See discussions, stats, and author profiles for this publication at: [Link]

net/publication/27774894

Age of the Corsica–Sardinia rotation and Liguro–Provençal Basin

spreading: new paleomagnetic and Ar/Ar evidence

Article · January 2002

Source: OAI

CITATIONS READS
109 902

20 authors, including:

Fabio Speranza Igor Maria Villa

National Institute of Geophysics and Volcanology Universität Bern
195 PUBLICATIONS 5,266 CITATIONS 414 PUBLICATIONS 16,273 CITATIONS

SEE PROFILE SEE PROFILE

Leonardo Sagnotti
National Institute of Geophysics and Volcanology
266 PUBLICATIONS 7,650 CITATIONS

SEE PROFILE

All content following this page was uploaded by Florindo Fabio on 31 May 2014.

The user has requested enhancement of the downloaded file.

 Tectonophysics 347 (2002) 231 – 251
[Link]/locate/tecto

Age of the Corsica–Sardinia rotation and Liguro–Provencßal Basin

spreading: new paleomagnetic and Ar/Ar evidence
F. Speranza a,*, I.M. Villa b, L. Sagnotti a, F. Florindo a, D. Cosentino c,
P. Cipollari c, M. Mattei c
a
Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, 00143 Rome, Italy
b
Isotopengeologie, Erlachstrasse 9a, 3012 Berne, Switzerland
c
Dipartimento di Scienze Geologiche, Università di Roma Tre, Largo S. L. Murialdo 1, 00146 Rome, Italy

Received 20 July 2001; accepted 14 January 2002

Abstract

The age of spreading of the Liguro – Provencßal Basin is still poorly constrained due to the lack of boreholes penetrating the
whole sedimentary sequence above the oceanic crust and the lack of a clear magnetic anomaly pattern. In the past, a consensus
developed over a fast (20.5 – 19 Ma) spreading event, relying on old paleomagnetic data from Oligo – Miocene Sardinian
volcanics showing a drift-related 30j counterclockwise (CCW) rotation. Here we report new paleomagnetic data from a 10-m-
thick lower – middle Miocene marine sedimentary sequence from southwestern Sardinia. Ar/Ar dating of two volcanoclastic
levels in the lower part of the sequence yields ages of 18.94 F 0.13 and 19.20 F 0.12 Ma (lower – mid Burdigalian).
Sedimentary strata below the upper volcanic level document a 23.3 F 4.6j CCW rotation with respect to Europe, while younger
strata rapidly evolve to null rotation values. A recent magnetic overprint can be excluded by several lines of evidence,
particularly by the significant difference between the in situ paleomagnetic and geocentric axial dipole (GAD) field directions.
In both the rotated and unrotated part of the section, only normal polarity directions were obtained. As the global magnetic
polarity time scale (MPTS) documents several geomagnetic reversals in the Burdigalian, a continuous sedimentary record
would imply that (unrealistically) the whole documented rotation occurred in few thousands years only. We conclude that the
section contains one (or more) hiatus(es), and that the minimum age of the unrotated sediments above the volcanic levels is
unconstrained. Typical back-arc basin spreading rates translate to a duration z 3 Ma for the opening of the Liguro – Provencßal
Basin. Thus, spreading and rotation of Corsica – Sardinia ended no earlier than 16 Ma (early Langhian). A 16 – 19 Ma, spreading
is corroborated by other evidences, such as the age of the breakup unconformity in Sardinia, the age of igneous rocks dredged
west of Corsica, the heat flow in the Liguro – Provencßal Basin, and recent paleomagnetic data from Sardinian sediments and
volcanics. Since Corsica was still rotating/drifting eastward at 16 Ma, it presumably induced significant shortening to the east,
in the Apennine belt. Therefore, the lower Miocene extensional basins in the northern Tyrrhenian Sea and margins can be
interpreted as synorogenic ‘‘intra-wedge’’ basins due to the thickening and collapse of the northern Apennine wedge. D 2002
Elsevier Science B.V. All rights reserved.

Keywords: Paleomagnetism; Corsica – Sardinia; Liguro – Provencal Basin; Back-arc spreading

*
Corresponding author.
E-mail address: speranza@[Link] (F. Speranza).

0040-1951/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 0 4 0 - 1 9 5 1 ( 0 2 ) 0 0 0 3 1 - 8
232 F. Speranza et al. / Tectonophysics 347 (2002) 231–251

1. Introduction times (Burrus, 1984). Basin spreading and simulta-

neous eastward migration of the adjacent Alpine belt
The Miocene-to-Present evolution of the Mediter- and Corsica – Sardinia – Calabria blocks were likely
ranean region is characterized by the fast opening of driven by the eastward retreat of a Ionian/Adriatic
several back-arc basins, generally floored by oceanic slab passively sinking into the mantle (Malinverno
crust, within the framework of the Africa – Eurasia and Ryan, 1986). Since the middle – late Miocene,
collision and Alpine orogenesis (e.g. Lonergan and further roll-back of the same slab caused the spreading
White, 1997; Jolivet et al., 1998). of the Tyrrhenian Sea, the southeastward drift of the
Constraining the age and spreading rate of those Calabrian block, and the orogenesis of the Apennines
basins is an important part to resolve the geodynamic (Patacca et al., 1990).
setting of the Africa – Eurasia boundary during Neo- The basin spreading chronology is fairly well
gene times. In the western Mediterranean, the Lig- known for the Tyrrhenian Sea, where several drill-
uro – Provencßal Basin, a triangular sea located between holes (ODP Leg 107) penetrated the entire sedimen-
the Provencßal – Catalan coasts and the Corsica– Sardi- tary cover and the underlying continental/oceanic crust
nia block (Fig. 1), opened during Oligo – Miocene (Kastens et al., 1986; Mascle et al., 1988; Sartori,

Fig. 1. Shaded relief map of the western Mediterranean Sea and surrounding regions, created by GMT software.
 F. Speranza et al. / Tectonophysics 347 (2002) 231–251 233

1989). Conversely, borehole data are lacking for the use of Ar/Ar dates allow us to overcome the problems
oceanic (and deepest) sector of the Liguro – Provencßal encountered by previous authors, who analyzed vol-
Basin. Although the onset of the syn-rift sedimentation canic rocks. Our data provide evidence for a post-19
is recorded on the Liguro –Provencß al conjugate mar- Ma 23j CCW rotation of the Corsica – Sardinia block.
gins (Réhault et al., 1984; Sowerbutts and Underhill,
1998), these data yield age estimates only for the oldest
rifting episodes, whereas the drifting age can be con- 2. Age constraints for the Liguro– Provencßal Basin
strained only by directly dating the sequences depos- spreading
ited above the oceanic crust.
The Liguro – Provencßal spreading took place simul- The timing of the Liguro –Provencßal Basin spread-
taneously with the eastward drift of the Corsica – ing has been extrapolated by (1) identification of the
Sardinia block, which rotated at least 30j counter- pre-, syn- and post-rift sequences in field studies over
clockwise (CCW) about a pole located north of the basin margins and from offshore seismic reflection
Corsica (e.g. Van der Voo, 1993 and references data, (2) analysis of the magnetic anomaly pattern in
therein). This drift/rotation link offers the opportunity the oceanic crust domain, (3) dating of dredged
to indirectly date the Liguro – Provencßal opening by igneous rocks, (4) dating of calc-alkaline volcanism
means of well-dated paleomagnetic directions. Mon- in western Sardinia, interpreted as the volcanic arc
tigny et al. (1981), by considering K/Ar ages and the located east of the Liguro – Provencßal back-arc basin,
synthesis of a large amount of previous paleomagnetic and (5) analysis of paleomagnetic data from Corsica
data on the Sardinian volcanics, proposed that the and Sardinia (Chamot-Rooke et al., 1999 and refer-
rotation occurred very rapidly between 20.5 and 19 ences therein).
Ma. This hypothesis was widely accepted afterwards, Extension-related sedimentation is recorded since
and the paleomagnetic rotation of Sardinia was late Oligocene both in the Catalan– Provencßal (Réhault
regarded as a key datum to constrain the Liguro – et al., 1984; Mauffret et al., 1992) and Sardinian
Provencßal spreading age (e.g. Burrus, 1984; Chamot- margins (Assorgia et al., 1997). Although these sedi-
Rooke et al., 1999). However, a critical analysis of the ments indicate the first Liguro – Provencßal rifting epi-
old paleomagnetic data set based on more modern and sodes, there is no general consensus on the age of the
stringent criteria demonstrated that coeval paleodecli- rift-to-drift transition. In the Valencia trough (Fig. 1),
nations from Sardinia are scattered, hindering a pre- extension continued during the Miocene and Plio –
cise definition of the age-related rotation evolution Pleistocene; thus, here, the post-rift sequences cannot
(Todesco and Vigliotti, 1993). be identified (Maillard et al., 1992). Burrus and
Inconsistencies in the old Sardinian paleomagnetic Foucher (1986), relying on offshore drilling and seis-
data likely arise from various sources: (1) few sites are mic data, suggest an upper Oligocene– Aquitanian age
dated, and the ages are obtained through K/Ar dating; of the syn-rift sediments and a lower Burdigalian
(2) a precise tilt correction is almost impossible to breakup unconformity in the Provencßal margin. Onland
perform for most of the calc-alkaline volcanics; (3) the in western Sardinia, the pre-upper Burdigalian deposits
volcanics recorded the directional scatter of the geo- are extensionally faulted, tilted, and unconformably
magnetic field secular variation; (4) the magnetic covered by virtually undeformed upper Burdigalian –
mineralogy is complex (as reported by many authors); lower Messinian shallow-marine post-rift successions
(5) it cannot be excluded that most of the sites dated at (Sowerbutts and Underhill, 1998). Lacking deep wells
ca. 19 Ma and exhibiting no rotation (and normal in the central Liguro – Provencßal Basin, the strata
magnetic polarity) were recently remagnetized. imaged by seismic reflection data upon the oceanic
In this paper, we report on new paleomagnetic data crust cannot be precisely constrained in age, and are
obtained from a 10-m-thick marine sedimentary sec- generically referred to the lower – middle Miocene
tion from southwestern Sardinia (Marmilla region). (Burrus, 1984; Mauffret et al., 1992).
Ar/Ar dating of two volcanoclastic levels from the The central part of the Liguro – Provencßal Basin is
lower third of the section gives absolute age tie points. characterized by a prominent positive magnetic anom-
The paleomagnetic study of sedimentary strata and the aly (Bayer et al., 1973). This feature is hardly datable
234 F. Speranza et al. / Tectonophysics 347 (2002) 231–251

as the correlation to the global magnetic polarity time Since the late 1960s and during the 1970s, several
scale (MPTS) strongly depends on other data sets that paleomagnetic studies were carried out on the Sardinian
can independently constrain the temporal window. rocks (De Jong et al., 1969, 1973; Zijderveld et al.,
Burrus (1984) associated the positive magnetic resid- 1970; Coulon et al., 1974; Manzoni, 1974; Edel and
uals to the polarity chron C6n (19.048 –20.131 Ma Lortscher, 1977; Edel, 1979; Horner and Lowrie,
according to Cande and Kent, 1995) relying on the 1981). Mainly the Oligo – Miocene and Permian vol-
approximately 20 Ma age of the Corsica– Sardinia canics were sampled. The data showed two f 30j
rotation then inferred from paleomagnetic data from CCW rotation events with respect to Europe occurring
Sardinia. This dating has been widely accepted for during late Mesozoic –early Tertiary and early – middle
quite some time, although other authors recently have Miocene times, respectively (e.g. Van der Voo, 1993).
tended to make the anomaly age younger. Chamot- The debatable coupling or decoupling between Corsica
Rooke et al. (1999) propose that the magnetic anom- and Sardinia during drifting (Westphal et al., 1976) was
aly pattern might be more coherently related to anom- resolved by Vigliotti et al. (1990), who showed that
aly C5Cn, showing a normal polarity triplet in the after the Permian the two islands rotated as one block.
time interval 16.014 – 16.726 Ma according to Cande The Tertiary paleomagnetic data set was completed
and Kent (1995). and synthesized by Montigny et al. (1981), who
A tristanite sample dredged from a ridge northwest proposed the 20.5 –19 Ma fast rotational event. More
of Corsica yielded an age of 18.0 F 0.5 Ma (Réhault recently, Todesco and Vigliotti (1993) showed that a
et al., 1984), but recent redating would indicate a 30 –40j scatter in declinations exists among coeval
f 17 Ma age (Chamot-Rooke et al., 1999). More paleomagnetic sites considered by Montigny et al.
recently, several volcanic edifices were sampled (1981), and that coeval normal and reverse polarity
northwest and southwest of Corsica, and Ar/Ar and directions are not antipodal, strongly reducing the
fission tracks dating indicate 16.1 F 04 and about 17.5 reliability of the whole data set. Moreover, after a
Ma ages, respectively (Deverchère, 1996). However, careful statistical analysis of the existing data, they
the use of dredged igneous rocks ages to constrain the confirmed a 20 – 21 Ma age for the beginning of
timing of basin opening remains controversial as (1) rotation, but showed that the termination of the
the volcanic cones west of Corsica display both arc- rotation is virtually unconstrained.
and rift-type signature, making unclear whether they Recently, paleomagnetic data from three Sardinian
are related to the Liguro – Provencß al Basin spreading upper Burdigalian –Langhian sedimentary sites (Vig-
or whether they represent the northern prolongation of liotti and Langenheim, 1995) seemed to indicate that at
the Sardinian volcanic arc, and (2) isolated volcanic 15 –18 Ma, the rotation was not yet completed. Finally,
eruptions may not be linked to and can occur after, the Gattacceca (2001) and Deino et al. (2001) have
basin spreading. reported new paleomagnetic data and Ar/Ar ages from
In western Sardinia a wealth of calc-alkaline vol- volcanics exposed in several central –northern Sardi-
canics (mainly domes and ignimbrites) crop out along nia localities, also resampling some units studied by
the N – S Sardinia Trough (Fig. 2) interbedded with Montigny et al. (1981). Their data confirm that Sardi-
upper Oligocene – upper Miocene sediments. Their nia started rotating at f 21 Ma, and indicate that the
age (13 –32 Ma according to Beccaluva et al., 1985) overall rotation was as high as 50j. A well-constrained
was used to approximate the time span comprising the tie point in the rotation history further documents that
whole Liguro – Provencß al rifting and drifting events 13j of rotation occurred after 18.18 F 0.03 Ma.
(Vigliotti and Langenheim, 1995). We note, however,
that arc volcanism is attributed to the activity of the
subduction process beneath Sardinia, which may have 3. Geological outline of the Marmilla region
continued also after the cessation of the Liguro –
Provencßal Basin spreading. Therefore, the end of the The earliest marine deposits of the late Paleogene–
Liguro – Provencßal spreading is not constrained by the Neogene sedimentary cycles in central Sardinia are
cessation of Sardinian calc-alkaline volcanism at 13 Chattian in age and belong to the Riu Su Rettore Fm.
Ma. (Farris, 1990). This latter has been correlated to the
 F. Speranza et al. / Tectonophysics 347 (2002) 231–251 235

Fig. 2. Simplified geological map of Sardinia and location of the sampling locality.
236 F. Speranza et al. / Tectonophysics 347 (2002) 231–251

uppermost part of the Ussana Fm. (Pecorini and 4. Geological features and biostratigraphy of the
Pomesano Cherchi, 1969), late Oligocene – earliest studied sedimentary section
Miocene in age. Despite the older age of these depo-
sits, all over the central and northern Sardinia the ma- The studied stratigraphic section has been sampled
rine transgression is widely extended only starting along a new (in 1995) N – S subvertical road cut, to the
from the earliest Aquitanian (Assorgia et al., 1992a,b, NE of the Collinas village (Lat. 39.6jN, Long. 8.8jE,
1997; Barca et al., 1996) when, in western Sardinia, Fig. 2). The section is about 10 m thick. Bed dip is
extensional tectonics gave rise to the so-called ‘‘Sar- towards the SE and decreases from 22j in the north-
dinia Trough’’. This is a N – S trending elongate zone ern (lower) part of the section to 11j in the southern
where the pre-Mesozoic crystalline rocks are covered (upper) part. The lower 5.3 m of section are mainly
by Tertiary continental and marine sediments inter- composed by calcareous marls interbedded with con-
bedded with mainly Miocene calc-alkaline volcanics glomeratic levels and by arenaceous and volcaniclas-
(Fig. 2). Both in the northern Iglesiente (Funtanazza tic horizons. In particular, at 1.8 and 3.3 m above the
section, Assorgia et al., 1988) and in Marmilla, the base of the section, we sampled for radiometric dating
basal portion of the marine succession, sedimented purposes two yellow volcaniclastic layers (LV1 and
above the continental deposits (Ussana Fm. and/or Sa LV2), 10 and 25 cm thick, respectively (Fig. 3). They
Tellura lacustrine limestones), is referable to the N4 are calc-alkaline volcanic sandstones, mostly consist-
Zone (Blow, 1967). ing of quartz, plagioclase and k-feldspar grains. The
In the Marmilla region, particularly between Sar- calcareous marls are normally blue-grey except above
dara and Villanovaforru villages, the lower portion of LV1, where they turn into yellow for a thickness of
the Marmilla Fm. has been referred to the lower about 40 cm (Fig. 3).
Burdigalian by means of calcareous nannofossil An abrupt lithologic change occurs at about 5.3 m
assemblages (Iaccarino et al., 1985; Cherchi et al., above the base of the section. Here, the marly thick
1985). In the same area, close to the Collinas village, strata pass upwards, through a highly irregular surface
silicized marly horizons and andesitic lava bodies with (disconformity), to blue-grey clays containing scat-
frequent pillow structures testify an intense submarine tered marly strata 10 – 30 cm thick. The clays are about
volcanic activity. In this region, the early Miocene 4 m thick and pass upwards to a chaotic clayey complex
sedimentary basin was also affected by strong insta- including slumps from marly beds. The marls of the
bility, probably linked to intense syn-sedimentary lower part of the section are cut by E – W syn-sedi-
tectonics. Here, the sedimentary succession shows mentary normal faults, producing maximum offsets of
angular unconformities, channellized coarse-grained some centimeters, which are sealed by the clays.
deposits and frequent slumped horizons. Moreover, A biostratigraphic analysis, based on calcareous
the Lower Miocene deposits (belonging to the 1st nannofossil assemblages, has been carried on all the
Miocene sedimentary cycle) are cut by growth-faults samples collected for paleomagnetic study. Despite the
and by faults that are sealed by younger sediments. very dense sampling (average spacing of 10 cm), the
The instability of the Marmilla sedimentary basin has scarce presence of nannofossils prevented a quantita-
to be related to the tectonic activity connected with the tive analysis. Moreover, strong etching affected the
spreading of the Liguro –Provencßal Basin. nannofossil assemblage, inducing the preservation
The Upper Oligocene– ?upper Burdigalian deposits exclusively of the more resistant species such as
are unconformably overlain by the 2nd Miocene Dictyococcites sp., D. scrippsae, D. productus, Cocco-
sedimentary cycle (Langhian – ?upper Serravallian) lithus pelagicus, Cyclicargolithus floridanus, Spheno-
through an irregular surface characterised by a strati- lithus moriformis. Unfortunately, these species are not
graphic gap and, sometimes, by continental deposits significant from a biostratigraphic point of view. How-
(central– southern Sardinia) and/or by calc-alkaline ever, samples collected close to the study section, in a
ignimbrites (central-northern Sardinia) (Assorgia et al., lower stratigraphic position, provided more constrain-
1997). In the Marmilla region, the 2nd Miocene sedi- ing data. Even if also in these samples the assemblages
mentary cycle is mainly characterised by marly and are mainly characterised by very resistant species (C.
silty – marly deposits of the Marne di Gesturi Fm. pelagicus, Dictyococcites sp., D. scrippsae, D. produc-
 F. Speranza et al. / Tectonophysics 347 (2002) 231–251
Fig. 3. Stratigraphic column of the studied section and correspondent magnetic parameters. The stratigraphic positions are referred to the bottom of the exposed section. Lithologies:
(1) chaotic clayey complex; (2) clays with subordinate marly levels; (3) blue-grey marls; (4) volcanoclastic levels; (5) yellow marls.

237
238 F. Speranza et al. / Tectonophysics 347 (2002) 231–251

tus, Cy. floridanus, S. gr. moriformis), delicate species the paleomagnetic laboratory of the Istituto Nazionale
have been detected (Helicosphaera carteri and Retic- di Geofisica e Vulcanologia (Roma). Susceptibility
ulofenestra < 7 Am). As H. carteri is commonly present values are generally < 200 ASI (Fig. 3), then they pro-
since the lowermost analyzed samples, this portion of bably reflect the main contribution of the paramag-
succession can be located above the First Common netic clay minerals (i.e. Sagnotti et al., 1998). The
Occurrence (FCO) of this species and then it should not lowest k values (30 – 50 ASI) are measured in the marls
be older than the top of MNN 1d Zone (Fornaciari and located between LV2 and the clays, suggesting a re-
Rio, 1996), which is early Burdigalian. This age is in duced input of the terrigenous (clay) fraction at such
agreement with the data coming from the biostrati- stratigraphic interval.
graphic analyses performed on a stratigraphic section The AMS was evaluated using Jelinek (1978)
of the Marmilla Fm., located a few kilometers to the statistics. In almost all samples, the magnetic foliation
south of the Collinas area, along the Sardara – Villano- plane is always well defined and parallel to the bed-
vaforru road. In this section, nannofossil assemblages ding plane (Fig. 4), as is commonly observed in
are generally better preserved. The species detected are undeformed to weakly deformed sediments. There-
the same, but they are much more abundant with fore, AMS data show a typical sedimentary fabric and
respect to those of Collinas. Furthermore, previous help to exclude both anomalous magnetic fabrics due
biostratigraphic analyses, based on planktonic foramin- to peculiar magnetic mineralogy (e.g. Gialanella et al.,
ifera, referred the Sardara –Villanovaforru section to 1994; Lehman et al., 1996) and significant tectonics-
the Catapsidrax dissimilis Subzone (Iaccarino et al., related fabric modification.
1985), confirming an early Burdigalian age.

5. Paleomagnetic sampling

We drilled 84 cores along the section (average

spacing of 10 cm) using an ASC 280E petrol-powered
portable drill, and oriented them in situ using a mag-
netic compass. We drilled 42 cores in the lower marls
(Fig. 3) and avoided the volcanoclastic levels as they
likely did not average the secular variation of the
geomagnetic field. The remaining 42 cores were
sampled from the upper clayey section. As a conse-
quence of the bedding attitude and the road cut ori-
entation, the lower section crops out in the north,
whereas the upper beds are mainly exposed in the
south. Therefore, we drilled the cores along two verti-
cal profiles, about 25 m apart: 62 samples (from both
the lower marls and the upper clays) were drilled in the
north (bedding dip here is 22j), while the remaining 22
samples (two from the marls and 20 from the clays)
were collected in the south, where the beds dip 11j.

6. Anisotropy of magnetic susceptibility Fig. 4. Anisotropy of magnetic susceptibility (AMS) data for all the
samples; Schmidt equal-area projection, lower hemisphere, geo-
graphic coordinates. The magnetic foliation (i.e. the plane con-
The low-field magnetic susceptibility (k) and ani- taining the maximum and intermediate susceptibility axes) of almost
sotropy of magnetic susceptibility (AMS) of all the 84 all the samples is subparallel to bedding, showing a typical sedi-
samples was measured with a KLY-2 bridge (Agico) in mentary fabric.
 F. Speranza et al. / Tectonophysics 347 (2002) 231–251 239

7. Paleomagnetism unblocked either below 120 jC (Fig. 5e), or between

600 and 650 jC (Fig. 5f), indicating the presence of
The natural remanent magnetization (NRM) of all goethite and hematite, respectively. An intermediate-
the 84 samples was measured, and the magnetic mi- coercivity phase with maximum unblocking temper-
neralogy of 14 representative samples was specifically atures of 300 –400 jC was noticed in most of the
investigated (about one sample every 65 cm). All the samples (Fig 4c –e). Furthermore, in some cases, a
measurements were done in the magnetically shielded low- to intermediate-coercivity phase with unblocking
room of the paleomagnetic laboratory at the Istituto temperature of about 180– 220 jC was also observed
Nazionale di Geofisica e Vulcanologia. (Fig. 5f). These magnetic properties are not uniquely
linkable to a specific magnetic mineral and may be due
7.1. Magnetic mineralogy to magnetic iron sulphides, maghemite or high-Ti
titanomagnetites.
For the selected 14 specimens, we studied the As a summary, the studied samples contain a com-
isothermal remanent magnetization (IRM) acquisition plex magnetic mixture in which magnetite carries the
curves up to 900 mT (Fig. 5a) and the coercivity of main part of remanence, but small amounts of goethite,
remanence (Hcr), evaluated by applying an increasing hematite and other magnetic minerals are present.
back-field to the IRM at 900 mT (Fig. 5b). The
inducing field was imparted stepwise by a 2G pulse 7.2. Demagnetization of the natural remanent magnet-
magnetizer, and the remanence was measured in-line ization
with a 4.5 cm access pass through 2G cryogenic
magnetometer. Seven out of fourteen samples (equally The NRM of all the 84 samples (one sample per
distributed through the section) are virtually saturated core) was thermally demagnetized in 10 –15 steps (13
at 100– 200 mT (samples SA0205B and SA0222B in in average) and the remanence was measured using a
Fig. 5a), and showed an Hcr ranging from 30 to 50 JR-5A spinner magnetometer (Agico). NRM inten-
mT (Fig. 5b). In the remaining samples, the ratio sities are higher (greater than 500 AA/m) in the marly
between the remanence acquired at 100 and 900 mT is levels around LV1 and in most of the upper clays,
60 – 70%, and the Hcr ranges from 50 to 80 mT. while lower values (between 100 and 200 AA/m) are
Therefore, in about 50% of the samples, the main measured in the marls located between LV2 and the
magnetic carriers are only low-coercivity minerals, clays (Fig. 3). The magnetization directions from each
while in the others, both low- and high-coercivity sample were analyzed in orthogonal diagrams and the
minerals coexist. characteristic remanence components were evaluated
In order to better define the nature of these low- using the principal component analysis (Kirschvink,
coercivity and high-coercivity magnetic fractions, we 1980).
thermally demagnetized a three-component IRM In all the samples, a viscous component sub-
(acquired at fields of 900, 500 and 120 mT applied parallel to the present-day field is eliminated by 250
in sequence along the three sample axes), according to to 310 jC (Fig. 6). In about 60% of the samples, a
the method of Lowrie (1990). In the samples contain- characteristic remanent magnetization (ChRM) is
ing only the low-coercivity fraction, the magnetization defined between 250 and 550 jC (Fig. 6a). In four
is unblocked between 550 and 600 jC (Fig. 5c), samples from the yellow marls above LV1, two differ-
indicating the presence of magnetite. In the specimens ent components are unblocked between 310 and 460
sampled in the yellow marls above LV1, a minor jC [low temperature (LT) component] and 460 –600
amount of high-coercivity fraction (not detected dur- jC [high temperature (HT) component], respectively
ing IRM acquisition) is demagnetized between 600 (Fig. 6b). As the HT component is subparallel to the
and 650 jC, showing that some hematite is associated present-day field and is likely carried by hematite, it
to magnetite (Fig. 5d). In the samples containing a probably arises from recent alteration of magnetite
mixture of magnetic minerals, the low-coercivity frac- and acquisition of a chemical remanent magnetization
tion is demagnetized between 550 and 600 jC (indi- by hematite. Therefore, this HT component was not
cating magnetite), while the high-coercivity fraction is considered to have tectonic implications.
240 F. Speranza et al. / Tectonophysics 347 (2002) 231–251
 F. Speranza et al. / Tectonophysics 347 (2002) 231–251 241

Fig. 6. Vector diagrams of typical thermal demagnetization data, in geographic coordinates. Open and solid symbols represent projections on the
vertical and horizontal planes, respectively. All the samples show a viscous component parallel to the present-day field unblocked at 250 – 310
jC. (a) Sample showing a single magnetization component (after the removal of the viscous component), (b) sample from the yellow marls
showing LT and HT components, (c) sample showing a linear decay not anchored to the origin in the interval 250 – 500 jC, (d) sample showing
a well defined ChRM below 600 jC and scattered directions above 600 jC.

In about 30% of the samples, the ChRM is well a characteristic component carrying 85– 90% of rema-
defined between 250 and 500 jC (Fig. 6c), but the best- nence is well resolvable between 250 and 600 jC, while
fit line does not pass through the origin of the orthog- above 600 jC, scattered directions of magnetization
onal diagram, implying that a small (and not resolvable) (carried by hematite) are observed (Fig. 6d). Finally,
HT component above 500 jC is also present. In a about 10% of the samples (mainly from the upper
few samples containing both magnetite and hematite, clays) yielded random changes of the paleomagnetic

Fig. 5. Results of rock magnetic measurements for some representative samples. Acquisition of normalized IRM up to 0.9 T (a), back field
demagnetization curve (b), and thermal demagnetization of a three-component IRM according to Lowrie (1990) (c – f ). (c) Sample containing
prevailing magnetite and an intermediate-coercivity phase with maximum unblocking temperature of ca. 350 – 400 jC, (d) sample from the
yellow marls containing prevailing magnetite, hematite and an intermediate-coercivity phase with maximum unblocking temperature of ca.
250 – 300 jC, (e) sample containing prevailing magnetite, goethite and an intermediate-coercivity phase with maximum unblocking temperature
of ca. 350 – 400 jC, (f ) sample containing magnetite, hematite and a low- to intermediate-coercivity phase with maximum unblocking
temperature of ca. 200 – 250 jC. See text for discussion.
242 F. Speranza et al. / Tectonophysics 347 (2002) 231–251

direction during heating and no ChRM could be calcu- ent detrital minerals (magnetite, maghemite?, titano-
lated. Overall, a well-defined ChRM could be isolated magnetite?).
for 73 out of 84 samples (Fig. 3).
All the identified remanence components have a 7.3. Paleomagnetic directions
normal polarity. This observation, coupled with the
recognition of a complex mineral magnetic mixture The 73 samples revealing reliable paleomagnetic
suggests that the present-day normal component results show normal polarity ChRMs (Fig. 7). Marly
removed at low-temperature may be due to recent samples below LV2 show systematic northwestward
remagnetization, whereas we interpret the ChRM declinations (Fig. 3), whereas the samples from the
component as a primary remanence carried by differ- upper clays display on average northward declinations.

Fig. 7. Equal-area projection (lower hemisphere) of the ChRMs from the studied section. The open circles and ellipses are the projections of the
mean paleomagnetic directions and the relative a95 cones, respectively. The GAD field direction for the locality (Lat. 39.6jN, Long. 8.8jE) is
also represented (stars). Lower section samples are from the marls below LV2, upper section samples are from the upper marls and clays (see
also Fig. 3).
 F. Speranza et al. / Tectonophysics 347 (2002) 231–251 243

The marls located above LV2 show scattered declina- (2000). Age spectra are internally discordant, which
tions rapidly passing upward from northwestward to means that no ‘‘plateau’’ is formed; however, progress
northward (on average) directions. in understanding Ar isotope systematics now allows to
When analyzed separately, the ChRMs from the obtain robust and reliable ages even in the absence of a
marls below LV2 and the upper marls/clays define ‘‘plateau’’ (e.g. Villa et al., 2000). The Ca/K vs. Cl/K
paleomagnetic mean directions, which are significant- chemical correlation diagram (for a discussion, see
ly different (Fig. 7). After tilt correction, they are de- Belluso et al., 2000) shows that the separates consist
fined by N = 24, D = 340.6j, I = 39.3j, k = 58.6, a95 = of several (at least three) Ar reservoirs, i.e. intergrown
3.9j for the lower stratigraphic interval (0– 310 cm), minerals. One low-temperature step showing a high
and N = 40, D = 359.6j, I = 46.7j, k = 57.0, a95 = 3.0j Cl/K ratio is likely to reflect an alteration product (clays
for the upper stratigraphic interval (465 –955 cm, or fluid inclusions), as feldspars are known not to
Fig. 7). Given the slight difference in bedding attitude incorporate Cl in their structure (Smith, 1974). In all
(11j) observed through the section, the fold test ap- other steps, the Cl/K ratio does not correlate in an
plied to the two intervals (according to McFadden, obvious way with age.
1990) is not significant: nin situ = 0.148; nunfolded = The Ca/K ratio also correlates very weakly with
0.266; n95% = 7.357. the apparent age. This is unusual and might point to the
There are several bits of evidence suggesting that fact that the various minerals in the separates are
the isolated ChRM is a primary (post) depositional intergrown at the 200-nm scale so that recoil of reac-
remanence: (1) The in situ paleomagnetic directions tor-produced Ar causes blurring of the isotope ratios.
from the whole section are far from the local geo- Isochrons were calculated for all samples. Invaria-
centric axial dipole (GAD) field direction (Fig. 7). (2) bly, they showed excess scatter (high values of the
The tilt-corrected mean inclinations from the lower MSWD parameter: Ludwig, 1999), which can be
(I = 39.3j) and upper section (I = 46.7j) are both attributed to the presence of several mineral subsystems
significantly shallower than both the local GAD field with different initial 40Ar/36Ar ratios. What is more
inclination (I = 58.8j) and the lower – middle Miocene important, however, is that all calculated intercepts of
inclination (53.9j) predicted for Europe (Besse and all reasonable subset fits are atmospheric within 1
Courtillot, in press), and similar paleomagnetic flat- standard deviation. This means that there is no evi-
tening in sediments are commonly due to compaction- dence for an unusual trapped Ar component. In turn,
related effects. In fact, AMS data have confirmed the
presence of an almost purely sedimentary fabric. (3)
The large scatter in inclination values (Fig. 3) suggests
a prolonged remanence acquisition (i.e. during sed-
imentation) rather than a single remagnetization epi-
sode that should produce less scattered data (see May
et al., 1986). Scatter is not due from inherent noise
due to weak remanence since the linear segments of
the demagnetization diagrams are well defined (Fig.
6), and inclination scatter is similar in both weakly
and strongly magnetized samples (Fig. 3). (4) The
presence of a multicomponent ChRM (Figs. 5 and 6)
instead of a single overprint-related component.

8. Ar/Ar dating of the volcanoclastic levels

Two feldspar separates (one from each volcanoclas-

tic level) were analyzed by 39Ar – 40Ar stepwise heat- Fig. 8. 40Ar/39Ar age spectra of two feldspar separates from the
ing. Experimental procedures followed Belluso et al. volcanoclastic level LV1 and LV2, respectively.
244 F. Speranza et al. / Tectonophysics 347 (2002) 231–251

this gives confidence that estimates based on step 0.03 Ma near the base of C6Cn.3n. Therefore, we
ages (calculated assuming an atmospheric trapped Ar, assumed a calibrated age of 24.250 Ma for the base
and shown in Fig. 8) are fairly reliable. For both of C6Cn.3n and performed a simple linear interpolation
samples, we calculate the weighted average of those between the two new tie points. We found that, within
steps most likely to correspond to pristine feldspar, the present inherent uncertainties, the samples LV1 and
i.e. those with lowest Cl/K ratios: for LV1, the six LV2 can be assumed to fall both into Chron C5En (with
steps having Cl/K < 0.0009 and a uniform Ca/K be- a revised age of 19.095 – 19.526 Ma, following calcu-
tween 5 and 7 account for 73% of the total Ar and lations described above).
give t = 18.94 F 0.13 Ma; for LV2, five steps having
Cl/K < 0.0006 and Ca/K < 1 account for 85% of the 9.2. Age of the studied section and of remanence
total Ar and give t = 19.20 F 0.12 Ma. Assuming that acquisition
the two levels are contemporaneous, their common
age is calculated as 19.08 F 0.09 Ma. The studied strata lack significant fossils, although
an early Burdigalian age can be tentatively inferred.
The only stringent clues for the section age are given by
9. Discussion Ar/Ar dates, yielding 18.94 F 0.13 and 19.20 F 0.12
Ma for the lower (LV1) and upper (LV2) volcanoclastic
9.1. Correlation with the global magnetic polarity level, respectively. These Ar/Ar dates indicate an age
time scale > 19.1 Ma for the lower part of the section. The two
radiometric ages suggest a high sedimentation rate for
The discussion of our results should make clear the lower part of the section (at least), which is con-
that there are two levels of interpretation. The first sistent with observation of a single normal magnetic
level is the novel, and consistent, young age of CCW polarity in the sequence. Conversely, no age clues are
rotated strata. The finer (‘‘second-order’’) details may available for the strata located above the volcanic
be debatable, but do not question the first-order levels.
interpretation. Our paleomagnetic data show solely normal polar-
A considerable difficulty when attempting to relate ities along all the f 10 m of studied section, and a
the isotopic ages to the magnetic data is the accuracy, rapid shift from about 20j to 0j paleodeclination
for the time interval of interest, of the existing magnetic values in the strata above the volcanics, suggestive to
polarity time scale (MPTS) (Cande and Kent, 1995). As represent the CCW rotation of the Corsica –Sardinia
the latter makes use of very distant tie points, it is clear block. However, according to our recalculation of
that the choice of an inaccurate tie point affects a large chron ages from the Cande and Kent (1995) scale, a
number of chrons. In the case of the Oligo – Miocene, normal to reverse transition (C5En –C5Dr) occurred at
uncertainties are particularly large because both tie 19.095 Ma. Therefore, by assuming that the sedimen-
points of relevance to our samples listed in Cande tary sequence is complete and all the sediments were
and Kent (1995) MPTS (C5Bn(0.0) at 14.8 Ma, and deposited during one normal chron, the CCW rotation
C6Cn.2r(0.0) at 23.8 Ma) are incorrect (Baksi, 1993; would have taken place in < 0.1 Ma, before the
McIntosh et al., 1992; Wei, 1995). According to Cande transition to the reverse chron. Clearly, this is geo-
and Kent (1995) MPTS, revised by Huestis and Acton logically unreasonable when considering that a large
(1997), our two volcanoclastic levels should fall within crustal block (both Corsica and Sardinia) rotated.
Chrons C5Er – C6n. We consider here two important We conclude that the sedimentary sequence is not
updates for radiometric ages of MPTS tie points: the complete although the stratigraphic evidence is very
base of C5Bn.2n is now dated at 16.4 F 0.1 Ma (Baksi, ambiguous and a depositional hiatus in the marls
1993, recalculated for the FCT monitor age of 28.02 between the rotated and unrotated strata is not visible
Ma (Renne et al., 1998) used in our calculations) and a (Fig. 3). However, this stratigraphic interval has
new integrated age model constrains Chron C6Cn.3n at reduced susceptibility and remanence, thus reduced
about 24.2 –24.3 Ma (Naish et al., 2001). This latter detrital input (Fig. 3) and is the most likely candidate
study provides a precise 40Ar/39Ar dating of 24.22 F to have a lower sedimentation rate and to host sedi-
 F. Speranza et al. / Tectonophysics 347 (2002) 231–251 245

mentary hiatuses. Another hiatus may also occur A considerable uncertainty still exists on the total
further upward, along the disconformity separating amount of the Corsica –Sardinia rotation, as well as on
the marls from the clays. the timing of this event. The old data from Sardinian
The unique normal polarity along the section could volcanics (Montigny et al., 1981) seemed to suggest a
also suggest a widespread remagnetization, although 30j rotation occurring between 20.5 and 19 Ma, but
not supported by the previously described magnetic there is growing evidence that the magnitude of total
characters. However, in order to explain by remagne- rotation may be significantly larger, and that the end
tization the in situ and tilt-corrected paleomagnetic of the rotation may be 3 –4 Ma younger (Todesco and
directions, as well as the CCW rotation in the lower Vigliotti, 1993; Vigliotti and Langenheim, 1995;
section (Figs. 3 and 7), a complicated process is Muttoni et al., 1998; Gattacceca, 2001; Deino et al.,
needed: a first remagnetization, followed by CCW 2001). Our data indicate that Sardinia rotated by 23j
rotation, a further remagnetization of the upper section after 19 Ma, but cannot help to constrain the amount
only, and final strata tilting. We reject this hypothesis and timing of the older rotation. By considering the
as highly unlikely. late Burdigalian age of the breakup unconformity
In conclusion, the Ar/Ar ages of the two volcano- observed in Sardinia (Sowerbutts and Underhill,
clastic levels LV1 and LV2 (18.94 F 0.13 and 19.20 F 1998), we argue that the pre-19 Ma rotation occurred
0.12 Ma, respectively) are the only age constraints that during rifting, while the post-19 Ma rotation is likely
we can retain from our study, and the age of the upper due to drifting (Fig. 9).
part of the section remains a matter of speculation. It Although our data show that the 23j rotation
may be argued that the upper clays are upper Burdiga- occurred after 19 Ma and seems to be recorded in less
lian – Langhian, as they seal the small syndepositional than 1 m of the stratigraphic sequence, the timing of the
extensional faults cutting the marls and the transition rotation itself is not clear. However, the oldest possible
from faulted to undeformed sediments in southern age for the end of the rotation (i.e. the minimum time
Sardinia is known to be late Burdigalian in age span required to complete the rotation) can be esti-
(Sowerbutts and Underhill, 1998), but this conclusion mated considering typical spreading rates in similar
remains weak. back-arc settings. Normally, back-arc basins spread at
average speeds of 2– 6 cm/year (e.g. Taylor, 1995),
9.3. Age of the Corsica –Sardinia rotation and Lig- although in the Lau Basin (above the Earth’s most
uro – Provencßal Sea spreading active zone of mantle seismicity) spreading rates of
6.5– 10 cm/year occurred during the Brunhes Chron
When compared to the paleodeclination expected (Taylor et al., 1996) and up to 16 cm/year are occurring
for Europe for the lower Burdigalian (at 20 Ma, Besse nowadays (Bevis et al., 1995). In the southern Tyr-
and Courtillot, in press), the samples below LV2 rhenian Sea, an average spreading rate of 6 cm/year and
define a 23.3 F 4.6j CCW tectonic rotation. Given a Messinian apex reaching 8 cm/year were inferred
the inferred sedimentary hiatuses, we cannot constrain (Patacca et al., 1990). Given these literature values, we
the rotation timing, but at least we show that it believe that it is unlikely that parts of the Liguro –
occurred after 19 Ma (lower –mid Burdigalian), as Provencßal Basin spread faster than 8 cm/year. By
proved by the Ar/Ar ages. Strike-slip activity is assuming a rotation pole north of Corsica (43.5jN,
recorded on NE – SW faults cutting all of northern – 9.0jE, Réhault et al., 1984) and the f 570 km N – S
central Sardinia, but their activity is pre-Burdigalian length of Corsica – Sardinia block undergoing rigid
(Carmignani et al., 1995) and predates the deposition drifting, a f 230 km eastward drift of the southern
of the sediments studied by us. Therefore, tectonic Sardinian edge during the 23j rotation can be calcu-
activity producing locally CCW rotations is not docu- lated. This drift was likely accompanied by the em-
mented in the regional geology, and the paleomag- placement of oceanic crust in the Liguro – Provencßal
netic results can be extrapolated to the whole Basin. A maximum drift rate of 8 cm/year for the
Corsica– Sardinia block (no differential rotation bet- southern Sardinian margin implies that the 23j rotation
ween Corsica and Sardinia was documented by Vig- was not completed in less than f 3 million years, i.e. at
liotti et al., 1990). about 16 Ma (early Langhian). This is only a lower
 246
F. Speranza et al. / Tectonophysics 347 (2002) 231–251
Fig. 9. Inferred tectonic evolution of the central – western Mediterranean from 20 to 7 Ma. The 10j and 23j CCW rotation of the Corsica – Sardinia block occurring during the 20 – 19
and 19 – 16 Ma time intervals, respectively, are with respect to Europe held fixed. GK and PK are the Grande Kabylie and Petite Kabylie blocks, respectively. Legend: (1) newly
formed oceanic crust; (2) areas affected by rifting; (3) emplacement of arc-related volcanics (or equivalent intrusive rocks) (e.g. Beccaluva et al., 1985; Serri et al., 1993; Lonergan
and White, 1997 and references therein); (4) Alpine accretionary wedge; (5) synorogenic ‘‘intra-wedge’’ extensional basins due to local collapse of the overthickened accretionary
wedge; (6) continental Adria – Africa lithosphere; (7) oceanic lithosphere of the ‘‘paleo-Ionian’’ Sea; (8) pole for the Corsica – Sardinia rotation (Lat. 43.5jN, Long. 9jE) in present-
day coordinates. The f 20 Ma Corsica – Sardinia position is inferred from the rotational timing proposed by Gattacceca (2001).
 F. Speranza et al. / Tectonophysics 347 (2002) 231–251 247

bound for the termination of the rotation, as obviously easily in the mantle causing a speedy trench retreat
smaller drift rates would make it even younger. This (and Liguro –Provencßal Basin spreading). Faster sub-
would imply that the main hiatus(es) should be in the duction beneath Sardinia than beneath Corsica, due to
80 –100 cm that record the gradual change from rotated the heterogeneous nature of the subducting plate, is a
to not-rotated declinations. As a matter of fact, this marl plausible reason to explain the triangular geometry of
interval has the lowest susceptibility and remanence the Liguro – Provencßal Basin and the CCW rotation of
values (Fig. 3), suggesting a reduced detrital input and Corsica – Sardinia.
sedimentation rate. At about 16 Ma (according to our model), the
Other independent evidence concurs for a f 16– Liguro – Provencßal basin stopped widening (and Cor-
19 Ma age for the Liguro –Provencßal oceanic spread- sica –Sardinia drifting), likely due to the locking of
ing: (1) the upper Burdigalian age of the breakup the subduction process as continental lithosphere
unconformity observed in Sardinia (Sowerbutts and continued colliding east of Corsica (Fig. 9). Roll-
Underhill, 1998), (2) the 16 –17 Ma age of the igneous back continued later only along two isolated slab
rocks dredged west of Corsica (Deverchère, 1996; fragments, observed today below the northern and
Chamot-Rooke et al., 1999), (3) the high heat flow southern Tyrrhenian Sea by seismic tomography
(>120 mW/m2) in the axial part of the basin, which is (Lucente et al., 1999). From middle – late Miocene
the theoretical value for a young oceanic crust (15 Ma onward, only the ‘‘paleo-Ionian’’ oceanic corridor
according to Chamot-Rooke et al., 1999), and (4) underwent rapid passive subduction, generating the
paleomagnetic data from three upper Burdigalian – southern Tyrrhenian Sea (partly floored by oceanic
Langhian (18 – 15 Ma) sites from Sardinia (Vigliotti crust) and the fast Calabrian drift (Malinverno and
and Langenheim, 1995), still showing a small (5– 10j) Ryan, 1986).
CCW rotation, and recent Ar/Ar dated paleomagnetic Our paleogeographic reconstruction of the nature
directions from a Sardinian volcanic sequence, docu- of the lithosphere subducting in central – western
menting a 13j rotation occurring after 18.18 F 0.03 Mediterranean during Tertiary and Quaternary times
Ma (Deino et al., 2001). is supported by data from Serri et al. (1993), who
analyzed the characters of the Neogene –Quaternary
9.4. From the Liguro –Provencßal to the Tyrrhenian magmatism of the Tyrrhenian Sea and margins. Serri
back-arc spreading et al. (1993) argued on petrological – geochemical
grounds that the ‘‘paleo-Ionian’’ and Adria subduction
The rejuvenation of the Liguro – Provencßal Basin were oceanic and continental, respectively, and that
spreading and Corsica – Sardinia rotation has signifi- east of Corsica the transition from oceanic to con-
cant implications for the timing of the tectonic events tinental subduction must have occurred in the late
in the adjacent Tyrrhenian – Apennine system. It is Eocene –early Miocene time interval.
believed that the Liguro – Provencßal and Tyrrhenian
seas are both back-arc basins opening in subsequent 9.5. Tectonic significance of the lower Miocene north-
times west of the same Adriatic/Ionian slab in pro- ern Tyrrhenian deposits
gressive eastward roll-back (Malinverno and Ryan,
1986). However, the question arises as to why the Our data give a key to interpret the oldest exten-
locus of back-arc spreading migrated eastward during sional phenomena observed in the northern Tyrrhenian
middle – late Miocene, as extension ceased in the domain. Widespread back-arc extension started in the
Liguro –Provenc ß al domain and jumped east of Cor- whole Tyrrhenian domain only since late Tortonian
sica – Sardinia, where the Tyrrhenian Sea started to ( f 8 Ma), as testified by the onset of syn-rift con-
spread. tinental sedimentation in several basins from Tuscany
We propose that the back-arc spreading migration (Martini and Sagri, 1993), transgressive continental
was due to the heterogeneity of the Adriatic/Ionian deposits observed over the western Calabrian margin
lithosphere undergoing passive subduction. We esti- (Di Nocera et al., 1979) and alluvial fan sequences
mate that at 16– 19 Ma, the lithosphere of a ‘‘paleo- transgressive over the continental basement drilled
Ionian’’ oceanic corridor east of Sardinia (Fig. 9) sunk east of Sardinia by the ODP Site 654 (Mascle et al.,
248 F. Speranza et al. / Tectonophysics 347 (2002) 231–251

1988). Given the clear onset of widespread back-arc Acknowledgements

extension in the Tyrrhenian realm since late Tortonian,
the question arises on the geologic significance of the We thank C. Faccenna for scientific advice and
lower Miocene extensional deposits locally observed help in the field during the 1995 sampling campaign,
in the northern Tyrrhenian Sea and margins that sig- and N. D’Agostino for providing magnificent ima-
nificantly predate the general late Tortonian extension ges of the Liguro – Provencßal Basin topography. A.F.
(Jolivet et al., 1990; Carmignani et al., 1995; Mauffret Glazner, M. Cosca, G. Ruffet and R. Van der Voo
et al., 1999; Cornamusini et al., 2000; Foresi et al., provided thoughtful comments on a preliminary ver-
2000). sion of the manuscript. We are also grateful to the
We note that our paleogeographic model suggests referees E. Platzman and L. Vigliotti, and the Editor
that the Corsica – Sardinia block did not stop rotating J.-P. Burg for carefully reviewing our paper. This
and colliding with the Adriatic lithosphere before work was partly founded by the European Union
f 16 Ma or possibly more recently, i.e. well after project ‘‘GeoModAp’’ (contract EV5V-CT94-0464).
the onset of the lower Miocene sedimentation in the
northern Tyrrhenian Sea (Fig. 9). Clearly, the 19 –16
Ma 23j drift-related CCW rotation of Corsica– Sardi-
nia induced strong shortening in the Apennine belt, References
overthickening the orogenic wedge and inducing
tectonic collapses (e.g. Platt, 1986, and references Assorgia, A., Barca, S., Casula, G., Spano, C., 1988. Le successioni
therein). Therefore, the lower Miocene deposits can sedimentarie e vulcaniche del Miocene nei dintorni di Giave e
Cossoine (Logudoro, Sardegna NW). Boll. Soc. Sarda Sci. Nat.
be interpreted as related to extension developing 26, 75 – 107.
inside the Apennines wedge (as also proposed by Assorgia, A., Barca, S., Rizzo, R., Spano, C., 1992a. The beginning
Jolivet et al., 1998), and not as precursors of the of the Lower Miocene marine sedimentation in western Sardinia
northern Tyrrhenian back-arc sediments (as suggested (Italy) and its implications with the Oligocene – Miocene rift
by Bartole, 1995). system. IXth Congress R.C.M.N.S. ‘‘Global Events and Neo-
gene Evolution of the Mediterranean’’ (Barcelona, 1990). Pale-
ontol. Evol. 24 – 25, 295 – 305.
Assorgia, A., Barca, S., Spano, C., 1992b. Upper Oligocene – Lower
Miocene sequences of the Arbus – Funtanazza Coast (South-
10. Conclusions Western Sardinia, Italy), IGCP N. 276, Newsletter, 5, 21 – 31.
Assorgia, A., Barca, S., Spano, C., 1997. A synthesis on the Cen-
ozoic stratigraphic, tectonic and volcanic evolution in Sardinia
New paleomagnetic and Ar/Ar data from Sardinia (Italy). Boll. Soc. Geol. Ital. 116, 407 – 420.
show that a 23j CCW rotation of Corsica– Sardinia and Baksi, A., 1993. A geomagnetic polarity time scale for the period
contemporaneous Liguro – Provencß al Basin spreading 0 – 17 Ma, based on 40Ar/39Ar plateau ages for selected field
occurred after 19 Ma. By considering normal extension reversals. Geophys. Res. Lett. 20, 1607 – 1610.
rates in other back-arc basins, we infer that the Liguro – Barca, S., Carmignani, L., Oggiano, G., Pertusati, P.C., Salvadori, I.,
1996. Geologic Map of Sardinia, Scale 1:200,000, 2 sheets,
Provencß al Basin kept spreading (and the Corsica – Servizio Geologico Nazionale-Regione Autonoma della Sar-
Sardinia block rotating) for at least 3 million years. degna.
Therefore, the rotation is estimated to end not before Bartole, R., 1995. The north Tyrrhenian – northern Apennines post-
f 16 Ma (early Langhian). A 16 –19 Ma age of the collisional system: constraints for a geodynamic model. Terra
Liguro – Provencal Basin, oceanic spreading is consis- Nova 7, 7 – 30.
Bayer, R., Le Mouel, J.-L., Le Pichon, X., 1973. Magnetic anomaly
tent with other independent data, such as (1) age of the pattern in the western Mediterranean. Earth Planet. Sci. Lett. 49,
break-up unconformity observed in Sardinia (Sower- 168 – 176.
butts and Underhill, 1998), (2) age of igneous rocks Beccaluva, L., Civetta, L., Maciotta, G., Ricci, C.A., 1985. Geo-
dredged west of Corsica (Deverchère, 1996), (3) ele- chronology in Sardinia: results and problems. Rend. Soc. Ital.
vated heat-flow in the Liguro – Provencßal Basin (Cha- Mineral. Petrol. 40, 57 – 72.
Belluso, E., Ruffini, R., Schaller, M., Villa, I.M., 2000. Electron-
mot-Rooke et al., 1999), and (4) recent paleomagnetic microscope and Ar isotope characterization of chemically het-
data from Sardinian sediments and volcanics (Vigliotti erogeneous amphiboles from the Palala Shear Zone, Limpopo
and Langenheim, 1995; Deino et al., 2001). Belt, South Africa. Eur. J. Mineral. 12, 45 – 62.
 F. Speranza et al. / Tectonophysics 347 (2002) 231–251 249

Besse, J., Courtillot, V., in press. Apparent and true polar wander Margins, Abstract of the meeting ‘‘The Mediterranean Basins’’,
and the geometry of the geomagnetic field in the last 200 million Cergy – Pontoise (Paris, France), December 11 – 13.
years, J. Geophys. Res. Di Nocera, S., Ortolani, F., Torre, M., Russo, B., 1979. Caratteris-
Bevis, M., Taylor, F.W., Schultz, B.E., et al., 1995. Geodetic ob- tiche stratigrafiche e paleoambientali dei depositi altomiocenici
servations of very rapid convergence and back-arc extension at nella zona di Falconara Albanese (Catena Costiera Calabra).
the Tonga arc. Nature 374, 249 – 251. Boll. Soc. Nat. Napoli 83, 1 – 29.
Blow, W.H., 1967. Late Middle Eocene to Recent planktonic fora- Edel, J.B., 1979. Paleomagnetic study of the Tertiary volcanics of
miniferal biostratigraphy. Proc. 1st. Intern. Conf. Plankt. Micro- Sardinia. J. Geophys. 45, 259 – 280.
fossils, Geneve, 1 – 422. Edel, J.B., Lortscher, A., 1977. Paléomagnétisme du volcanisme
Burrus, J., 1984. Contribution to a geodynamic synthesis of the tertiaire de Sardaigne. Nouveaux résultats et synthèse. Bull.
Provencßal Basin (North-Western Mediterranean). Mar. Geol. Soc. Geol. Fr. 19, 815 – 824.
55, 247 – 269. Farris, M., 1990. Studio geo-vulcanologico e stratigrafico del settore
Burrus, J., Foucher, J.P., 1986. Contribution to the thermal regime di Monastir – Ussana – Nuraminis, Univ. Cagliari, unpublished
of the Provencßal Basin based on flumed heat flow surveys and thesis, 1 – 217.
previous investigations. Tectonophysics 128, 303 – 334. Foresi, L.M., Cornamusini, G., Bossio, A., Ferrandini, M., Mazzei,
Cande, S.C., Kent, D.V., 1995. Revised calibration of the geomag- R., Salvatorini, G., Argenti, P., 2000. The Miocene sedimentary
netic polarity time scale for the Late Cretaceous and Cenozoic. succession of the Pianosa island, northern Tyrrhenian Sea paper
J. Geophys. Res. 100, 6093 – 6095. presented at the meeting. Evoluzione geologica e geodinamica
Carmignani, L., Decandia, F.A., Disperati, L., Fantozzi, P.L., Laz- dell’Appennino, Foligno (Italy) February 16 – 18, 155 – 157 (ab-
zarotto, A., Liotta, D., Oggiano, G., 1995. Relationships between stract).
the Tertiary structural evolution of the Corsica – Sardinia – Fornaciari, E., Rio, D., 1996. Latest Oligocene to early middle
Provencß al Domain and the northern Apennines. Terra Nova 7, Miocene quantitative calcareous nannofossil biostratigraphy in
128 – 137. the Mediterranean region. Micropaleontology 42, 1 – 36.
Chamot-Rooke, N., Gaulier, J.-M., Jestin, F., 1999. Constraints on Gattacceca, J., 2001. Cinématique du bassin liguro – provencal entre
Moho depth and crustal thickness in the Liguro – Provencßal basin 30 et 12 Ma. Implications géodynamiques, PhD thesis, Ecole
from a 3D gravity inversion: geodynamic implications. In: Du- des mines de Paris, 299 pp.
rand, B., Jolivet, L., Horvath, F., Seranne, M. (Eds.), The Medi- Gialanella, P.R., Heller, F., Incoronato, A., 1994. Rock magnetism
terranean Basins: Tertiary Extension Within the Alpine Orogen. of deformed upper Triassic limestones from the Lagonegro Ba-
Geological Society of London, Special Publications, vol. 156, sin (southern Apennines Italy). Geophys. Res. Lett. 21, 2665 –
pp. 37 – 61. 2668.
Cherchi, A., Corradini, D., D’Onofrio, S., Iaccarino, S., Martini, E., Horner, F., Lowrie, W., 1981. Paleomagnetic evidence from Meso-
Murru, M., Russo, A., 1985. Sardara – Villanovaforru section. zoic carbonate rocks for the rotation of Sardinia. J. Geophys. 49,
In: Cherchi, A. (Ed.), Micropaleontological Researches in Sar- 11 – 19.
dinia, 19th European Micropaleontological Colloquium Guide- Huestis, S.P., Acton, G.D., 1997. On the construction of geomag-
book, Sardinia October, vol. 1 – 10, pp. 234 – 249. netic timescales from non-prejudicial treatment of magnetic
Cornamusini, G., Lazzarotto, A., Merlini, S., Pascucci, V., 2000. anomaly data from multiple ridges. Geophys. J. Int. 129, 176 –
Eocene – Quaternary evolution of the Corsica basin-Tuscan shelf 182.
area (north Tyrrhenian Sea) through CROP-MARE seismic Iaccarino, S., D’Onofrio, S., Murru, M., 1985. Miocene foramini-
analysis paper presented at the meeting. Evoluzione geologica fera of several sections of the Marmilla area (central western
e geodinamica dell’Appennino, Foligno (Italy) February 16 – 18, Sardinia). Boll. Soc. Paleontol. Ital. 23, 395 – 412.
108 – 110 (abstract). Jelinek, V., 1978. Statistical processing of anisotropy of magnetic
Coulon, C., Demant, A., Bobier, C., 1974. Contribution du paléo- susceptibility measured on a group of specimens and its appli-
magnétisme a l’étude des séries volcaniques cénozoiques et cations. Stud. Geophys. Geod. 22, 50 – 62.
quaternaires de Sardaigne nord-occidentale. Tectonophysics Jolivet, L., Dubois, R., Fournier, M., Goffé, B., Michard, A., Jour-
22, 59 – 82. dan, C., 1990. Ductile extension in Alpine Corsica. Geology 18,
Deino, A., Gattacceca, J., Rizzo, R., Montanari, A., 2001. 40Ar/39Ar 1007 – 1010.
dating and paleomagnetism of the Miocene volcanic succession Jolivet, L., Faccenna, C., Goffé, B., Mattei, M., Rossetti, F., Brunet,
of Monte Furru (western Sardinia): implications for the rotation C., Storti, F., Funiciello, R., Cadet, J.-P., D’Agostino, N., Parra,
history of the Corsica – Sardinia microplate. Geophys. Res. Lett. T., 1998. Midcrustal shear zones in postorogenic extension:
28 (17), 3373 – 3376. example from the northern Tyrrhenian Sea. J. Geophys. Res.
De Jong, K.A., Manzoni, M., Zijderveld, J.D.A., 1969. Paleomag- 103, 12123 – 12160.
netism of the Alghero Trachyandesites. Nature 224, 67 – 69. Kastens, K., Mascle, J., Auroux, C., et al., 1986. A microcosm of
De Jong, K.A., Manzoni, M., Stavenga, T., Van Dijk, F., Van der ocean basin evolution in the Mediterranean. Nature 321, 383 –
Voo, R., Zijderveld, J.D.A., 1973. Palaeomagnetic evidence for 384.
rotation of Sardinia during the early Miocene. Nature 243, 281 – Kirschvink, J.L., 1980. The least-square line and plane and the
283. analysis of paleomagnetic data. Geophys. J. R. Astron. Soc.
Deverchère, J., et al., 1996. Geodynamics of the Ligurian Basin and 62, 699 – 718.
250 F. Speranza et al. / Tectonophysics 347 (2002) 231–251

Lehman, B., Sagnotti, L., Winkler, A., Lo Cascio, C., 1996. Mag- Dunn, A.G., Fielding, C.R., Florindo, F., Hannah, M.J., Har-
netic mineralogy changes in the Pleistocene marine sequence of wood, D.M., Watkins, D.K., Henrys, S.A., Krissek, L.A., Lav-
Montalto di Castro (central Italy) and influences on the magnetic elle, M., van der Meer, J., McIntosh, W.C., Niessen, F.,
anisotropy. Geophys. J. Int. 127, 529 – 541. Passchier, S., Powell, R.D., Roberts, A.P., Sagnotti, L., Scherer,
Lonergan, L., White, N., 1997. Origin of the Betic-Rif mountain R.P., Strong, C.P., Talarico, F., Verosub, K.L., Villa, G., Wonik,
belt. Tectonics 16 (3), 504 – 522. T., 2001. Orbitally induced oscillations in the East Antarctic Ice
Lowrie, W., 1990. Identification of ferromagnetic minerals in a rock Sheet at the Oligocene – Miocene boundary. Nature 413, 719 –
by coercivity and unblocking temperature properties. Geophys. 723.
Res. Lett. 17, 159 – 162. Patacca, E., Sartori, R., Scandone, P., 1990. Tyrrhenian basin and
Lucente, F.P., Chiarabba, C., Cimini, G.B., Giardini, D., 1999. To- Apenninic arcs: kinematic relations since late Tortonian times.
mographic constraints on the geodynamic evolution of the Ital- Mem. Soc. Geol. Ital. 45, 425 – 451.
ian region. J. Geophys. Res. 104, 20307 – 20327. Pecorini, G., Pomesano Cherchi, A., 1969. Ricerche geologiche e
Ludwig, K.R., 1999. User’s Manual for Isoplot/Ex Version 2.2, A biostratigrafiche sul Campidano meridionale (Sardegna). Mem.
Geochronological Toolkit for Microsoft Excel: Berkeley Geo- Soc. Geol. Ital. 8, 421 – 451.
chronology Center Special Publication No. 1a, 53 pp. Platt, J.P., 1986. Dynamic of orogenic wedges and the uplift of high-
Maillard, A., Mauffret, A., Watts, A.B., Torné, M., Pascal, G., Buhl, pressure metamorphic rocks. Geol. Soc. Am. Bull. 97, 1037 –
P., Pinet, B., 1992. Tertiary sedimentary history and structure of 1053.
the Valencia trough (western Mediterranean). Tectonophysics Réhault, J.-P., Boillot, G., Mauffret, A., 1984. The western Medi-
203, 57 – 75. terranean basin geological evolution. Mar. Geol. 55, 447 – 477.
Malinverno, A., Ryan, W.B.F., 1986. Extension in the Tyrrhenian Renne, P.R., Swisher, C.C., Deino, A.L., Karner, D.B., Owens, T.L.,
Sea and shortening in the Apennines as result of arc migration DePaolo, D.J., 1998. Intercalibration of standards, absolute ages
driven by sinking of the lithosphere. Tectonics 5, 227 – 245. and uncertainties in 40Ar/39Ar dating. Chem. Geol. 145, 117 –
Manzoni, M., 1974. Un’interpretazione dei dati paleomagnetici del 152.
terziario della Sardegna ed alcuni nuovi esempi. Rend. Semin. Sagnotti, L., Speranza, F., Winkler, A., Mattei, M., Funiciello, R.,
Fac. Sci. Univ. Cagliari 43, 163 – 167. 1998. Magnetic fabric of clay sediments from the external north-
Martini, I.P., Sagri, M., 1993. Tectono-sedimentary characteristics ern Apennines (Italy). Phys. Earth Planet. Inter. 105, 73 – 93.
of Late Miocene – Quaternary extensional basins of the northern Sartori, R., 1989. Evoluzione neogenico – recente del bacino tirre-
Apennines, Italy. Earth Sci. Rev. 34, 197 – 233. nico e suoi rapporti con la geologia delle aree circostanti. G.
Mascle, J., Kastens, K., Auroux, C., et al., 1988. A land-locked Geol. 3 (51/2), 1 – 39.
back-arc basin: preliminary results from ODP Leg 107 in the Serri, G., Innocenti, F., Manetti, P., 1993. Geochemical and petro-
Tyrrhenian Sea. Tectonophysics 146, 149 – 162. logical evidence of the subduction of delaminated Adriatic con-
Mauffret, A., Maillard, A., Pascal, G., Torné, M., Buhl, P., Pinet, B., tinental lithosphere in the genesis of the Neogene – Quaternary
1992. Long-listening multichannel seismic profiles in the Valen- magmatism of central Italy. Tectonophysics 223, 117 – 147.
cia trough (Valsis 2) and the Gulf of Lions (ECORS): a compar- Smith, J.V., 1974. Feldspar Minerals, vol. 2. Springer, Berlin, 690
ison. Tectonophysics 203, 285 – 304. pp.
Mauffret, A., Contrucci, I., Brunet, C., 1999. Structural evolution of Sowerbutts, A.A., Underhill, J.R., 1998. Sedimentary response to
the Northern Tyrrhenian Sea from new seismic data. Mar. Pet. intra-arc extension: controls on Oligo – Miocene deposition, Sar-
Geol. 16, 381 – 407. cidano sub-basin, Sardinia. J. Geol. Soc. (London) 155, 491 –
May, S.R., Butler, R.F., Shafiqullah, M., Damon, P.E., 1986. Pale- 508.
omagnetism of Jurassic rocks in the Patagonia Mountains, Taylor, B. (Ed.), 1995. Backarc Basins, Tectonics and Magmatism.
southeastern Arizona: implications for the North American Plenum, New York, 524 pp.
170 Ma reference pole. J. Geophys. Res. 91, 11545 – 11555. Taylor, B., Zellmer, K., Martinez, F., Goodliffe, A., 1996. Sea-floor
McFadden, P.L., 1990. A new fold test for paleomagnetic studies. spreading in the Lau back-arc basin. Earth Planet. Sci. Lett. 144,
Geophys. J. Int. 103, 163 – 169. 35 – 40.
McIntosh, W.C., Geismann, J.W., Chapin, C.E., Kunk, M.J., Henry, Todesco, M., Vigliotti, L., 1993. When did Sardinia rotate? Statis-
C.D., 1992. Calibration of the latest Eocene – Oligocene geo- tical evaluation of the paleomagnetic data. Ann. Geofis. 36,
magnetic polarity time scale using 40Ar/39Ar dated ignimbrites. 119 – 134.
Geology 20, 459 – 463. Van der Voo, R., 1993. Paleomagnetism of the Atlantic, Thethys,
Montigny, R., Edel, J.B., Thuizat, R., 1981. Oligo – Miocene rota- and Iapetus Oceans. Cambridge Univ. Press, Cambridge, 411
tion of Sardinia: K – Ar ages and paleomagnetic data of Tertiary pp.
volcanics. Earth Planet. Sci. Lett. 54, 261 – 271. Vigliotti, L., Langenheim, V.E., 1995. When did Sardinia stop rotat-
Muttoni, G., Argnani, A., Kent, D.V., Abrahamsen, N., Cibin, U., ing? New paleomagnetic results. Terra Nova 7, 424 – 435.
1998. Paleomagnetic evidence for Neogene tectonic rotations in Vigliotti, L., Alvarez, W., Mc Williams, M., 1990. No relative mo-
the northern Apennines, Italy. Earth Planet. Sci. Lett. 154, 25 – tion detected between Corsica and Sardinia. Earth Planet. Sci.
40. Lett. 98, 313 – 318.
Naish, T.R., Woolfe, K.J., Barrett, P.J., Wilson, G.S., Atkins, C., Villa, I.M., Hermann, J., Müntener, O., Trommsdorff, V., 2000.
39
Bohaty, S.M., Bücker, C.J., Claps, M., Davey, F.J., Dunbar, G., Ar – 40Ar dating of multiply zoned amphibole generations
 F. Speranza et al. / Tectonophysics 347 (2002) 231–251 251

(Malenco, Italian Alps). Contrib. Mineral. Petrol. 140, 363 – so-sarde, sa position initiale: données paléomagnétiques et rac-
381. cords géologiques. Tectonophysics 30, 141 – 157.
Wei, W., 1995. Revised age calibration points for the geomagnetic Zijderveld, J.D.A., De Jong, K.A., Van der Voo, R., 1970. Rotation
polarity time scale. Geophys. Res. Lett. 22 (8), 957 – 960. of Sardinia: palaeomagnetic evidence from Permian rocks. Na-
Westphal, M., Orsini, J., Vellutini, P., 1976. Le microcontinent cor- ture 226, 933 – 934.

View publication stats