in pelagic carbonate sediments. (a–c) Middle Eocene marly lime-cene nannofossil ooze sample from ODP Hole 690B, Maud Rise, Southern Ocean (Florindoteau, Southern Ocean (Florindo et al., in press). An expanded view of the curves in (e) ismatite (Néel temperature at 680 °C), occasional maghemite (inversion to hematite at up

121A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

andmarl samples from the Contessa Highway section, Italy (Jovane et al.,2007), the most marked susceptibility decrease is associated with theCurie temperature of magnetite at 580°C (Fig. 6). The susceptibility gen-erally increases to a broad Hopkinson peak immediately below the Curietemperature (Fig. 6a). This increase can be obscured in some samples(Fig. 6c) where there is a marked susceptibility decrease at 250–350 °C,which is attributed to the thermally induced conversion of maghemiteto hematite (e.g., Deng et al., 2001; Liu et al., 2005). Evidence for thepresence of maghemite is most marked in more hematite-rich reddish-coloured sediments (Jovane et al., 2007). Persistence of a significant sus-ceptibility signal above 580 °C and up to 680 °C is indicative of the pres-ence of hematite (Fig. 6a–c). In the case of Fig. 6c, some of the hematitemay have resulted from thermal conversion of maghemite to hematite,but it is still likely that much of the hematite was originally present be-fore heating, as indicated by the red colour of the sediment. In other sam-ples, there is no clear evidence for thermal conversion of maghemite tohematite (Fig. 6a, b), which indicates that the hematite is a significantcontributor to the primary magnetic mineral assemblage in these sedi-ments. The presence of magnetite and hematite is also consistently indi-cated for nannofossil ooze samples (Fig. 6d–f), as exemplified by resultsfromOligocene sediments fromODPHole 690B,MaudRise (Florindo andRoberts, 2005) and from Middle Miocene sediments from ODP Hole744A, Kerguelen Plateau (Florindo et al., in press). The fact that hematitemakes such a noticeable contribution to χ−T curves (Fig. 6), despite thefact that it has susceptibility values that are 0.1% that of magnetite(O'Reilly, 1984), means that it is present in significant abundances.

In addition to magnetite, hematite and occasional maghemite, weonly occasionally find evidence for goethite in the pelagic carbonatesthat we have studied. This contrasts with the findings of Heller (1977)and Lowrie and Heller (1982), which is probably explained by the factthat they observed goethite most frequently as a weathering product insubaerial weathering environments. Nevertheless, occasional hints ofgoethite dehydration are suggested by small susceptibility decreasesbelow 100 °C in some samples e.g., Fig. 6a. The presence of detrital goe-thite is expected in sediments that contain significant products of soilerosion, including maghemite. It is, therefore, puzzling that we do notobservemore evidence for the presence of goethite in our studied pelagiccarbonate samples. This conundrum is probably best explained by thefact that magnetic methods will underestimate the presence of goethitein sediments because its magnetisation is strongly dependent on Al sub-stitution within the crystal lattice (Al substitution is common in pedo-genic goethite). For example, only goethite with Al contents of ~5–12mol% is likely to be magnetically important in sediments (Liu et al.,2007). We, therefore, suggest that the presence of detrital goethite isgenerally under-recognised in studies of pelagic carbonates because itis not magnetically significant enough to be detected.

Overall, thermomagnetic results consistently point to the presenceof magnetite, hematite, occasional maghemite, and possible goethitein pelagic carbonate sediments. Thesemeasurements, however, provideno strong evidence for the origin of the respective minerals, whether itis detrital, biogenic, or diagenetic. All of these origins are potentiallypossible. For example, maghemite is a common product of pedogenesis(e.g., Zhou et al., 1990; Verosub et al., 1993; Maher, 1998; Deng et al.,2001; Liu et al., 2005) and is expected to be a common detrital mineralin sediments sourced from soil erosion (with transport by fluvial andeolian mechanisms both being possible). It can also occur via oxidationof magnetite particles, with maghemite occurring as a skin on largermagnetite particles (e.g., Cui et al., 1994; van Velzen and Zijderveld,1995). Surficial oxidation of magnetite can occur at ambient tempera-tures when particles are in transit in oxic environments, but magnetitehas also been reported to oxidise in oxic diagenetic sedimentary envi-ronments (e.g., Torii, 1997; Smirnov and Tarduno, 2000; Yamazakiand Solheid, 2011; Kawamura et al., 2012). Likewise, hematite canoccur as both a detrital and authigenic phase in pelagic sediments(e.g., Channell et al., 1982). Magnetite can occur as a detrital or biogenicphase (e.g., Tarduno, 1994; Abrajevitch and Kodama, 2009; Roberts

et al., 2011), and diagenetic magnetite has been widely reported tocause remagnetisations in pelagic carbonates (e.g., Jackson, 1990; Suket al., 1990a,b; Channell and McCabe, 1994; Gong et al., 2009a,b;Jackson and Swanson-Hysell, 2012; van der Voo and Torsvik, 2012).Goethite has been reported widely in Tethyan pelagic carbonates as alater weathering product that gives rise to important secondarymagnetisations (e.g., Heller, 1977; Lowrie and Heller, 1982), but it canalso have a detrital origin. These observations underline the importanceof understanding the environmental origin and significance of eachmineral within these mixed magnetic assemblages in pelagic carbon-ates in order to makemeaningful paleomagnetic and environmental in-terpretations. Such considerations are discussed in more detail inSection 6.

5.3. IRM acquisition, hysteresis and unmixing analysis

Unmixing procedures can provide information that is diagnostic ofmagnetic mineralogy from IRM acquisition (Kruiver et al., 2001; Heslopet al., 2002; Egli, 2004; Heslop and Dillon, 2007) and hysteresis(Heslop and Roberts, 2012b) data. IRM acquisition and backfielddemagnetisation curves for representative Eocene pelagic carbonatesamples from ODP Hole 738B have variable approaches to magnetic sat-uration that reflect variable coercivity distributions (Fig. 7a, b). Likewise,hysteresis parameters, as plotted in a Day diagram (Day et al., 1977),have a range of values that are consistent with variablemixtures ofmag-netic particles with different grain sizes. The data lie between mixingcurves (Dunlop, 2002) for SD+MD and SD+ superparamagnetic (SP)mixtures (Fig. 7c). The well-known ambiguities of the Day diagram(e.g., Roberts et al., 2000; Muxworthy et al., 2003), however, mean thatmore detailed analyses are needed to explain the data distribution inFig. 7c. IRM acquisition curves (Fig. 7a) for a collection of pelagic carbon-ate samples from the Southern Ocean were unmixed to determine endmember coercivity components following the procedure of Heslop andDillon (2007). Results from a three-component unmixing model areshown for two contrasting samples in Fig. 8; the same colour coding isused for hysteresis unmixing results shown in Fig. 9, where blue repre-sents detrital magnetite (end member 1), green represents “biogenicsoft” (BS) magnetite (end member 2), and red represents “biogenichard” (BH) magnetite (end member 3; see Egli (2004) for discussion ofBS and BH components). End members are defined by variation withina given data set and represent the smallest parameter space in a 3-component mixing system that bounds the data (Heslop and Dillon,2007). The end members, therefore, do not necessarily correspond topure mineral magnetic components. For example, a high coercivity (de-trital hematite) component is evident in thermomagnetic (Fig. 6) andIRM acquisition (Fig. 7a) curves, but is not identified as a separate endmember in our unmixing analyses shown in Figs. 8 and 9. Detrital hema-tite gives rise to a high coercivity component in end member 3 in IRM(Fig. 8a) and hysteresis (Fig. 9c) unmixing results, which indicates thatend member 3 represents a mixture of BHmagnetite and detrital hema-tite. The detrital hematitewithin endmember 3 is interpreted to have aneolian origin (Roberts et al., 2011). The fact that the endmembers do notrepresent puremineral magnetic componentsmeans that the dispersionparameter (DP) of Robertson and France (1994) is not as readily diag-nostic of mineralogy as used, for example, by Egli (2004). The IRMunmixing results illustrate samples with dominant BH (Fig. 8a) and BS(Fig. 8b) components.

A three-component solution is also obtained using a mathematicalunmixing procedure for magnetic hysteresis data (Heslop and Roberts,2012b), which explains 97% of the variance in hysteresis data for thesame pelagic carbonate samples (Fig. 9). Slightly different interpreta-tions were provided for these sediments by Roberts et al. (2011) andHeslop and Roberts (2012b); for this paper, we carefully reanalyzedthe data to ensure consistent identification of magnetic componentswith different unmixing approaches. The principal difference surroundsour failure to previously identify the BS and BH components of Egli

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 1

M/M

r

-1 -0.8 -0.6 -0.4 -0.2 0

Field (T)

n = 12

-1

-0.5

0

0.5

1

M/M

r

(a)

(b)

(c)

0

0.1

0.2

0.3

0.4

0.5

0.6

1 1.5 2 2.5 3 3.5 4 4.5 5

Hole 689DHole 690CHole 738BHole 738C

SD+MD1

SD+MD2

SD+SP (10 nm)

n = 86

Mr/

Ms

Bcr/Bc

Fig. 7. IRMacquisition and backfield demagnetisation curves andhysteresis parameters forEocene–Oligocene pelagic carbonate samples from the Southern Ocean. (a) IRM acquisi-tion and (b) backfield demagnetisation curves for Eocene samples from ODP Hole 738B(from Roberts et al., 2011). Variable approaches to magnetic saturation reflect variablemixtures of high and low coercivity magnetic minerals (see Fig. 8 for results of IRMunmixing analyses). (c) Day plot (cf. Day et al., 1977) of magnetic hysteresis parametersfor samples from ODP Holes 689D, 690C, 738B, and 738C. Mixing lines for single domain(SD)+multi-domain (MD) and SD+ (10 nm) superparamagnetic (SP) magnetite parti-cles are shown (Dunlop, 2002) (see Fig. 9 for results of hysteresis unmixing analyses).

122 A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

(2004); the overall environmental interpretations remain the samewith identification of detritalmagnetite andhematite, butwenow iden-tify two biogenic magnetite components (BS and BH) instead of one.The three components identified with hysteresis unmixing (Fig. 9) arethe same as discussed above for the IRM unmixing.

In summary, using IRM and hysteresis unmixing approaches, wehave identified three magnetic mineral end members, with highly vari-able relative contributions to the magnetisation of pelagic carbonates(Fig. 9d): detrital magnetite and hematite (whichwill have an eolian or-igin inmany distal pelagic settings) and biogenic soft and hardmagnetite(cf. Egli, 2004). The biogenic magnetite component is usually significantand can dominate the magnetic properties (e.g., Abrajevitch andKodama, 2009; Roberts et al., 2011; Chang et al., 2012a; Larrasoañaet al., 2012; Yamazaki, 2012; Yamazaki and Ikehara, 2012; Channellet al., 2013; Yamazaki et al., 2013).

5.4. FORC diagrams

FORC diagrams (Pike et al., 1999; Roberts et al., 2000) have become astandard technique for assessing the magnetic mineralogy of naturalsamples because any magnetic particle assemblage that makes a signif-icant contribution to themagnetisation of a sample can be detected anddiscriminated using FORC diagrams. Egli et al. (2010) provided evidencethat intact magnetosome chains produce a characteristic signature dueto their uniaxial non-interacting SD properties that give rise to a “centralridge” on FORC diagrams. This signature arises because, even though in-dividual magnetic particles in magnetosome chains interact strongly(e.g., Dunin-Borkowski et al., 1998; Muxworthy and Williams, 2009),the overall magnetic behaviour is like that of a single elongated SD par-ticlewith strongmagneticmoment (Penninga et al., 1995; Hanzlik et al.,2002). Lack of magnetic interactions among magnetosome chains (i.e.,through lack of chain collapse), iswhat gives rise to the central ridge sig-nature in FORC diagrams. Central ridge FORC signatures have now beenwidely reported from magnetotactic bacteria (Pan et al., 2005; Chenet al., 2007; Fischer et al., 2008; Carvallo et al., 2009; Li et al., 2009;Jovane et al., 2012; Roberts et al., 2012) and from pelagic carbonates(Yamazaki, 2008; Abrajevitch and Kodama, 2009; Yamazaki, 2009;Roberts et al., 2011; Larrasoaña et al., 2012; Roberts et al., 2012;Yamazaki, 2012; Yamazaki and Ikehara, 2012; Channell et al., 2013;Yamazaki et al., 2013). These results are leading to re-evaluation ofthe role of magnetofossils as carriers of paleomagnetic signals (e.g.,Roberts et al., 2012, 2013).

Representative FORC diagrams from ODP Holes 738B, 738C, 689Dand 690C (Roberts et al., 2011, 2012; Larrasoaña et al., 2012) areshown in Fig. 10. Variable FORC signatures that are not associated withthe central ridge are evident, as discussed in the original studies and assummarised in Fig. 10. Nevertheless, all FORC diagrams from all pelagiccarbonate samples studied from these ODP holes have a strong centralridge signature, which is statistically significant at the 0.05 level as indi-cated by the thick contour line about the central ridge (calculated follow-ing the method of Heslop and Roberts (2012a)). Vertical profiles of theinteraction field distributions through the peak of the FORC distributionsare sharply peaked, which reflects the lack of magnetostatic interactionsamong intact magnetosome chains that behave effectively as isolatedelongated SD particles. Horizontal profiles of coercivity distributionsalong Bu=0mT are highly variable. These profiles contain signals fromall particles that contribute to the magnetisation, including detrital andbiogenic particles. The variable shape of these coercivity distributionsand their variable peak positions is indicative of a mixedmagnetic parti-cle system. If the central ridge contribution is separated from this profileso that it represents only the non-interacting SD component, followingthe procedure of Egli et al. (2010), the coercivity distribution of the cen-tral ridge can be represented by binary mixtures (Fig. 11) that accountfor 94% of the signal variance. In Fig. 11a, the relative abundances ofthe two endmembers are plotted on the vertical axes versus themediancoercivity for the central ridge on the horizontal axis. The median coer-civity is ~27mT for the soft component and ~43mT for the hard compo-nent (Fig. 11b). As is the case for hysteresis unmixing, an end membermay not represent a pure mineral magnetic component. The negativeskewness of the central ridge profiles is attributed to the impurity ofthe end members, which include non-interacting low coercivity detrital

(a) 738B-5H-2-145

101 102 103 101 102 103

(b) 738C-11R-2-108

Nor

mal

ized

gra

dien

t

Field (mT) Field (mT)

0

0.2

0.4

0.6

0.8

1.0

1.2

Fig. 8. Results of IRM end member unmixing for contrasting samples from ODP Holes (a) 738B (from Roberts et al., 2011) and (b) 738C (from Larrasoaña et al., 2012). Black open circlesrepresent the derivative of the IRMacquisition curves. Different coloured curves represent the IRMendmembers (following themethod of Heslop andDillon (2007)). The sumof the threefitted components (black line) enables comparison of the fit to the measured data. The blue curve represents detrital (probably eolian) magnetite, green represents “biogenic soft”mag-netite (cf. Egli (2004)), and red represents “biogenic hard”magnetitewith a high coercivity hematite contribution. The same endmembers are identified fromhysteresis unmixing (Fig. 9).

123A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

magnetite (Egli, 2004). The modes of the end members at 32 and 45mTare interpreted to correspond to the BS and BH components, respective-ly,which is consistentwith results from IRMandhysteresismixing as de-scribed above. The scatter of values along a linear trend reflects thevariable binarymixing of BS and BH components (Fig. 11a). For example,as can be seen in Figs. 9d and 10, samples fromHoles 689D and 690C aredominated by the BS component, while those from Hole 738B (Eocene)

Nor

mal

ized

mag

netiz

atio

nN

orm

aliz

ed m

agne

tizat

ion

Field (T)

Field (T)0 0.5-0.5

0 0.5-0.5

End-member 1(Detrital

magnetite)

End-member 3 (Biogenic hard

magnetite + detrital

hematite)

0

-1.0

1.0

0

-1.0

1.0

M /M = 0.26r sB /B = 2.1rh cB = 12 mTc

M /M = 0.49r sB /B = 1.5rh cB = 33 mTc

(a) (b

(c) (d

Fig. 9. Results of a 3 end-member unmixing model for 30 hysteresis loops for Southern OceanHeslop and Roberts, 2012b). (a–c) Hysteresis loops for the 3 end-members, with their respectivmagnetisation (Mr/Ms) and themedian field of the remanent component of the loop (Fabian anmembers can be mixtures, and are defined by the data distribution without extrapolating to pmember 2 is “biogenic soft” SDmagnetite, and endmember 3 appears to be a mixture of “biogetation of the relative abundance of each end-member. The 3 end-member model accounts for

are dominated by BH magnetite with mixtures of significant BS magne-tite. In contrast, samples from Hole 738C (Paleocene–Eocene ThermalMaximum; PETM) are dominated by BS magnetite, with variable mix-tures of BH magnetite. The ability to routinely identify the relative pro-portion of BH and BS components is a welcome advance that canpotentially enable extraction of environmental signals from biogenicmagnetite in pelagic sediments. This possibility has been indicated by

Nor

mal

ized

mag

netiz

atio

nR

elat

ive

abun

danc

e (%

)

Field (T)0 0.5-0.5

End-member 2(Biogenic

soft magnetite)

0

100

80

60

40

20

0

-1.0

1.0

689D 690C 738B 738C

Hole

M /M = 0.48r sB /B = 1.5rh cB = 20 mTc

EM1

EM2

EM3

)

)

pelagic carbonate samples from ODP Holes 689D, 690C, 738B, and 738C (modified frome parameters, including the ratios of the saturation remanent magnetisation to saturationd von Dobeneck, 1997) to the coercivity (Brh/Bc), and Bc. In this unmixing procedure, end-ure end member cases. End-member 1 is detrital pseudo-single domain magnetite, end-nic hard” SDmagnetite and a high coercivity detrital SD hematite. (d) Graphical represen-97% of the variance in the hysteresis data set.

ρ / ρ

max

00.20.40.60.81

ODP-738B-4H-6-130

ODP-738B-5H-7-5

ODP-738C-11R-1-28

ODP-738C-11R-1-78

ODP-738C-11R-2-51

ODP-689D-8H-4-71

ODP-690C-7H-3-22

Bc [mT]

Bu

[mT

]

0 10 20 30 40 50 60 70 80 90 100 110-15-10-505

1015

ODP-690C-9H-6-76

0

0.2

0.4

0.6

0.8

1.0

1.2

-10 0 100

0.2

0.4

0.6

0.8

1

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

Bc [mT]

Bu

[mT

]

0 10 20 30 40 50 60 70 80 90 100 110-15-10-505

1015

Bc [mT]

Bu

[mT

]

0 10 20 30 40 50 60 70 80 90 100 110-15-10-505

1015

Bc [mT]

Bu

[mT

]

0 10 20 30 40 50 60 70 80 90 100 110-15-10-505

1015

Bc [mT]

Bu

[mT

]

0 10 20 30 40 50 60 70 80 90 100 110-15-10-505

1015

Bc [mT]

Bu

[mT

]

0 10 20 30 40 50 60 70 80 90 100 110-15-10-505

1015

Bc [mT]

Bu

[mT

]

0 10 20 30 40 50 60 70 80 90 100 110-15-10-505

1015

Bc [mT]

Bu

[mT

]

0 10 20 30 40 50 60 70 80 90 100 110-15-10-505

1015

Bc [mT]0 10 20 30 40 50 60 70 80 90 100 110

Bc [mT]0 10 20 30 40 50 60 70 80 90 100 110

Bc [mT]0 10 20 30 40 50 60 70 80 90 100 110

Bc [mT]0 10 20 30 40 50 60 70 80 90 100 110

Bc [mT]0 10 20 30 40 50 60 70 80 90 100 110

Bc [mT]0 10 20 30 40 50 60 70 80 90 100 110

Bc [mT]0 10 20 30 40 50 60 70 80 90 100 110

Bc [mT]0 10 20 30 40 50 60 70 80 90 100 110

ρ / ρ

max

Bu [mT]

ρ / ρ

max

00.20.40.60.81

0

0.2

0.4

0.6

0.8

1.0

1.2

-10 0 100

0.2

0.4

0.6

0.8

1

ρ / ρ

max

Bu [mT]

ρ / ρ

max

00.20.40.60.81

0

0.2

0.4

0.6

0.8

1.0

1.2

-10 0 100

0.2

0.4

0.6

0.8

1

ρ / ρ

max

Bu [mT]ρ

/ ρm

ax

00.20.40.60.81

0

0.2

0.4

0.6

0.8

1.0

1.2

-10 0 100

0.2

0.4

0.6

0.8

1

ρ / ρ

max

Bu [mT]

ρ / ρ

max

00.20.40.60.81

0

0.2

0.4

0.6

0.8

1.0

1.2

-10 0 100

0.2

0.4

0.6

0.8

1

ρ / ρ

max

Bu [mT]

ρ / ρ

max

00.20.40.60.81

0

0.2

0.4

0.6

0.8

1.0

1.2

-10 0 100

0.2

0.4

0.6

0.8

1

ρ / ρ

max

Bu [mT]

00.20.40.60.81

0

0.2

0.4

0.6

0.8

1.0

1.2

-10 0 100

0.2

0.4

0.6

0.8

1

ρ / ρ

max

ρ / ρ

max

Bu [mT]

ρ / ρ

max

00.20.40.60.81

0

0.2

0.4

0.6

0.8

1.0

1.2

-10 0 100

0.2

0.4

0.6

0.8

1

ρ / ρ

max

Bu [mT]

ρ / ρ

max

ρ / ρ

max

ρ / ρ

max

ρ / ρ

max

ρ / ρ

max

ρ / ρ

max

ρ / ρ

max

ρ / ρ

max

124 A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

25 30 35 4540

Central ridge median coercivity (mT)30 4010 20 50 60 90

Field (mT)

EM1EM2

1.0

0.2

0

1.0

0.5

0

1.5

2.0

Nor

mal

ized

der

ivat

ive

ODP 689DODP 690CODP 738BODP 738C

0.4

0.6

0.8

1.0

0.2

0

0.4

0.6

0.8

EM

1 re

lativ

e ab

unda

nce E

M2 relative abundance

AB

A = 738B-5H-2-145B = 738C-11R-2-108

27 mT

43 mT

(a) (b)

32 mT45 mT

Fig. 11. Unmixing of central ridge coercivity spectra from FORC distributions. The central ridge component from pelagic carbonate samples (as seen in profiles on the right-hand side ofFig. 10),whichwas separated from the rest of the FORC distribution following Egli et al. (2010), has been fittedwith a simple 2-component system that explains 94% of the signal variance.The two endmembers correspondmainly to (a) the biogenic soft (EM1) and biogenic hard (EM2) components of Egli (2004),with a non-interacting SDdetrital component contributing to(b) a skew of the distributions to lower coercivities. In (a), the relative abundances of the EM1 and EM2 components are plotted on the vertical axes versus the median coercivity of thecentral ridge (horizontal). The coercivity spectra shown in (b) are for the extreme BS (blue) and BH (green) end-members indicated in (a). Themodal coercivities are also indicated witharrows in (b), whose values are close to the Brh values indicated in Fig. 9. The labels “A” and “B” in (a) correspond to the samples illustrated in Fig. 8a and b, respectively.

125A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

Hesse (1994), Yamazaki and Kawahata (1998) and Lean and McCave(1998) who reported stratigraphic variations between the relative pro-portions of elongated and equant particles associated with time-varying delivery of organic matter to the seafloor. Such possibilitieshave yet to be tested with FORC diagrams, but appear to be a highlypromising avenue for future investigations.

In addition to FORC diagrams for Southern Ocean ODP holes (Fig. 10),results from a range of Tethyan pelagic carbonates (Fig. 1c) are shown inFig. 12. In all cases, a clear central ridge signature is evident. Collectively,our data demonstrate the widespread preservation of biogenic magnetitein ancient pelagic carbonates.

5.5. Ferromagnetic resonance (FMR) spectroscopy

FMR spectroscopy is nowwidely used to detect the narrow size dis-tribution of SD particles and the strong magnetic anisotropy associatedwith magnetosome chain structures within bulk sediment samples(Weiss et al., 2004; Kopp et al., 2006a,b; Fischer et al., 2008;Mastrogiacomo et al., 2010; Charilaou et al., 2011; Kind et al., 2011;Gehring et al., 2011; Roberts et al., 2011, 2012; Chang et al., 2012b).FMR spectra from representative Eocene to Pleistocene pelagic carbon-ate samples (Fig. 13) often contain 6 intense and 10 weak lines due toMn2+ in calcite (e.g., Boughriet et al., 1992; Otamendi et al., 2006).These high-frequency signals were filtered using a Fast Fourier Trans-form e.g., Fig. 13a. The spectra all contain two clear maxima at lowfields, are asymmetrical toward lower fields, and have a large negativepeak at high fields (Fig. 13c–h). The asymmetry and the shift of thespectra to low fields compared to non-biogenic magnetite are charac-teristic of intact magnetosome chains (Weiss et al., 2004; Kopp et al.,2006a,b; Fischer et al., 2008; Mastrogiacomo et al., 2010; Charilaouet al., 2011; Kind et al., 2011; Gehring et al., 2011; Roberts et al., 2011,2012; Chang et al., 2012b), as are parameters (Fig. 13b) derived fromthe spectra (Weiss et al., 2004; Kopp et al., 2006a,b). FMR parameters

Fig. 10. FORC results for Paleocene–Oligocene pelagic carbonate samples from the Southern Owith vertical (interaction field; middle) and horizontal (coercivity; right) profiles through thearound the central ridge defines the 0.05 significance level (Heslop and Roberts, 2012a), whilefor the FORC distributions. The dominance of the central ridge signature (left), that lacks vertic2010). The coercivity profiles (right) are different from those presented by Roberts et al. (2011which fits a skewed Gaussian distribution to the FORC distribution, whereas that of Heslop andmakes it possible to unmix the profiles along the central ridge to determine the respective con

for some localities discussed here are listed elsewhere (Roberts et al.,2011, 2012; Larrasoaña et al., 2012).

5.6. Low-temperature magnetic measurements

Magnetic structures within minerals can be highly dynamic at lowtemperatures, which make low-temperature measurements useful foridentifyingmagnetic minerals and grain size variations. In particular, dif-ferences between FC and ZFC low-temperature SIRMwarming curves areused to detect intact magnetofossil chains (Moskowitz et al., 1993). TheδFC/δZFC ratio is the normalised difference between themagnetisation be-fore and after warming through the Verwey transition at ~110K for eachcurve. Moskowitz et al. (1993) argued that this ratio will exceed 2 for in-tact unoxidised magnetite magnetosome chains and that chain disrup-tion or oxidation of magnetosome surfaces to maghemite will reducethe ratio to ~1. Such values have been observed in several studies (e.g.,Moskowitz et al., 1993; Smirnov and Tarduno, 2000; Passier andDekkers, 2002; Housen and Moskowitz, 2006), although Carter-Stiglitzet al. (2004) concluded from numerical models that the δFC/δZFC ratiowill have a maximum value of ~1–6 for intact magnetite magnetosomechains, and that it can exceed 2 for slightly oxidised intact chains. InFig. 14, we show typical results for samples that are known to contain in-tact magnetofossil chains. Samples for which there is clear evidence of aVerwey transition have distinctly different ZFC and FC magnetisationcurves with δFC/δZFC ratios that approach 2 (Fig. 14a, b). In contrast, sim-ilar samples give rise to low-temperature magnetisation curves with δFC/δZFC values that approach 1 (Fig. 14c,d), which reflects surficial oxidationof magnetosomes (Roberts et al., 2012). The fact that low-temperaturemeasurements simultaneously provide positive and inconclusive evi-dence for magnetofossils (Fig. 14) verifies the well-known limitationsof the test of Moskowitz et al. (1993). Smirnov and Tarduno (2000) con-cluded that surficial maghemitisation (cf. Özdemir et al., 1993; Cui et al.,1994; Özdemir and Dunlop, 2010) of magnetosomes and reductive

cean (reprocessed from Roberts et al., 2011, 2012). High-resolution FORC diagrams (left),peak of the central ridge that dominates the FORC distributions. The black contour linethe grey shading around the profiles represents the respective 95% confidence intervalsal spread (middle), indicates a dominant contribution from magnetofossils (cf. Egli et al.,, 2012) who used the algorithm of Egli et al. (2010) to calculate their FORC distributions,Roberts (2012a)makes no such assumptions. The variable shape of the coercivity profilestributions of biogenic hard and soft components (see Fig. 11).

(a)

(b)

(c)

ρ / ρ

max

00.20.40.60.81

0

0.2

0.4

0.6

0.8

1.0

1.2

-10 0 100

0.2

0.4

0.6

0.8

1

Bc [mT]

Bu

[mT

]

0 10 20 30 40 50 60 70 80 90 100 110-15-10

-505

1015

Bc [mT]0 10 20 30 40 50 60 70 80 90 100 110

ρ / ρ

max

Bu [mT]

ρ / ρ

max

Dolomites, Italy (P3)

Monte Cagnero, Italy (63.5 m)

Argolis Peninsula, Greece (AR1-3)

Bc [mT]

Bu

[mT

]

0 10 20 30 40 50 60 70 80 90 100 110-15-10

-505

1015

ρ / ρ

max

00.20.40.60.81

ρ / ρ

max

00.20.40.60.81

-10 0 100

0.2

0.4

0.6

0.8

1

Bu [mT]

ρ / ρ

max

0

0.2

0.4

0.6

0.8

1.0

Bc [mT]0 10 20 30 40 50 60 70 80 90 100 110

ρ / ρ

max

-10 0 100

0.2

0.4

0.6

0.8

1

Bu [mT]

ρ / ρ

max

0

0.2

0.4

0.6

0.8

1.0

Bc [mT]0 10 20 30 40 50 60 70 80 90 100 110

ρ / ρ

max

Bc [mT]

Bu

[mT

]

0 10 20 30 40 50 60 70 80 90 100 110-15-10

-505

1015

Fig. 12. FORC results for pelagic carbonate samples from the Tethys Ocean (see Fig. 10 caption for details). The dominance of the central ridge signature (left), with lack of vertical spread ofthe FORC distribution (middle) indicates a dominant contribution frommagnetofossils (cf. Egli et al., 2010). (a) Upper Triassic carbonate (sample AR1-3) from Argolis Peninsula, Greece,(b) carbonate sample (from the 63.5m level) from within the Middle Eocene Climate Optimum at Monte Cagnero, Italy, and (c) Cretaceous limestone (sample P3) from Puez, the Dolo-mites, Italy.

126 A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

diagenesis control the low-temperature magnetic properties of pelagiccarbonates. Under oxic conditions, surface oxidation dominates, whileburial of oxidised magnetofossils under iron-reducing diagenetic condi-tions causes reduction of the oxidised maghemite skin to leave puremagnetite (e.g., Torii, 1997; Smirnov and Tarduno, 2000; Yamazaki andSolheid, 2011). Low-temperature results, therefore, have been arguedtoprovidemore information about sedimentary oxidation/reductionpro-cesses than about the distribution of biogenic magnetite (Smirnov andTarduno, 2000).

While low-temperature measurements are often made for pelagiccarbonate samples, they are often inconclusive. FORC and FMR analysesare, therefore, more routinely useful for detecting fossil magnetosomechains. However, Chang et al. (in press) reported that LTC measure-ments of a RTSIRM for magnetofossil-bearing samples is less affectedby surface oxidation; the strong magnetic anisotropy of intactmagnetofossil chains gives rise to reversible LTC curves (Fig. 15) thatprovide a more diagnostic test for biogenic magnetite than that ofMoskowitz et al. (1993). Humped LTC curves (Fig. 15k, l) are knownto occur for oxidised magnetite (Özdemir and Dunlop, 2010); however,LTC-RTSIRM curves for samples containing intact magnetofossil chainsare reversible, a phenomenon that is attributed to the strong uniaxialanisotropy of the chain arrangement in biogenic magnetite (Changet al., in press). The magnitude of the hump in these curves varieswith degree of oxidation, but the curves are reversible even for slightlyto heavily oxidised samples (Fig. 15h–n). Numerical models predict thesamebehaviour for any uniaxial SDmagnetite, whether it is inorganic orbiogenic, in chains or isolated (Carter-Stiglitz et al., 2004). The value ofthis method for recognizing biogenic magnetite, therefore, lies in thefact that SD magnetite within sediments usually has a biogenic origin.It should produce ambiguous results when other types of SD magnetiteare present. Further testing of the diagnostic value of the method ofChang et al. (in press) is needed. However, at face value, their test isnot compromised by oxidation, which is a key factor that limits the di-agnostic value of the Moskowitz et al. (1993) test. Crucially, the test of

Chang et al. (in press) appears to be useful for a wide range of degreesof oxidation of biogenic magnetite within pelagic carbonates.

5.7. Transmission electron microscope observations

TEM observations enable direct imaging to confirm the presence ofthe distinct morphologies associated with biogenic magnetite particles.Under ideal conditions, it is also possible to confirm the alignment of bio-genicmagnetite particles in chains (although these distinctive structuresare often destroyed during sample preparation). Characteristic featuresof biogenic magnetite particles from TEM observations of Middle Eocenepelagic carbonate sediments from ODP Hole 711A, northern IndianOcean, are illustrated in Fig. 16. They include a collapsed chain(Fig. 16a), short chain fragments (Fig. 16b, c), and aggregates of biogenicmagnetite particles with variable morphologies from collapsed chains(Fig. 16d). TEM images for other localities discussed in this paper are re-ported elsewhere (e.g., Roberts et al., 2011, 2012; Larrasoaña et al., 2012;Chang et al., 2012a). Collectively, these observations confirm the in-terpretation that the central ridge feature in FORC diagrams (Egliet al., 2010) in these sediments is due to intact magnetofossilchains. We, therefore, use FORC diagrams to detect the widespreadpresence of magnetite magnetosome chains in pelagic carbonates.As suggested above in relation to central ridge coercivity spectra(Figs. 10–12), there is considerable potential for assessing the rela-tive contributions of BH and BS components in pelagic sediments.Such assessments need to be confirmed by extensive TEM observa-tions, but, if successful, they could lead to magnetic identification ofenvironmental signals associated with different magnetosomemorphologies.

In addition to conventional magnetofossils, Chang et al. (2012a) re-ported the presence of giant magnetofossils in pelagic PETM sedimentsfrom the Southern Ocean (Fig. 17a–c). The only previous report of theseexotic magnetofossil morphologies, which include rod-shaped, spindle-shaped and spearhead forms, is from shallow marine sediments from

0

8x108

1x109

1.2x109

1.4x109

Abs

orpt

ion

geff = h BBeff

B = Bhigh + Blow

A = Bhigh / Blow

Blow Bhigh

Beff

Field (mT)

4x108

6x108

2x108

(a) (b)

Der

ivat

ive

of a

bsor

ptio

n0

1x106

-1x106

-2x106

2x106

Field (mT)

geff = 2.04BFWHM = 174 mT

A = 0.70 = 0.29

0

ODP 758C-1H-1-15.5

geff = 1.99BFWHM = 148 mT

A = 0.67 = 0.26

0

ODP 711A-20-30-040

Der

ivat

ive

of a

bsor

ptio

n

121.44 mbsf

ODP Hole 689D

82.26 mbsf

Der

ivat

ive

of a

bsor

ptio

n ODP Hole 690C

22.20 mbsf

Der

ivat

ive

of a

bsor

ptio

n ODP Hole 738B

284.06 mbsf

ODP Hole 738C

Der

ivat

ive

of a

bsor

ptio

n

geff = 1.99BFWHM = 142 mT

A = 0.87 = 0.29

geff = 2.01BFWHM = 153 mT

A = 0.68 = 0.27

geff = 2.00BFWHM = 136 mT

A = 0.78 = 0.27

geff = 2.01BFWHM = 142 mT

A = 0.61 = 0.24

(c) (d)

(e) (f)

(g) (h)

Der

ivat

ive

of a

bsor

ptio

n

Der

ivat

ive

of a

bsor

ptio

n

0 100 200 300 400 500 600 700 0 100 200 300 400 500 600 700

0 100 200 300 400 500 600 700

0 100 200 300 400 500 600 700

0 100 200 300 400 500 600 700

0 100 200 300 400 500 600 700

0 100 200 300 400 500 600 7000 100 200 300 400 500 600 700

Field (mT)

Field (mT)

Field (mT)

Field (mT)

Field (mT)Field (mT)

Fig. 13. FMR spectra for representative samples from a range of pelagic carbonate samples (see Fig. 1c for locations). (a) Typical spectrum for a sample containing magnetitemagnetofossils. The shaded line is the measured FMR spectrum; the black curve fit is after filtering of the high frequency signal using a Fast Fourier Transform (which has not distortedthe signal). (b) Idealised (smoothed) FMR absorption spectrumwith definition of parameters commonly used to assess the presence ofmagnetitemagnetofossils (fromWeiss et al., 2004;Kopp et al., 2006a). FMR spectra for samples from ODP Hole (c) 689D and (d) 690C, both fromMaud Rise, Southern Ocean (Florindo and Roberts, 2005; Roberts et al., 2012); (e) 711A,northern Indian Ocean (Savian et al., 2013); (f) 738B and (g) 738C, Kerguelen Plateau, Southern Ocean (Roberts et al., 2011; Larrasoaña et al., 2012); and (h) 758C, Ninety East Ridge,Indian Ocean (unpublished). In all cases, FMR parameters are consistent with those of intact magnetosome chains (see tables in Roberts et al. (2011, 2012), Larrasoaña et al. (2012)for details). Each spectrum is normalised by sample mass.

127A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

New Jersey, USA (Schumann et al., 2008). Demonstration that thesegiant magnetofossils occur in both hemispheres and in contrasting ma-rine environments suggests a wide geographic, probably global, rangefor the organisms that produced these biogenic magnetite particles.Chang et al. (2012a) also reported previously undescribed giant

bullet-shaped magnetofossils from the Middle Eocene Climatic Opti-mum in the Indian Ocean (Fig. 17d) and from the PETM in the SouthernOcean. Together, these observations suggest an association betweenwarm hyperthermal climates and magnetofossil gigantism (Changet al., 2012a). A link between increased temperatures and increased

(a) (b)

(c) (d)

1.4 x 10-4

1.6 x 10-4

1.8 x 10-4

2.0 x 10-4

2.2 x 10-4

2.4 x 10-4 690C-7H-3-22 cm(57.82 mbsf)

0 50 100 150 200 250 300

1.5 x 10-3

2.0 x 10-3

2.5 x 10-3

3.0 x 10-3

3.5 x 10-3

689D-6H-2-65 cm(68.55 mbsf)

738C-11R-1-135 cm(284.75 mbsf)

1.8 x 10-3

1.6 x 10-3

1.4 x 10-3

1.2 x 10-3

1.0 x 10-3

1.8 x 10-3

1.6 x 10-3

1.4 x 10-3

1.2 x 10-3

1.0 x 10-3

738C-11R-1-91 cm(284.31 mbsf)

Temperature (K)

δFC/δZFC = 1.85 δFC/δZFC = 1.73

δFC/δZFC = 1.09 δFC/δZFC = 1.07

0 50 100 150 200 250 300

Temperature (K)

Mr(

Am

2 kg-1

)M

r(A

m2 k

g-1)

Mr(

Am

2 kg-1

)M

r(A

m2 k

g-1)

Fig. 14. Representative FC/ZFC low-temperature spectra for pelagic carbonates from the Southern Ocean (from Roberts et al., 2012). (a, b) PETM samples from ODP Hole 738C, KerguelenPlateau. A strong Verwey transition due to magnetite and δFC/δZFC ratios close to 2 are consistent with the test of Moskowitz et al. (1993) for identifying magnetofossils. (c, d) Eocene–Oligocene samples from ODP Holes 689D and 690C, Maud Rise. Lack of a Verwey transition and a difference between FC and ZFC curves mean that these samples fail the Moskowitzet al. (1993) test despite having other properties that indicate the presence of magnetofossils (Figs. 9–12). This failure is probably due to surficial oxidation of biogenic magnetite tomaghemite.

128 A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

abundance of magnetotactic bacteria, but with decreased species diver-sity, has recently been suggested (Lin et al., 2012). Other environmentalchanges associated with global warming could also be important forbiomineralisation of giant magnetofossils, so further work is needed totest this possible relationship in the geological record. Micromagneticmodelling indicates that all of the giant magnetofossil morphologies,except the rods, would have non-ideal magnetic structures and are,therefore, unlikely to have been used for magnetotaxis (navigation)(Chang et al., 2012a). Future TEMwork in search of giantmagnetofossilsin hyperthermals and through the background greenhouse climates ofthe Paleogene has the potential to address many questions surroundingthe stratigraphic extent, biological function and origin of these exoticmagnetic particles.

6. Origin of magnetic minerals in pelagic marine carbonates

When discussing the origin of magnetic minerals in pelagic carbon-ates, it is worth noting that similarmagnetic properties to those report-ed here are observed inmany biosiliceous pelagic sediments, which areoften intercalated with carbonates (e.g., Yamazaki, 2012; Yamazakiet al., 2013). Pelagic biosiliceous sediments are widespread in the equa-torial Pacific, North Pacific and Southern Oceans. The observationsmade in this paper are, therefore, likely to also apply to other pelagic en-vironments with moderate, but not high, primary productivity. Sedi-ment magnetic properties of high productivity zones are likely to becontrolled by early diagenetic magnetic mineral dissolution (seeSection 4). The same range of magnetic minerals has also been reportedfrom pelagic red clays (Yamazaki and Ioka, 1997), which occur widelyin the lowest productivity zones of the North and South Pacific Oceans,and are devoid of carbonate and biogenic opal.

6.1. Biogenic magnetite

After magnetotactic bacteria were discovered (Blakemore, 1975), itwas expected that biogenic magnetite would be widely responsiblefor sedimentary paleomagnetic signals (Kirschvink, 1982; Petersenet al., 1986; Stoltz et al., 1986; Chang et al., 1987; Vali et al., 1987). Itwas later recognised thatmagnetotactic bacteria are gradient organismsthat live within highly chemically stratified environments around theoxic–anoxic transition zone (OATZ), which can occur within the watercolumn or sediment (e.g., Bazylinski and Frankel, 2004; Faivre andSchüler, 2008). Recognition that magnetite dissolves when buriedbelow the OATZ (Karlin and Levi, 1983; Canfield and Berner, 1987;Karlin, 1990), and the resultant poor prognosis for magnetofossil pres-ervation in the geological record (e.g., Vali and Kirschvink, 1989), wasonly appreciated at about the same time. Therefore, while there havebeen occasional reports over the last 30 years of well-preservedmagnetofossils in ancient shallow water and pelagic carbonate sedi-ments (e.g., Petersen et al., 1986; Chang et al., 1987; McNeill et al.,1988; Vali and Kirschvink, 1989; Aissaoui et al., 1990; Hounslow andMaher, 1996; Belkaaloul and Aissaoui, 1997; Montgomery et al.,1998), it was only relatively recently that Kopp and Kirschvink (2008)concluded that the pre-Quaternary magnetofossil record is sparse andthat more magnetofossil-bearing localities have been identified in theQuaternary than in the rest of the geological record. However, with re-cent development and widespread use of magnetic techniques thatare well suited to detecting the distinctive magnetic properties of intactmagnetofossil chains (e.g., Weiss et al., 2004; Kopp et al., 2006a,b; Egliet al., 2010), the catalogue of ancient magnetofossil occurrences, partic-ularly in pelagic carbonates, has expanded significantly (e.g., Yamazaki,2008; Abrajevitch and Kodama, 2009; Yamazaki, 2009; Roberts et al.,2011, 2012; Yamazaki, 2012; Chang et al., 2012a; Larrasoaña et al.,

1.0

0.5

0.6

0.7

Nor

mal

ized

rem

anen

ce

(d) 738B-5H1-130 (33.80 m)

FCZFC

0.8

0.9deriv . FC

1.0

0.5

0.6

0.7

0.8

0.9

1.0

0.5

0.6

0.7

0.8

0.9

1.0

0.5

0.6

0.7

0.8

0.9

(a) 738B-3H5-115 (20.65 m)

(b) 689D-6H2-65 (68.55 m)

(c) 738C-11H1-40 (283.80 m)

Derivative of F

C curve

0.000

0.002

0.004

0.000

0.002

0.004

0.000

0.002

0.004

0.000

0.002

0.004

1.12

1.00

1.04

1.08

Nor

mal

ized

rem

anen

ce 1.12

1.00

1.04

1.08

1.12

1.00

1.04

1.08

1.12

1.00

1.04

1.08

coolingwarming

(k) 738B-5H1-130 (33.80 m)

(h) 738B-3H5-115 (20.65 m)

(i) 689D-6H2-65 (68.55 m)

(j) 738C-11H1-40 (283.80 m)

0 50 100 150 200 250 300

Temperature (K)

Maghemite(γ -Fe2O3)

Magnetite(Fe3O4)

Ca

tion

defic

ien

t m

agne

tite

(Fe

3-z

ΔzO

4)

Incr

easi

ngo

xida

tion

(z )

(o)

(f) Magnetobacterium bavaricum MYR-1

(g) Magnetospirillum gryphiswaldense MSR-1 (n) Magnetospirillum gryphiswaldense MSR-1

1.0

0.4

0.6

0.8 0.004

0.002

0.000

0.006(e) Magnetospirillum magneticum AMB-1 (l) Magnetospirillum magneticum AMB-1

1.04

0.92

0.96

1.00

0.006

0.004

0.000

0.008

0.002

1.0

0.5

0.6

0.7

0.8

0.9

1.10

0.98

1.02

1.06

Tv = 96 K

Tv = 99 K

Tv = 101 K

1.10

0.98

1.02

1.06

1.0

0.5

0.6

0.7

0.8

0.9

0.02

0.01

0.00

0.03

0 50 100 150 200 250 300

Temperature (K)

Tv = 103 K

(m) Magnetobacterium bavaricum MYR-1

Fig. 15. Systematic illustration of the low-temperature magnetic properties of pelagic carbonates and fresh cells from magnetotactic bacteria. (a–g) Low-temperature SIRM warming curvesmeasured in zero field after ZFC and FC treatments. (h–n) Low-temperature cycling of RTSIRM in zero field. Pelagic carbonate samples are from ODP Holes 738B (Roberts et al., 2011),689D (Roberts et al., 2012), and738C (Larrasoaña et al., 2012). Samples that containwholemagnetotactic bacteria cellswith biogenicmagnetite chains include (e, l)wild-typeMagnetospirillummagneticum AMB-1 (Kopp et al., 2006b), (f, m) uncultivatedMagnetobacterium bavaricumMYR-1 (Li et al., 2010), and (g, n) culturedMagnetospirillum gryphiswaldenseMSR-1 (Fischer et al.,2008; Scheffel et al., 2008). The low-temperature properties can be divided into fourmain groups depending on the degree of oxidation: (a, b, h, i) fully oxidised, (c, j) highly oxidised, (d, e, k, l)moderately or slightly oxidised, and (f, g, m, n) nearly fresh biogenic magnetite within whole magnetotactic bacteria cells. Images on the right are inferred degrees of oxidation of magnetiteparticles. After Chang et al. (2013).

129A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

200 nm 200 nm

100 nm 200 nm

(a) (b)

(c) (d)

Fig. 16. Transmission electron microscope (TEM) images of magnetofossils extracted from Eocene sediments at ODP Hole 711A, northern Indian Ocean. In many cases (a–c), short fossilmagnetosome chain segments, or collapsed chain segments, have been preserved despite the extraction process. (d) In other cases, chain structures were destroyed by the sample prepara-tion process. TEM images from other sequences discussed in this paper are presented elsewhere (e.g., Roberts et al., 2011; Chang et al., 2012a; Larrasoaña et al., 2012; Roberts et al., 2012).

130 A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

2012; Yamazaki and Ikehara, 2012; Channell et al., 2013). Inmany cases,these occurrences have been validated by TEM imaging of morphologi-cally distinctive bacterial magnetite particles (e.g., Roberts et al., 2011,2012; Chang et al., 2012a; Larrasoaña et al., 2012; Yamazaki, 2012;Yamazaki and Ikehara, 2012), which give rise to high scores in themagnetofossil scoring system of Kopp and Kirschvink (2008).Magnetofossils can now, therefore, be considered awidespread contrib-utor to the magnetisation of pelagic marine carbonates rather than arare magnetic mineral component.

Widespreadmagnetitemagnetofossil preservationmeans that therewas an ancient redox gradient in the respective pelagic environments,which supported magnetotactic bacterial populations. However, thatgradient could not have been sharp enough to cause subsequent burialunder anoxic conditions in which biogenic magnetite will dissolve.Roberts et al. (2011) suggested that magnetofossil abundances arehigher when organic carbon delivery to the seafloor is sufficient to trig-germild iron reduction that releases iron from themost reactive detritalminerals so that it becomes bioavailable for magnetotactic bacterialbiomineralisation. These conditions can be met by supply of detritaliron from eolian dust, which stimulates oceanic primary productivityin iron-limited oceanic environments and increases delivery of organiccarbon and reactive iron to the seafloor, both of which are required formagnetite biomineralisation by magnetotactic bacteria (Roberts et al.,2011). Various paleoredox indicators from the studied bulk sedimentsindicate that they never became anoxic (Chambers and Cranston,1991) and hence corrosive to the biogenicmagnetite produced throughthismechanism. The potential conundrumassociatedwith preservationversus destruction of biogenic magnetite is discussed further inSection 9.

6.2. Detrital magnetic minerals

Detrital magnetic minerals, particularly magnetite and hematite,have long been assumed to be the dominant carriers of paleomagneticsignals in pelagic marine carbonates (Lowrie and Heller, 1982;Freeman, 1986). It has also long been known that pelagic carbonatesoften have magnetic properties that are dominated by SD magnetite(Lowrie and Heller, 1982). This observation is now best explained bythe widespread occurrence of biogenic magnetite, as described above.However, detrital magnetite and hematite are still routinely observedin analyses that enable identification of magnetic mineral componentswithinmixedmagneticmineral assemblages, particularly in IRM compo-nent analysis (Fig. 8), hysteresis unmixing analysis (Fig. 9), and FORC di-agrams (Figs. 10, 12). This result is not surprising. However, along withrecent recognition of the importance ofmagnetitemagnetofossils in con-tributing to themagnetisation of pelagic carbonates, the concentration ofdetrital magnetic minerals often undergoes stratigraphic variations thatcoincide with variations in the concentration of biogenic magnetite(e.g., Roberts et al., 2011). This raises questions about whether there isa relationship between the concentrations of detrital and biogenic min-erals. Roberts et al. (2011) suggested that supply of detrital iron from eo-lian dust stimulates oceanic primary productivity in iron-limited surfacewaters and increases delivery of organic carbon and reactive iron to theseafloor, both of which are required for magnetite biomineralisation bymagnetotactic bacteria, and which can explain coincident variations indetrital and biogenic magnetic minerals (see also Larrasoaña et al.,2012). Glacial–interglacial variations in concentration-dependent mag-netic properties have been known of for many years in different oceanbasins (e.g., Amerigian, 1974; Kent, 1982; Robinson, 1986; Bloemendal

Fig. 17. TEM images of giantmagnetitemagnetofossils from the PETM(Hole 738C) in the SouthernOcean and from theMiddle Eocene Climatic Optimum(MECO; Hole 711A) in the IndianOcean (from Chang et al., 2012a). (a) Spindle morphology from the pre-PETM interval, (b) an irregularly shaped magnetite crystal from the PETM interval, (c) rod-shaped morphologyfrom after the PETM interval, and (d) giant bullet morphology from MECO.

131A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

and deMenocal, 1989). In the Southern Ocean, coincidences betweensedimentarymagnetic susceptibility records and eolian dust flux recordsfrom Antarctic ice cores (Lambert et al., 2008) have often been observedat millennial as well as glacial–interglacial timescales (e.g., Petit et al.,1990; Bareille et al., 1994; Pugh et al., 2009; Weber et al., 2012). How-ever, dust fluxes recorded in ice cores are too low to explain the mag-netic susceptibility values documented in Southern Ocean sediments.Yamazaki and Ikehara (2012) demonstrated that the eolian ironfertilisation hypothesis (Roberts et al., 2011) can explain observedmag-netic variations in Southern Ocean sediments and the co-varying con-centrations of biogenic and detrital magnetic minerals. Thus, despitethe common assumption that detrital magnetic minerals are responsi-ble for the magnetisation of pelagic carbonates, and recognition thatmagnetic mineral concentration variations are often paleoclimaticallymodulated, deeper understanding of the drivers of magnetic mineralvariations is opening new avenues of investigation in understandingenvironmental signals in pelagic carbonates.

6.3. Authigenic magnetic minerals

Authigenic magnetic minerals have been reported to form in pelagiccarbonates during early burial, later diagenesis, and subaerialweathering.For example, authigenic hematite or goethite formation can be importantin oxic diagenetic environments. Early diagenetic dissolution of detritalmagnetic minerals is common in reducing environments with relativelyhigh organic carbon fluxes (e.g., Tarduno, 1992, 1994). Pyrite is paramag-netic and does not carry a remanent magnetisation, but it is the mostcommon authigenic mineral to form in anoxic sediments (Fig. 5); itforms at the expense of detrital magnetic minerals and has been com-monly reported in pelagic carbonates (e.g., Lowrie and Heller, 1982).

When such limestones are tectonically uplifted and subjected to subaerialweathering, the pyrite often oxidises to form goethite that gives rise to apersistent secondary magnetisation that is extremely resistant to AFdemagnetisation (Lowrie and Heller, 1982). Monoclinic pyrrhotite isstronglymagnetic, has been observed to occur in somepelagic carbonates(Lowrie and Heller, 1982), and is often problematic in paleomagneticstudies of sediments. Formation of monoclinic pyrrhotite is extremelyslow at ambient temperatures, so its diagenetic formation will mean, bydefinition, that it records a late diagenetic remagnetisation (e.g.,Weaver et al., 2002; Horng and Roberts, 2006; Larrasoaña et al., 2007).Pyrrhotite is a stable carrier of thermal remagnetisations in pelagic car-bonates that have been subjected to low-grade regional metamorphismin orogenic belts (e.g., Appel et al., 2012), largely as a result of thermal al-teration of pyrite to pyrrhotite (Rochette, 1987). Late diagenetic magne-tite formation also gives rise to widespread remagnetisation in pelagiccarbonates, as discussed in Section 8. Regardless of the mechanism thatcauses post-depositional authigenic magnetic mineral formation, pelagiccarbonates often provide outstanding records of ancient geomagneticfield behaviour. In this paper, we focus on the magnetic properties of pe-lagic carbonates that have not been extensively diagenetically modified,while recognizing that authigenic magnetic mineral formation can beimportant.

6.4. Exotic magnetic particles

In addition to thewidespread presence of biogenic and detritalmag-neticminerals, and the occasional presence of authigenicmagneticmin-erals, exotic magnetic minerals have been demonstrated or inferred tobe presentwithin pelagic carbonates. For example, global extraterrestri-al fallout causes deposition of tons of material per day, which is unlikely

132 A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

to be magnetically detectable except in environments with weak back-groundmagnetisations such as ice (e.g., Lanci et al., 2012) or carbonates.Freeman (1986) reported cosmic microspherules in pelagic carbonates,which demonstrates that such exotic particles can be present. Volcanicparticles are a less exotic, but more likely, source of magnetic particlesin pelagic carbonate environments. Heider et al. (1993) and Touchardet al. (2003) inferred the presence of fine-grained detrital magneticminerals from discrete or disseminated volcanic ash layers in regionsdown-wind of explosive volcanoes. While volcanism is highly likely tocontribute to the magnetisation of many pelagic carbonates, the postu-lated presence of SDmagnetite from disseminated volcanic ash in sedi-ments surrounding Kerguelen Plateau in the Southern Ocean (Heideret al., 1993) has been demonstrated to be due to biogenic magnetite(Roberts et al., 2011). Likewise, claims that SD magnetite particles pro-duced by fallout of comet impact condensates during the PETM (Kentet al., 2003) have also been argued to result frommisidentified biogenicSDmagnetite (Lippert and Zachos, 2007; Kopp et al., 2007; Chang et al.,2012a). These cases demonstrate the importance of direct imaging todiagnose the processes that produce SD magnetite occurrences in sedi-ments. Nevertheless, the concentration of exotic magnetic particles islikely to be low, and would normally go undetected in most sedimenttypes, including pelagic carbonates. Efforts to image exotic biogenicmagnetite particles in the coming years are likely to improve our under-standing of giant magnetofossils and their relationship to ancienthyperthermal events.

7. Paleomagnetic recording in pelagic marine carbonates

The conventional paradigm concerning the origin of sedimentary pa-leomagnetic signals involves the depositional remanent magnetisation(DRM) or post-depositional remanent magnetisation (PDRM) concepts(e.g., Johnson et al., 1948; King, 1955; Irving and Major, 1964; Kent,1973; Verosub, 1977; Tauxe et al., 2006). A DRM is invoked when mag-netic particles rotate freely in alignment with the geomagnetic field inaqueous solution and settle on the sedimentary substrate and, on aver-age, retain a record of the ambient field at the time of deposition(Fig. 18a). In contrast, a PDRM is invoked when the sedimentary

export to b

Input of magnetic m

dilutesuspension

0 100

co

(a) DRM (b) PDRM

water

slurry

consolidatedsediment

Particleconcentration (%)

nepheloidlayer

benthicboundary

layer

Percentage PDRM locked in

100 0

no lock-in

PDRM lock-inzone

Geomagnetic field direction

B

Fig. 18. Schematic illustrationof sedimentary remanence acquisition through (a)depositional remaPDRM alongside a biogeochemical remanent magnetisation (BRM). (a) DRM acquisition is morflocculating environments (Katari and Tauxe, 2000), such as lakes,where particles act essentially inalignment ofmagnetic particles, andwhere themagnetisation is progressively locked-in over someWinklhofer, 2004). Modified from Roberts et al. (2013). Flocculation of particles will occur formagnetotactic bacteria live in the water column or within the surface mixed layer, their inorga1998) is envisaged when magnetotactic bacteria live within the sediment, but below the surface(1993). Detrital particles are envisaged to contribute to a PDRM, while the biogenic magnetite cthat a BRMwill only occur undermicroaerobic conditions (Blakemore et al., 1985) rather than in twere suitable for magnetotactic bacteria to live over expanded thicknesses of microaerobic sedimnutrient availability with increasing depth are inferred to constrain magnetotactic bacteria to livemagnetite is likely to dissolve within the anoxic part of the sediment column, which will lead to p

environment is disturbed, for example, by bioturbation, and any DRMis disrupted, with magnetic particles realigned post-depositionally bygeomagnetic torques within the relatively unconsolidated uppermostsediment (Fig. 18b). PDRM is argued to progressively lock in withcompaction (Irving and Major, 1964; Kent, 1973; Verosub, 1977;deMenocal et al., 1990; Roberts and Winklhofer, 2004; Suganuma et al.,2011). As discussed by Tauxe et al. (2006) and Roberts et al. (2013),major uncertainties remain in our understanding of sedimentary rema-nence acquisition despite the fact that this phenomenon has beensubjected to experimental, numerical and theoretical investigation forover 60 years. These uncertainties include issues related to the effectsof inter-particle flocculation, the lack of definitive and widespread con-firmation of the role of any sedimentary remanence acquisition mecha-nism, and lack of understanding about how signals related to theintensity of the geomagnetic field are recorded.

Given the evidence provided above concerning the widespread pres-ervation of biogenic magnetite in ancient sediments (e.g., Roberts et al.,2012), and that these particles can dominate the magnetic propertiesofmarine pelagic carbonates, it is important to consider their role in sed-imentary paleomagnetic recording. Tarduno et al. (1998) interpreted de-tailed rock magnetic and geochemical data from Ontong–Java Plateausediments as indicating the presence of magnetotactic bacteria that livewell below the surface mixed layer of carbonate sediments. In thiscase, the biogenic magnetite that they produce will not give rise to aDRM or PDRM. Tarduno et al. (1998), therefore, proposed the existenceof a new remanence acquisition mechanism named a biogeochemicalremanent magnetisation (BRM; Fig. 18c). Rigorous assessment of theBRM concept is now urgently required (Roberts et al., 2012, 2013). Thisis important because a BRM will be acquired at greater depths withinthe sediment than a conventionally assumed DRM or PDRM (Fig. 18c).This is likely to cause major paleomagnetic smoothing, loss of geomag-netic information, and potential biasing of information gleaned aboutthe workings of the geodynamo.

When considering the role of biogenicmagnetite in sedimentary pa-leomagnetic recording, the depth habitat of magnetotactic bacteria iscrucially important. If biogenic magnetite is produced in the watercolumn or within the bioturbated sedimentary surface mixed layer,

B

turbulent waterflocculation/pellitization

ottom

inerals ( ), clays ( )

surface mixedlayer

nsolidating sediment

(c) PDRM + BRM

B

surface mixed layer

consolidating sediment

magnetic/hydrodynamic

torques

resuspension/reflocculation/

pellitizationbioturbation

compactionearly diagenesis

export to bottom

Input of magnetic minerals ( ), clays ( )

Percentage PDRM + BRM locked in

100 0

no lock-in

PDRM + BRMlock-inzone

benthicboundarylayer

0 100

Particleconcentration (%)

dilutesuspension

nepheloidlayer

nentmagnetisation (DRM), (b) post-depositional remanentmagnetisation (PDRM), and (c) ae likely to occur where bioturbation does not mix the sediment and in low salinity, non-dependently. (b) PDRMacquisition is envisagedwhen bioturbation disrupts any depositionaldepth range,which is represented herewith an exponential lock-in function (cf. Roberts andthe range of salinities encountered in marine environments (Katari and Tauxe, 2000). Ifnic post-mortem remains will contribute to a PDRM. (c) BRM acquisition (Tarduno et al.,mixed layer. A schematic depth habitat distribution is shown following Petermann and Bleilontributes to a BRM, which is acquired later than the PDRM. In this schematic, we assumehe presence of an OATZ (e.g., Bazylinski and Frankel, 2004). Even if environmental conditionsent, high-fidelity paleomagnetic records from pelagic carbonates and reduced porosity andat relatively shallow depths so that they contribute to a coherent BRM. If an OATZ is present,oor preservation potential for biogenic magnetite.

133A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

themagnetite could be subjected to the same randomizing sedimentaryprocesses and (re-) aligning geomagnetic torques as detrital particles,and should, therefore, contribute to a DRM or PDRM. If magnetotacticbacteria live within the sediment below the surface mixed layer, theirpost-mortem remains will contribute to a BRM (Tarduno et al., 1998).However, as discussed in Section 4, magnetite will dissolve in anoxicsettings and a BRM due to magnetite magnetofossils will not be pre-served. We observe that when biogenic magnetite is preserved at suffi-cient depths in pelagic carbonates to be paleomagnetically important,there is no evidence that the diagenetic environments were ever anoxic(Roberts et al., 2011).We, therefore, assume that a BRMwill only be im-portant undermicroaerobic conditions (Blakemore et al., 1985) that didnot subsequently become anoxic rather than in sediments in which anOATZ was present (e.g., Bazylinski and Frankel, 2004).

Few studies have assessed the BRMconcept, sowe still do not under-stand whether this remanence acquisition mechanism is widely impor-tant. Recent environmentalmagnetic studies of pelagic carbonates havereported that magnetofossil concentrations vary in phase with changesin concentration of detrital magnetic minerals (Larrasoaña et al., 2012;Yamazaki and Ikehara, 2012). In this case, paleomagnetic signals carriedby the biogenic magnetite would not be expected to be offset signifi-cantly from signals carried by detrital magnetic minerals. In contrast,Tarduno et al. (1998) suggested the possibility of major delays in

0.1

1.0

1 10

Laytonville Limestone (Tarduno and Myers, 1994)Maiolica Limestone (Channell and McCabe, 1994)

Southern Ocean Paleogene nannofossil ooze (Roberts et al., 2012)Tahiti Pleistocene Reef (Ménabréaz et al., 2010)

NE USA Paleozoic carbonates (Jackson, 1990)Varied remagnetized carbonates (McCabe and Channell, 1994)Leadville carbonates (Xu et al., 1998)Ordovician/Devonian limestone (Elmore et al., 2006)Neoproterozoic Bitter Springs carbonate, Australia (Swanson-Hysell et al., 2012)

Bcr/Bc

SP saturation envelope

20 nm

15 nm

10 nm

SD+MDmixing curve

20 302 5

Mr/

Ms

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%

70%

60%

50%

40%30%

20%10%

80%

60%

50%

40%

30%40%

30%

Remagnetized carbonates

Carbonates with primary magnetizations

Fig. 19. Comparison of magnetic hysteresis data for carbonates with primarymagnetisations (solid symbols) with those for remagnetised carbonates (open symbols)on a Day plot (Day et al., 1977). Themixing curves are fromDunlop (2002), with percent-ages indicating the contributions from the MD or SP components, respectively. TheSD+ SP mixing curves have sizes indicated for 10, 15 and 20 nm SP magnetite. Sourcesof the respective data sets are indicated in the key (see Jackson and Swanson-Hysell,2012). Carbonates with primary magnetisations: Tarduno and Myers (1994), ChannellandMcCabe (1994), Ménabréaz et al. (2010), Roberts et al. (2012); remagnetised carbon-ates: Jackson (1990), McCabe and Channell (1994), Xu et al. (1998), Elmore et al. (2006),and Swanson-Hysell et al. (2012).

recording of a BRM associated with magnetotactic bacteria that livebelow the surface mixed layer of sediment with respect to that carriedby detrital particles. Lack of clarity about the importance of BRM acqui-sition underlines the urgency of understanding how widely BRMs arepreserved in the geological record.

8. Remagnetisations in pelagic marine carbonates

While we focus on pelagic carbonates that carry what is interpretedto be a primary magnetic signal, it is important to note thatremagnetisations have been widely reported, particularly in ancient(Paleozoic and Mesozoic) carbonates (see reviews by McCabe andElmore (1989), Jackson and Swanson-Hysell (2012) and van der Vooand Torsvik (2012)). Remagnetised carbonates often contain high con-centrations of authigenic fine-grained (SP)magnetite. Although SP parti-cles do not contribute to the natural remanent magnetisation, theirpresence in large concentrations in remagnetised carbonates, in combi-nation with SD particles, produces characteristic hysteresis behaviourthat contrasts with that of non-remagnetised carbonates (Fig. 19). Hys-teresis loops for remagnetised carbonates are often wasp-waisted dueto the mixture of magnetic responses from magnetically soft (SP) andmagnetically hard (SD) minerals (Jackson 1990; Roberts et al., 1995;Tauxe et al., 1996). Hysteresis signatures (Fig. 19) have been argued toprovide a fingerprint for remagnetisation in carbonates (Jackson, 1990;McCabe and Channell, 1994; Channell and McCabe, 1994; Borradaileand Lagroix, 2000; Jackson and Swanson-Hysell, 2012; van der Voo andTorsvik, 2012). Multiple mechanisms have been proposed for carbonateremagnetisations, including continent-scale orogenic fluid flow (Oliver,1986), thermoviscous remanent magnetisation acquisition (Kent,1985), burial to temperatures that cause thermal alteration of pyrite tomagnetite (e.g., Suk et al., 1990a,b; Banerjee et al., 1997) or transforma-tion of iron-rich smectite to iron-free illite+magnetite (McCabe et al.,1989; Katz et al., 1998; Tohver et al., 2008), and pressure solution(Zegers et al., 2003). Exploring these mechanisms is beyond the scopeof this paper; we simply note that remagnetised carbonates usuallyhave distinctive magnetic properties compared to other rocks, includingpelagic carbonates (Fig. 19). Regardless, various processes can producemagnetic hysteresis properties that overlap with those of other samplesin any such simple test, which makes it important to apply a range oftests for remagnetisations.

9. Outstanding questions concerning the magnetisation of pelagicmarine carbonates

As discussed throughout this paper, recent developments haveprompted reevaluation of our understanding of themagnetic propertiesof pelagic carbonates. This reevaluation raises important questions,some of which we explore below. The unresolved nature of these ques-tions indicates that, instead of being a well-worn subject that has beeninvestigated for over 60 years, further work is needed to exploit themagnetic information carried by pelagic marine carbonates.

9.1. Are magnetotactic bacteria always gradient-organisms?

A key question concerning the observed long-term geological pres-ervation of magnetite magnetofossils (Roberts et al., 2012) is how dosuchmagnetic nanoparticles avoid dissolution if their presence is inher-ently linked to the presence of an OATZ? Such particles ought to dis-solve soon after they experience anoxic diagenetic conditions oncethey are buried below the OATZ (cf. Canfield and Berner, 1987). Evi-dence is emerging that some magnetite magnetofossil morphologiesare more resistant than others to dissolution, although the relict con-centration of magnetofossils appears to be small (Kodama et al., inpress). Thus, it is unlikely that the widespread geological occurrenceof magnetofossils is purely a signature of preferential preservation. Itismore likely thatmany such sediments have never become completely

134 A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

anoxic and, therefore, that their constituent magnetic mineral as-semblages have not undergone reductive diagenesis (Roberts et al.,2011). How is this possible if magnetotactic bacteria are gradient or-ganisms that live around the OATZ? With strong recent emphasis onmagnetotactic bacterial habitats being associated with highly stratifiedredox environments near the OATZ (e.g., Bazylinski and Frankel, 2004;Faivre and Schüler, 2008), it is useful to note that Aquaspirillummagnetotacticum (Blakemore et al., 1985) and Magnetospirillumgryphiswaldense (Schüler and Baeuerlein, 1998) require microaerobicconditions for magnetite biomineralisation that are not necessarilylinked to the presence of anOATZ. In pelagic environments in particular,organic carbon supply and sedimentation rate can be so low that reac-tive organic carbon is effectively oxidised by oxic and then suboxic pro-cesses and that diagenetic conditions never become anoxic. Suchsituations have been argued to explain the extensive preservation ofmagnetite magnetofossils in pelagic carbonates (Roberts et al., 2011).This supports the conclusion of Blakemore et al. (1985) that preserva-tion of bacterial magnetite in sediments can provide evidence that sed-imentation (and subsequent diagenesis) occurred under microaerobicconditions. We take the persistence of magnetite magnetofossils overconsiderable stratigraphic thicknesses in pelagic carbonates to indicatethat they have never undergone reductive diagenesis. Furthermore, it islikely that the species preserved in such settings prefer microaerobicconditions rather than the more highly chemically stratified environ-ments around an OATZ.

9.2. Are biogeochemical remanent magnetisations globally important?

Recognition of the widespread presence ofmagnetofossils in pelagiccarbonates and other sediment types (Roberts et al., 2012) means thatthe possibility of sediments acquiring a paleomagnetic signal via a dis-tinct BRM mechanism needs to be thoroughly tested. If a distinct BRMproves to be widespread, for which there is currently only modest evi-dence (Tarduno and Wilkison, 1996; Tarduno et al., 1998; Abrajevitchand Kodama, 2009), it will be necessary to develop a new paradigmfor sedimentary remanence acquisition. The fact that a detailed under-standing of sedimentary remanence acquisition has not been developedin over 60years of sedimentary paleomagnetic research underlines thedifficulty of the problem. However, it has not been until the last decadethatwe have had the tools needed to identify in detail the contributionsof different magnetic mineral components to the magnetisation of sed-iment, and, when used, these tools have largely been applied to envi-ronmental rather than paleomagnetic applications. The time is nowripe to use these tools to assess the relative contributions of biogenicand detrital magnetic minerals to the measured paleomagnetic signalof sediments.

9.3. At what depths do magnetotactic bacteria live?

Determining the redox conditions underwhich ancientmagnetotacticbacteria lived is useful for assessing the potential depth range over whicha BRMmight have been acquired. Natural mineral compositions can varyconsiderably, with different cations substituting for each other within thecrystal lattice. The nature of cation substitution is indicative of the ambi-ent chemical environment. For example, trace metals can be highly solu-ble andmobile under varying redox conditions. Tracemetal incorporationwithinminerals that form in different parts of a zoned redox environmentcan, therefore, provide a proxy measure of the redox state of that part ofthe environment (e.g., Morford and Emerson, 1999; Algeo and Lyons,2006). Uptake of trace metals into biogenic magnetic minerals bymagnetotactic bacteria can then provide a direct link to the part of theredox zonation inwhich the biomineralisation occurred. Testing whetherredox sensitive tracemetals havebeen incorporated into biogenicmagne-tite would help to develop paleoredox proxies that could be useful to pa-leomagnetism. This is best done using magnetotactic bacteria frommodern environments with directly measurable redox stratification.

Synchrotron analysis is ideal for addressing such problems because it pro-vides high-resolution analysis of all relevant elements, even in nanoparti-cles, including determinationof the oxidation state of cations of interest. Italso enables unambiguous assessment of the environmental significanceof trace metal substitution in magnetic biominerals. Synchrotron analysishas not been widely used in environmental magnetism (e.g., Liu et al.,2012), but the few such studies that have been made on magneticbiominerals indicate that highly useful environmental information canbe obtained. For example, synchrotron analyses demonstrate that extra-cellular magnetite produced by iron-reducing bacteria has similar Fe cat-ion site occupancies to abiogenic magnetite, but with oxygen deficiencydue to formation of the magnetite in oxygen-poor environments aroundthe OATZ (Coker et al., 2007). Other studies indicate that both extracellu-larly produced magnetite and magnetite magnetosomes produced bymagnetotactic bacteria contain a higher concentration of Fe2+ than abio-genicmagnetite (Carvallo et al., 2008; Lamet al., 2010), which is also like-ly to indicate the relatively reducing environment inwhich themagnetitewas biomineralised. So far, no studies have assessed the possible presenceof other redox sensitive trace metals in magnetic biominerals. It is, there-fore, timely to undertake synchrotron analyses on magnetic biomineralsto assess whether meaningful paleoredox information can be extractedfrom these minerals.

9.4. Can magnetofossils provide useful paleoproductivity orpaleoenvironmental information?

It has been suggested that sedimentarymagnetofossil concentrationsreflect ancient variations in marine primary productivity (Roberts et al.,2011; Larrasoaña et al., 2012; Yamazaki and Ikehara, 2012). It has alsobeen suggested that the morphology of magnetofossils varies in relationto paleoclimatically-controlled changes in sediment oxygen content,which is controlled by organic carbon delivery to the seafloor, and,hence, by paleoproductivity (Hesse, 1994; Yamazaki and Kawahata,1998; Lean and McCave, 1998). Resolving the question of whethermagnetofossil assemblages can provide information about oceanicpaleoproductivity (in the affirmative) will provide easily measurable,non-destructive magnetic parameters to constrain past variations of amajor component of the global carbon cycle, by using results of thetype shown in Figs. 8–12 that enable discrimination of biogenic “soft”and “hard” (Egli, 2004) magnetofossil components. Rigorous andstraightforward testing of whether meaningful paleoenvironmental sig-nals can be extracted frommagnetofossils needs to be done by determin-ing the concentration and morphological variation of magnetofossilsfrom regions of the global oceanwhere detailed, high-quality sedimenta-ry paleoproductivity records already exist. Proof of concept will requirepainstakingmagnetic extraction of magnetofossils and detailed TEM ob-servations to ground-truth magnetic interpretations. The potential ofthis approach and its implications for ocean biogeochemistry havebeen suggested in various studies over the last 20 years (Hesse, 1994;Yamazaki and Kawahata, 1998; Lean and McCave, 1998; Roberts et al.,2011; Larrasoaña et al., 2012; Yamazaki and Ikehara, 2012). The time isripe to resolve this question.

10. Conclusions

Limestones have been subjected to paleomagnetic analyses sincethe study of Graham (1949). Pelagic carbonates have provided out-standing records of geomagnetic field behaviour that have been widelyuseful in the Earth Sciences, particularly since the 1970s. Nevertheless,recent developments have significantly improved our understandingof themagnetic properties of pelagic carbonates and the paleomagneticand environmental information that they record, and point tomany im-portant questions that remain unanswered. The possible presence ofbiogeochemical remanent magnetisations has the potential to revolu-tionise our understanding of how pelagic carbonates record paleomag-netic signals; we are only in the early stages of understanding this

135A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

important issue. Much more work is needed to test whether this rema-nence acquisition mechanism is globally widespread. Recently devel-oped techniques provide the tools needed to test how differentmagnetotactic bacterial species, which produce different magnetitemorphologies, respond to changing nutrient and oxygenation condi-tions in pelagic environments. Future work needs to test whether it ispossible to develop proxies for ancient nutrient conditions from well-calibrated modern magnetotactic bacterial occurrences. A tantalizinglink between giant magnetofossils and Paleogene hyperthermal eventsneeds to be tested; much remains to be learned about the relationshipbetween climate and the organisms that biomineralised these largeand novel magnetite morphologies. Thus, rather than being a well-worn subject that has been intensively investigated for over 60 years,the magnetic properties of pelagic carbonates hold many secrets thatawait discovery.

Acknowledgements

Weacknowledge use of SeaWiFS data (Fig. 1b) provided by the God-dard Space Flight Center, NASA, which are used in accordance with theSeaWiFS Research Data Use Terms and Conditions Agreement. Wethank Xiang Zhao for making the measurements shown in Fig. 12 andNick Swanson-Hysell for providing much of the data shown in Fig. 19.We are grateful to our home institutions and to several laboratorieswhose facilities we have used as visitors, including the Institute forRock Magnetism, University of Minnesota, USA (supported by the USNational Science Foundation Earth Sciences Division, the Keck Founda-tion and the University ofMinnesota), the Kochi Core Center, Universityof Kochi, Japan, and the Electron Paramagnetic Resonance National Ser-vice Centre, University ofManchester (supported by theUK Engineeringand Physical Sciences Research Council), UK. Our work has benefittedfrom the support of the Australian Research Council (through grantDP120103952), the European Community Marie Curie Actions (FP7-PEOPLE-IEF-2008, no. 236311), and FAPESP (project JP no. 11/22018-3).

We are grateful to Mike Jackson and Toshi Yamazaki for insightfulreview comments, Paul Wignall for efficient editorial handling, andTim Horscroft for inviting this contribution.

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.

Aissaoui, D.M., McNeill, D.F., Kirschvink, J.L., 1990. Magnetostratigraphic dating ofshallow-water carbonates from Mururoa Atoll, French Polynesia—implications forglobal eustasy. Earth Planet. Sci. Lett. 97, 102–112.

Algeo, T.J., Lyons, T.W., 2006. Mo-total organic carbon covariation in modern anoxic ma-rine environments: implications for analysis of paleoredox and paleohydrographicconditions. Paleoceanography 21, PA1016. http://dx.doi.org/10.1029/2004PA001112.

Alvarez, W., Arthur, M.A., Fischer, A.G., Lowrie, W., Napoleone, G., Premoli Silva, I.,Roggenthen, W.M., 1977. Upper Cretaceous–Paleocene magnetic stratigraphy atGubbio, Italy, V. Type section for the Late Cretaceous–Paleocene geomagnetic reversaltime scale. Geol. Soc. Am. Bull. 88, 383–389.

Alvarez, L.W., Alvarez, W., Asaro, F., Michel, H.V., 1980. Extraterrestrial cause for the Cre-taceous–Tertiary extinction. Science 208, 1095–1108.

Amerigian, C., 1974. Sea-floor dynamic processes as the possible cause of correlation be-tween paleoclimate and paleomagnetic indices in deep-sea sedimentary cores. EarthPlanet. Sci. Lett. 21, 321–326.

Appel, E., Crouzet, C., Schill, E., 2012. Pyrrhotite remagnetizations in the Himalaya: a re-view. Geol. Soc. Lond. Spec. Publ. 371, 163–180.

Archer, D.E., 1996. An atlas of the distribution of calcium carbonate in sediments of thedeep sea. Glob. Biogeochem. Cycles 10, 159–174.

Archer, D.E., Winguth, A., Lea, D., Mahowald, N., 2000.What caused the glacial/interglacialatmospheric pCO2 cycles? Rev. Geophys. 38, 159–189.

Banerjee, S., Elmore, R.D., Engel, M.H., 1997. Chemical remagnetization and burial diagen-esis: testing the hypothesis in the Pennsylvanian Belden Formation, Colorado.J. Geophys. Res. 102, 24825–24842.

Bareille, G., Grousset, F.E., Labracherie, M., Labeyrie, L.D., Petit, J.-R., 1994. Origin of detritalfluxes in the southeast Indian Ocean during the last climatic cycles. Paleoceanography9, 799–819.

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

Beitler, B., Parry, W.T., Chan, M.A., 2005. Fingerprints of fluid flow: chemical diagenetichistory of the Jurassic Navajo Sandstone, southern Utah, U.S.A. J. Sediment. Res. 75,547–561.

Belkaaloul, N.K., Aissaoui, D.M., 1997. Nature and origin of magnetic minerals within theMiddle Jurassic shallow-water carbonate rocks of the Paris Basin, France: implicationsfor magnetostratigraphic dating. Geophys. J. Int. 130, 411–421.

Berner, R.A., 1980. Early Diagenesis: A Theoretical Approach. Princeton University Press,New Jersey (241 pp.).

Berner, R.A., 1981. A new geochemical classification of sedimentary environments.J. Sediment. Petrol. 51, 359–365.

Berner, R.A., 1984. Sedimentary pyrite formation: an update. Geochim. Cosmochim. Acta48, 605–615.

Blakemore, R.P., 1975. Magnetotactic bacteria. Science 190, 377–379.Blakemore, R.P., Short, K.A., Bazylinski, D.A., Rosenblatt, C., Frankel, R.B., 1985. Microaerobic

conditions are required for magnetite formation within Aquaspirillum magnetotacticum.Geomicrobiol. J. 4, 53–71.

Blanchet, C.L., Thouveny, N., Vidal, L., 2009. Formation and preservation of greigite (Fe3S4)in sediments from the Santa Barbara basin: implications for paleoenvironmentalchanges during the past 35 ka. Paleoceanography 24, PA2224. http://dx.doi.org/10.1029/2008PA001719.

Blatt, H., Middleton, G., Murray, R., 1980. Origin of Sedimentary Rocks. Prentice-Hall, NewJersey (782 pp.).

Bloemendal, J., DeMenocal, P., 1989. Evidence for a change in the periodicity of tropicalclimate cycles at 2.4 Myr from whole-core magnetic susceptibility measurements.Nature 342, 897–900.

Borradaile, G.J., Lagroix, F., 2000. Magnetic characterization using a three-dimensionalhysteresis projection, illustrated with a study of limestones. Geophys. J. Int. 141,213–226.

Boughriet, A., Ouddane, B., Wartel, M., 1992. Electron spin resonance investigations of Mncompounds and free radicals in particles from the Seine River and its estuary. Mar.Chem. 37, 149–169.

Boyd, P.W., Ellwood, M.J., 2010. The biogeochemical cycle of iron in the ocean. Nat. Geosci.3, 675–682.

Briden, J.C., Irving, E., 1964. Palaeolatitude spectra of sedimentary palaeoclimatic indica-tors. In: Nairn, A.E.M. (Ed.), Problems in Palaeoclimatology. Interscience, New York,pp. 199–250.

Canfield, D.E., Berner, R.A., 1987. Dissolution and pyritization of magnetite in anoxic ma-rine sediments. Geochim. Cosmochim. Acta 51, 645–659.

Carter-Stiglitz, B., Moskowitz, B., Jackson, M., 2004. More on the low-temperature magne-tism of stable single domain magnetite: reversibility and non-stoichiometry.Geophys. Res. Lett. 31, L06606. http://dx.doi.org/10.1029/2003GL019155.

Carvallo, C., Sainctavit, P., Arrio, M.-A., Menguy, N., Wang, Y., Ona-Nguema, G., Brice-Profeta, S., 2008. Biogenic vs. abiogenic magnetite nanoparticles: a XMCD study.Am. Mineral. 93, 880–885.

Carvallo, C., Hickey, S., Faivre, D., Menguy, N., 2009. Formation of magnetite inMagnetospirillum gryphiswaldense studied with FORC diagrams. Earth, Planets Space61, 143–150.

Chambers, S.R., Cranston, R.E., 1991. Interstitial-water geochemistry of Kerguelen Plateausediments. Proc. Ocean Drill. Program Sci. Results 119, 347–374.

Chan, M.A., Parry, W.T., Bowman, J.R., 2000. Diagenetic hematite and manganese oxidesand fault-related fluid flow in Jurassic sandstones, southeastern Utah. Am. Assoc.Pet. Geol. Bull. 84, 1281–1310.

Chang, S.B.R., Kirschvink, J.L., Stolz, J.F., 1987. Biogenic magnetite as a primary remanencecarrier in limestone deposits. Phys. Earth Planet. Inter. 46, 289–303.

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

Chang, L., Winklhofer, M., Roberts, A.P., Dekkers, M.J., Horng, C.S., Hu, L., Chen, Q.W.,2012b. Ferromagnetic resonance characterization of greigite (Fe3S4), monoclinic pyr-rhotite (Fe7S8) and non-interacting titanomagnetite (Fe3−xTixO4). Geochem.Geophys. Geosyst. 13, Q05Z41. http://dx.doi.org/10.1029/2012GC004063.

Chang, L., Winklhofer, M., Roberts, A.P., Heslop, D., Florindo, F., Dekkers, M.J., Krijgsman,W., Kodama, K., Yamamoto, Y., 2013. Low-temperature magnetic properties of pelag-ic carbonates: oxidation of biogenic magnetite and identification of magnetosomechains. J. Geophys. Res. http://dx.doi.org/10.1002/2013JB010381R (in press).

Channell, J.E.T., Hawthorne, T., 1990. Progressive dissolution of titanomagnetites at ODPSite 653 (Tyrrhenian Sea). Earth Planet. Sci. Lett. 96, 469–480.

Channell, J.E.T., McCabe, C., 1994. Comparison of magnetic hysteresis parameters ofunremagnetized and remagnetized limestones. J. Geophys. Res. 99, 4613–4623.

Channell, J.E.T., Stoner, J.S., 2002. Plio-Pleistocene magnetic polarity stratigraphies anddiagenetic magnetite dissolution at ODP Leg 177 Sites (1089, 1091, 1093 and1094). Mar. Micropaleontol. 45, 269–290.

Channell, J.E.T., Tarling, D.H., 1975. Paleomagnetism and the rotation of Italy. Earth Planet.Sci. Lett. 25, 177–188.

Channell, J.E.T., Freeman, R., Heller, F., Lowrie, W., 1982. Timing of diagenetic haematitegrowth in red pelagic limestones from Gubbio (Italy). Earth Planet. Sci. Lett. 58,189–201.

Channell, J.E.T., Galeotti, S., Martin, E.E., Billups, K., Scher, H.D., Stoner, J.S., 2003. Eocene toMiocene magnetostratigraphy, biostratigraphy, and chemostratigraphy at ODP Site1090 (sub-Antarctic South Atlantic). Geol. Soc. Am. Bull. 115, 607–623.

Channell, J.E.T., Ohneiser, C., Yamamoto, Y., Kesler, M.S., 2013. Oligocene–Miocene mag-netic stratigraphy carried by biogenic magnetite at Sites U1334 and U1335 (equato-rial Pacific Ocean). Geochem. Geophys. Geosyst. 14, 265–282.

Charilaou, M., Winklhofer, M., Gehring, A.U., 2011. Simulation of ferromagnetic resonancespectra of linear chains of magnetite nanocrystals. J. Appl. Phys. 109, 093903. http://dx.doi.org/10.1063/1.3581103.

136 A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

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.

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, S., 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.

Coker, V.S., Pearce, C.I., Lang, C., van der Laan, G., Pattrick, R.A.D., Telling, N.D., Schüler, D.,Arenholz, E., Lloyd, J.R., 2007. Cation site occupancy of biogenic magnetite comparedto polygenic ferrite spinels determined by X-ray magnetic circular dichroism. Eur.J. Mineral. 19, 707–716.

Cui, Y.L., Verosub, K.L., Roberts, A.P., 1994. The effect of maghemitization on large multi-domain magnetite. Geophys. Res. Lett. 21, 757–760.

D'Hondt, S., Spivack, A.J., Pockalny, R., Fischer, J.P., Kallmeyer, J., Ferdelman, T., Abrams, L.,Smith, D.C., Graham, D., Hasiuk, F., Schrumm, H., Stancin, A.M., 2009. Subseafloor sedi-mentary life in the South Pacific Gyre. Proc. Natl. Acad. Sci. U. S. A. 106, 11651–11656.

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

de Boor, C., 1994. A practical guide to splines. Applied Mathematical Sciences, 27. Spring-er, New York (346 pp.).

deMenocal, P.B., Ruddiman, W.F., Kent, D.V., 1990. Depth of post-depositional remanenceacquisition in deep-sea sediments—a case-study of the Brunhes–Matuyama reversaland oxygen isotopic stage-19.1. Earth Planet. Sci. Lett. 99, 1–13.

Deng, C., Zhu, R.X., Jackson, M.J., Verosub, K.L., Singer, M.J., 2001. Variability of thetemperature-dependent susceptibility of the Holocene eolian deposits in the Chineseloess plateau: a pedogenesis indicator. Phys. Chem. Earth 26, 873–878.

Dunin-Borkowski, R.E., McCartney, M.R., Frankel, R.B., Bazylinski, D.A., Pósfai, M., Buseck,P.R., 1998. Magnetic microstructure of magnetotactic bacteria by electron hologra-phy. Science 282, 1868–1870.

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, B2056.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. Newbiostrat-igraphic,magnetostratigraphic and isotopic insights into theMiddle Eocene Climatic Op-timum in low latitudes. Palaeogeogr. Palaeoclimatol. Palaeoecol. 297, 670–862.

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.

Elmore, R.D., Foucher, J.L.-E., Evans, M., Lewchuk, M., 2006. Remagnetization of theTonoloway Formation and the Helderberg Group in the Central Appalachians: testingthe origin of syntilting magnetizations. Geophys. J. Int. 166, 1062–1076.

Fabian, K., von Dobeneck, T., 1997. Isothermal magnetization of samples with stablePreisach function: a survey of hysteresis, remanence, and rock magnetic parameters.J. Geophys. Res. 102, 17659–17677.

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

Fischer, H., Mastrogiacomo, G., Löffler, J.F., Warthmann, R.J., Weidler, P.G., Gehring, A.U.,2008. Ferromagnetic resonance and magnetic characteristics of intact magnetosomechains in Magnetospirillum gryphiswaldense. Earth Planet. Sci. Lett. 270, 200–208.

Fischer, J.P., Ferdelman, T.G., D'Hondt, S., Røy, H., Wenzhöfer, F., 2009. Oxygenpenetration deep into the sediment of the South Pacific gyre. Biogeosciences6, 1467–1478.

Florindo, F., Roberts, A.P., 2005. Eocene–Oligocene magnetobiostratigraphy of ODP sites689 and 690, Maud Rise, Weddell Sea, Antarctica. Geol. Soc. Am. Bull. 117, 46–66.

Florindo, F., Farmer, R.K., Harwood, D.M., Cody, R.D., Levy, R., Bohaty, S.M., Carter, L.,Winkler, A., 2013. Palaeomagnetism and biostratigraphy of sediments from SouthernOcean ODP Site 744 (southern Kerguelen Plateau): implications for early-to-middleMiocene climate in Antarctica. Glob. Planet. Change. http://dx.doi.org/10.1016/j.gloplacha.2013.05.004 (in press).

Freeman, R., 1986. Magnetic mineralogy of pelagic limestones. Geophys. J. R. Astron. Soc.85, 433–452.

Froelich, P.N., Klinkhammer, G.P., Bender, M.L., Luedtke, N.A., Heath, G.R., Cullen, D.,Dauphin, P., Hammond, D., Hartman, B., Maynard, V., 1979. Early oxidation of organicmatter in pelagic sediments of the eastern equatorial Atlantic: suboxic diagenesis.Geochim. Cosmochim. Acta 43, 1075–1090.

Fuller, M., Molina-Garza, R., Touchard, Y., Kidane, T., 2006. Paleomagnetic records fromcarbonate legs in the Southern Oceans and attendant drilling and coring related ef-fects. Phys. Earth Planet. Inter. 156, 242–260.

Gehring, A.U., Kind, J., Charilaou, M., García-Rubio, I., 2011. The detection of magnetotacticbacteria and magnetofossils by means of magnetic anisotropy. Earth Planet. Sci. Lett.309, 113–117.

Gong, Z., van Hinsbergen, D.J.J., Dekkers, M.J., 2009a. Diachronous pervasive remagnetizationin northern Iberian basins during Cretaceous rotation and extension. Earth Planet. Sci.Lett. 284, 292–301.

Gong, Z., Dekkers, M.J., Heslop, D., Mullender, T.A.T., 2009b. End-member modelling ofisothermal remanent magnetization (IRM) acquisition curves: a novel approach todiagnose remagnetization. Geophys. J. Int. 178, 693–701.

Goree, W.S., Fuller, M.D., 1976. Magnetometers using RF-driven SQUIDs and their applica-tions in rock magnetism and paleomagnetism. Rev. Geophys. Space Phys. 14, 591–608.

Graham, J.W., 1949. The stability and significance of magnetism in sedimentary rocks.J. Geophys. Res. 54, 131–167.

Guidry, E.P., Richter, C., Acton, G.D., Channell, J.E.T., Evans, H.F., Ohneiser, C., Yamamoto, Y.,Yamazaki, T., 2013. Oligocene–Miocene magnetostratigraphy of deep-sea sedimentsfrom the equatorial Pacific (IODP Site U1333). Geol. Soc. Lond. Spec. Publ. 373,13–27. http://dx.doi.org/10.1144/SP373.7.

Hanzlik, M., Winklhofer, M., Petersen, N., 2002. Pulsed-field-remanence measurementson individual magnetotactic bacteria. J. Magn. Magn. Mater. 248, 258–267.

Heider, F., Körner, U., Bitschene, P., 1993. Volcanic ash particles as carriers of remanentmagnetization in deep-sea sediments from the Kerguelen Plateau. Earth Planet. Sci.Lett. 118, 121–134.

Heller, F., 1977. Palaeomagnetism of Upper Jurassic limestones from southern Germany.J. Geophys. 42, 475–488.

Henshaw, P.C., Merrill, R.T., 1980. Magnetic and chemical change in marine sediments.Rev. Geophys. Space Phys. 18, 483–504.

Heslop, D., Dillon, M., 2007. Unmixing magnetic remanence curves without a prioriknowledge. Geophys. J. Int. 170, 556–566.

Heslop, D., Roberts, A.P., 2012a. 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., Roberts, A.P., 2012b. A method for unmixing magnetic hysteresis loops.J. Geophys. Res. 117, B03103. http://dx.doi.org/10.1029/2011JB008859.

Heslop, D., Dekkers, M.J., Kruiver, P.P., van Oorschot, I.H.M., 2002. Analysis of isothermalremanent magnetisation acquisition curves using the expectation–maximisation al-gorithm. 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.

Horng, C.S., Roberts, A.P., 2006. Authigenic or detrital origin of pyrrhotite in sediments?:resolving a paleomagnetic conundrum. Earth Planet. Sci. Lett. 241, 750–762.

Hounslow, M.W., Maher, B.A., 1996. Quantitative extraction and analysis of carriers ofmagnetization in sediments. Geophys. J. Int. 124, 57–74.

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.

Hrouda, F., 1994. A technique for the measurement of thermal changes of magnetic sus-ceptibility of weakly magnetic rocks by the CS-2 apparatus and KLY-2 Kappabridge.Geophys. J. Int. 118, 604–612.

Irving, E., 1964. Paleomagnetism and its Application to Geological and Geophysical Prob-lems. John Wiley, New York (399 pp.).

Irving, E., Major, A., 1964. Post-depositional detrital remanentmagnetization in a synthet-ic sediment. Sedimentology 3, 135–143.

Jackson, M., 1990. Diagenetic sources of stable remanence in remagnetized Paleozoic cra-tonic carbonates: a rock magnetic study. J. Geophys. Res. 95, 2753–2761.

Jackson, M., Swanson-Hysell, N.L., 2012. Rock magnetism of remagnetized carbonaterocks: another look. Geol. Soc. Lond. Spec. Publ. 371, 229–251.

Johnson, E.A., Murphy, T., Torreson, O.W., 1948. Pre-history of the Earth's magnetic field.Terr. Magn. Atmos. Electr. 53, 349–372.

Johnson, H.P., Lowrie, W., Kent, D.V., 1975. Stability of anhysteretic remanent magnetiza-tion in fine and coarse magnetite and maghemite particles. Geophys. J. R. Astron. Soc.41, 1–10.

Jovane, L., Florindo, F., Dinarès-Turell, J., 2004. Environmental magnetic record ofpaleoclimate change from the Eocene–Oligocene stratotype section, Massignano,Italy. Geophys. Res. Lett. 31, L15601. http://dx.doi.org/10.1029/2004GL020554.

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, 423–427.

Jovane, L., Coccioni, R., Marsili, A., Acton, G., 2009. The late Eocene greenhouse–icehousetransition: observations from the Massignano global stratotype section and point(GSSP). In: Koeberl, C., Montanari, A. (Eds.), The Late Eocene Earth—Hothouse, Ice-house, and Impacts. Geol. Soc. Am. Spec. Papers, 452, pp. 149–168.

Jovane, L., Florindo, F., Bazylinski, D.A., Lins, U., 2012. Prismatic magnetite magnetosomesfrom cultivated Magnetovibrio blakemorei strain MV-1: a magnetic fingerprint in ma-rine sediments? Environ. Microbiol. Rep. 4, 664–668. http://dx.doi.org/10.1111/1758-2229.12000.

Jovane, L., Savian, J.F., Coccioni, R., Frontalini, F., Bancalà, G., Catanzariti, R., Luciani, V.,Bohaty, S.M., Wilson, P.A., Florindo, F., 2013. Integrated magnetobiostratigraphy ofthemiddle Eocene–lower Oligocene interval from theMonte Cagnero section, centralItaly. Geol. Soc. Lond. Spec. Publ. 373, 79–95. http://dx.doi.org/10.1144/SP373.13.

Kao, S.J., Horng, C.S., Roberts, A.P., Liu, K.K., 2004. Carbon–sulfur–iron relationships in sed-imentary rocks from southwestern Taiwan: influence of geochemical environmenton greigite and pyrrhotite formation. Chem. Geol. 203, 153–168.

Karlin, R., 1990. Magnetic mineral diagenesis in suboxic sediments at Bettis site W–N, NEPacific Ocean. J. Geophys. Res. 95, 4421–4436.

Karlin, R., Levi, S., 1983. Diagenesis of magnetic minerals in Recent haemipelagic sedi-ments. Nature 303, 327–330.

Kasten, S., Zabel, M., Heuer, V., Hensen, C., 2003. Processes and signals of nonsteady-statediagenesis in deep-sea sediments and their pore waters. In: Wefer, G., Mulitza, S.,Ratmeyer, V. (Eds.), The South Atlantic in the Late Quaternary: Reconstruction of Ma-terial Budgets and Current Systems. Springer-Verlag, Berlin, pp. 431–459.

Katari, K., Tauxe, L., 2000. Effects of pH and salinity on the intensity of magnetization inredeposited sediments. Earth Planet. Sci. Lett. 181, 489–496.

Katz, B., Elmore, R.D., Engel, M.H., 1998. Authigenesis of magnetite in organic-rich sedi-ment next to a dike: implications for thermoviscous and chemical remagnetizations.Earth Planet. Sci. Lett. 163, 221–234.

Kawamura, N., Ishikawa, N., Torii, M., 2012. Diagenetic alteration of magnetic minerals inLabrador Sea sedments (IODP Sites U1305, U1306, and U1307). Geochem. Geophys.Geosyst. 13, Q08013. http://dx.doi.org/10.1029/2012GC004213.

137A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

Kent, D.V., 1973. Post-depositional remanent magnetisation in deep sea sediment. Nature246, 32–34.

Kent, D.V., 1982. Apparent correlation of palaeomagnetic intensity and climate records indeep-sea sediments. Nature 299, 538–539.

Kent, D.V., 1985. Thermoviscous remagnetization in some Appalachian limestones.Geophys. Res. Lett. 12, 805–808.

Kent, D.V., Cramer, B.S., Lanci, L., Wang, D., Wright, J.D., van der Voo, R., 2003. A case for acomet impact trigger for the Paleocene/Eocene thermal maximum and carbon iso-tope excursion. Earth Planet. Sci. Lett. 211, 13–26.

Kind, J., Gehring, A.U., Winklhofer, M., Hirt, A.M., 2011. Combined use of magnetometryand spectroscopy for identifying magnetofossils in sediments. Geochem. Geophys.Geosyst. 12, Q08008. http://dx.doi.org/10.1029/2011GC003633.

King, R.F., 1955. The remanent magnetism of artificially deposited sediments. Mon. NotesR. Astron. Soc. Geophys. Suppl. 7, 115–134.

Kirschvink, J.L., 1980. The least-squares line and plane and the analysis of palaeomagneticdata. Geophys. J. R. Astron. Soc. 62, 699–718.

Kirschvink, J.L., 1982. Paleomagnetic evidence for fossil biogenic magnetite in WesternCrete. Earth Planet. Sci. Lett. 59, 388–392.

Kodama, K., Moeller, R.E., Bazylinski, D.A., Kopp, R.E., Chen, A.P., 2013. The mineral mag-netic record of magnetofossils in recent lake sediments of Lake Ely, PA. Glob. Planet.Change. http://dx.doi.org/10.1016/j.gloplacha.2013.03.012 (in press).

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

Kopp, R.E., Weiss, B.P., Maloof, A.C., Vali, H., Nash, C.Z., Kirschvink, J.L., 2006a. Chains,clumps, and strings: magnetofossil taphonomy with ferromagnetic resonance spec-troscopy. Earth Planet. Sci. Lett. 247, 10–25.

Kopp, R.E., Nash, C.Z., Kobayashi, A., Weiss, B.P., Bazylinski, D.A., Kirschvink, J.L., 2006b.Ferromagnetic resonance spectroscopy for assessment of magnetic anisotropy andmagnetostatic interactions: a case study of mutant magnetotactic bacteria.J. Geophys. Res. 111, B12S25. http://dx.doi.org/10.1029/2006JB004529.

Kopp, R.E., Raub, T.D., Schumann, D., Vali, H., Smirnov, A.V., Kirschvink, J.L., 2007.Magnetofossil spike during the Paleocene–Eocene thermal maximum: ferromagneticresonance, rock magnetic, and electron microscopy evidence from Ancora, New Jersey,United States. Paleoceanography 22, PA4103. http://dx.doi.org/10.1029/2007PA001473.

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.

Lam, K.P., Hitchcock, A.P., Obst, M., Lawrence, J.R., Swerhone, G.D.W., Leppard, G.G.,Tyliszczak, T., Karunakaran, C., Wang, J., Kaznatcheev, K., Bazylinski, D.A., Lins, U.,2010. Characterizing magnetism of individual magnetosomes by X-ray circular di-chroism in a scanning transmission X-ray microscope. Chem. Geol. 270, 110–116.

Lambert, F., Delmonte, B., Petit, J.R., Bigler, M., Kaufmann, P.R., Hutterli, M.A., Stocker, T.F.,Ruth, U., Steffensen, J.P., Maggi, V., 2008. Dust-climate couplings over the past800,000 years from the EPICA Dome C ice core. Nature 452, 616–619.

Lanci, L., Pares, J.M., Channell, J.E.T., Kent, D.V., 2004. Miocene magnetostratigraphy fromequatorial Pacific sediments (ODP Site 1218, Leg 199). Earth Planet. Sci. Lett. 226,207–224.

Lanci, L., Pares, J.M., Channell, J.E.T., Kent, D.V., 2005. Oligocenemagnetostratigraphy fromEquatorial Pacific sediments (ODP Sites 1218 and 1219, Leg 199). Earth Planet. Sci.Lett. 237, 617–634.

Lanci, L., Delmonte, B., Kent, D.V., Maggi, V., Biscaye, P.E., Petit, J.R., 2012. Magnetization ofpolar ice: a measurement of terrestrial dust and extraterrestrial fallout. Quat. Sci. Rev.33, 20–31.

Larrasoaña, J.C., Roberts, A.P., Stoner, J.S., Richter, C., Wehausen, R., 2003. A new proxy forbottom-water ventilation based on diagenetically controlled magnetic properties ofeastern Mediterranean sapropel-bearing sediments. Paleogeogr. Paleoclimatol.Paleoecol. 190, 221–242.

Larrasoaña, J.C., Roberts, A.P., Musgrave, R.J., Gràcia, E., Piñero, E., Vega, M., Martínez-Ruiz,F., 2007. Diagenetic formation of greigite and pyrrhotite in marine sedimentary sys-tems containing gas hydrates. Earth Planet. Sci. Lett. 261, 350–366.

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 dustsupply during the Palaeocene–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.

Li, J., Pan, Y., Chen, G., Liu, Q., Tian, L., Lin,W., 2009. Magnetite magnetosome and fragmen-tal chain formation of Magnetospirillum magneticum AMB-1: transmission electronmicroscopy and magnetic observations. Geophys. J. Int. 177, 33–42.

Li, J., Pan, Y., Liu, Q., Zhang, K., Menguy, N., Che, R., Qin, H., Lin, W., Wu, W., Petersen, N.,Yang, X., 2010. Biomineralization, crystallography and magnetic properties ofbullet-shaped magnetite magnetosomes in giant rod magnetotactic bacteria. EarthPlanet. Sci. Lett. 293, 368–376.

Lin, W., Wang, Y.Z., Pan, Y.X., 2012. Short-term effects of temperature on the abundanceand diversity of magnetotactic cocci. Microbiol. Open 1, 53–63. http://dx.doi.org/10.1002/mbo3.7.

Lippert, P.C., Zachos, J.C., 2007. A biogenic origin for anomalous fine-grained magneticmaterial at the Paleocene–Eocene boundary at Wilson Lake, New Jersey.Paleoceanography 22, PA4104. http://dx.doi.org/10.1029/2007PA001471.

Liu, Q.S., Deng, C.L., Yu, Y.J., Torrent, J., Jackson, M.J., Banerjee, S.K., Zhu, R.X., 2005. Temper-ature dependence of magnetic susceptibility in an argon environment: implicationsfor pedogenesis of Chinese loess/paleosols. Geophys. J. Int. 161, 102–112.

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.S., 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.

Lowrie, W., Alvarez, W., 1975. Paleomagnetic evidence for the rotation of the Italian Pen-insula. J. Geophys. Res. 80, 1579–1592.

Lowrie, W., Alvarez, W., 1977a. Late Cretaceous geomagnetic polarity sequence: detailedrock and palaeomagnetic studies of the Scaglia Rossa limestone at Gubbio, Italy.Geophys. J. R. Astron. Soc. 51, 561–581.

Lowrie, W., Alvarez, W., 1977b. Upper Cretaceous–Paleocene magnetic stratigraphy atGubbio, Italy, III. Upper Cretaceous magnetic stratigraphy. Geol. Soc. Am. Bull. 88,374–377.

Lowrie, W., Alvarez, W., 1981. One hundred million years of geomagnetic polarity history.Geology 9, 392–397.

Lowrie, W., Fuller, M., 1971. On the alternating field demagnetization characteristics ofmultidomain thermoremanent magnetization in magnetite. J. Geophys. Res. 76,6339–6349.

Lowrie, W., Heller, F., 1982. Magnetic properties of marine limestones. Rev. Geophys.Space Phys. 20, 171–192.

Lowrie, W., Channell, J.E.T., Alvarez, W., 1980. A review of magnetic stratigraphy investi-gations in Cretaceous pelagic carbonate rocks. J. Geophys. Res. 85, 3597–3605.

Lowrie,W., Alvarez, W., Napoleone, G., Perch-Nielsen, K., Premoli Silva, I., Toumarkine, M.,1982. Paleogene magnetic stratigraphy in Umbrian pelagic carbonate rocks: theContessa sections, Gubbio. Geol. Soc. Am. Bull. 93, 414–432.

Maher, B.A., 1998. Magnetic properties of modern soils and Quaternary loessic paleosols:paleoclimatic implications. Palaeogeogr. Palaeoclimatol. Palaeoecol. 137, 25–54.

Martin, J.H., Gordon, R.M., Fitzwater, S.E., 1991. The case for iron. Limnol. Oceanogr. 36,1793–1802.

Mastrogiacomo, G., Fischer, H., García-Rubio, I., Gehring, A.U., 2010. Ferromagnetic reso-nance spectroscopic response of magnetite chains in a biological matrix. J. Magn.Magn. Mater. 322, 661–663.

McCabe, C., Channell, J.E.T., 1994. Late Paleozoic remagnetization in limestones of the Cra-ven Basin (northern England) and the rock magnetic fingerprint of remagnetizedsedimentary carbonates. J. Geophys. Res. 99, 4603–4612.

McCabe, C., Elmore, R.D., 1989. The occurrence and origin of Late Paleozoic remagnetizationin the sedimentary rocks of North America. Rev. Geophys. 27, 471–494.

McCabe, C., Jackson,M., Saffer, B., 1989. Regional patterns ofmagnetite authigenesis in theAppalachian Basin: implications for themechanism of late Paleozoic remagnetization.J. Geophys. Res. 94, 10429–10443.

McNeill, D.F., Ginsburg, R.N., Chang, S.B.R., Kirschvink, J.L., 1988. Magnetostratigraphicdating of shallow-water carbonates from San-Salvador, Bahamas. Geology 16, 8–12.

Mead, G., Tauxe, L., LaBrecque, J.L., 1986. Oligocene paleoceanography of the South Atlan-tic. Paleoceanography 1, 273–284.

Ménabréaz, L., Thouveny, N., Camoin, G., Lund, S.P., 2010. Paleomagnetic record of the latePleistocene reef sequence of Tahiti (French Polynesia): a contribution to the chronol-ogy of the deposits. Earth Planet. Sci. Lett. 294, 58–68.

Montgomery, P., Hailwood, E.A., Gale, A.S., Burnett, J.A., 1998. The magnetostratigraphy ofConiacian late Campanian chalk sequences in southern England. Earth Planet. Sci.Lett. 156, 209–224.

Moore, C.M., Mills, M.M., Achterberg, E.P., Geider, R.J., LaRoche, J., Lucas, M.I., McDonagh,E.L., Pan, X., Poulton, A.J., Rijkenberg, M.J.A., Suggett, D.J., Ussher, S.J., Woodward,E.M.S., 2009. Large-scale distribution of Atlantic nitrogen fixation controlled by ironavailability. Nat. Geosci. 2, 867–871.

Morford, J.L., Emerson, S., 1999. The geochemistry of redox sensitive trace metals in sed-iments. Geochim. Cosmochim. Acta 63, 1735–1750.

Morse, J.W., 2005. Formation anddiagenesis of carbonate sediments. In:Mackenzie, F.T. (Ed.),Sediments, Diagenesis, and Sedimentary Rocks. Treatise on Geochemistry 7, pp. 67–85.

Moskowitz, B.M., Frankel, R.B., Bazylinski, D.A., 1993. Rockmagnetic criteria for the detec-tion of biogenic magnetite. Earth Planet. Sci. Lett. 120, 283–300.

Muxworthy, A.R., Williams, W., 2009. Critical superparamagnetic/single domain grainsizes in interacting magnetite particles: implications for magnetosome crystals. J. R.Soc. Interface. 6, 1207–1212.

Muxworthy, A.R., Williams, W., Virdee, D., 2003. Effect of magnetostatic interactions onthe hysteresis properties of single-domain and pseudo-single domain grains.J. Geophys. Res. 108 (B11), 2517. http://dx.doi.org/10.1029/2003JB002588.

Napoleone, G., Premoli Silva, I., Heller, F., Cheli, P., Corezzi, S., Fischer, A.G., 1983. Eocenemagnetic stratigraphy at Gubbio, Italy, and its implications for Paleogene geochronol-ogy. Geol. Soc. Am. Bull. 94, 181–191.

O'Reilly, W., 1984. Rock and Mineral Magnetism. Blackie, Glasgow (230 pp.).Oliver, J., 1986. Fluids expelled tectonically from orogenic belts: their role in hydrocarbon

migration and other geological phenomena. Geology 14, 99–102.Opdyke, N.D., Henry, K.W., 1969. A test of the dipole hypothesis. Earth Planet. Sci. Lett. 6,

139–151.Otamendi, A.M., Díaz, M., Costanzo-Álvarez, V., Aldana, M., Pilloud, A., 2006. EPR stratigra-

phy applied to the study of two marine sedimentary sections in southwesternVenezuela. Phys. Earth Planet. Inter. 154, 243–254.

Özdemir, Ö., Dunlop, D.J., 2010. Hallmarks of maghemitization in low-temperature rema-nence cycling of partially oxidized magnetite nanoparticles. J. Geophys. Res. 115,B02101. http://dx.doi.org/10.1029/2009JB006756.

Özdemir, Ö., Dunlop, D.J., Moskowitz, B.M., 1993. The effect of oxidation on the Verweytransition in magnetite. Geophys. Res. Lett. 20, 1671–1674.

Pan, Y.X., Petersen, N.,Winklhofer, M., Davila, A.F., Liu, Q.S., Frederichs, T., Hanzlik, M., Zhu,R.X., 2005. Rock magnetic properties of uncultured magnetotactic bacteria. EarthPlanet. Sci. Lett. 237, 311–325.

Passier, H.F., Dekkers, M.J., 2002. Iron oxide formation in the active oxidation front abovesapropel S1 in the eastern Mediterranean Sea as derived from low-temperature mag-netism. Geophys. J. Int. 150, 230–240.

138 A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

Penninga, I., de Waard, H., Moskowitz, B.M., Bazylinski, D.A., Frankel, R.B., 1995. Rema-nence measurements on individual magnetotactic bacteria using a pulsed magneticfield. J. Magn. Magn. Mater. 149, 279–286.

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.

Petit, J.R., Mounier, L., Jouzel, J., Korotkevich, Y.S., Kotlyakov, V.I., Lorius, C., 1990.Palaeoclimatological and chronological implications of the Vostok core dust record.Nature 343, 56–58.

Pike, C.R., Roberts, A.P., Verosub, K.L., 1999. Characterizing interactions in fine magneticparticle systems using first order reversal curves. J. Appl. Phys. 85, 6660–6667.

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.

Pugh, R.S., McCave, I.N., Hillenbrand, C.-D., Kuhn, G., 2009. Circum-Antarctic age model-ling of Quaternary marine cores under the Antarctic Circumpolar Current: ice-coredust-magnetic correlation. Earth Planet. Sci. Lett. 284, 113–123.

Reynolds, R.L., Rosenbaum, J.G., van Metre, P., Tuttle, M., Callender, E., Goldin, A., 1999.Greigite as an indicator of drought—the 1912–1994 sediment magnetic record fromWhite Rock Lake, Dallas, Texas, USA. J. Paleolimnol. 21, 193–206.

Roberts, A.P., Turner, G.M., 1993. Diagenetic formation of ferrimagnetic iron sulphideminerals in rapidly deposited marine sediments, South Island, New Zealand. EarthPlanet. Sci. Lett. 115, 257–273.

Roberts, A.P., Weaver, R., 2005. Multiple mechanisms of remagnetization involving sedi-mentary greigite (Fe3S4). Earth Planet. Sci. Lett. 231, 263–277.

Roberts, A.P., Winklhofer, M., 2004. Why are geomagnetic excursions not always recordedin sediments? Constraints from post-depositional remanent magnetization lock-inmodelling. Earth Planet. Sci. Lett. 227, 345–359.

Roberts, A.P., Cui, Y., Verosub, K.L., 1995. Wasp-waisted hysteresis loops: mineral magnet-ic characteristics and discrimination of components in mixed magnetic systems.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.

Roberts, A.P., Bicknell, S., Byatt, J., Bohaty, S.M., Florindo, F., Harwood, D.M., 2003.Magnetostratigraphic calibration of Southern Ocean diatom biostratigraphic datumsfrom the Eocene–Oligocene of Kerguelen Plateau (ODP sites 744 and 748).Paleogeogr. Paleoclimatol. Paleoecol. 198, 145–168.

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 singledomainmagnetite in the ‘pseudo-single-domain’ sedimentary haystack: implicationsof biogenic magnetite preservation for sediment magnetism and relativepaleointensity determinations. J. Geophys. Res. 117, B08104. http://dx.doi.org/10.1029/2012JB009412.

Roberts, A.P., Tauxe, L., Heslop, D., 2013. Magnetic paleointensity stratigraphy and high-resolution Quaternary geochronology: successes and future challenges. Quat. Sci.Rev. 61, 1–16.

Robertson, D.J., France, D.E., 1994. Discrimination of remanence-carryingminerals inmix-tures, using isothermal remanent magnetisation acquisition curves. Phys. Earth Plan-et. Inter. 84, 223–234.

Robinson, S.G., 1986. The late Pleistocene palaeoclimatic record of North Atlantic deep-sea sediments revealed by mineral magnetic measurements. Phys. Earth Planet.Inter. 42, 22–57.

Robinson, S.G., Sahota, J.T.S., Oldfield, F., 2000. Early diagenesis in North Atlantic abyssalplain sediments characterized by rock-magnetic and geochemical indices. Mar.Geol. 163, 77–107.

Rochette, P., 1987. Metamorphic control of the magnetic mineralogy of black shales in theSwiss Alps: toward the use of “magnetic isogrades”. Earth Planet. Sci. Lett. 84, 446–456.

Roggenthen, W.M., Napoleone, G., 1977. Upper Cretaceous–Paleocene magnetic stratigra-phy at Gubbio, Italy, IV. Upper Maastrichtian–Paleocene magnetic stratigraphy. Geol.Soc. Am. Bull. 88, 378–382.

Rowan, C.J., Roberts, A.P., Broadbent, T., 2009. Reductive diagenesis, magnetite dissolu-tion, greigite growth and paleomagnetic smoothing in marine sediments: a newview. Earth Planet. Sci. Lett. 277, 223–235.

Sarthou, G., Baker, A.R., Blain, S., Achterberg, E.P., Boye, M., Bowie, A.R., Croot, P., Laan, P.,de Baar, H.J.W., Jickells, T.D., Worsfold, P.J., 2003. Atmospheric iron deposition andsea-surface dissolved iron concentrations in the eastern Altantic Ocean. Deep-SeaRes. I 50, 1339–1352.

Satolli, S., Besse, J., Speranza, F., Calamita, F., 2007. New 125–150Ma high-resolution ap-parent polar wander path for Adria from magnetostratigraphic sections in Umbria–Marche (Northern Apennines, Italy): timing and duration of the global Jurassic–Cre-taceous hairpin turn. Earth Planet. Sci. Lett. 257, 329–342.

Savian, J.F., Jovane, L., Bohaty, S.M., Wilson, P.A., 2013. Middle Eocene to early Oligocenemagnetostratigraphy of ODP Hole 711A (Leg 115), western equatorial IndianOcean. Geol. Soc. Lond. Spec. Publ. 373, 97–110. http://dx.doi.org/10.1144/SP373.16.

Scheffel, A., Gärdes, A., Grünberg, K., Wanner, G., Schüler, D., 2008. The majormagnetosome proteins MamGFDC are not essential for magnetite biomineralizationin Magnetospirillum gryphiswaldense, but regulate the size of magnetosome crystals.J. Bacteriol. 190, 377–386.

Schneider, D.A., Backman, J., Chaisson, W.P., Raffi, I., 1997. Miocene calibration for calcar-eous nannofossils from low-latitude Ocean Drilling Program sites and the Jamaicanconundrum. Geol. Soc. Am. Bull. 109, 1073–1079.

Schüler, D., Baeuerlein, E., 1998. Dynamics of iron uptake and Fe3O4 biomineralizationduring aerobic and microaerobic growth of Magnetospirillum gryphiswaldense.J. Bacteriol. 180, 159–162.

Schumann, D., Raub, T.D., Kopp, R.E., Guerquin-Kern, J.L., Wu, T.D., Rouiller, I., Smirnov,A.V., Sears, S.K., Lücken, U., Tikoo, S.M., Hesse, R., Kirschvink, J.L., Vali, H., 2008. Gigan-tism in unique biogenic magnetite at the Paleocene–Eocene ThermalMaximum. Proc.Natl. Acad. Sci. U. S. A. 105, 17648–17653.

Smirnov, A.V., Tarduno, J.A., 2000. Low-temperature magnetic properties of pelagic sedi-ments (Ocean Drilling Program Site 805C): tracers of maghemitization and magneticmineral reduction. J. Geophys. Res. 105, 16457–16471.

Speranza, F., Satolli, S.,Mattioli, E., Calamita, F., 2005.Magnetic stratigraphy of Kimmeridgian–Aptian sections from Umbria–Marche (Italy): new details on the M polarity sequence.J. Geophys. Res. 110, B12109. http://dx.doi.org/10.1029/2005JB003884.

Stoltz, J.F., Chang, S.B.R., Kirschvink, J.L., 1986. Magnetotactic bacteria and single-domainmagnetite in hemipelagic sediments. Nature 321, 849–851.

Stumm,W., Morgan, J.J., 1996. Aquatic Chemistry: Chemical Equilibria and Rates in Natu-ral Waters, 3rd ed. John Wiley & Sons, New York (1022 pp.).

Suganuma, Y., Okuno, J., Heslop, D., Roberts, A.P., Yamazaki, T., Yokoyama, Y., 2011. Post-depositional remanent magnetization lock-in for marine sediments deduced from10Be and paleomagnetic records through the Matuyama–Brunhes boundary. EarthPlanet. Sci. Lett. 311, 39–52.

Suk, D., Peacor, D.R., van der Voo, R., 1990a. Replacement of pyrite framboids bymagnetite in limestone and implications for palaeomagnetism. Nature 345,611–613.

Suk, D.W., van der Voo, R., Peacor, D.R., 1990b. Scanning and transmission electron-microscope observations of magnetite and other iron phases in Ordovician carbon-ates from east Tennessee. J. Geophys. Res. 95, 12327–12336.

Swanson-Hysell, N.L., Maloof, A.C., Kirschvink, J.L., Halverson, G.P., Hurtgen, M.T., 2012.Constraints on Neoproterozoic paleogeography and Paleozoic orogenesis from paleo-magnetic records of the Bitter Springs Formation, Amadeus Basin, central Australia.Am. J. Sci. 312, 817–884.

Tarduno, J.A., 1992. Magnetic susceptibility cyclicity and magnetic dissolution in Creta-ceous limestones of the Southern Alps (Italy). Geophys. Res. Lett. 19, 1515–1518.

Tarduno, J.A., 1994. Temporal trends ofmagnetic dissolution in the pelagic realm: gaugingpaleoproductivity? Earth Planet. Sci. Lett. 123, 39–48.

Tarduno, J.A., Myers, M., 1994. A primary magnetization fingerprint from the CretaceousLaytonville Limestone: further evidence for rapid oceanic plate velocities.J. Geophys. Res. 99, 21691–21703.

Tarduno, J.A., Wilkison, S., 1996. Non-steady state magnetic mineral reduction, chemicallock-in, and delayed remanence acquisition in pelagic sediments. Earth Planet. Sci.Lett. 144, 315–326.

Tarduno, J.A., Tian, W., Wilkison, S., 1998. Biogeochemical remanent magnetization in pe-lagic sediments of the western equatorial Pacific Ocean. Geophys. Res. Lett. 25,3987–3990.

Tauxe, L., Hartl, P., 1997. 11 million years of Oligocene geomagnetic field behaviour.Geophys. J. Int. 128, 217–229.

Tauxe, L., Tucker, P., Petersen, N., LaBrecque, J.L., 1983. The magnetostratigraphy of Leg 73sediments. Paleogeogr. Paleoclimatol. Paleoecol. 42, 65–90.

Tauxe, L.,Mullender, T.A.T., Pick, T., 1996. Potbellies,wasp-waists and superparamagnetism inmagnetic hysteresis. J. Geophys. Res. 101, 571–583.

Tauxe, L., Steindorf, J.L., Harris, A., 2006. Depositional remanentmagnetization: toward an im-proved theoretical and experimental foundation. Earth Planet. Sci. Lett. 244, 515–529.

Tohver, E., Weil, A.B., Solum, J.G., Hall, C.M., 2008. Direct dating of carbonateremagnetization by 40Ar/39Ar analysis of the smectite–illite transformation. EarthPlanet. Sci. Lett. 274, 524–530.

Torii, M., 1997. Low-temperature oxidation and subsequent downcore dissolution ofmagnetite in deep-sea sediments, ODP Leg 161 (western Mediterranean).J. Geomagn. Geoelectr. 49, 1233–1245.

Touchard, Y., Rochette, P., Aubry, M.P., Michard, A., 2003. High-resolutionmagnetostratigraphic and biostratigraphic study of Ethiopian traps-relatedproducts in Oligocene sediments from the Indian Ocean. Earth Planet. Sci. Lett.206, 493–508.

Valet, J.-P., Meynadier, L., 1993. Geomagnetic field intensity and reversals during the pastfour million years. Nature 366, 234–238.

Vali, H., Kirschvink, J.L., 1989. Magnetofossil dissolution in a palaeomagnetically unstabledeep-sea sediment. Nature 339, 203–206.

Vali, H., Förster, O., Amarantidis, G., Petersen, N., 1987. Magnetotactic bacteria and theirmagnetofossils in sediments. Earth Planet. Sci. Lett. 86, 389–400.

van der Voo, R., Torsvik, T.H., 2012. The history of remagnetization of sedimentary rocks:deceptions, developments and discoveries. Geol. Soc. Lond. Spec. Publ. 371, 23–53.

van Velzen, A.J., Zijderveld, J.D.A., 1995. Effects of weathering on single-domainmagnetitein Early Pliocene marine marls. Geophys. J. Int. 121, 267–278.

VandenBerg, J., Klootwijk, C.T., Wonders, A.A.H., 1978. The Late Mesozoic and Cenozoicmovements of the Umbrian Peninsula: further paleomagnetic data from the Umbriansequence. Geol. Soc. Am. Bull. 89, 133–150.

Verosub, K.L., 1977. Depositional and post-depositional processes in the magnetization ofsediments. Rev. Geophys. Space Phys. 15, 129–143.

Verosub, K.L., Fine, P., Singer, M.J., TenPas, J., 1993. Pedogenesis and paleoclimate: inter-pretation of the magnetic susceptibility record in Chinese loess/paleosol sequences.Geology 21, 1011–1014.

Weaver, R., Roberts, A.P., Barker, A.J., 2002. A late diagenetic (synfolding) magnetizationcarried by pyrrhotite: implications for paleomagnetic studies from magnetic ironsulphide-bearing sediments. Earth Planet. Sci. Lett. 200, 371–386.

Weber, M.E., Kuhm, G., Sprenk, D., Rolf, C., Ohlwein, C., Ricken, W., 2012. Dust transportfrom Patagonia to Antarctica—a new stratigraphic approach from the Scotia Seaand its implications for the last glacial cycle. Quat. Sci. Rev. 36, 177–188.

139A.P. Roberts et al. / Earth-Science Reviews 127 (2013) 111–139

Weeks, R., Laj, C., Endignoux, L., Fuller, M., Roberts, A., Manganne, R., Blanchard, E., Goree,W., 1993. Improvements in long-core measurement techniques: applications inpalaeomagnetism and palaeoceanography. Geophys. J. Int. 114, 651–662.

Weiss, B.P., Kim, S.S., Kirschvink, J.L., Kopp, R.E., Sankaran, M., Kobayashi, A., Komeili, A.,2004. Ferromagnetic resonance and low-temperature magnetic test for biogenicmagnetite. Earth Planet. Sci. Lett. 224, 73–89.

Williams, P.W., Ford, D.C., 2006. Global distribution of carbonate rocks. Z. Geomorphol.Suppl. 147, 1–2.

Xu,W., van der Voo, R., Peacor, D.R., 1998. Electronmicroscopic and rockmagnetic study ofremagnetized Leadville carbonates, central Colorado. Tectonophysics 296, 333–362.

Yamazaki, T., 2008. Magnetostatic interactions in deep-sea sediments inferred from first-order reversal curve diagrams: implications for relative paleointensity normalization.Geochem. Geophys. Geosyst. 9, Q02005. http://dx.doi.org/10.1029/2007GC001797.

Yamazaki, T., 2009. Environmental magnetism of Pleistocene sediments in the North Pacificand Ontong–Java Plateau: temporal variations of detrital and biogenic components.Geochem. Geophys. Geosyst. 10, Q07Z04. http://dx.doi.org/10.1029/2009GC002413.

Yamazaki, T., 2012. Paleoposition of the Intertropical Convergence Zone in the eastern Pa-cific inferred from glacial–interglacial changes in terrigenous and biogenic magneticmineral fractions. Geology 40, 151–154.

Yamazaki, T., Ikehara, M., 2012. Origin of magnetic mineral concentration variation in theSouthernOcean. Paleoceanography27, PA2206. http://dx.doi.org/10.1029/2011PA002271.

Yamazaki, T., Ioka, N., 1997. Environmental rock-magnetism of pelagic clay: implicationsfor Asian eolian input to the North Pacific since the Pliocene. Paleoceanography 12,111–124.

Yamazaki, T., Kawahata, H., 1998. Organic carbon flux controls the morphology ofmagnetofossils in sediments. Geology 26, 1064–1066.

Yamazaki, T., Solheid, P., 2011. Maghemite-to-magnetite reduction across the Fe-redoxboundary in a sediment core from the Ontong–Java Plateau: influence on relativepalaeointensity estimation and environmental magnetic application. Geophys. J. Int.185, 1243–1254.

Yamazaki, T., Yamamoto, Y., Acton, G., Guidry, E.P., Richter, C., 2013. Rock magnetic arti-facts on long-term relative paleointensity variations in sediments. Geochem.Geophys. Geosyst. 14, 29–43.

Zegers, T.E., Dekkers, M.J., Bailly, S., 2003. Late Carboniferous to Permianremagnetization of Devonian limestones in the Ardennes: role of temperature,fluids, and deformation. J. Geophys. Res. 108, 2357. http://dx.doi.org/10.1029/2002JB002213.

Zhou, L.P., Oldfield, F., Wintle, A.G., Robinson, S.G., Wang, J.T., 1990. Partly pedogenic ori-gin of magnetic variations in Chinese loess. Nature 346, 737–739.

Zijderveld, J.D.A., 1967. AC demagnetization of rocks: analysis of results. In: Collinson,D.W., Creer, K.M., Runcorn, S.K. (Eds.), Methods in Paleomagnetism. Elsevier,Amsterdam, pp. 254–286.