A continuous palaeosecular variation record of the last four millennia from the Augusta Bay (Sicily,...

Geophys. J. Int. (2011) 184, 191–202 doi: 10.1111/j.1365-246X.2010.04860.x

GJI

Geo

mag

netism

,ro

ckm

agne

tism

and

pala

eom

agne

tism

A continuous palaeosecular variation record of the last fourmillennia from the Augusta Bay (Sicily, Italy)

Leonardo Sagnotti,1 Alessandra Smedile,1 Paolo Marco De Martini,1 Daniela Pantosti,1

Fabio Speranza,1 Aldo Winkler,1 Paola Del Carlo,2 Luca Giorgio Bellucci3

and Luca Gasperini31Istituto Nazionale di Geofisica e Vulcanologia, Sede di Roma, Via di Vigna Murata 605, 00143 Roma, Italy. E-mail: [email protected] Nazionale di Geofisica e Vulcanologia, Sezione di Pisa, Via della Faggiola 32, 56126 Pisa, Italy3Istituto di Scienze Marine, CNR, Sede di Bologna – Geologia Marina, Via P. Gobetti 101, 40129 Bologna, Italy

Accepted 2010 October 20. Received 2010 October 20; in original form 2010 July 13

S U M M A R YWe present a high-resolution palaeomagnetic and rock magnetic study of two cores, MS06and MS06-SW (6.7 and 1.1 m long, respectively), collected at 72 m below sea level in theAugusta Bay shelf (Eastern Sicily, Ionian Sea, Italy) about 2.3 km from the coastline. Geophys-ical surveying carried out in the sampling area highlighted the presence of a homogeneoussedimentary sequence that most likely was deposited after the Last Glacial Maximum andwas not affected by anthropogenic disturbances. The two cores penetrated a monotonous mudsedimentary sequence, interrupted at ∼3 m depth by a 3–4-cm-thick volcanic sandy layer thatis correlated with the tephra fallout deposit produced by the 122 BC plinian eruption of MtEtna. This tephra, along with radiocarbon dating of nine marine shells and with radioactivetracers for the uppermost 0.3 m (210Pb and 137Cs), provide the chronological constraints forthe stratigraphic sequence that resulted younger than 4500 yr BP.

Palaeomagnetic and rock magnetic data show that the sample sequence is magnetically ho-mogeneous. A single peak of high magnetic mineral concentration is present and correspondsto the volcanic sandy layer. Palaeomagnetic data allowed the identification of a well-definedcharacteristic remanent magnetization that provides a high-resolution record of palaeosecularvariation (PSV) at the sampling site. The reconstructed PSV curve is in good agreement withthe available regional reference PSV curves and with the prediction from recent PSV mod-elling for Europe. The palaeomagnetic data obtained in this study on the one hand supportand refine the age model for the cores, derived from other independent constraints, and on theother hand provide an original high-resolution PSV curve that can serve as a reference for thecentral Mediterranean over the last 4 ka.

Key words: Palaeointensity; Palaeomagnetic secular variation; Marine magnetics andpalaeomagnetics; Europe.

I N T RO D U C T I O N

Historic measurements of the geomagnetic field have documentedits primarily dipolar spatial structure at the Earth’s surface and itstemporal variability, and direct observations of geomagnetic fieldelements have been collected to construct models for the evolu-tion of the Earth’s magnetic field over the last four hundred years(Bloxham & Jackson 1992; Jackson et al. 2000). Variations ontimescales longer than ca. 5 yr are collectively defined as sec-ular variation (SV) (Thompson & Barraclough 1982; Bloxham &Gubbins 1985). Geomagnetic variations with timescales longer than22 yr (i.e. the duration of the turn-over of solar magnetic field) areof internal origin and are caused by the dynamo processes in theEarth’s core. SV results from the effect of advection of the mag-

netic field by the flow in the fluid outer core and from effects ofmagnetic field diffusion in the core and the mantle (e.g. Holme &Olsen 2006). Palaeomagnetic studies provide the main source ofinformation to extend the direct observations of the geomagneticfield variations to recent geological past. Palaeomagnetic studiesalso make it clear that during intervals of stable magnetic polaritythe Earth’s magnetic field has undergone a wider range of tem-poral and spatial variability than has been seen in historic times(see review by Lund 2007). Palaeosecular variation (PSV) stud-ies therefore are a fundamental source of information for a deeperunderstanding of the natural variability of the geomagnetic fieldand for the modelling of the geodynamo processes and outer core’sfluid dynamics. In addition to their implications for geomagnetism,the detailed reconstruction of palaeomagnetic records retracing the

C© 2010 The Authors 191Geophysical Journal International C© 2010 RAS

Geophysical Journal International

192 L. Sagnotti et al.

Earth’s magnetic field variations have a valuable stratigraphic poten-tial, since continuous PSV records may provide an original tool forhigh-resolution correlation and dating of Holocene sedimentary andvolcanic sequences. In fact, PSV studies are carried out from thepalaeomagnetic study of archaeological materials, volcanic rocksand lacustrine and marine sediments.

Sediments, in particular, offer the possibility to reconstruct con-tinuous high-resolution PSV records, provided that a series of con-straints are met (e.g. Verosub 2000): (1) the sedimentation is contin-uous and the sedimentation rate is high; (2) the record has a robustchronological control and the age of the primary magnetization isunequivocally established; (3) the sediments are not affected bysignificant diagenetic changes in the primary magnetic mineralogy,that is, they are free from dissolution of detrital magnetic miner-als and/or formation of secondary magnetic minerals under anoxicconditions (e.g. Leslie et al. 1990a,b).

Additional inherent limitations often prevent a successful use ofsedimentary sequences for PSV studies. These limitations includetechnical issues, such as physical disturbance that may occur in situor during the coring and sampling procedure of soft sediments, andthe lack of azimuthal orientation for most sediment cores. As well,there are natural issues, such as depositional remanent magnetiza-tion (DRM) acquisition processes, which may induce a deviation ofthe palaeomagnetic direction from the palaeofield directions, andremanence acquisition smoothing. In particular, two sedimentaryeffects may disturb the fidelity with which sediments are capable of

recording past geomagnetic variations. The first is that the DRM isoften characterized by an inclination shallowing due to the sedimen-tary processes and burial compaction effects (King 1955; Verosub1977; Blow & Hamilton 1978; Anson & Kodama 1987). The secondis that sediments may be affected by a magnetic realignment dur-ing compaction, which results in a post-depositional magnetization(pDRM) and introduces a time lag between the sediment and rema-nence ages, known as lock-in time, that depends on the sedimentaryenvironment (Verosub 1977; Hyodo 1984; Bleil & von Dobeneck1999; Roberts & Winklhofer 2004; Sagnotti et al. 2005).

The integration of different coeval PSV records from multiplesedimentary or volcanic sequences in the same region lead to thedefinition of regional PSV master curves (Turner & Thompson1981, 1982; Hagstrum & Champion 2002; St-Onge et al. 2003;Snowball et al. 2007), which are fundamental tools for the strati-graphic use of new individual PSV records. In any case, PSV datafrom oceans and major seas are still relatively rare and there is aclear need to collect more widely distributed PSV data to improvetime-varying global geomagnetic field models (see Nilsson et al.2010).

In this study, we present a palaeomagnetic and rock magneticstudy of two piston cores that recovered a 6.7-m-thick marine strati-graphic sequence seaward of the Augusta harbour, in eastern Sicily(Fig. 1). The sequence is characterized by excellent palaeomag-netic properties and provided a continuous stratigraphic record thatallows a high-resolution reconstruction of the local geomagnetic

Figure 1. Bathymetric map of Augusta Bay and location of cores MS06 and MS06-SW. Yellow pattern indicates lateral extent of the ‘acoustic anomaly’represented by reflector R1 (see text for details). (a) Close-up view of the location of the cores and seismic line AU107. (b) High-resolution seismic profileAU-107 collected close to coring stations; a scaled projection of core MS06 on the section is also indicated.

C© 2010 The Authors, GJI, 184, 191–202

Geophysical Journal International C© 2010 RAS

Palaeosecular variation from Augusta Bay 193

PSV for the past 4.2 ka, both in terms of PSV directions and ofrelative palaeointensity (RPI). The obtained PSV curves are cor-related to the other available reference European PSV records andstacks. These data, integrated and supported by nine acceleratormass spectrometry (AMS) radiocarbon dates and the identificationof a precisely dated tephra layer from Mt Etna, allow us to estab-lish a high-resolution age model for the sequence. This record mayserve as reliable PSV curve for the central Mediterranean (SicilyChannel) and provides valuable input for the detailed reconstructionof the history of the Earth’s magnetic field variations over southernEurope during the last 4 millennia.

L O C AT I O N O F T H E C O R E S A N DS A M P L I N G

The two study cores were collected in the Augusta Bay offshore(eastern Sicily), about 2.3 km from the coastline close to an indus-trial area (Lat: 37◦12.69′N; Long: 15◦15.19′E), at a depth of 72 mbelow sea level (bsl) and at some metres distance from each other(Fig. 1). The coring location was chosen following an analysis of aclosely spaced grid of seismic reflection chirp-sonar profiles carriedout across the Augusta Bay (Smedile et al. 2010). The study coreswere collected from the northern part of the bay, in an area character-ized by a continuous sequence of sediments not affected by anthro-pogenic or local disturbance. Seismic profiles image a sequence oftransparent, well-layered fine-grained sediments deposited after thelast sea level rise and alternating zones of hard seafloor shaped byerosional processes. The most prominent seismostratigraphic fea-ture observed in the Holocene deposits is a high-amplitude reflector(R1) observed in a bathymetric range between 50 and 140 m bsl.In this range, R1 constitutes an ‘acoustic anomaly’ relative to otherreflectors observed in the sequence. For this reason it was mappedusing a semi-automatic picking function contained in the softwarepackage SeisPrho (Gasperini & Stanghellini 2009). The result ofthis work is a map of lateral distribution for this reflector, as shownin Fig. 1.

Core MS06 was collected in 2007 by the R/V Urania using a CP20piston corer, that allowed us to retrieve cores of large diameter witha high recovery ratio. A 6.7-m-thick sequence of homogeneousmud, interrupted only by a 3–4-cm-thick volcanic sandy layer at∼3 m below the seafloor (bsf), was recovered. Core MS06-SWwas collected in 2008 by a crane pontoon, using a SW-104 corerspecifically designed to sample the water-sediment interface and toobtain a section of sediment up to 1.2 m in length with minimizedcoring disturbance (Magagnoli & Mengoli 1999). Core MS06-SWrecovered a 1.12-m-thick sequence of homogeneous mud and it wascollected specifically and measured to replace the upper part of coreMS06 that appeared much softer and seemed not to be suitable forgeomagnetic PSV studies.

Sediments were cored into plastic liners, of 94 mm (MS06) and104 mm (MS06-SW) inner diameter. After dividing the cores intoworking and archive halves, each core section of about 1 m lengthwas sub-sampled with u-channel holders for continuous palaeomag-netic and rock magnetic measurements in the laboratory.

S E D I M E N T O L O G Y O F T H E C O R E S A N DA G E C O N S T R A I N T S

Seismic lines indicate that the cores penetrated a nearly transparenthomogeneous sedimentary sequence, that was deposited after the

last glacial maximum and includes R1 (Fig. 1). The distribution ofthe R1 reflector in the Augusta Bay appears to be controlled by twofactors: (1) bathymetric range and (2) distance from the shoreline. Infact, R1 is not visible at depths shallower than about 50 m bsl due tothe lack of fine-grained deposits at such shallow depths. Moreover,the amplitude of the signal from the reflector is significantly reducedat depths greater than about 140 m bsl, approaching the shelf break,where R1 still appears conformable with the underlying sequence,but it cannot be clearly distinguished from other reflectors withinthe Holocene sequence.

The two cores were collected with the aim to highlight anysubtle anomaly in the sediments, fauna assemblages, physicalproperties, etc., that could represent a proxy for tsunami occur-rence (Smedile et al. 2010). The recovered sedimentary sequenceconsists of homogeneous mud deposited in a low-energy envi-ronment (see lithologic log in Fig. 2) as pointed out by grainsize laser analysis (Smedile et al. 2010). In fact, sedimentolog-ical analysis confirmed that the sampled sequence is generallydominated by fine-grained sediments with prevailing silt content(79 per cent) followed in order of abundance by clay (19 per cent)and fine to very fine sand (2 per cent). Grain-size data were notcollected in the 15 cm core section centred on the dark volcanicmedium-coarse sandy layer to prevent any damage to the laserparticle counter, which is characterized by a measuring range of0.3–300 µm (Smedile et al. 2010).

Macroscopic evidence and X-ray images pointed out localizedconcentration of mollusc shells and thin darker layers enrichedin concentration of organic material, mostly vegetal remnants ofPosidonia oceanica. Moreover, the multidisciplinary study of coreMS06 resulted in identifying at least 12 anomalous intervals charac-terized by the presence of displaced epiphytic benthic foraminiferaand a relative increase of grain size in the sediments, clearly causedby high-energy events, with tsunamis as best candidates (Smedileet al. 2010).

The main lithological feature is, however, the volcanic sandyinterval identified at ∼3 m bsf in core MS06. It is a 3–4-cm-thicktephra layer consisting of black fine lapilli and coarse ash, whichon the basis of morphoscopic and petrochemical analyses of itsclasts and glass composition has been identified as a product of the122 BC Etna plinian eruption (Smedile et al. 2010) and correlatedto analogous tephra deposits cropping out on land on the flanks ofEtna Volcano and at the Priolo Reserve area (Coltelli et al. 1998; DeMartini et al. 2010). Combined analysis of seismic images and core-logs carried out using a synthetic seismogram software (ChirCor;developed by Del Forno & Gasperini 2008) indicates that the tephralayer found in the MS06 core correlates with reflector R1.

The presence of mollusk shells allowed us to estimate the radio-metric ages for nine distinct layers of the MS06 core (Table 1).Ages were determined by AMS 14C analysis carried out atthe AMS facility of the Poznan Radiocarbon Laboratory. Rawradiocarbon ages were dendrochronologically corrected accord-ing to the radiocarbon calibration program OxCal 4.1 (Ramsey2009), using a marine calibration curve (Reimer et al. 2009)with an appropriate �R offset (124 ± 60) selected from theChrono Marine reservoir Database (Reimer & Reimer 2001). Theconcentration-depth profile of short-lived radioactive tracers, suchas 210Pb and 137Cs, was also investigated to obtain the chronologyof the uppermost 0.3 m of the sampled sequence (e.g. Bellucci et al.2007). As a whole, the available set of independent chronostrati-graphic constraints indicate that the sampled sequence depositedcontinuously during the past 4.2 ± 0.2 ka (Smedile et al. 2010).

C© 2010 The Authors, GJI, 184, 191–202



0 100 200 300 400 500 600

k (10 SI)-5

k (10 SI)-5

k (10 SI)-5

0 5 10 15 20 25 30 35 400

10

20

30

40

50

60

70

80

90

100

0 5 10 15 20 25 30

mud

organic layer

shells concentration

tephra layer

Core legend

MS-06

2

1

0

6

3

4

5

MS-06-SW

1

0

Dep

th(m

)

Dep

th(c

m)

MS06 MS06-SW

a) c)b)

Figure 2. Stratigraphic trend of the magnetic susceptibility (k) for cores MS06 and MS06-SW. The data show a sharp peak to values higher than 500 × 10−5

SI corresponding to the tephra layer (a). Apart from this peak, the data are characterized by a substantial homogeneity, with values fluctuating between 6 and 28(×10−5 SI) (b). A magnification of the uppermost 1 m (c) shows the comparison of the k values for the MS06 and MS06-SW cores. Apart for the uppermost10–15 cm, the two data sets support a close correlation between the two cores down to ca. 0.8 m bsf.

Table 1. Ages of 9 AMS 14C analysis and 210Pb/137Cs data (modified after Smedile et al. 2010). The 14C measured ages were dendrochronologically correctedusing a marine calibration curve (Reimer et al. 2009) with an appropriate �R offset (124 ± 60) (Reimer & Reimer 2001).

Sample code Depth (m bsf) Sample type δ 13C Measured age (years BP) Cal. curve Calibrated age 2 σ

MS-06 14 0.14 Pb–Cs profile 150 ± 20MS-06 48–49 0.48 Marine shell (Turritella) 0.8 1495 ± 30 Marine 09 AD 885–1200MS-06 82–84 0.84 Marine shell (Venus) −0.1 1655 ± 35 Marine 09 AD 715–1015MS-06 125–127 1.27 Marine shell (Nucula) 6.8 1820 ± 30 Marine 09 AD 570–855MS-06 179 1.79 Marine shell (Turritella) −1.2 1960 ± 30 Marine 09 AD 420–690MS-06 276–278 2.78 Marine shell (Venus) −1.7 2330 ± 30 Marine 09 BC 20–AD 325MS-06 355–357 3.57 Marine shell (Nucula) −1.6 2680 ± 35 Marine 09 BC 435–80MS-06 467 4.67 Marine shell (Nucula) 5.5 2925 ± 35 Marine 09 BC 755–400MS-06 562–563 5.63 Marine shell (Nucula) 6.4 3270 ± 35 Marine 09 BC 1185–810MS-06 -1 6.63 Marine shell (Turritella) 3.4 4245 ± 35 Marine 09 BC 2440–2030

M E T H O D S A N D M E A S U R E M E N T S

Palaeomagnetic and rock magnetic measurements were carriedout at the palaeomagnetic laboratory of the Istituto Nazionale diGeofisica e Vulcanologia, in Rome, on a small access (45 mmdiameter) automated pass through ‘2G Enterprises’ DC 755 su-perconducting rock magnetometer (SRM), installed in a magnet-ically shielded room and equipped with a standard air-cooledsolenoid and a transverse split pair, for in-line orthogonal alter-nating field (AF) demagnetization. For each u-channel, we mea-sured at 1 cm spacing both the low-field magnetic susceptibil-ity (k), using a Bartington magnetic susceptibility probe MS2Cin-line with the SRM system, and the natural remanent mag-netization (NRM). To investigate the stability of the NRM and

to reveal possible secondary overprints, all u-channels were thenstepwise AF demagnetized in peak fields of 5–10-15–20-25–30-40–50-60–80-100 mT, with remanence vectors measured aftereach demagnetization step. After each demagnetization cycle ananhysteretic remanent magnetization (ARM) was imparted in a0.05 mT bias field, applied with an in-line single-axis dc coil,combined with an axial peak AF of 100 mT and translating theu-channel through the AF and dc coil system at a translation speedof 10 cm s−1. The imprinted ARM was then measured and stepwiseAF demagnetized in fields of 5–10-20–30-40–60-80–100 mT.

The raw magnetic moment data measured by the three orthog-onal SQUID sensors of the SRM system were corrected, directlyby the measuring software, by compensating for the different shapeand widths of the response function curves of the three SQUID

C© 2010 The Authors, GJI, 184, 191–202



pick-up coils. This correction compensates for the effects of thenegative regions on the edge of the SQUID response functions forthe transverse axes and for the broader width of the SQUID re-sponse function along the long axis of the u-channel. The obtainedpalaeomagnetic data are therefore free from fictitious effects, suchas inclination shallowing (or steepening) that may result from un-compensated raw magnetic moment data (Roberts 2006). Moreover,we took particular care in avoiding eventual disturbance effects thatmay be introduced during the coring, cutting and sampling pro-cedures and could result in remanence deflections due to plasticdeformation of the soft sediments. In this study, notwithstandingthat disturbance was not noted during the entire sampling processof the u-channel sub-sections, we adopted a conservative approach,and disregarded the palaeomagnetic data for ∼5 cm at both ends ofeach u-channel and stratigraphic gap.

R E S U LT S

Long-core magnetometers are explicitly designed for measurementof homogeneously magnetized materials. Large contrast in mag-netization may produce spurious artefacts (see Roberts 2006) andthe influence of thin, very magnetic strata on the results maybeunderestimated due to smoothing. In the studied cores, apart fromthe tephra layer, which causes a pronounced magnetic suscepti-bility peak (up to 182 × 10−5 SI) between 267 and 300 cm bsf(Fig. 2a), the magnetic susceptibility values oscillate between 6 and28 (×10−5 SI), with an average value of 14.8 (×10−5 SI) (Fig. 2b).Palaeomagnetic and rock magnetic results from the whole strati-graphic interval (267–300 cm bsf) whose magnetic properties areaffected by the 3–4-cm-thick tephra layer were excluded from theanalysis. The magnetic susceptibility trends of the cores may beeasily correlated (Fig. 2c) for the uppermost 1 m of the sampledsedimentary sequence (apart from the topmost 10–15 cm, wherethe two trends diverge), indicating essentially a one-by-one matchbetween the stratigraphic depths of the two cores.

The sedimentary sequence is characterized by excellent palaeo-magnetic properties. Each measured interval showed an almostsingle-component NRM: after the removal of a low-coercivity com-ponent at 5–10 mT, the palaeomagnetic direction remains remark-ably stable up to 60 mT, with demagnetization vectors alignedalong linear paths towards the origin in orthogonal vector diagrams(Fig. 3). Therefore, a clear and well-defined characteristic remanentmagnetization (ChRM) has been isolated for each measurement in-terval and its direction computed by principal component analysis(Kirschvink 1980) using at least seven consecutive steps at AF peakvalues between 5 and 60 mT. As mentioned before, the top of coreMS06 proved to be unsuitable for PSV studies, with a progressiveupward large deflection of the remanence due to deformation of thesoft sediments during sampling. Therefore, we used only the upper78 cm of core MS06-SW to estimate palaeomagnetic SV in thatstratigraphic interval.

For NRM and ARM, which are both almost single-componentmagnetic remanences, we computed the median destructive field(MDF) from the AF demagnetization curves. The MDF is definedas the value of the peak AF necessary to reduce the remanence inten-sity to half of its initial value (Fig. 3) and it is a coercivity-dependentmagnetic parameter. The stratigraphic trends of the rock magneticparameters are shown in Fig. 4, excluding the data from the strati-graphic interval whose magnetic properties affected by the tephralayer (267–300 cm bsf). The low-field magnetic susceptibility (k)and the NRM and ARM intensities mostly depend on the concen-

tration of magnetic minerals. However, all rock-forming mineralscontribute to k in proportion to their relative abundance and specificmagnetic susceptibility, so that the measured k value can be biasedby diamagnetic and paramagnetic contributions, if the concentra-tion of ferrimagnetic minerals is low. ARM is primarily sensitive tothe concentration of fine, single-domain (SD), ferrimagnetic grains(King et al. 1982; Maher 1988), and NRM may also reflect thestrength of the geomagnetic field during the time of the acquisitionof the remanence. The MDF of the NRM and ARM mostly de-pends on the composition and/or grain size of the minerals carryingthose remanences; the ARM/k ratio depends on the grain size offerrimagnetic minerals, with higher values for finer grained (SD)ferrimagnetic particles and lower values for larger (multidomain,MD) grains. In agreement with the overall lithological homogene-ity, the rock magnetic parameters indicate a substantial magnetichomogeneity of the mud sequence as well. Neither the thin darkerlayers enriched in organic material with the localized concentrationof mollusc shells nor the 12 layers characterized by the presenceof displaced epiphytic benthic foraminifera and by a relative in-crease of grain size in the sediments show distinctive rock magneticfeatures, suggesting that no significant changes occurred in theirmagnetic mineralogy. Throughout the stratigraphic sequence, theconcentration-dependent magnetic parameters show a limited rangeof variability, with maximum oscillations comprised within a factor2 of their mean value. The coercivity-dependent magnetic param-eters show very limited oscillations too (within 15–30 mT with amean value of 26 mT for MDFNRM and within 21–28 mT with amean value of 24 mT for MDFARM), with mean values and range ofvariability typical for magnetite grains (Maher 1988). The ARM/kratio increases from ca. 200 to 400 A m−1 in the uppermost 0.5 mbsf and then shows a rather steady decrease to a minimum valueof ca. 100 A m−1 at the bottom of the sequence. We conclude that,apart from the interval affected by the tephra layer, the entire set ofrock magnetic data indicate limited variations in the concentration,composition and grain size of the magnetic minerals, which consistof magnetite, with a tendency to a downward decrease in the contentof fine (SD) particles, mostly evident below ca. 4.75 m bsf (Fig. 4).

The main palaeomagnetic results for cores MS06 and MS06-SWare shown in Fig. 5.

Because the cores were not azimuthally oriented, the observedChRM declinations (shown in the ‘ChRM Decl measured’ plot ofFig. 5) for each u-channel section have been arbitrarily rotated, toalign the mean value of the uppermost u-channel section with truenorth and to align the declination trends and values across consecu-tive u-channel sections. The resulting data are shown in the ‘ChRMDecl rotated’ plot (Fig. 5). The palaeomagnetic inclination trend ofthe Augusta Bay cores is affected by some 10◦–20◦ shallowing withrespect to the expected values at the core sites (56.6◦ for a geocentricdipolar axial field, indicated by a dashed line in the ‘ChRM Incl’plot). The maximum angular deviation (MAD) for each determinedChRM direction is lower than 2◦ in most of the measured inter-vals and always is lower than 4◦ throughout the sampled sequence.RPI curves were obtained by normalizing the NRM intensity withconcentration-dependent rock magnetic parameters (k and ARM).The two RPI curves match closely each other for the upper 4.75 mbsf (boundary between u-channel sections III and IIB). Below thisdepth, the two curves diverge and the NRM/k curve is deflectedtowards lower values (Fig. 5). We think that this effect is due tothe decrease in the concentration of fine (SD) ferrimagnetic grainsand by a corresponding increase in the relative proportion of large(MD) magnetic grains, as pointed out by the rock magnetic param-eters (Fig. 4). Since SD ferrimagnetic grains are the more effective

C© 2010 The Authors, GJI, 184, 191–202



Figure 3. Representative NRM demagnetization plots for selected specimens subjected to AF demagnetization. The demagnetization data have been visualizedand analysed using the Remasoft program (Chadima & Hrouda 2006). In the orthogonal vector diagrams (Zijderveld 1967), open and closed symbols representprojections onto vertical and horizontal planes, respectively. The plots showing the decay of the NRM intensity are shown as insets in the upper right-hand sideof each Zijederveld plot. The median destructive field of the NRM is indicated by the dashed line.

recorders of the geomagnetic field, whereas MD grains are less ef-ficient in the magnetic field recording and mostly contribute to themagnetic susceptibility, we conclude that the NRM/ARM curve isthe proxy that better represents the RPI variation of the geomagneticfield throughout the sampled sequence.

D I S C U S S I O N

The palaeomagnetic data of the Augusta Bay cores are of excellentquality and are very promising for a high-resolution reconstructionof the PSV for the last four millennia. The homogeneity of the sam-pled sequence warrants for an almost continuous deposition, withan average sedimentation rate of over 150 cm ka−1. The sedimen-tation rate has obvious effects on the resolution which it is possibleto achieve in PSV reconstruction. As stated, palaeomagnetic androck magnetic measurements were taken at 1 cm spacing. However,it should be pointed out that since the half-width of the response

function of the SQUID pick-up coils is ca. 4.5 cm, all remanencemeasurements have to be considered truly independent every ca.5 cm only. This implies that the PSV record reconstructed for theMS06 and MS06-SW cores has, on average, a resolution of about30–35 yr.

Even if there is no guarantee that the corer penetrated absolutelyvertically into the sedimentary sequence, we think that the observedChRM inclination shallowing of 10◦–20◦ cannot be entirely due toan unrecognized core tilting during sampling, but it rather is acharacteristic feature of the studied sequence. Palaeomagnetic in-clination shallowing may arise as an effect due in part to gravity(i.e. by factors related to the size and shape of the magnetic parti-cles) and to the horizontality of the depositional surface, so that thefinal alignment of the NRM carriers is not controlled solely by thedirection of the magnetic field. Any realignment of the magneticparticles taking place after a certain critical level of sediment dewa-tering and compaction cause the acquisition of a pDRM, which may

C© 2010 The Authors, GJI, 184, 191–202



mud

organic layershells concentration

tephra layer

Core legend

0.02 0.04 0.06 0.08 0.1

NRM (A/m)0.05 0.1 0.15 0.2

ARM (A/m)5 10 15 20 25 30 35

MDF_NRM (mT)5 10 15 20 25 30 35

MDF_ARM (mT)100 200 300 400 500

ARM/k (mT)0 20 40 60 80 100

k (10 SI)-5

MS06 MS06-SW

MS-06

2

1

0

6

3

4

5

MS-06-SW

1

0

Dep

th(m

)

Figure 4. Downcore variation of the main rock magnetic parameters measured for the MS06 and MS06-SW cores. See text for discussion.

mud

organic layer

shells concentration

tephra layer

Core legend

2020 12040 22060 320 42080

ChRM Decl_measured (°)ChRM Incl (°)-50 0 50

ChRM Decl_rotated (°)0 2 4

MAD (°)

MS06 MS06-SW

MS-06

2

1

0

6

3

4

5

MS-06-SW

1

0

Dep

th(m

)

IVII

VI

V

IV

III

I

IIB

IIA

NRM /k0mT NRM /k0mT

NRM/ARM0mT NRM/ARM0mT

0 0.2 0.4 0.6 0.8 1 1.2

RPI

MS06MS06-SW

Figure 5. Palaeomagnetic properties of cores MS06 and MS06-SW. ChRM declination and inclination, maximum angular deviation (MAD) and the normalizedrelative palaeointensity (RPI) curves NRM/k and NRM /ARM. The dashed lines in the ChRM inclination and declination plots indicate the expected valuesfor a geocentric axial dipole field at the sampling site. The u-channel sections (I–VII) are indicated to the right of the lithostratigraphic column.

produce a more faithful recording of the geomagnetic inclination.The acquisition of a pDRM may be favoured by bioturbation. How-ever, in the case of a prevalent pDRM, the palaeomagnetic recordwill be lagging the sediment age, with a delay related to the depthat which the remanent magnetization is ultimately fixed, defined aspalaeomagnetic lock-in depth (Verosub 1977; Hyodo 1984; Bleil &von Dobeneck 1999). In Holocene sequences with relatively highsedimentation rates (10–100 cm ka−1) palaeomagnetic lock-in cancause spreading of the reconstructed PSV ages on a century scale(Sagnotti et al. 2005).

The occurrence of inclination shallowing indicates that pDRMrealignment of magnetic grains has not been significant and suggestsa minor to negligible lock-in depth. Moreover, we point out that arecent numerical modelling of the relationship between inclinationshallowing and RPI in sediments indicates that the occurrence of

inclination shallowing is a pre-requisite for a linear palaeointensityrecording mechanism and therefore for reliable RPI studies (Mitra& Tauxe 2009).

P S V H I G H - R E S O LU T I O N DAT I N GA N D C O R R E L AT I O N

Taking into account the available chronostratigraphic constraints(Table 1) and the position of the 122 BC Etna tephra layer, thesediment sequence spans the last 4.2 ± 0.2 ka. We therefore com-pared the PSV curves obtained for cores MS06 and MS06-SW, withthe available reference PSV curves for the European region overthe same time interval and tried to maximize their correlation, bymatching largest PSV oscillations, to improve the original age modelbased on the AMS 14C ages and their associated uncertainties.

C© 2010 The Authors, GJI, 184, 191–202



20

30

40

50

60

70

80

Incl

(°)

MS06-MS06SW UK-lakes Fennostack SCHA.DIF.3K gufm CALS7K.2 CALS3K.3

-30

0 500 1000 1500 2000 2500 3000 3500 4000 4500

Age BP (ka)

0

0.2

0.4

0.6

0.8

1

RP

Io

rV

AD

M_

no

rmm

ax

-20

-10

0

10

20

30

40

Decl

(°)

MS06-MS06SW UK-lakes Fennostack SCHA.DIF.3K gufm CALS7K.2 CALS3K.3

MS06_MS06SW VADM FENNORPIS CALS3K.3CALS7K.2

’a)

a

b

c

d

e

f

b)

c)

SCHA.DIF.3K

Figure 6. (a) Inclinations, (b) declinations and (c) relative palaeointensity of cores MS06 and MS06-SW (this study), as a function of age, comparedwith PSV results from the United Kingdom and FENNOSTACK PSV stacks from sedimentary lacustrine cores, the SCHA.DIF.3k model based on Europeanarchaeomagnetic data, the global geomagnetic main field models CALS7k.2 and CALS3k.3, the gufm model based on direct historical geomagnetic observationsand the virtual axial dipole moment (VADM) reconstruction (see text for references). Nomenclature of inclination and declination minima and maxima followsthat originally proposed for the UK master PSV curve of Turner and Thompson (1981).

In particular, the PSV data from the Augusta Bay have beencompared with the UK Holocene PSV master curve (Turner &Thompson 1981; 1982), the Fennostack Holocene PSV master curvefor Fennoscandia (Snowball et al. 2007), the SCHA.DIF.3K modelfor Europe of Pavon-Carrasco et al. (2009), the global spherical har-monic models CALS7k.2 (Korte & Constable 2005) and CALS3k.3(Donadini et al. 2009; Korte et al. 2009), the virtual axial dipole

moment variations reconstructed by Knudsen et al. (2008), and, forthe four last centuries, with the gufm model of Jackson et al. (2000)based on direct geomagnetic observations (Fig. 6).

For correlation purposes, the palaeomagnetic directions cor-responding to the reference sedimentary PSV curves have beenrelocated to the geographic location of the core site (latitude37.1◦N, longitude 15.1◦E) via the virtual geomagnetic pole (VGP)

C© 2010 The Authors, GJI, 184, 191–202



Table 2. Tie-points used for PSV correlation.

Depth (cm bsf) Age PSV (yr BP∗) Source

7 140 Incl: gufm; UK15 240 Incl: gufm; UK (α); SCHA.DIF.3k33 560 Incl:UK (β); SCHA.DIF.3k77 1040 Decl: UK (d); CALS3k.3112 1350 Incl: UK (γ ); SCHA.DIF.3k; CALS3k.3; FENNOSTACK223 1680 Incl: UK (δ); CALS3k.3; CALS7k.2266 2000 Incl UK296 2072 Volcanic layer316 2120 Decl: UK (e); SCHA.DIF.3k; CALS3k.3;473 2680 Decl: UK (f); SCHA.DIF.3k; FENNOSTACK568 3160 Decl : UK; FENNOSTACK629 3600 Decl UK664 4200 AMS dating

∗BP = before 1950 AD.

method (Noel & Batt 1990). According to this method, the palaeo-magnetic direction observed at one site is first converted to aVGP and the VGP position is then used to calculate the cor-responding geomagnetic direction at the location of the secondsite.

The UK Holocene master curve represents the stack of palaeo-magnetic directions gathered from 10 sedimentary cores drilled inthree lakes from Great Britain (Turner & Thompson 1981; 1982)and has been used as a reliable reference PSV curve for Europeover the last 30 yr. It still appears as the most useful PSV record forpalaeomagnetic correlation-dating of Holocene Italian sequences(Iorio et al. 2004; Rolph et al. 2004; Speranza et al. 2004; Vigliotti2006; Speranza et al. 2008).

The Fennostack is a recent Holocene master curve fromFennoscandia, representing a stack of 14 cores from seven lakes,accurately dated by annual varve chronology (Snowball et al. 2007).In FENNOSTACK, inclination is expressed as deviations from thecore average and 95 per cent confidence limits. We added a con-stant value of 67.6◦ to all FENNOSTACK relative inclination data,and relocated the FENNOSTACK master curve to the geographiccoordinates of Augusta via the VGP method. A constant value of10◦ was then added to the relative declination curve to bring theyoungest declination data (dated at ca. 400 yr BP) close the coevalvalues predicted from all models (Fig. 6b).

SCHA.DIF.3K is a new regional model for the geomagnetic fieldvariation in Europe for the last 3000 yr developed by the integrationof the available European archeomagnetic and instrumental data(from 1000 BC to 1900 AD, connecting with the epoch coveredby the IGRF models) (Pavon-Carrasco et al. 2009). According tothis model, geomagnetic field values (declination, inclination andintensity) have been computed for the geographic coordinates ofAugusta Bay (latitude 37.1◦N, longitude 15.1◦E).

CALS7k.2 and CALS3k.3 are global continuous spherical har-monic models describing the geomagnetic main field variationfor the past 7000 yr (Korte & Constable 2005) and for the past3000 yr (Donadini et al. 2009; Korte et al. 2009), respectively.The PSV prediction at the geographic coordinates of Augusta,according to both models, have been computed by using thesoftware provided in the EarthRef.org Digital Archive (ERDA,http://www.earthref.org).

The gufm model is based on the historical direct measurements,mostly from maritime observations, of the geomagnetic field com-ponents since 1590 AD (Jackson et al. 2000).

The data from Knudsen et al. (2008) refer to a reconstruction ofthe virtual geomagnetic dipole moment variation on a global scale

during the past 50 ka, based on a database of absolute palaeointen-sity data published in peer-review journals.

The PSV data of cores MS06 and MS06-SW (declination, in-clination, RPI as computed by the NRM/ARM ratio scaled to itsmaximum value) are shown as a function of age in Fig. 6, witha high-resolution age model developed by using the available ra-diometric constraints and a visual correlation of palaeomagnetictrends with the discussed available PSV reference curves, based on12 tie-points (Table 2).

Prominent declination and inclination features for the UK mastercurve are indicated by Latin and Greek letters, respectively, follow-ing the nomenclature by Turner & Thompson (1981).

The PSV record of the Augusta Bay cores clearly shows the ma-jor maxima and minima in inclination (α–η) and declination (a–g)as seen in the other PSV records and models, indicating the samesources of the Earth’s magnetic field variation. The palaeomagneticdeclination trend of the Augusta Bay core, in particular, appearsto match closely the UK master curve (Fig. 6b). Notably, an al-most perfect match is observed for the large magnetic declinationswing at ∼2.0–2.6 ka BP, shown by all records. This prominentwesterly swing in declination (‘f–e event’), originally noted in UKPSV records by Turner and Thompson (1981) and recently reportedalso in lake sequences from Scandinavia (Snowball et al. 2007;Haltia-Hovi et al. 2010), seems to be a large-scale geomagneticfeature because it has also been reported in several other stud-ies extending from northern Siberia (Frank et al. 2002) to Iceland(Kristjansdottir et al. 2007) to eastern Canada (Barletta et al. 2010).The simultaneousness of the ‘f–e’ geomagnetic feature over wideareas throughout the Northern Hemisphere, within the errors of thedifferent chronologies for different sites supports the fidelity of theindividual PSV records.

A few features do not find a close match between the various PSVrecords and models. In particular, the sharp peak in palaeointensityat ca. 1–1.2 ka BP indicated by the various geomagnetic models,and mostly evident from the SCHA.DIF.3k model, was not found inthe MS06 record, whereas there is an excellent match between theMS06 RPI record and the palaeointensity variations predicted fromthe SCHA.DIF.3k model for the interval between 1.3 and 2.5 ka BP(Fig. 6c). In any case, The MS06 RPI record matches the generaltrend for the global virtual axial dipole moment, indicating that thedipole field decreased continuously during the last 1 ka and reacheda maximum peak value at about 2.6–2.7 ka BP (Yang et al. 2000;Knudsen et al. 2008).

A refined, ‘PSV-calibrated’ age model for cores MS06 and MS06-SW is shown in Fig. 7 and compared to the age model constrained

C© 2010 The Authors, GJI, 184, 191–202



Figure 7. A refined age model for cores MS06 and MS06-SW. The palaeomagnetic record provides a high-resolution age model through correlation of PSVfeatures with reference PSV stack and models (see text and Fig. 6). In the PSV calibrated age model, the age estimates generally follow the older rangeindicated by the uncertainties in the radiometric ages. All data suggest two significant changes in the sedimentation rate at ca. 0.7–1.0 and 5.7 m bsf (i.e. at ca.1200–1500 and 3200 yr BP, respectively).

by AMS 14C ages only. The correlation of PSV records provides ahigh-resolution age model, with a trend that generally follows theolder part of the time window associated to the 2 σ calibrated radio-metric ages. In fact, the data indicate that the mean calibrated 14Cages are for the most part younger than the ages transferred by PSVcorrelation: this may indicate either that the adopted constant �Roffset value is not appropriate and/or that the ages of the referencePSV features used as tie-points are too old. In any case, all data sug-gest two significant changes in the sedimentation rate at ca. 0.7–1.0and 5.7 m bsf (i.e. at ca. 1200–1500 and 3200 yr BP, respectively),the estimated mean sedimentation rate increases from a value of ca.75 cm ka−1 above 1 m bsf to ca. 250 cm ka−1 between 1 and 5.7 mbsf and then decreases again below 5.7 m bsf to ca. 90 cm ka−1.

C O N C LU S I O N

The palaeomagnetic data obtained from the Augusta Bay coresare of high quality and thus can provide a reliable high-resolutionand continuous central Mediterranean PSV record for the last 4.2millennia. This PSV record matches the main trends recognizedin reference PSV stack curves for Europe and PSV models basedon compilation of archeomagnetic, sedimentary and volcanic data.The available age constraints for the studied cores, mostly basedon radiometric dating of mollusc shells and the recognition of aprecisely dated tephra layer from Etna, provide a well-defined time-framework for the palaeomagnetic data. In turn, the correlation ofdetailed PSV features with regional PSV reference curves allows arefinement of the age model within the uncertainties implicit in theradiocarbon dating method.

Notwithstanding that a convincing correlation was observed withall the other considered PSV curves (Fig. 6), the best fit is observedwith the UK PSV stack, in agreement with the other Holocene PSVrecords from the Italian region.

Geomagnetic field models need input from all over the globe toprovide accurate estimates on dipolar as well as multipolar vari-ability. Cores MS06 and MS06-SW from the Augusta Bay havehigh potential for improving geomagnetic field models, since theyrefer to a region that is central to the Mediterranean and close tothe main European volcano (Mt Etna), which is another valuablesource of information for PSV reconstructions (Tanguy et al. 1985;Rolph & Shaw 1986), despite the discontinuous character of thevolcanic record and the limitations implicit in the palaeomagneticrecording of lava flows (Speranza et al. 2006). Of course, the PSVdata presented in this study refer to a single sequence and they arenot enough to establish a reference PSV master curve. Further care-fully measured and dated sequences from the same region couldeventually be used to construct such a regional PSV master curve.

As a whole, the correlation between the PSV data from theAugusta Bay cores and the other European PSV records and modelssuggests that, at least during the last four millennia, the same sourceis responsible for the geomagnetic SV recorded over the whole ofEurope.

The most prominent PSV feature appears to be the large declina-tion swing at ca. 2000–2600 cal BP, known as ‘f–e event’ which isclearly expressed in all records (Fig. 6b).

The revised age model for the MS06 core has implication forthe age calibration of given stratigraphic features, including the 12possible tsunami signatures found by Smedile et al. (2010), andindicates that the age of the core in the interval 1400–3200 yr BPis close to the older part of the age range indicated by radiocarbondating.

A C K N OW L E D G M E N T S

This work was partially funded by the Italian Civil ProtectionDepartment in the frame of the 2004–2006 agreement with Is-

C© 2010 The Authors, GJI, 184, 191–202



tituto Nazionale di Geofisica e Vulcanologia–INGV (Seismolog-ical Project S2) and by E.C. project 196 TRANSFER (con-tract number 037058). We are indebted with R. Civico and M.Mancini for the help provided during the sampling operations. Manythanks also to E. Lipparini and G. Marozzi for the technical sup-port in the ISMAR-CNR labs. The manuscript was improved bythe constructive comments of Joe Liddicoat and an anonymousreviewer.

R E F E R E N C E S

Anson, G.L. & Kodama K.P., 1987. Compaction-induced shallowing ofthe post-depositional remanent magnetization in a synthetic sediment,Geophys. J. R. astr. Soc., 88, 7–23.

Barletta, F., St-Onge, G., Stoner, J.S., Lajeunesse, P. & Locat, J., 2010. Ahigh-resolution Holocene paleomagnetic secular variation and relativepaleointensity stack from eastern Canada, Earth planet. Sci. Lett., 298,162–174.

Bellucci, L.G., Frignani, M., Cochran, J.K., Albertazzi, S., Zaggia, L.,Cecconi, G. & Hopkins, H., 2007. 210Pb and 137Cs as chronometersfor salt marsh accretion in the Venice Lagoon: links to flooding frequencyand climate change, J. Environ. Radioact., 97, 85–102.

Bleil, U. & von Dobeneck, T., 1999. Geomagnetic events and relative pale-ointensity records: clues to high-resolution paleomagnetic chronostrati-graphies of Late Quaternary marine sediments?, in Use of Proxies inPaleoceanography: Examples from the South Atlantic, pp. 635–654, edsFischer G. & Wefer G., Springer, New York.

Blow, R.A. & Hamilton, N., 1978. Effect of compaction on the acquisitionof a detrital remanent magnetization in fine-grained sediments, Geophys.J. R. astr. Soc., 52, 13–23.

Bloxham, J. & Gubbins, D., 1985. The secular variation of Earth’s magneticfield, Nature, 317, 777–781.

Bloxham, J. & Jackson, A., 1992. Time-dependent mapping of the magneticfield at the core-mantle boundary, J. geophys. Res., 97, 19 537–19 563.

Chadima, M. & Hrouda, F., 2006. Remasoft 3.0–a user-friendly paleomag-netic data browser and analyzer, Travaux Geophysiques, XXVII, 20–21.

Coltelli, M., Del Carlo, P. & Vezzoli, L. 1998. The discovery of a Plinianbasaltic eruption of Roman age at Mt. Etna, Geology, 26, 1095–1098.

Dal Forno, G. & Gasperini, L., 2008. ChirCor: a new tool for generatingsynthetic chirp-sonar seismograms, Comput. Geosci., 34, 103–114.

De Martini, P.M., Barbano, M.S., Smedile, A., Gerardi, F., Pantosti, D., DelCarlo, P. & Pirrotta, C., 2010. A unique 4000 year long geological recordof multiple tsunami inundations in the Augusta Bay (eastern Sicily, Italy),Marine Geol., 276, 42–57.

Donadini, F., Korte, M. & Constable, C.G., 2009. Geomagnetic field for0–3 ka: 1. New data sets for revised global models, Geochem. Geophys.Geosyst., 10, Q06007, doi:10.1029/2008GC002295.

Frank, U., Nowaczyk, N.R., Negendank, J.F.W. & Melles, M., 2002. A pale-omagnetic record from Lake Lama, Northern Central Siberia, Phys. Earthplanet. Int., 133, 3–20.

Gasperini, L. & Stanghellini, G., 2009. SEISPRHO: an interactive computerprogram for processing and interpretation of high-resolution seismic re-flection profiles, Comput. Geosci., 35, 1497–507.

Hagstrum, J.T. & Champion, D.E., 2002. A Holocene paleosecular varia-tion record from 14C-dated volcanic rocks in western North America, J.geophys. Res., 107(B1), 2025, doi:10.1029/2001JB000524.

Haltia-Hovi, E., Nowaczyk, N. & Saarinen, T., 2010. Holocene palaeomag-netic secular variation recorded in multiple lake sediments cores fromeastern Finland, Geophys. J. Int., 180, 609–622.

Holme, R. & Olsen, N., 2006. Core surface flow modelling from high-resolution secular variation, Geophys. J. Int., 166, 518–528.

Hyodo, M., 1984. Possibility of reconstruction of the past geomagnetic fieldfrom homogeneous sediments, J. Geomagn. Geoelectr., 36, 45–62.

Iorio, M., Sagnotti, L., Angelino, A., Budillon, F., D’Argenio, B., Dinares-Turell, J., Macrı, P., & Marsella, E., 2004. High resolution petrophysicaland palaeomagnetic study of Late Holocene Shelf Sediments, SalernoGulf, Tyrrhenian Sea, Holocene, 14, 417–425.

Jackson, A., Jonkers, R.T. & Walker, M.R., 2000. Four centuries of geomag-netic secular variation from historical records, Phil. Trans. R. Soc. Lond.A, 358, 957–990.

King, R.F., 1955. The remanent magnetism of artificially deposited sedi-ments, Mon. Not. R. astr. Soc., 7, 115–134.

King, J.W., Banerjee, S.K., Marvin, J.A. & Ozdemir, O., 1982. A comparisonof different magnetic methods for determining the relative grain size ofmagnetite in natural materials: some results from lake sediments, Earthplanet. Sci. Lett., 59, 404–419.

Kirschvink, J.L., 1980. The least-squares line and plane and the analysis ofpaleomagnetic data, Geophys. J. R. astr. Soc., 62, 699–718.

Knudsen, M.F., Riisager, P., Donadini, F., Snowball, I., Muscheler, R., Ko-rhonen, K. & Pesonen, L.J., 2008. Variations in the geomagnetic dipolemoment during the Holocene and the past 50 kyr, Earth planet. Sci. Lett.,272, 319–329.

Korte, M., & Constable C.G., 2005. Continuous geomagnetic field modelsfor the past 7 millennia: 2. CALS7K, Geochem. Geophys. Geosyst., 6,Q02H16, doi:10.1029/2004GC000801.

Korte, M., Donadini, F. & Constable, C.G., 2009. Geomagnetic field for 0–3ka: 2. A new series of time-varying global models, Geochem. Geophys.Geosyst., 10, Q06008, doi:10.1029/2008GC002297.

Kristjansdottir, G.B., Stoner, J.S., Jennings, A.E., Andrews, J.T. & Gronvold,K., 2007. Geochemistry of Holocene cryptotephras from the North Ice-land Shelf (MD99–2269): intercalibration with radiocarbon and palaeo-magnetic chronostratigraphies, Holocene, 17, 155–176.

Leslie, B.W., Lund, S.P. & Hammond, D.E., 1990a. Rock magnetic evi-dence for the dissolution and authigenic growth of magnetic mineralswithin anoxic marine sediments of the California continental borderland,J. geophys. Res., 95, 4437–4452.

Leslie, B.W., Hammond, D.E., Berelson, W.M. & Lund, S.P., 1990b. Diage-nesis in anoxic sediments from the California continental boreland andits influence on Iron, Sulfur, and Magnetite behavior, J. geophys. Res., 95,4453–4470.

Lund, S., 2007. Paleomagnetic secular variation, in Encyclopedia of Ge-omagnetism and Paleomagnetism, pp. 766–776, eds Gubbins, D. &Herrero-Bervera, E., Springer, Dordrecht.

Magagnoli, A. & Mengoli, M., 1999. Un brevetto CNR per lacampionatura dei fondali marini, Ricerca & Futuro, 14, 67–68(http://www.area.fi.cnr.it/r&f/n14/magnoli.htm)

Maher, B., 1988. Magnetic properties of some synthetic submicron mag-netites, Geophys. J., 94, 83–86.

Mitra, R. & Tauxe, L., 2009. Full vector model for magnetization in sedi-ments, Earth planet. Sci. Lett., 286, 535–545.

Nilsson, A., Snowball, I., Muscheler, R. & Uvo, C. B., 2010, Holocene geo-centric dipole tilt model constrained by sedimentary paleomagnetic data,Geochem. Geophys. Geosyst., 11, Q08018, doi:10.1029/2010GC003118.

Noel, M. & Batt, C.M., 1990. A method for correcting geographically sep-arated remanence directions for the purpose of archeomagnetic dating,Geophys. J. Int., 102, 753–756.

Pavon-Carrasco, F. J., Osete, M.L., Torta, J.M. & Gaya-Pique, L.R., 2009.A regional archeomagnetic model for Europe for the last 3000 years,SCHA.DIF.3K: applications to archeomagnetic dating, Geochem. Geo-phys. Geosyst., 10, Q03013, doi:10.1029/2008GC002244.

Ramsey, B.C., 2009. Bayesian analysis of radiocarbon dates, Radiocarbon,51(1), 337–360.

Reimer, P.J. & Reimer, R.W., 2001. A marine reservoir correctiondatabase and on-line interface, Radiocarbon, 43, 461–463. [suppl.mat.URL:http://www.calib.org)].

Reimer, P.J. et al., 2009. IntCal09 and Marine09 radiocarbon age calibrationcurves, 0–50,000 years cal BP, Radiocarbon, 51(4), 1111–1150.

Roberts, A.P., 2006. High-resolution magnetic analysis of sedimentcores: strengths, limitations and strategies for maximizing the valueof long-core magnetic data, Phys. Earth planet. Int., 156, 162–178.

Roberts, A.P. & Winklhofer, M., 2004. Why are geomagnetic excursionsnot always recorded in sediments? Constraints from post-depositionalremanent magnetization lock-in modelling, Earth planet. Sci. Lett., 227,345–359.

C© 2010 The Authors, GJI, 184, 191–202



Rolph, T.C. & Shaw, J., 1986. Variations of the geomagnetic field in Sicily,J. Geomagn. Geoelectr., 38, 1269–1277.

Rolph, T.C., Vigliotti, L. & Oldfield, F., 2004. Mineral magnetism andgeomagnetic secular variation of marine and lacustrine sediments fromcentral Italy: timing and nature of local and regional Holocene environ-mental change, Quat. Sci. Rev., 23, 1699–1722.

Sagnotti, L., Budillon, F., Dinares-Turell, J., Iorio, M. & Macrı, P.,2005. Evidence for a variable paleomagnetic lock-in depth in theHolocene sequence from the Salerno Gulf (Italy): implications for‘high-resolution’ paleomagnetic dating, Geochem. Geophys. Geosyst., 6,Q11013, doi:10.1029/2005GC001043.

Smedile, A. et al., 2010. Possible tsunami signatures from an integratedstudy in the Augusta Bay offshore (Eastern Sicily – Italy), Mar. Geol.,submitted.

Snowball, I., Zillen, L., Ojala, A., Saarinen, T. & Sandgren, P., 2007.FENNOSTACK and FENNORPIS: varve dated Holocene paleomagneticsecular variation and relative palaeointensity stacks for Fennoscandia,Earth planet. Sci. Lett., 255, 106–116.

Speranza, F., Pompilio, M. & Sagnotti, L., 2004. Paleomagnetism of spat-ter lavas from Stromboli volcano (Aeolian Islands, Italy): implicationsfor the age of paroxysmal eruptions, Geophys. Res. Lett., 31, doi:10.1029/2003GL018944, 2004.

Speranza, F., Branca, S., Coltelli, M., D’Ajello Caracciolo, F. & Vigliotti,L., 2006. How accurate is ‘‘paleomagnetic dating’’? New evidencefrom historical lavas from Mount Etna, J. geophys. Res., 111, B12S33,doi:10.1029/2006JB004496.

Speranza, F., Pompilio, M., D’Ajello Caracciolo, F. & Sagnotti, L.,2008. Holocene eruptive history of the Stromboli volcano: con-straints from paleomagnetic dating, J. geophys. Res., 113, B09101,doi:10.1029/2007JB005139.

St-Onge, G., Stoner, J.S. & Hillaire-Marcel, C., 2003. Holocene paleomag-netic records from the St. Lawrence Estuary, eastern Canada: centennial tomillennial-scale geomagnetic modulation of cosmogenic isotopes, Earthplanet. Sci. Lett., 209, 113–130.

Tanguy, J.-C., Bucur, I. & Thompson J.F.C., 1985. Geomagnetic secularvariation in Sicily and revised ages of historic lavas from Mount Etna,Nature, 318, 453–455.

Thompson, R. & Barraclough, D.R., 1982. Geomagnetic secular varia-tion based on Spherical Harmonic and Cross Validation analyses ofhistorical and archaeomagnetic data, J. Geomagn. Geoelectr., 34, 245–263.

Turner, G.M. & Thompson, R., 1981. Lake sediment record of the geomag-netic secular variation in Britain during Holocene times, Geophys. J., 65,703–725.

Turner, G.M. & Thompson, R., 1982. Detransformation of the British ge-omagnetic secular variation record for Holocene times, Geophys. J., 70,789–792.

Verosub, K.L., 1977. Depositional and postdepositional processes in themagnetization of sediments, Rev. Geophys., 15, 129–143.

Verosub, K.L., 2000. Paleomagnetic dating, in Quaternary Geochonology:Methods and Applications, pp. 339–356, eds Noller, J., Sower, J. & Lettis,W.R., AGU Reference Shelf 4, Washington, D.C.

Vigliotti, L., 2006. Secular variation record of the Earth’s magnetic field inItaly during the Holocene: constraints for the construction of a mastercurve, Geophys. J. Int., 165, 414–429.

Yang, S., Odah, H. & Shaw, J., 2000. Variations in the geomagnetic dipolemoment of the last 12 000 years, Geophys. J. Int., 140, 158–162.

Zijderveld, J.D.A., 1967. A. C. demagnetization of rocks: analysis of results,in Methods in Palaeomagnetism, pp. 254–286, eds Collinson, D.W., Creer,K.M. & Runcorn, S.K., Elsevier, Amsterdam.

C© 2010 The Authors, GJI, 184, 191–202


A continuous palaeosecular variation record of the last four millennia from the Augusta Bay (Sicily,...

Documents

Transcript of A continuous palaeosecular variation record of the last four millennia from the Augusta Bay (Sicily,...