1

Late Quaternary sediments from deep-sea sediment drifts on the Antarctic 1

Peninsula Pacific margin: chronostratigraphic framework and magnetic 2

mineralogy conundrum 3

4

Alessandra Venuti (1), Fabio Florindo (1), Andrea Caburlotto (2), Mark W. Hounslow (5), Claus-5

Dieter Hillenbrand (4), Eleonora Strada (1,3), Franco M. Talarico (3), Andrea Cavallo (1) 6

7

(1) Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, 00143 Roma, Italy 8

(2) Istituto Nazionale di Oceanografia e di Geofisica Sperimentale, Borgo Grotta Gigante 42/c, 34010 Sgonico (TS), 9

Italy 10

(3) Dipartimento di Scienze della Terra, Università di Siena, Via del Laterino 8, 53100 Siena, Italy 11

(4) British Antarctic Survey, High Cross, Madingley Road, Cambridge CB3 0ET, United Kingdom 12

(5) Lancaster Environmnent Centre, Lancaster University, Farrer Ave., Lancaster LA1 4YQ, United Kingdom 13

14

Keywords: Antarctic Peninsula, Pacific margin, sediment drift, late Pleistocene, Mineral 15

magnetism, relative palaeointensity 16

17

Abstract We present results of detailed paleomagnetic investigations on deep-sea cores from 18

sediment drifts located along the Pacific continental margin of the Antarctic Peninsula. High-19

resolution magnetic measurements on u-channel samples provide detailed age models for three of 20

these cores collected from Drift 7, documenting an age of 122 ka for the oldest sediments recovered 21

near the drift crest at site SED-07 and a high sedimentation rate (11 cm/kyr) at site SED-12 located 22

close to the Alexander Channel system. Low- and high-temperature magnetic measurements in 23

conjunction with microscopic and mineralogic observations from drifts 4, 5 and 7 indicate that 24

pseudo-single domain detrital titanomagnetite (partially oxidized and with limited Ti-substitution) 25

is the dominant magnetic mineral in the drift sediments. The titanomagnetite occurs in two 26

magnetic forms: i) a low coercivity form similar to laboratory-synthesized titanomagnetite, and ii) a 27

2

high coercivity form (Bcr > 60 mT), which is probably present as inclusions or lamellar 28

intergrowths in composite silicate or Fe-Ti-oxide grains. These two forms vary in amount and 29

stratigraphic distribution across the drifts. We did not find evidence for the presence of diagenetic 30

magnetic iron sulfides as has been previously suggested for these drift deposits. The observed 31

change of magnetic mineralogy in sediments deposited during Heinrich-events on Drift 7 probably 32

relates to warming periods, which temporarily modified the normal glacial transport pathways of 33

glaciogenic detritus to and along the continental rise and thus resulted in deposition of sediments 34

with a different provenance. Understanding this sediment provenance delivery signature at a wider 35

spatial scale should provide information about ice-sheet dynamics in West Antarctica over the last 36

~100 kyr. 37

38

1. Introduction 39

A system of twelve giant deep-sea sediment drifts that are separated by large channels, eroded by 40

turbidity currents, characterizes the continental rise west of the Antarctic Peninsula (Figure 1). 41

These drifts provide the most proximal continuous sedimentary records of late Neogene to 42

Quaternary ice-sheet dynamics in this part of West Antarctica [Barker et al., 1999b; 2002]. The 43

sediment drifts are up to 300 km long, 100 km wide and 1 km thick, with the main axis elongated 44

perpendicular to the continental margin [e.g., Rebesco et al., 1996; 1997; 2002; 2007; Barker et al., 45

1999a; Amblas et al., 2006; Uenzelmann-Neben, 2006]. The drifts are part of a complex glacial 46

sedimentary feeder-dispersal system composed of lobes and troughs on the outer shelf, a steep 47

continental slope and deep-sea channels that separate the drifts on the upper rise [e.g., Rebesco et 48

al., 1998; Hernández-Molina et al., 2006]. 49

During the 1990’s, these drifts were intensely investigated by the SEDANO program (SEdiment 50

Drift of the ANtarctic Offshore) of the Italian Programma Nazionale di Ricerche in Antartide 51

(PNRA), the British Antarctic Survey (BAS), and the Ocean Drilling Program (ODP) Leg 178 52

[Barker et al., 1999b; 2002]. During two SEDANO cruises with the R/V OGS-Explora (1995/1997-53

3

98), seismic surveys, gravity coring, and moorings with current meters and sediment traps yielded 54

information about the morphology and mechanisms of deposition along the Pacific continental 55

margin [e.g., Camerlenghi et al., 1997a,b; Rebesco et al., 1997, 2002, 2007; Pudsey and 56

Camerlenghi, 1998; Harland and Pudsey, 1999; Lucchi et al., 2002; Giorgetti et al., 2003; Villa et 57

al., 2003; Lucchi and Rebesco, 2007]. The SEDANO program collected 17 gravity cores from Drift 58

7 and an additional two cores at Drift 4 [Lucchi et al., 2002], while BAS recovered 11 piston cores 59

from the drifts located to the NE of Drift 7 on cruise JR19 with RRS James Clark Ross [Pudsey, 60

2000]. 61

Paleomagnetic and mineral magnetic studies were conducted by Sagnotti et al. [2001] and Macrì et 62

al. [2006] on seven gravity cores (SED-02, SED-04, SED-06, SED-14, SED-15, SED-16, and SED-63

17) collected on Drift 7 (Figure 1). The ages of these sequences were determined using relative 64

paleomagnetic intensity (RPI) [Sagnotti et al., 2001; Macrì et al., 2006], with biostratigraphic and 65

chemostratigraphic information providing additional age constraints [Pudsey and Camerlenghi, 66

1998; Lucchi et al., 2002; Villa et al., 2003]. The RPI-based high-resolution age models established 67

by Sagnotti et al. [2001] and Macrì et al. [2006] suggested a link between the occurrence of 68

intervals with magnetic coercivity minima in Drift 7 and the simultaneous deposition of layers rich 69

in iceberg-rafted debris (IRD) in the North Atlantic [so-called ‘Heinrich layers’, e.g., Heinrich, 70

1988; Hemming, 2004] during major rapid cooling events between ~12 kyr and ~60 kyr ago. In 71

cores from Drift 7, a variable mixture of detrital pseudo-single domain (PSD) magnetite (Fe3O4) 72

and fine-grained monoclinic pyrrhotite (Fe7S8) was inferred by Sagnotti et al. [2001], with 73

magnetite being the main magnetic mineral [cf. Macrì et al., 2006]. Contrary to this mineralogical 74

interpretation, Hawkes et al. [2003] identified in sediment cores from drifts 3, 4, 4a and 5 only 75

magnetite. A full understanding of the magnetic mineral origin of the remanence is important, 76

because the diagenetic formation of an iron sulfide component may have delayed the magnetic field 77

recording process to some time after deposition [e.g., Florindo and Sagnotti, 1995; Roberts et al., 78

2005; Sagnotti et al., 2005; Florindo et al., 2007; Roberts et al., 2010a]. 79

4

The main objectives of our study are to: (1) provide robust identification of the main magnetic 80

carriers in the drift sediments, which is important for understanding the paleomagnetic recording 81

process, and (2) construct a robust chronostratigraphic framework for cores SED-7, SED-12, and 82

SED-13, which is which is important for reconstructing Late Pleistocene paleoenvironments of the 83

Antarctic Peninsula. The Antarctic Peninsula may represent the best natural laboratory to 84

investigate the interaction of atmosphere, hydrosphere and cryosphere and the response of Earth's 85

climate system to rapid regional warming [e.g., Florindo and Siegert, 2009]. Over the past 50 years, 86

the Antarctic Peninsula has experienced the greatest atmospheric temperature increase on Earth, 87

rising by nearly 3°C [e.g., King et al., 2003; Turner et al., 2005], which is approximately 10 times 88

the mean rate of present global warming [IPCC, 2007]. 89

We have investigated the paleomagnetic properties of two gravity cores (SED-12 and -13) located 90

near the Alexander Channel system and of two cores (SED-07 and -14) located near the crest of 91

Drift 7 (Figure 1). We also examined the spatial distribution of magnetic minerals in the Drift 7 92

sediments, and compared these results to the magnetic mineralogy in sediments recovered from 93

drifts 4 and 5 located to the NE of Drift 7. 94

95

2. Oceanographic setting, site locations and core lithostratigraphy 96

The Pacific continental margin of the Antarctic Peninsula is under the influence of the Southern 97

Boundary (SB) of the Antarctic Circumpolar Current (ACC), which is a massive eastward flowing 98

wind-driven current. The ACC plays an important role in the thermohaline circulation and heat 99

budget of the world ocean (Figure 1) [e.g., Barker et al., 2007; Carter et al., 2009]. Surface and 100

deep-water masses north of the SB are located within the Antarctic Zone, i.e. within the clockwise 101

flowing ACC. Nevertheless, Hillenbrand et al. [2008a] showed that in the area of investigation the 102

ACC only affects ocean circulation above about 1000 m water depth, whereas today, and for most 103

of the late Neogene and Quaternary, a generally southwestward flowing bottom current affected the 104

deposition on the drifts. Surface water currents on the continental shelf to the south of our study 105

5

area are driven by the westward flowing Antarctic Coastal (or Polar) Current, where the direction of 106

flow is mainly controlled by the orientation and shape of the Antarctic coast (Figure 1). 107

The large hemipelagic sediment drifts on the Pacific margin of the Antarctic Peninsula are 108

characterized by an asymmetric cross-section, with a steep southwestern side and a gently sloping 109

northeastern side (Figure 1). Based on both multi-channel seismic profiles [Rebesco et al., 1996; 110

1997; 2007] and core analysis [Pudsey and Camerlenghi, 1998; Pudsey, 2000; Lucchi et al., 2002; 111

Hillenbrand and Ehrmann, 2005], this geometry is interpreted to result from the interplay between 112

turbidity and bottom currents: fine-grained particles supplied by turbidity currents running through 113

the channels spill over the channel banks (in particular the western banks), and are entrained in the 114

SW-flowing bottom contour currents and so are redeposited over the continental rise, thereby 115

contributing to the growth of inter-channel sediment drifts [e.g., Rebesco et al., 2007]. The 116

sedimentary successions in cores from the gentle NE slope of Drift 7 facing the Alexander Channel 117

system (Figure 1) were therefore directly influenced by both turbidity and contour currents [Lucchi 118

et al., 2002]. 119

Up to nine lithostratigraphic units, named A to I, have been identified within the mid to late 120

Pleistocene succession of Drifts 1 to 7 [Pudsey and Camerlenghi, 1998; Pudsey, 2000; Lucchi et 121

al., 2002]. The lithostratigraphic units were correlated by magnetic susceptibility profiles, clay 122

mineral assemblages, sediment facies and textural characteristics, and were assigned to marine 123

isotope stages (MIS) 1 to 11 on the basis of bio- and chemostratigraphy and excess 230Thorium 124

activity [Pudsey and Camerlenghi, 1998; Pudsey, 2000; Lucchi et al., 2002; Villa et al., 2003]. 125

Interglacial units consist of brown bioturbated hemipelagic muds with microfossils (diatoms, 126

foraminifera, radiolaria, calcareous nannofossils) and IRD, while glacial units consist of olive-grey 127

to grey, laminated terrigenous muds [Pudsey and Camerlenghi, 1998; Pudsey, 2000; Lucchi et al., 128

2002; Villa et al., 2003]. Tephra layers with a trachytic composition occur within lithostratigraphic 129

units D and C (latest MIS 6 and early MIS 5) [Pudsey and Camerlenghi, 1998; Pudsey, 2000; 130

6

Lucchi et al., 2002], and provide important time-markers for correlating and dating cores from the 131

West Antarctic continental margin [Hillenbrand et al., 2008b]. 132

133

3. Material and methods 134

From the original set of 17 gravity cores from Drift 7, which were collected during the SEDANO-I 135

and -II projects, we selected four cores (SED-07, -12, -13, and -14) for paleomagnetic and magnetic 136

mineralogy investigations. These cores were collected near the crest of Drift 7 and in the proximity 137

of the Alexander Channel, respectively (Figure 1). The sediment cores do not contain sharp 138

boundaries between different lithological units, which suggests that major erosion did not affect the 139

recovered sequences [Lucchi et al., 2002]. However, lithostratigraphic Unit A (corresponding to 140

MIS 1) is missing in both cores SED-12 and SED-14 and appears to be condensed in core SED-07, 141

which suggests that, at least during the Holocene, erosion or non-deposition played an important 142

role at these sites. Alternatively, Unit A may have been lost or only partially recovered at these sites 143

during coring operations [Lucchi et al., 2002]. The cores were logged for magnetic susceptibility 144

using a Bartington Instruments MS2F point sensor on split cores [Pudsey and Camerlenghi, 1998] 145

and a Bartington Instruments MS2C loop sensor on the whole core [Lucchi et al., 2002]. 146

A suite of mineral magnetic measurements was also carried out on BAS piston cores PC102 and 147

PC111 from Drift 4 and PC107 and PC108 from Drift 5 (see inset in Figure 1 for location of the 148

drifts). Magnetic susceptibility had been measured at 2 cm intervals on the split surface of the 149

archive halves of these cores by Pudsey [2000] and Ó Cofaigh et al. [2001], who used a Bartington 150

Instrument susceptibility meter with MS2F probe (for a compilation of magnetic susceptibility data 151

from the SEDANO and BAS cores, see supplementary information in Hillenbrand et al. [2008b]. 152

All metadata for the SEDANO and BAS cores studied here are given in Table 1 and in Pudsey 153

[2000] and Lucchi et al. [2002]. 154

155

3.1. Sampling 156

7

Cores SED-07, SED-12, and SED-13 from Drift 7 were sampled using u-channels (1.2 m in length) 157

at the core repository of the Italian Museo Nazionale dell’Antartide in Trieste (u-channels are open-158

sided 2 cm x 2 cm square cross-sectioned liners made of non-magnetic plastic [Weeks et al., 1993]). 159

After removing the sediment-filled u-channel from the core, an airtight cover was clipped over the 160

u-channel to hold the sediment in place and to prevent it from drying. The u-channel ends were 161

sealed using plastic film and tape to minimize moisture loss. The u-channels were taken from the 162

center of the archive-halves of cores SED-7 and SED-13 and from the working halves of core SED-163

12 and the working half of one section of core SED-14. This section was sampled to investigate the 164

anomalously inclination values of the characteristic remanent magnetization (ChRM) observed on 165

the corresponding archive-half [Macrì et al., 2006]. 166

A suite of mineral magnetic measurements was carried out on sediment samples from 167

lithostratigraphic units A to C in cores PC102 and PC111 from Drift 4 and cores PC107 and PC108 168

from Drift 5. Discrete samples for magnetic analyses were collected by inserting ~10 cm3 plastic 169

boxes into the sediment, and sealing the resulting encapsulated specimens. 170

171

3.2. Magnetic analyses 172

All magnetic analyses on the SEDANO cores were performed at the paleomagnetism laboratory of 173

the Istituto Nazionale di Geofisica e Vulcanologia (INGV) in Rome. Natural and artificial 174

magnetizations were measured using a narrow-access pass-through 2-G Enterprises cryogenic 175

magnetometer (model 755R) with an internal diameter of 4.2 cm [Weeks et al., 1993], equipped 176

with three DC SQUID sensors (noise level 3x10-9 Am2kg-1), housed in a field-free room. The 177

natural remanent magnetization (NRM) was measured at 1 cm intervals, although smoothing occurs 178

due to the Gaussian shape of the response curve of the magnetometer pick-up coils [Weeks et al., 179

1993] (the half-power width suggests smoothing across 4.8 cm for the radial x and y directions and 180

5.9 cm for the axial z direction). Data from the upper and lower 5 cm of each u-channel were not 181

used, because these data are affected by “edge effects” due to the width of the magnetometer 182

8

response function. The NRM was then stepwise demagnetized using an in-line static demagnetizer 183

at peak alternating fields (AF) of 10, 20, 30, 40, 50, 60, 80, and 100 milliTesla (mT), with 184

measurements after each step. The ChRM directions and the maximum angular deviation (MAD) 185

[Kirschvink, 1980] were computed using the software of Mazaud [2005]. The studied cores lack 186

azimuthal orientation. The geocentric axial dipole field at the latitude of the coring sites (about 187

67°S; Table 1) has an inclination of -78°. 188

After demagnetization of the NRM, mineral magnetic analyses were carried out on both u-channel 189

and discrete samples. Low-field volume specific magnetic susceptibility (κ) was measured at 1-cm 190

intervals using a Bartington Instruments MS2C loop sensor (frequency of 0.565 kHz) with 40 mm 191

diameter, mounted in line with the 2-G Enterprises magnetometer system. An anhysteretic 192

remanent magnetization (ARM) was imparted using a solenoid that is mounted in-line with the 193

demagnetization coils on the INGV magnetometer system (a DC bias field of 0.1 mT was used in 194

conjunction with a peak AF of 100 mT) and was then measured at 1-cm intervals. 195

Additional mineral magnetic analyses were carried out on discrete samples (<50 mg) collected from 196

the u-channels of the Drift 7 cores. Isothermal remanent (IRM) acquisition curves, in 16 steps up to 197

0.5 T, and hysteresis properties, were determined using an alternating gradient magnetometer 198

(AGM) (Princeton Measurements Corporation Model 2900 Micromag). The hysteresis parameters 199

measured include the saturation magnetization (Ms), saturation remanent magnetization (Mrs), and 200

coercive force (Bc). Backfield demagnetization of the IRM was used to determine the coercivity of 201

remanence (Bcr). The ratios of Mrs/Ms and of Bcr/Bc are useful indicators of the domain state of 202

magnetic grains, as outlined by Day et al. [1977]. Continuously monitored temperature dependence 203

of κ (up to 700°C), was measured using an AGICO CS-3 furnace equipped Kappabridge KLY-3 204

(noise level 2x10−8 SI volume units), both in air and with argon gas flow (to prevent oxidation) in 205

order to identify Curie temperatures of the magnetic minerals [Hrouda, 1994]. 206

All magnetic measurements on the BAS sediment cores recovered from drifts 4 and 5 were carried 207

out at the Centre for Environmental Magnetism and Palaeomagnetism in Lancaster (UK). The 208

9

mineral magnetic analyses included measurement of low-frequency (χlf) and high-frequency mass-209

specific magnetic susceptibility using a Bartington Instruments MS2 meter, ARM acquisition in DC 210

fields of 0.08 mT (80 mT AF field), and backfield DC demagnetization at 20, 50, 100 and 300 mT 211

of a 1 T saturation IRM (SIRM). These remanences were measured on a Minispin fluxgate 212

magnetometer. Orthogonally applied IRM’s (using inducing fields of X mT, Y mT and Z mT) for 213

selected specimens were thermally demagnetized [Lowrie, 1990] to identify unblocking 214

characteristics of different coercivity fractions. This was performed using a Magnetic 215

Measurements thermal demagnetizer, with remanences measured on a CCL 3-axis SQUID 216

magnetometer. Thermal unblocking (in zero magnetic field) of a 2.5 T IRM applied at 10 K was 217

measured on a Quantum Design magnetic property measurement system, to examine possible low- 218

temperature magnetic transitions associated with titanomagnetite and pyrrhotite. 219

220

3.3. Microscopic analyses and mineral extraction 221

Scanning electron microscopy and energy dispersive spectrometry (SEM-EDS) were performed to 222

qualitatively identify opaque minerals in selected samples from glacial and interglacial 223

lithostratigraphic units in Drift 7 cores SED-12 (close to the Alexander Channel system) and SED-224

14 (drift crest). We used a Philips 515 scanning electron microscope with an EDAX 9100/70 X-ray 225

energy dispersive detector, running at 15 kV and with an emission current of 20 nA at the 226

Dipartimento di Scienze della Terra, Università di Siena. Additional analyses were carried out on a 227

set of resin impregnated polished sections from cores SED-07, SED-12, and SED-14 using a field 228

emission scanning electron microscope (FESEM) operated at 10 kV. In addition, a subset of 229

samples from the same cores was gently ground into a fine powder and magnetic minerals were 230

extracted using a Nd-Fe-B-REE (Rare Earth Element) magnet housed in a plastic sheath. Elemental 231

analyses were performed on individual minerals using a JEOL JSM-6500F thermal FESEM, which 232

is a high performance analytical FESEM that integrates with the JEOL energy dispersive X-ray 233

analyzer. 234

10

Magnetic mineral extracts from selected specimens taken from the cores recovered on drifts 4 and 5 235

were performed using both a quick-method (<1 hour) of extraction (Me) and the magnetic probe 236

(Mp) method detailed in Hounslow and Maher [1999]. The former method tends to limit oxidation 237

of sulphides, but strongly biases magnetic minerals to larger grain sizes and the more strongly 238

particles. In contrast, the latter method is more representative and removes also the smaller 239

magnetic particles, but the longer extraction times (7 days) tend to oxidize any greigite that may be 240

present [Hounslow and Maher, 1996]. The extracts were then examined optically under transmitted 241

and reflected light at magnifications up to 1000, and were subjected to X-ray diffraction to evaluate 242

their mineralogy. The magnetite lattice spacing was determined with reference to the quartz peaks 243

[e.g., Hounslow and Maher, 1999]. 244

245

3.4. Chronology 246

Age models were developed for cores SED-12, SED-13, and SED-07 by peak-to-peak correlation 247

of their physical properties with those of the SEDANO cores that have been dated using RPI 248

correlation [Sagnotti et al., 2001] using the ‘linage’ function of the AnalySeries 2.0.4.2 software 249

package [Paillard et al., 1996] (see paragraph 4.4 below). We correlated the κ profile of core SED-250

12 with that of SED-02 and the κ profile of core SED-07 with those of cores SED-04 and SED-06, 251

respectively (for core locations see Figure 1). The ARM/κ ratio was then used for correlating core 252

SED-13 to core SED-12. In this case, we preferred the ARM/κ rather than the magnetic 253

susceptibility because the former has a clearer correspondence between the two profiles. 254

255

4. Results and interpretation 256

4.1. Paleomagnetism 257

The NRM intensity of the studied specimens is between 0.1 and 0.2 A/m and is characterized by 258

stable demagnetization behavior with MAD values mostly < 5° and never > 10°. Typical 259

demagnetization behavior is shown in Figure 2. Most of the studied intervals have a low-stability, 260

11

high-angle normal polarity remanence acquired during coring and/or storage that was successfully 261

removed at 10-20 mT. The ChRM was clearly identified as a single, stable and well-defined 262

remanence component above 20 mT for most of the analyzed sediments (Figure 2). For some 263

intervals, samples were not completely demagnetized at a peak AF of 100 mT, which indicates the 264

presence of a high-coercivity magnetic phase. The ChRM inclination record for these cores always 265

indicates normal polarity (negative inclination), but the inclination records have occasional swings 266

to inclinations of about -30° (Figure 3), possibly through the smoothed recording of geomagnetic 267

excursions caused by post-depositional remanent magnetization (PDRM) acquisition [e.g., Roberts 268

and Winklhofer, 2004] and/or because of the presence of small and highly magnetic pebbles within 269

the sediment [e.g., Venuti and Florindo, 2004; Florindo et al., 2005]. The latter hypothesis is 270

supported by petrographic analysis on granule- to pebble-sized clasts from these cores, which 271

documents the presence of rhyolite, basalt, cataclastic gabbro, quartzodiorite, volcanoclastic 272

microbreccia with fine-grained biotite-muscovite gneiss and siltstone. Both the occurrence of clasts 273

with highly magnetic petrologies and the effects of granule and pebble contents on magnetic 274

susceptibility records have been previously reported for sediment cores from Antarctic Peninsula 275

drifts [Pudsey and Camerlenghi, 1998; Pudsey, 2000; Ó Cofaigh et al., 2001; Macrì et al., 2006; 276

Cowan et al., 2008]. An interval with unusually shallow ChRM inclinations, which Macrì et al. 277

[2006] had previously observed between ca 3.1 and 4.1 mbsf in the archive half of core SED-14, is 278

also evident from our measurements on the working half (Figure 3b). Thus, our finding supports the 279

conclusion of Macrì et al. [2006] that the relatively coarse-grained lithology in the corresponding 280

section of core SED-14 is responsible for the shallow ChRM inclinations. We argue also that the 281

shallow inclinations observed in some intervals of the other SEDANO cores (Figure 3) may be 282

caused by higher contents of coarser-grained material. 283

284

4.2 Mineral-magnetic analyses 285

4.2.1 Magnetic mineralogy of sediments from Drift 7 286

12

In the cores from Drift 7, variations of concentration-dependent magnetic parameters (κ and ARM) 287

(Figure 3) remain within about the same order of magnitude (κ ranges between 13 and 135 x 10-5 SI 288

and ARM ranges between 0.1 and 0.8 A/m). The largest variability occurs in silty layers, which are 289

mainly present within glacial lithostratigraphic Unit B of cores SED-12 and SED-13 near 290

Alexander Channel [Lucchi et al., 2002]. Down-core variations in NRM intensity are not associated 291

with changes in ARM and κ. 292

Continuous monitoring of κ-T changes indicates Curie temperatures between 586 and 605°C 293

(average = 597°C), which indicates the ubiquitous presence of Fe-spinels with a composition 294

similar to magnetite (Figure 4). A few samples (e.g., figure 4h) have κ remaining at 640-680°C 295

which indicates the presence of hematite. The cooling curves are often above the heating curve 296

which indicates the production of new magnetic phases during the continuous heating cycle. For 297

some samples, the continuously monitored κ-T data contain a slight decrease in susceptibility at 298

280-300°C, a feature also observed by Sagnotti et al. [2001], which they attributed to the presence 299

of fine-grained monoclinic pyrrhotite. For such samples, we repeated the thermomagnetic 300

experiment (Figure 4d) on a new sample from the same level. The κ-T runs were first stopped at 301

350°C (red line in Figure 4d) to detect the pyrrhotite Curie temperature [Dekkers, 1989] which 302

should be evident during both heating and cooling cycles [Horng and Roberts, 2006]. The heating-303

cooling cycle was then repeated from room temperature up to 700°C (black line in Figure 4d). The 304

major decrease observed above 300°C is not reversible during the cooling cycle, and it is not 305

present during the second heating run. If the monoclinic pyrrhotite were stable to heating, this 306

experiment should give an indication of a Curie temperature at 325°C [Dekkers, 1989] during both 307

heating and cooling cycles [e.g., see Figure 4c,d in Horng and Roberts, 2006]. The peak in the κ-T 308

runs (Figure 4) and the low temperature (< 400°C) thermal instability in both κ-T runs may be 309

caused by the presence of greigite (Fe3S4) [e.g., Roberts, 1995; Roberts et al., 1999; Roberts et al., 310

2010b]. 311

13

Hysteresis loop data for samples from Drift 7 have regular shapes, with no evidence of wasp-312

waisted features that might indicate an important contribution from grains near the super-313

paramagnetic (SP) threshold grain size [Roberts et al., 1995; Tauxe et al., 1996]. The hysteresis 314

ratios are characterized by Mrs/Ms values between 0.10 and 0.26 and Bcr/Bc ratios between 2.72 and 315

5.76 that plot in the PSD grain-size range for pure magnetite (0.5 < d < 10 µm) [Day et al., 1977] 316

(Figure 5). However, the hysteresis data do not fall on the SD-multidomain (MD) mixing line for 317

magnetite, but have higher Bcr/Bc values, which might result from a contribution from SP particles 318

[Dunlop, 2002]. IRM acquisition curves on sediment chips, in fields up to 0.5 T, indicate that 319

saturation is nearly reached at fields of about 0.4 T for all samples (Figure 6). This suggests the 320

dominance of a low to moderate coercivity mineral with no significant amounts of anti-321

ferrimagnetic phases, such as hematite or goethite. 322

Macrì et al. [2006] showed that in Drift 7 sediments the median destructive field of the NRM 323

(MDFNRM, i.e., the peak AF at half the initial NRM intensity) is related to Bcr. The MDFNRM is 324

generally higher (75 to 90 mT) in the glacial lithostratigraphic units and lower (30 to 60 mT) in 325

both the interglacial units and in some thin (a few tens of centimeters thick) intervals within glacial 326

unit B (Figure 3). These thin, low-MDFNRM intervals have also been observed at sites SED-02 and 327

SED-04 by Sagnotti et al. [2001], who correlated them to Heinrich events 1 to 6 in the North 328

Atlantic. Sagnotti et al. [2001] argued that this magnetic response reflects a dominance of detrital 329

magnetite (for MDFNRM < 60 mT) and pyrrhotite (for MDFNRM > 60 mT), respectively. According 330

to Macrì et al. [2006], this dual magnetic behavior corresponds to values of Bcr < ~ 65-70 mT and 331

Bcr > ~70 mT, respectively. 332

333

4.2.2 Magnetic mineralogy of sediments from drifts 4 and 5 334

As for Drift 7 samples, a dual mineral magnetic behavior is apparent on drifts 5 and 4, where larger 335

Bcr values are also associated with large SIRM/χlf values (Figure 7). However, here the response is 336

not related to the lithostratigraphic units, but to the geographical location of the cores, with some 337

14

Unit B sediments dominated by high coercivity behaviour (Bcr 60-85 mT, e.g., core PC108) and 338

others dominated by lower coercivity behaviour (Bcr 35-50 mT) (Figure 7). Interglacial 339

lithostratigraphic units A and C possess high and intermediate Bcr values. When compared to the 340

data from synthetic and sized magnetic oxides of Peters and Dekkers [2003], the Bcr and SIRM/χlf 341

data of the cores from drifts 4 and 5 suggest a behaviour incompatible with magnetic sulphides, but 342

compatible with magnetite (Figure 7). Moreover, our samples display a larger range of Bcr values 343

that do not overlap these laboratory prepared materials. 344

The thermal demagnetisation of orthogonal IRMs likewise shows the importance of the magnetite 345

Curie temperature around 600°C, but the heating-cooling cycles have a dramatically different effect 346

on the susceptibility, which shows both low (< 400°C) and high (> 500°C) temperature 347

mineralogical alteration (Figure 8a-d). The low temperature IRMs show the magnetite Verwey 348

transition at ~111 K (Figure 8e). X-ray diffraction of the magnetic mineral extracts shows 349

magnetite dominants, with lattice constants of 8.388Å to 8.405 Å (Figure 9). These values, in 350

combination with the Curie temperatures, indicate that the titanomagnetites are slightly oxidised (z 351

< 0.3), with low ulvospinel contents (x < 0.1) according to the synthesised data of Nishitani and 352

Kono [1983]. One of the magnetic extracts (from 356 cm depth in core PC108; Figure 9) shows 353

possible small amounts of haematite. Mean percentages of frequency-dependent magnetic 354

susceptibility (%χfd) are 1.8% (σ = 0.7, n = 303), with an average percentage of IRM acquired 355

between 0.3 and 1 Tesla of 1.2% (σ = 0.6, n = 303), and thus indicate very low SP magnetite 356

contributions and low contents of anti-ferrimagnetic minerals. 357

No pyrrhotite is apparent in the X-ray diffraction of the magnetic extracts (Figure 9), and the 358

pyrrhotite low temperature transition near 34 K [Rochette et al., 1990] is not present in the IRM 359

unblocking data (Figure 8e, f). The apparent unblocking up to ~300°C observed in the orthogonal 360

IRM data (Figure 8) is common in greigite dominated magnetisations, but is by no means a 361

phenomenon restricted to greigite. No greigite is detected in the magnetic extracts (Figure 9), which 362

implies that it is either absent or present in very low concentrations, and therefore could not account 363

15

for the large drop (from ~20 to 300°C) in IRM intensity, in all coercivity fractions of the thermal 364

demagnetisation data (Figure 8). 365

366

4.3 Microscopy 367

Characterisation of opaque minerals in Drift 7 samples, based on incident light polarized 368

microscopy and completed by means of EDS analysis of selected grains, indicate that detrital Fe 369

oxides (mostly titanomagnetite with low Ti substitutions), and Fe-Ti oxides (ilmenite, often Mn-370

rich) are abundant. Moreover, occasional hexagonal Fe-oxide grains (inferred to be hematite) were 371

observed in a sample from the SED-12 core (from 4.40 mbsf). SEM and FESEM analysis revealed 372

the occurrence of rare (<1%) Fe sulfides with irregular shape dispersed within a clayey matrix in 373

samples from glacial lithostratigraphic units of cores SED-07 and SED-12 (from 2.77-2.78 mbsf 374

and 4.40 mbsf, respectively) and both interglacial and glacial units of core SED-14 (from 1.89 and 375

3.80 mbsf). Sulfides occur as <1 µm aggregates in the glacial unit of core SED-07. These features 376

are typical of authigenic Fe sulfide phases that grow during diagenesis [Rowan and Roberts, 2005; 377

2006]. The size of the aggregates is smaller than the interaction volume of the FESEM electron 378

beam (1 µm), so that other elemental peaks (e.g., Si, Al, Mg) are recorded in addition to Fe and S 379

peaks, which indicates contamination (or scattering X-rays) with adjacent matrix minerals (e.g., 380

quartz, chlorite). It has therefore not been possible to measure the chemistry of the sulfides. 381

Optical microscopy of the magnetic extracts from drifts 4 and 5 shows that ~ 40% of the volume is 382

composed of angular detrital opaque phases up to about 20 µm in size. The remainder is mostly 383

composed of quartz and feldspar, and smaller amounts of mica (chlorite), diatom fragments and 384

trace amounts of Cr-spinels. The silicate particles often contain opaque inclusions. In reflected 385

light, a search was made for iron sulphides (having brassy-coloured reflections) in the extracts, but 386

neither nodules, framboids or polyframboidal aggregates were found, indicating the absence of 387

sulphides greater than ~ 2µm in size. 388

16

These investigations indicate that low-Ti- substituted, partially oxidixed titanomagnetite dominate 389

the magnetic properties of all drift sediments. Magnetic sulphides make no significant contribution 390

to the magnetic properties. The rare < 1µm sulphides detected by EDAX in the Drift 7 sediments 391

are probably derived from rock particles. Haematite probably makes a very small contribution to 392

the magnetic properties, but cannot account for the larger than normal Bcr values observed in most 393

drift cores. 394

395

4.4 Age models 396

The magnetic susceptibility record from core SED-12 was optimally matched with the 397

susceptibility record from core SED-02, which had been dated using its RPI curve [Sagnotti et al., 398

2001]. After this procedure (Figure 10a), the two susceptibility records are significantly correlated 399

(Pearson correlation coefficient r2=0.80), and the resulting age model indicates an age of 77 ka for 400

the sediments at the base of core SED-12. The correlation of the magnetic susceptibility record of 401

core SED-07 with those of RPI-dated cores SED-04 and SED-06 indicates an age of 122 ka for the 402

oldest sediments at site SED-07 (Figure 10b). Our age model for core SED-07 is consistent with the 403

age constraints provided by the tephra layers in Unit C and at the top of Unit D [Hillenbrand et al., 404

2008b]. The correlation of the ARM/κ ratios of cores SED-12 and SED-13 indicates that core SED-405

13 spans the last 76 kyr (Figure 10c). Our age models for the Drift 7 cores imply that sedimentation 406

rates are highest near the Alexander Channel (12 cm/kyr in core SED-12 and 9 cm/kyr in core SED-407

13, respectively), where sedimentation is directly influenced by both turbidity and contour currents, 408

compared to the values recorded toward the crest of Drift 7 (5 cm/kyr in core SED-07) (see also 409

figure 6 in Macrì et al., 2006). 410

411

5. Discussion and conclusions 412

Previous investigations have demonstrated the huge potential of sedimentary sequences recovered 413

from giant deep-sea drifts on the Pacific continental rise of the Antarctic Peninsula in 414

17

reconstructing the late Neogene and Pleistocene paleoenvironmental history of West Antarctica 415

[e.g., Pudsey and Camerlenghi, 1998; Pudsey, 2000; Lucchi et al., 2002; Hillenbrand, and 416

Ehrmann, 2005]. Detailed reconstructions of the sedimentary evolution of the drifts and its 417

relationship to climate change and glacial history of the Antarctic Peninsula can only be carried out, 418

if a high-resolution and robust chronostratigraphic framework can be established for the sediment 419

cores. Calcareous (micro-)fossils are either absent or not continuously present in the terrigenous, 420

glaciomarine sediments from the Antarctic continental margin, which prevents the applications of 421

reliable AMS 14C dating or oxygen isotope stratigraphy. Therefore, age models with the potential to 422

resolve millennial time scales can only be established for these sequences by measuring their RPI 423

records and correlating the RPI records to well-dated reference curves [Guyodo et al., 2001; 424

Sagnotti et al., 2001; Macrì et al., 2006]. This dating technique, however, can only be successfully 425

applied, if it is convincingly demonstrated that the magnetic mineralogy of the drift sediments is 426

suitable for reliable recording of the original geomagnetic field. 427

Our new microscopic observations and measurements on sediments and magnetic extracts, taken 428

from late Pleistocene sediment cores recovered from the drifts on the western Antarctic Peninsula 429

rise, demonstrate that these sequences have a magnetic mineralogy suitable for the application of 430

RPI dating. Our results indicate that detrital PSD titanomagnetite (z <0.3; x<0.1) represents the 431

main magnetic mineral in the drift sediments. Based on thermomagnetic, low temperature, X-ray 432

diffraction and microscopy data we find no evidence for either pyrrhotite or greigite, which 433

contrasts with the inferences of Sagnotti et al. [2001]. There are rare Fe-sulphides present, and 434

these appear to be restricted to the internal parts (i.e. inclusion) of clastic particles. Haematite 435

appears to be present in minor amounts. Instead, we consider the presence of “magnetically hard 436

titanomagnetite” as the most likely explanation of the observed higher coercivities, which are in 437

excess of those known from synthetic titanomagnetites [Peters and Dekkers, 2003]. Similarly, 438

interacting single-domain magnetite of bacterial origin (still in chains) has remanent coercivities of 439

45-50 mT [Pan et al., 2005], which are lower than those found in some of our sediments. The 440

18

remanent coercivities in excess of 60 mT are similar to those reported from silicate-hosted 441

titanomagnetites oxides [Feinberg et al., 2005]. In such particles, high elongation ratio and internal 442

sub-division (resulting in magnetostatically interacting single-domain particles) are key factors that 443

dictate the high remanent coercivities. Unlike the data of Brachfeld et al. [2001] from ODP Site 444

1096 near the crest of Drift 7 [Barker et al., 1999b; 2002], we find no evidence for defect or 445

substantial cation substitution, which could account for such large Bcr values, if these behave like 446

synthetic titanomagnetites. In Drift 7, the supply of “hard titanomagnetite” increased significantly 447

during glacial periods, when the Antarctic Peninsula Ice Sheet reached its maximum extent across 448

the adjacent shelf, resulting in a higher terrigenous input to the continental rise [Pudsey, 2000; 449

Lucchi et al., 2002; Hillenbrand and Ehrmann, 2005]. The same appears not to be the case on drifts 450

4 and 5, where this hard-magnetite component is not restricted to the glacial lithostratigraphic units 451

of the corresponding sediment cores. We attribute these differences in magnetic mineralogy 452

between the drifts to regional differences in the provenance of the terrigenous detritus supplied to 453

the continental rise. A regional variability in provenance has previously been inferred from the 454

anorganic chemical and clay mineral composition of drift sediments along the margin [Pudsey, 455

2000; Lucchi et al., 2002] and from the petrological composition of their coarse-grained fractions 456

[Cowan et al., 2008]. 457

We have provided new high-resolution age models by analysing magnetic properties in three 458

sediment cores recovered near the Alexander Channel and near the crest of Drift 7, respectively. 459

High-resolution measurements on u-channels facilitated the correlation of these cores with RPI- 460

dated cores from the same drift. The resulting age models document a relatively long history of 461

sedimentation in core SED-07, which has an age of 122 ka at its base. Conversely, the 462

sedimentation rates are much higher near the Alexander Channel (cores SED-12 and SED-13), 463

where sedimentation is directly influenced by turbidity currents and bottom currents. Consequently, 464

the oldest sediments recovered at these sites are younger than 80 ka. 465

19

Similar to cores from Drift 7 studied by Sagnotti et al. [2001], the MDFNRM curves in cores SED-466

07, SED-12 and SED-13 reveal distinct minima within intervals of glacial lithostratigraphic Unit B 467

(Figure 2d) that they correlate to Heinrich events 1 to 4 in the Northern Hemisphere (Figure 3). 468

Sagnotti et al. [2001] argued that, in contrast to the other glacial sediments of Unit B, the diagenetic 469

formation of pyrrhotite was drastically or completely suppressed in these layers due to a major 470

reduction in the supply of organic matter to the seafloor, which resulted from significantly 471

decreased biological productivity, caused by an intensification of sea-ice coverage in response to 472

cooling. This explanation, is in conflict with observations that Heinrich events seem to coincide 473

with i) periods of atmospheric warming in Antarctica [e.g., EPICA, 2006], and ii) enhanced 474

plankton productivity in the Southern Ocean, both to the north [Sachs and Anderson, 2005] and to 475

the south [Anderson et al., 2009] of the Antarctic Polar Front. Our more comprehensive magnetic 476

mineralogical data indicate the absence of pyrrhotite in the drift sediments from the Antarctic 477

Peninsula margin which is likewise inconsistent with the hypotheses of Sagnotti et al. [2001]. 478

Moreover, our magnetic mineralogical findings indicate only trace amounts of iron sulphides (if 479

any) in the glacial drift sediments and thus do not confirm the suggestion of Lucchi and Rebesco 480

[2007] that oxygen-reduced deep and bottom waters bathed the Antarctic Peninsula continental rise 481

during the glacial periods of the late Pleistocene. 482

Our data suggest that the MDFNRM minima recorded in the Drift 7 sediments, which have been 483

associated with Heinrich events, are related to changes in the delivery of Fe-oxides with differing 484

microstructural characteristics. These short-term temporal shifts in sediment provenance were 485

probably caused by the warming events that modified the delivery of glaciogenic detritus to the 486

drifts by affecting either the bottom-current transport of fine-grained detritus along the continental 487

rise or the glacial drainage pattern on the shelf and in the hinterland. 488

489

Acknowledgements 490

20

This work was supported by the Italian PNRA grants for the SEDANO-I and -II projects and the 491

BAS Palaeo-Ice Sheets and SAGES-500K projects. We thank Carol Pudsey, Ian Hawkes and 492

Barbara Maher. Furthermore, we gratefully acknowledge the support of the captains, officers, crew 493

and scientists who participated in the various cruises. 494

495

496

21

References 496

Amblas, D., Urgeles, R., Canals, M., Calafat, A.M., Rebesco, M., Camerlenghi, A., Estrada, F., De 497

Batist, M., Hughes-Clarke, J.E., 2006. Relationship between continental rise development and 498

palaeo-ice sheet dynamics, Northern Antarctic Peninsula Pacific margin. Quat. Sci. Rev. 25, 499

933-944. 500

Anderson, R.F., Ali, S., Bradtmiller, L.I., Nielsen, S.H.H., Fleisher, M.Q., Anderson, B.E., Burckle, 501

L.H., 2009. Wind-driven upwelling in the Southern Ocean and the deglacial rise in atmospheric 502

CO2. Science 323, 1443-1448. 503

Barker, P.F., Barrett, P.J., Cooper, A.K., Huybrechts, P., 1999a. Antarctic glacial history from 504

numerical models and continental margin sediments. Palaeogeogr. Palaeoclimatol. Palaeoecol. 505

150, 247-267. 506

Barker, P.F., Camerlenghi, A., Acton, G.D., et al., 1999b. Proc. ODP, Init. Repts. 178: College 507

Station, TX (Ocean Drilling Program). doi:10.2973/odp.proc.ir.178.1999. 508

Barker, P.F., Camerlenghi, A., Acton, G.D., Ramsay, A.T.S. (Eds.), 2002. Proc. ODP, Sci. Results, 509

178: College Station, TX (Ocean Drilling Program). doi:10.2973/odp.proc.sr.178.2002. 510

Barker, P.F., Filippelli, G.M., Florindo, F., Martin, E.E., Scher, H.D., 2007. Onset and role of the 511

Antarctic Circumpolar Current, Deep-Sea Research II 54, 2388-2398. 512

Brachfeld, S.A., Guyodo, Y., Acton, G.D., 2001. Data report: the magnetic mineral assemblage of 513

hemipelagic drifts, ODP Site 1096. In Barker, P.F., Camerlenghi, A., Acton, G.D., and Ramsay, 514

A.T.S. (Eds.), Proc. ODP, Sci. Results. 178: College Station, TX (Ocean Drilling Program), 1–515

10. doi:10.2973/odp.proc.sr.178.233.2001. 516

Camerlenghi, A., Crise, A., Pudsey, C.J., Accerboni, E., Laterza, R., Rebesco, M., 1997a. Ten-517

month observation of the bottom current regime across a sediment drift of the Pacific margin of 518

the Antarctic Peninsula. Antarct. Sci. 9, 426-433. 519

Camerlenghi, A., Rebesco, M., Pudsey, C.J., 1997b. High resolution terrigenous sedimentary record 520

of a sediment drift on the Antarctic Peninsula Pacific Margin (Initial Results of the ‘SEDANO’ 521

22

Program). In: Ricci, C.A. (Ed.), The Antarctic Region: Geological Evolution and Processes. 522

Terra Antartica Publication, Siena, pp. 705-710. 523

Carter, L., McCave, I.N., Williams, M.J.M., 2009. Circulation and Water Masses of the Southern 524

Ocean: A review. In: Florindo, F., Siegert, M. (Eds.), Antarctic Climate Evolution, 525

Developments, Earth and Environmental Sciences, vol. 8, Elsevier, pp. 85-114. 526

Chapman, M.R., Shackleton N.J., Duplessy, J.C., 2000. Sea surface temperature variability during 527

the last glacial-interglacial cycle: assessing the magnitude and pattern of climate change in the 528

North Atlantic. Palaeogeogr. Palaeoclimatol. Palaeoecol. 157, 1-25. 529

Cowan, E.A., Hillenbrand, C.-D., Hassler, L.E., Ake, M.T., 2008. Coarse-grained terrigenous 530

sediment deposition on continental rise drifts: A record of Plio-Pleistocene glaciation on the 531

Antarctic Peninsula. Palaeogeogr. Palaeoclimatol. Palaeoecol. 265, 275-291. 532

Day, R., Fuller, M., Schmidt, V.A., 1977. Hysteresis properties of titanomagnetites: grain-size and 533

compositional dependence. Phys. Earth Planet. Inter. 13, 260–267. 534

Dekkers, M.J., 1989. Magnetic properties of natural pyrrhotite. II. High- and low-temperature 535

behaviour of Jrs and TRM as function of grain size. Phys. Earth Planet. Inter. 57, 266-283. 536

Dunlop, D.J., 2002. Theory and application of the Day plot (Mrs/Ms versus Hcr/Hc): 1. Theoretical 537

curves and tests using titanomagnetite data. J. Geophys. Res. 107, 2056, 538

doi:10.1029/2001JB000486. 539

EPICA Community Members, 2006. One-to-one hemispheric coupling of millennial polar climate 540

variability during the Last Glacial. Nature 444, 195-198. 541

Feinberg, J.M Scott, G.R. Renne, P.R., Wenk, H-R., 2005. Exsolved magnetite inclusions in 542

silicates: Features determining their remanence behaviour. Geology 33, 513-516. 543

Florindo, F., Sagnotti, L., 1995. Palaeomagnetism and rockmagnetism in the upper Pliocene Valle 544

Ricca (Rome, Italy) section. Geophys. J. Int. 123, 340-354. 545

Florindo, F., Wilson, G.S., Roberts, A.P., Sagnotti, L., Verosub, K.L., 2005. Magnetostratigraphic 546

chronology of a Late Eocene to early Miocene glacimarine succession from the Victoria Land 547

23

Basin, Ross Sea, Antarctica, Global and Planetary Changes, 548

doi:10.1016/j.gloplacha.2004.09.009, 207-236. 549

Florindo, F., Karner, D.B., Marra, F., Renne, P.R., Roberts, A.P., Weaver, R., 2007. Radioisotopic 550

age constraints for Glacial Terminations IX and VII from aggradational sections of the Tiber 551

River delta in Rome, Italy. Earth Planet. Sci. Lett. 256, 61-80. 552

Florindo, F., and Siegert, M. (Eds), 2009. Antarctic Climate Evolution, in: Developments in Earth 553

and Environmental Sciences series, Elsevier, 8, pp. 1-593. 554

Giorgetti, A., Crise, A., Laterza, R., Perini, L., Rebesco, M., Camerlenghi, A., 2003. Water masses 555

and bottom boundary layer dynamics above a sediment drift of the Antarctic Peninsula Pacific 556

Margin. Antarc. Sci. 15, 537-546. 557

Guyodo, Y., Acton, G. D., Brachfeld, S., Channell, J.E.T., 2001. A sedimentary paleomagnetic 558

record of the Matuyama chron from the Western Antarctic margin (ODP Site 1101). Earth 559

Planet. Sci. Lett. 191, 61-74. 560

Harland, R., Pudsey, C.J., 1999. Dinoflagellate cysts from sediment traps deployed in the 561

Bellingshausen, Weddell and Scotia seas, Antarctica. Mar. Micropaleontol. 37, 77-99. 562

Hawkes, I.J., Hounslow, M.W., Pudsey, C.J., Maher, B.A., 2003. Sediment dispersal pathways 563

from the western Antarctic Peninsula during the last glacial period and the significance of 564

temporal fluctuations in magnetic properties. Geophys. Res. Abstr., 5, 11099. 565

Heinrich, H., 1988. Origin and consequences of cycling ice rafting in the northeast Atlantic Ocean 566

during the past 130 000 years. Quat. Res. 29, 143-152. 567

Hemming, S.R., 2004. Heinrich events: Massive late Pleistocene detritus layers of the North 568

Atlantic and their global climate imprint. Rev. Geophys. 42, RG1005, 569

doi:10.1029/2003RG000128. 570

Hernández-Molina, F.J., Larter, R.D., Maldonado, A., Rodríguez-Fernández, J., 2006. Evolution of 571

the Antarctic Peninsula Pacific margin offshore from Adelaide Island since the late Miocene: an 572

example of a glacial passive margin. Terra Antart. Rep. 12, 81-90. 573

24

Hillenbrand, C.-D., Ehrmann, W., 2005. Late Neogene to Quaternary environmental changes in the 574

Antarctic Peninsula region: evidence from drift sediments. Global and Planetary Change 45, 575

165-191. 576

Hillenbrand, C.-D., Camerlenghi, A., Cowan, E.A., Hernandez-Molina, F.J., Lucchi, R.G., 577

Rebesco, M., Uenzelmann-Neben, G., 2008a. The present and past bottom-current flow regime 578

around the sediment drifts on the continental rise west of Antarctic Peninsula. Mar. Geol. 255, 579

55-63. 580

Hillenbrand, C.D., Moreton, S.G., Caburlotto, A., Pudsey, C.J., Lucchi, R.G., Smellie, J.L., Benetti, 581

S., Grobe, H., Hunt, J.B., Larter, R.D., 2008b. Volcanic time-markers for Marine Isotopic Stages 582

6 and 5 in Southern Ocean sediments and Antarctic ice cores: implications for tephra 583

correlations between palaeoclimatic records. Quat. Sci. Rev. 27, 518-540. 584

Horng, C.S., Roberts, A.P., 2006. Authigenic or detrital origin of pyrrhotite in sediments?: 585

Resolving a paleomagnetic conundrum. Earth Planet. Sci. Lett. 241, 750-762. 586

Hounslow, M.W., Maher, B.A., 1996. Quantitative extraction and analysis of carriers of 587

magnetisation in sediments. Geophys. J. Int. 124, 57-74. 588

Hounslow, M.W., Maher, B.A. 1999. Laboratory procedures for quantitative extraction and analysis 589

of magnetic minerals from sediments. In: Walden, J., Oldfield, F., Smith, J. (Eds.) 590

Environmental magnetism, a practical guide. Quaternary Research Association, Technical Guide 591

No. 6., pp.139-164. 592

Hrouda, F., 1994. A technique for the measurements of thermal changes of magnetic susceptibility 593

of weakly magnetic rocks by the CS-2 Apparatus and KLY-2 Kappabridge, Geophys. J. Int. 118, 594

604-612. 595

Hu, S., Stephenson, A., Appel, E., 2002. A study of gyroremanent magnetisation (GRM) and 596

rotational remanent magnetisation (RRM) carried by greigite from lake sediments. Geophys. J. 597

Int. 151, 469-474. 598

25

IPCC, 2007. Summary for policymakers. In: Solomon, S., Qin, D., Manning, M., Chen, Z., 599

Marquis, M., Averyt, K.B., Tignor, M., Miller, H.L., (Eds.), Climate change 2007: The physical 600

science basis. Contribution of working group I to the fourth assessment report of the 601

intergovernmental panel on climate change. Cambridge University Press, Cambridge, United 602

Kingdom and New York, NY, USA, 996 pp. 603

King, J.C., Turner, J., Marshall, G.J., Connolley, W.M., Lachlan-Cope, T.A., 2003. Antarctic 604

Peninsula climate variability and its causes as revealed by analysis of instrumental records. In: E. 605

Domack, A. Leventer, A. Burnett, R. Bindschadler, P. Convey, M. Kirby (Eds). Antarctic 606

Peninsula Climate Variability: Historical and Paleoenvironmental Perspectives. Antarctic 607

Research Series 79, American Geophysical Union, Washington, DC, pp. 17-30. 608

Kirschvink, J.L., 1980. The least-square line and plane and the analysis of palaeomagnetic data. 609

Geophys. J. R. Astron. Soc. 62, 699-718. 610

Lowrie, W. 1990. Identification of ferromagnetic minerals in a rock by coercivity and unblocking 611

temperature profiles. Geophys. Res. Lett. 17, 159-162. 612

Lucchi, G.R., Rebesco, M., 2007. Glacial contourites on the Antarctic Peninsula margin: insight for 613

palaeoenvironmental and palaeoclimatic conditions. Geol. Soc., London, Special Publications, 614

276, 111-127. 615

Lucchi, G.R., Rebesco, M., Camerlenghi, A., Busetti, M., Tomadin, L., Villa, G., Persico, D., 616

Morigi, C., Bonci, M.C., Giorgetti, G., 2002. Mid-late Pleistocene glacimarine sedimentary 617

processes of a high-latitude deep-sea sediment drift (Antarctic Peninsula margin). Mar. Geol., 618

189, 343-370. 619

Macrì, P., Sagnotti, L., Lucchi, R.G., Rebesco, M., 2006. A stacked record of relative geomagnetic 620

paleointensity for the past 270 kyr from the western continental rise of the Antarctic Peninsula. 621

Earth Planet. Sci. Lett. 252, 162-179. 622

Mazaud, A., 2005. User-friendly software for vector analysis of the magnetization of long sediment 623

cores. Geochem. Geophys. Geosyst., 6, Q12006, doi:10.1029/2005GC001036. 624

26

Nishitani, T., Kono, M., 1983. Curie temperatures and lattice constants of oxidized titanomagnetite. 625

Geophysics Journal of the Royal Astronomical Society 74, 585-600. 626

Ó Cofaigh, C., Dowdeswell, J. A., Pudsey, C. J., 2001. Late Quaternary iceberg rafting along the 627

Antarctic Peninsula continental rise and in the Weddell and Scotia Seas. Quat. Res. 56, 308-321. 628

Paillard, D., Labeyrie, L., Yiou, P., 1996. Macintosh program performs time-series analysis. Eos 629

Trans. AGU 77, 379. 630

Pan, Y, Petersen, N., Davila, A.F., Zhang, L., Winklhofer, M., Liu, Q., Hanzlik, M., Zhu, R., 2005. 631

The detection of bacterial magnetite in recent sediments of Lake Chiemsee (southern Germany). 632

Earth Planet. Sci. Lett. 232, 109-123. 633

Peters, C., Dekkers, M.J., 2003. Selected room temperature magnetic parameters as a function of 634

mineralogy, concentration and grain size, Phys. Chem. Earth, 28, 659-667. 635

Pudsey, C.J., 2000. Sedimentation on the continental rise west of the Antarctic Peninsula over the 636

last three glacial cycles. Mar. Geol. 167, 313-338. 637

Pudsey, C.J., Camerlenghi, A., 1998. Glacial-interglacial deposition on a sediment drift on the 638

Pacific margin of the Antarctic Peninsula. Antar. Sci. 10, 286-308. 639

Rebesco, M., Larter, R.D., Camerlenghi, A., Barker, P.F., 1996. Giant sediment drifts on the 640

continental rise west of the Antarctic Peninsula. Geo-Mar. Lett., 16, 65-75. 641

Rebesco, M., Larter, R.D., Barker, P.F., Camerlenghi, A., Vanneste, L.E., 1997. The history of 642

sedimentation on the continental rise west of the Antarctic Peninsula. In: Barker, P.F., Cooper, 643

A.K. (Eds.), Geology and Seismic Stratigraphy of the Antarctic Margin, Part 2. American 644

Geophysical Union, Antarct. Res. Ser. 71, Washington, pp. 29-49. 645

Rebesco, M., Camerlenghi, A., Zanolla, C., 1998. Bathymetry and morphogenesis of the 646

continental margin west of the Antarctic Peninsula. Terra Antart. 5, 715-725. 647

Rebesco, M., Pudsey, C.J., Canals, M., Camerlenghi, A., Barker, P.F., Estrada, F., Giorgetti, G., 648

2002. Sediment drifts and deep-sea channel systems, Antarctic Peninsula Pacific Margin. In: 649

Stow, D.A.V., Pudsey, C.J., Howe, J.A., Faugères, J.-C., Viana, A.R. (Eds.), Deep-Water 650

27

Contourite Systems: Modern Drifts and Ancient Series, Seismic and Sedimentary 651

Characteristics. Geol. Soc. London Mem. 22, pp. 353-371. 652

Rebesco, M., Camerlenghi, A., Volpi, V., Neagu, C., Accettella, D., Lindberg, B., Cova, A., Zgur, 653

F., the MAGICO party, 2007. Interaction of processes and importance of contourites: insights 654

from the detailed morphology of sediment drift 7, Antarctica. In: Viana, A.R., Rebesco, M. 655

(Eds.), Economic and Palaeoceanographic Significance of Contourite Deposits. Geol. Soc. 656

London Special Publ. 276, pp. 95-110. 657

Roberts, A.P., 1995. Magnetic properties of sedimentary greigite (Fe3S4). Earth Planet. Sci. Lett. 658

134, 227-236. 659

Roberts, A.P., Cui, Y.L., and Verosub, K.L., 1995, Waspwaisted hysteresis loops: Mineral 660

magnetic characteristics and discrimination of components in mixed magnetic systems: Journal 661

of Geophysical Research, 100, 17 909-17 924, doi: 10.1029/95JB00672. 662

Roberts, A.P., Winklhofer, M., 2004. Why are geomagnetic excursions not always recorded in 663

sediments? Constraints from post-depositional remanent magnetization lock-in modelling. Earth 664

Planet. Sci. Lett. 227, 345-359. 665

Roberts, A.P., Stoner, J.S., Richter, C., 1999. Diagenetic magnetic enhancement of sapropels from 666

the eastern Mediterranean Sea. Marine Geology, 103-116. 667

Roberts, A.P., Jiang, W.T., Florindo, F., Horng, C.S., Laj, C., 2005. Assessing the timing of greigite 668

formation and the reliability of the Upper Olduvai polarity transition record from the Crostolo 669

River, Italy. Geophys. Res. Lett. 32, L05307. doi:10.1029/2004GL022137. 670

Roberts, A.P., Florindo, F., Larrasoaña, J.C., O'Regan, M.A., Zhao, X., 2010a. Complex polarity 671

pattern at the former Plio-Pleistocene global stratotype section at Vrica (Italy): Remagnetization 672

by magnetic iron sulphides. Earth Planet. Sci. Lett. 292, 98-111. 673

Roberts, A.P., Chang, L., Rowan, C., Horng, C.-S., Florindo, F., 2010b. Magnetic properties of 674

sedimentary greigite (Fe3S4): An update. Reviews of Geophysics, doi:10.1029/2010RG000336, 675

in press. 676

28

Rochette, P., Fillion, G., Mattéi J-L. and Dekkers, M.J. 1990. Magnetic transition at 30-34 Kelvin 677

in pyrrhotite: insight into a widespread occurrence of this mineral in rocks. Earth Planet. Sci. 678

Lett. 98, 319-328. 679

Rowan, C.J., Roberts, A.P., 2005. Tectonic and geochronological implications of variably timed 680

magnetizations carried by authigenic greigite in marine sediments from New Zealand. Geology 681

33, 553-556. 682

Rowan, C.J., Roberts, A.P., 2006. Magnetite dissolution, diachronous greigite formation, and 683

secondary magnetizations from pyrite oxidation: unravelling complex magnetizations in 684

Neogene marine sediments from New Zealand. Earth Planet. Sci. Lett. 241, 119–137. 685

Sachs, J.P., Anderson, R.F., 2005. Increased productivity in the subantarctic ocean during Heinrich 686

events. Nature 434, 1118-1121. 687

Sagnotti, L., Macrì, P., Camerlenghi, A., Rebesco, M., 2001. Environmental magnetism of late 688

Pleistocene sediments from the Pacific margin of the Antarctic Peninsula and interhemispheric 689

correlation of climatic events. Earth Planet. Sci. Lett. 192, 65-80. 690

Sagnotti, L., Roberts, A.P., Weaver, R., Verosub, K.L., Florindo, F., Pike, C.R., Clayton, T., 691

Wilson, G.S., 2005. Apparent magnetic polarity reversals due to remagnetization resulting from 692

late diagenetic growth of greigite from siderite. Geophys. J. Int. 160, 89-100. 693

Stoner, J.S., Channell, J.E.T., Hillaire-Marcel, C., 1998. A 200 ka geomagnetic chronostratigraphy 694

for the Labrador Sea: Indirect correlation of the sediment record to SPECMAP. Earth Planet. 695

Sci. Lett. 159, 165-181. 696

Tauxe, L., Mullender, T.A.T., Pick, T., 1996. Potbellies, wasp-waists, and superparamagnetism in 697

magnetic hysteresis. J. Geophys. Res. 101, 571-583. 698

Turner, J., Colwell, S.R., Marshall, G.J., Lachlan-Cope, T.A., Carleton, A.M., Jones, P.D., Lagun, 699

V., Reid, P.A., Iagovkina, S., 2005. Antarctic climate change during the last 50 years. Int. J. 700

Climatol. 25, 279-294. 701

Uenzelmann-Neben, G., 2006. Depositional patterns at Drift 7, Antarctic Peninsula: Along-slope 702

29

versus down-slope sediment transport as indicators for oceanic currents and climatic conditions. 703

Mar. Geol. 233, 49-62. 704

Venuti, A., and Florindo, F. 2004. Magnetostratigraphy and environmental magnetism of two 705

Quaternary deep-sea gravity cores from the west Pacific Southern Ocean, Geochem. Geophys. 706

Geosyst., 5(12), doi:10.1029/2004GC000810. 707

Villa, G., Persico, D., Bonci, M.C., Lucchi, R.G., Morigi, C., Rebesco, M., 2003. Biostratigraphic 708

characterization and Quaternary microfossil palaeoecology in sediment drifts west of the 709

Antarctic Peninsula-implications for cyclic glacial-interglacial deposition. Palaeogeogr., 710

Palaeoclimatol., Palaeoecol. 198, 237-263. 711

Weeks, R., C. Laj, L. Endignoux, M. Fuller, A. Roberts, R. Manganne, E. Blanchard, and W. Goree 712

(1993), Improvements in long-core measurement techniques: Applications in palaeomagnetism 713

and palaeoceanography, Geophys. J. Int., 114, 651-662 714

715

30

Figure Captions 715

Figure 1. Colour-shaded relief bathymetric map of sediment Drift 7 off the Antarctic Peninsula 716

Pacific margin with location of SEDANO gravity cores mentioned in the text and ODP Leg 178 717

drill sites. The inset shows schematically the present-day surface and bottom water circulation 718

along the western margin of the Antarctic Peninsula with the hatched blue arrow illustrating the 719

episodic intrusion of Circumpolar Deep Water onto the shelf [Hillenbrand et al., 2008a, and 720

references therein]. 721

722

Figure 2. Typical orthogonal vector diagrams and intensity decay curves for cores SED-07, SED-723

12, SED-13, and SED-14. a), c), and g) samples are from glacial lithostratigraphic units; b) and e) 724

samples are from interglacial units; d) sample is from an interval of the glacial unit that exhibits a 725

relative minimum of MDFNRM (H2 in Figure 3b). Projections onto the vertical plane are represented 726

by open circle, while projections onto the horizontal plane are denoted by green circles. Decl. = 727

declination; Incl. = inclination. 728

729

Figure 3. Lithology and main magnetic parameters in cores a) SED-07 and SED-12 and b) SED-13 730

and SED-14. Parameters for core SED-14 were taken from Macrì et al. [2006], apart from the new 731

data from the working halves (see 3.1. Sampling). Low-field volume specific magnetic 732

susceptibility (κ), NRM20 intensity, ChRM inclination, MDF of the NRM, ARM intensity, and 733

ARM/κ are plotted at the same scale, if appropriate. The grey shaded areas indicate interglacial 734

lithostratigraphic units and the black lines volcanic ash layers/lenses after Hillenbrand et al. 735

[2008b] and Lucchi et al. [2002]. H1 to H4 indicate the four intervals with minima in MDFNRM that 736

are correlated to the Heinrich layers deposited in the North Atlantic between 14 and 36 kyr BP 737

[e.g., Stoner et al., 1998; Chapman et al., 2000; Sagnotti et al., 2001]. 738

739

31

Figure 4. (a-h) Temperature-dependence (up to 700°C) of low-field magnetic susceptibility for 740

representative samples from Drift 7. Tc values for heating curves, a=605, b=590, c=599, d=602, 741

e=594, f=603, g=586, h=600; average=597°C; d) heating-cooling cycle up to 350°C (red-line) 742

followed by a standard cycle up to 700°C (black-line). See text for discussion. 743

744

Figure 5 Mr/Ms and Bcr/Bc data for representative samples from the sediment cores recovered from 745

Drift 7 (solid circles) plotted on the theoretical Day plot [Day et al., 1977] curves calculated for 746

magnetite by Dunlop [2002]. The data for these samples can be interpreted to indicate the 747

dominance of pseudo-single domain (PSD) magnetite. Numbers along the curves are volume 748

fractions of the soft component (SP or MD) in mixtures with SD grains [after Dunlop, 2002]. 749

750

Figure 6. IRM acquisition curves for representative samples from SED-07, SED-13, and SED-14. 751

752

Figure 7. Bcr and SIRM/χlf values for specimens from drifts 4 and 5. Symbols correspond to 753

lithostratigraphy with colours for core number. Samples from lithostratigraphic units A and C fall in 754

overlapping fields indicated by box outline. Parts of the fields for synthetic magnetite and 755

titanomagnetite are indicated by inclined and vertical hash respectively, based on data in Peters and 756

Dekkers [2003]. Their data is based on remanent acquisition coercivity (Bcr’), rather than DC 757

demagnetisation remanent coercivity (Bcr), and has been overlain using a Bcr/Bcr’ ratio of 0.75 [see 758

Peters and Dekkers, 2003]. Fields for synthetic greigite and pyrrhotite plot with SIRM/χlf at >20 759

x103 A/m. 760

761

Figure 8. a) to d). Thermal demagnetisation of 3-component IRM using axis increments of 0 to 0.1 762

T, 0.1 to 0.3 T and 0.3-1T for samples from sediment cores recovered from drifts 4 and 5. In each 763

case the samples represent specimens, in which the ‘hard titanomagnetite’ fraction is important, and 764

the Bcr (in mT) and SIRM1T/χlf (in x103 A/m) values are indicated in […] respectively. e) and f)- 765

32

demagnetisation of an IRM acquired at 10 K, of two of the specimens shown above, showing the 766

transition temperature (Tv) interpreted as the Verwey transition. 767

768

Figure 9. X-ray diffraction data for magnetic extracts of samples from Drift 5 (cores PC108 and 769

PC107). In each case, the diffraction profiles represent samples, in which the ‘hard titanomagnetite’ 770

fraction is important. The upper three extracts were produced using a quick extraction method, 771

flagged as the Me method [Hounslow and Maher, 1999] to limit oxidation of potential sulphides in 772

the samples. For the lowest trace from the sample at 296 cm depth in core PC108 the magnetic 773

probe (Mp) method was applied [Hounslow and Maher, 1999]. Magnetic mineral extraction 774

efficiencies of χlf, ARM and SIRM1T are shown in {..}. a=magnetic lattice constant. Bcr (in mT) 775

and SIRM1T/χlf (in x103 A/m) values are indicated in […] respectively. 776

777

Figure 10. Establishment of age models for the studied cores from Drift 7 by correlation of the 778

magnetic susceptibility profiles of cores (a) SED-12 and (b) SED-7, and of the ARM/κ profile of 779

core (c) SED-13 with the corresponding parameters of RPI-dated cores from Drift 7. 780