Chronology of the transition from a spreading ridge to anaccretional seamount in the Marsili backarc basin (Tyrrhenian Sea)

L. Cocchi,1 F. Caratori Tontini,1 F. Muccini,1,2 M. P. Marani,3 G. Bortoluzzi3 and C. Carmisciano1

1Istituto Nazionale di Geofisica e Vulcanologia, Via Pezzino Basso 2, 19025, Fezzano, Italy; 2Universita di Bologna, Dipartimento di Scienze

della Terra e Geologico – Ambientali, Piazza Porta S. Donato 1, 40125 Bologna, Italy; 3ISMAR-CNR, Sezione di Geologia Marina, Via

Gobetti 101, 40129 Bologna, Italy

Introduction

The Tyrrhenian Sea (Italy) is aTortonian–Pleistocene oceanic back-arc basin related to the flexural retreatof the Apennine–Maghrebid subduc-tion system (Royden, 1988; Pataccaet al., 1990; Doglioni, 1991).Extensional tectonics affected the

entire western Mediterranean Seastarting in Oligocene time, stretchingcontinental crust and forming basins inthe Valencia and Alghero-Provencalzone and then in the Tyrrhenian Sea(Malinverno andRyan, 1986;Gueguenet al., 1998; Rosenbaum and Lister,2004). Crustal extension because of theroll-back of the NW-dipping Ionianslab caused ocean spreading in theMarsili basin at 2 Ma, as establishedby results of ODP leg 107-site 650(Kastens et al., 1988, 1990). The back-arc spreading features of the Marsilibasin are similar to structures else-where. The East Scotia Ridge in theSouth Atlantic Ocean, for example, ischaracterized by an average E–W fullspreading rate of 6–7 cm yr)1 since15 Ma (Larter et al., 2003; Thomaset al., 2003). On the other hand, theMarsili seamount could be considereda morphological anomaly of a slow

spreading ridge similar to the SpiessRidge, a swollen segment of the SouthWest Indian Ridge generated by amagmatic pulse related to the Bouvethot-spot (Ligi et al., 1999).Kastens et al. (1988, 1990) and

Marani and Trua (2002) describedthe Marsili basin as floored by oceaniccrust. Nicolosi et al. (2006) proposedan ultra fast spreading rate (full19 cm yr)1) for the opening of thebasin during the Olduvai subchron(C2n1, 1.77–1.95 Ma, Cande andKent, 1995), followed by a periodof quiescence. This interpretation as-sumes a rapid increase followed imme-diately by a rapid decrease in thespreading rate over a very short time-interval.The magnetic anomaly map pro-

posed by Nicolosi et al. (2006) differsfrom our representation. The maincauses are the low resolution of thedata used by Nicolosi et al. (2006) andthe directional filtering procedureused to emphasize the signal of thepresumed oceanic stripes, filtering outthe seamount anomaly. Conversely,our model used high-resolution unfil-tered data (see line-path in Fig. 1)indicating a continuity of the spread-ing during the late Matuyama (C1r1r,1.07–1.77 Ma) and Brunhes chron(C1n, 0–0.78 Ma).We propose and discuss the time-

space evolution of the Marsili systemduring the last 1.77 Myr aimed atclarifying its transition from spreadingridge to a huge volcanic seamount.

Geological framework of theMarsili basin

The Marsili basin has a roughly ellip-tic shape with a crustal thickness of7–8 km (Sartori et al., 2004) and a30-km thick lithosphere (Scarasciaet al., 1994). In addition, the model-ling of satellite gravity data indicatesthat the Marsili basin crust is oceanicin nature (Caratori Tontini et al.,2008). The basin is bounded by theAeolian Arc to the south, the PalinuroVolcanic Complex to the north andthe Issel Bridge, a thicker crustalsaddle, to the northwest (Fig. 1).The flat 3500 m deep seafloor of the

central part of the Marsili basin isinterrupted by the Marsili seamountwhich rises to a minimum depth of�500 m. The volcanic edifice is 60-kmlong and 20-km wide, elongatedmainly NNE–SSW and characterizedby steep flanks (�20�) (Marani andTrua, 2002). At depths shallower than1000 m, a 25-km long narrow belt ofvolcanic cones represents the summitrift zone of the volcano. Two sets ofhorst and graben structures generatedby NNE–SSW faults develop symmet-rically in the basin floor bordering themajor axis of the seamount.Lava compositions range from bas-

alts with medium-to low-K affinityfound on the volcano flanks, to high-Kandesites that crop out in the summitzone (Trua et al., 2002). The lavashave mostly Island Arc Basalt (IAB)affinities, similar to those of the

ABSTRACT

Inversion of new high-resolution magnetic data from the Marsiliseamount and the surrounding basin in the Tyrrhenian Seareveals NNE–SSW magnetization stripes ranging from theMatuyama chron to the Brunhes chron, including the shortpositive Jaramillo subchron. The detailed magnetic chronologyshows that from the late Matuyama (1.77 Ma), the average halfspreading rate was about 1.5 cm yr)1, with a slight decreasebetween the Jaramillo and the Brunhes events, when thegrowth of the volcanic edifice overcame lateral spreading.Analysis of spreading rate and volume of erupted lava indicates

that at the beginning of the Jaramillo subchron (1.07 Ma), theMarsili basin evolved from pure horizontal spreading to asuperinflated seamount as a consequence of tearing of theIonian slab. Our data give us a snapshot of the geodynamictransition from an active backarc spreading phase to thevertical accretion of the seafloor because of a radical change inthe subduction dynamics.

Terra Nova, 21, 369–374, 2009

Correspondence: Cocchi Luca, Istituto

Nazionale di Geofisica e Vulcanologia,

Geophysics and Marine Technology Unit,

Via Pezzino Basso, 2 19025, Fezzano (SP),

Italy. Tel.: +39 0187 794408; fax: +39

0187 766400; e-mail: cocchi@ingv.it

� 2009 Blackwell Publishing Ltd 369

doi: 10.1111/j.1365-3121.2009.00891.x

Aeolian arc immediately to the South(Trua et al., 2003). Sparse chronolog-ical data indicate that the summitandesitic lavas formed 0.1–0.2 Ma(Selli et al., 1977).During the Pleistocene, robust

volcanism affected the southernTyrrhenian Sea and parts of Sicily(Beccaluva et al., 1982; Trua et al.,2003). The vertical accretion of theMarsili seamount was associated witha deep magmatic pulse that occurredin a thermally constricted area throughtears in the Ionian slab (Marani andTrua, 2002). The only basement dataof the Marsili basin region derive fromODP Leg 107 (site 650, see Fig. 1),where a few metres of vesicular basaltswere drilled below a 500 m thicksediment layer and subsequently datedabout 1.8–2 Ma (Kastens et al., 1988).

Magnetic modelling

A new high-resolution ship-bornemagnetic survey of Marsili basin wascarried out in 2006 (Paltrinieri et al.,

2006). Data were collected at 1 Hzusing a Geometrics Cesium magneto-meter along ESE–WNW parallel linesfor a total length of 1400 km (see insetin Fig. 1).The magnetic anomaly map (grid

cell size 500 m, Fig. 2a) was computedby subtracting the IGRF-05 modelfrom the Total Magnetic Field. TheMarsili volcano displays a wide posi-tive anomaly (1500 nT), flanked by aset of negative anomalies symmetri-cally distributed (A and A¢ in Fig. 2a)with respect to the volcano axis.Between the wide positive N–S mag-netic anomaly and two negative lobes,the magnetic gradient shows localvariations, which define two symmet-ric narrow stripes (B and B¢ inFig. 2a). These anomalies, not visiblein previously published data, may beascribed to the Jaramillo subchron(C1r1n, 0.99–1.07 Ma).The magnetic anomaly field was

inverted to map in detail the distribu-tion of magnetization in the Marsilibasin that we assumed being confined

to oceanic layer 2A with top surface atthe seafloor and bottom surface at6 km below the sea-level. We sub-tracted, in the area of the basin, asediment infill of 500 m thickness(Kastens et al., 1988). The magneticdata were inverted sampling every250 m the crustal portion (top andbottom surfaces) following the inverseapproach by Caratori Tontini et al.(2008). The results show a set ofalternating positive and negativestripes symmetrically distributed rela-tive to the volcano axis (Fig. 2b).The number and width of these

reversal stripes can be correlated topolarity inversions from C1r2 sub-chron (1.07–1.77) to Present, includ-ing the short Jaramillo event. Thecircular anomaly on the top of thevolcano (Fig. 2b, latitude 39�15¢N)should not be confused with reversedmagnetic polarity, as it is locatedwithin the youngest portion of theseamount where recent high-K ande-sites crop out and extensive hydro-thermal alteration was reported(Marani et al., 1999). This intriguinganomaly may be caused by hydrother-mal alteration of volcanic lavas, asdescribed for the Mid Atlantic Ridge(Tivey et al., 1993, 1996).

Discussion of results

The magnetic anomalies of the Marsiliregion point to an episode of seafloorspreading between the late Matuyamaand Brunhes chron. A positive, NNE–SSW trending magnetization rangingfrom 0 to 11 A ⁄m is centred onthe Marsili volcano with two max-ima coinciding with the northernand southern flanks of the edifice(Fig. 2b). We interpret the positivemagnetization as correlated to theBrunhes chron, in agreement withthe age of the summit lavas. Thecentral positive magnetization is con-fined to the upper part of the se-amount from )2550 m to )495 m.The lower portions of the edifice arecharacterized by two reversely mag-netized stripes that we interpret as theJaramillo and C1r1r (0.78–0.99 Ma)subchrons (Figs 2b and 3). Theseevents allow us to date the initiationof seamount evolution at 1.07 Ma atthe end of C1r2r subchron, in contrastto the 0.7 Myr age (at the beginning ofthe Brunhes chron) previously pro-posed by Faggioni et al. (1995). The

Fig. 1 Geographical setting of the Marsili basin, in the framework of the SouthernTyrrhenian Sea (illumination from 45�N, data from ETOPO1 (Amante and Eakins,2008, http://www.ngdc.noaa.gov/mgg/global/etopo1sources.html), showing the majorfeatures discussed in the text: Vavilov Basin (VB), Marsili Basin (MB), Issel Bridge(IB), Palinuro Volcanic Complex (PVC), Aeolian Arc (AA), Ustica Island (U), Etnavolcano (E) and Sant�Arcangelo basin (SAB). We report, as inset, the tracks of theacquired magnetic data during the geophysical survey (the yellow box defines the areaof interest) superimposed to the high resolution bathymetry of the Marsili seamount(ISMAR-CNR data, Marani et al., 2004, contour every 100 m).

From spreading ridge to huge seamount in the Marsili Basin • L. Cocchi et al. Terra Nova, Vol 21, No. 5, 369–374

.............................................................................................................................................................

370 � 2009 Blackwell Publishing Ltd

negative magnetization pattern()6A ⁄m, Figs 2b and 3) that charac-terizes the basin area, is correlated tothe reverse C1r2r subchron. The pro-file of Fig. 4a (track no. 7 in Fig. 3)

shows a complete geomagnetic se-quence from the end of the Olduvaisubchron to Present.The Jaramillo and C1r1r subchrons

are symmetrically distributed and

located entirely within the volcanicedifice (Fig. 4a). Given its position atthe border of the seamount, the Jara-millo subchron marks the transitionfrom pure lateral spreading to thevertical growth of the seamount. Thefirst evidence of the super-inflation ofthe seamount occurred just before thebeginning of the Jaramillo subchron,whereas the phase of maximum accre-tion is visible in the Brunhes chronsuggesting a coexisting phase of bothprocesses lasting about 0.3 Ma (1.07–0.7 Ma).Magnetic stripes correlated to the

Jaramillo event have been detectedonly close to the central part of theseamount (Figs 3 and 4a). In thenorthernmost sector of the volcanicedifice, the magnetization distributionshows mostly positive values (Fig. 2band profiles no. 10, 11 and 12 inFig. 3) as correlated to the phase ofgrowth of the volcano. In the southernsector of Marsili, the short Jaramilloevent is not detected and the magne-tization pattern concerns only thesignature related to the C1r2r sub-chron and to the Brunhes chron(Fig. 4b). The geometry of the stripescorrelated to the Jaramillo subchron

Fig. 2 (a) Total intensity magnetic anomaly field, cell size 500 m, contour interval 50 nT. Labels A and A¢ indicate negativemagnetic anomalies related to the basin seafloor; labels B and B¢ are narrow magnetic stripes within the Marsili seamount;(b) apparent magnetization map (500 m grid size, contour interval 1A ⁄m) obtained by the inverse approach.

Fig. 3 Wiggle profiles of the magnetization pattern (A ⁄m) along 12 tracks orthogonalto the volcanic edifice superimposed to the high resolution bathymetry (contour every100 m). The black and red colours represent the positive and negative values ofmagnetization.

Terra Nova, Vol 21, No. 5, 369–374 L. Cocchi et al. • From spreading ridge to huge seamount in the Marsili Basin

.............................................................................................................................................................

� 2009 Blackwell Publishing Ltd 371

may be interpreted as because of theN–S and E–W distribution of themagmatic material produced duringthe seafloor opening. We suggest thatduring the Jaramillo, this process tookplace in a small delimitated areawithin the actual volcanic summitarea, generating the characteristicwidth and length of the magneticstripe strictly correlated to the dura-tion of the subchron similar to thatobserved in mature oceanic spreadingridges (e.g. the Carlsberg ridge,Kamesh Raju et al., 2008).We computed spreading rates by

averaging the results of a set of 12WNW–ESE profiles crossing ortho-gonally the inferred axis of spreading(Fig. 3). Figure 5a shows the age ofreversal boundaries plotted as a func-tion of their distance from the axisfor both sectors using as startingpoints the age of the youngest lavas(0.1 Myr). A constant full spreadingrate of 2.9 cm yr)1 was obtained bylinear fit of Fig. 5a. In detail, the fullspreading rate of Marsili basin was3.1 cm yr)1 from 1.77 to 1.07 Ma(C1r2r), decreasing to 2.4 cm yr)1,2.15 cm yr)1 and then to 1.8 cm yr)1

during the Jaramillo subchron, C1r1rsubchron and Brunhes chron respec-tively (Fig. 5a). This decreasing trend

is accompanied by the superinflationof the Marsili volcano that started atthe Jaramillo event, to the detrimentof the lateral opening.These results mismatch the model

provided by Nicolosi et al. (2006) thatindicates an ultra-fast spreadingrate of 19 cm yr)1 during the Olduvaisubchron. In addition, these authorssuggested a continuity of the ultrafastspreading during the first 0.1–0.2 Myrof the late Matuyama followed by atotal cessation lasting about 1.0 Myrand a reactivity of the spreadingduring the last geomagnetic chron.This interpretation is in contrast withour chronological reconstructionbased on a time-space analysis of themagnetic pattern, which highlightscontinuous seafloor spreading alsoduring the last 1.77 Myr. In thistime-interval, we record a maximumfull spreading rate of 3.1 cm yr)1

which slightly decreases in a uniformway from C1r2r subchron to Brunheschron (see misfit between linear trendand full spreading) without any evi-dence of cessation of the seafloorspreading activity.Extrapolating the linear decreasing

trend (�1.3 · 10)6 cm yr)1) to thebeginning of Olduvai (1.95 Ma), weestimate a full spreading rate of about

3.4 cm yr)1, that is a value very farfrom the 19 cm yr)1 proposed byNicolosi et al. (2006). We speculatethat the ultra-fast spreading phase, if

Fig. 4 Profiles no. 2 (Fig. 4b, right panel) and 7 (Fig. 4a, left panel) across Marsili seamount. See Figure 3 for their location. Forboth profiles, we have reported: upper side – morphological profile showing the position of the crest and the steep flanks; middleside – observed anomaly field (red) sampled from the 2D grid of Fig. 2a; the computed anomaly field (dashed-blue) recovered byusing the magnetization distribution of Fig. 2b; the residual (green) is the difference from observed and computed fields (for theprofile no. 2, mean value )10.95 nT and standard deviation 19.17 nT; in the case of profile no. 7, mean )31.34 nT and standarddeviation 26.12 nT); lower side – 1D apparent magnetization distribution obtained through the inverse approach; below we alsoreport the geomagnetic polarity timescale (age in Ma, from Cande and Kent, 1995) from Olduvai subchron to Brunhes chron.

Fig. 5 (a) Time-space diagram: the redand black lines represent the half spread-ing rate of the eastern and westernsectors respectively (mostly overlap-ping), the full rate points are shown inblue with associated error bars com-puted between the estimated averagevalue and the observed data; the greenline represents the linear fit of the fullspreading rate; (b) volume of lavaerupted vs. time diagram: we have plot-ted the half volume rate lines in red(eastern sector) and black (western sec-tor) colour, the blue line indicates thefull volume rate.

From spreading ridge to huge seamount in the Marsili Basin • L. Cocchi et al. Terra Nova, Vol 21, No. 5, 369–374

.............................................................................................................................................................

372 � 2009 Blackwell Publishing Ltd

it occurred, should be related to anepisodic and instantaneous openingprocess during the Olduvay subchronthat did not leave traces in the lateMatuyama.We also estimated the volume of the

erupted lava based on the crustal con-straints assumed in the inverse mag-neticmodel. The diagramof volume vs.time (Fig. 5b) shows a clear negativetrend. During the late Matuyama andJaramillo, the volumeof accretedmate-rial ranged between 2.1 · 10)3 km3

yr)1 and 3.34 · 10)3 km)3 yr)1, whileduring the C1r1r subchron and Brun-hes chron, this rate increased to3.6 · 10)3 km)3 yr)1 and 4.21 · 10)3

km)3 yr)1, respectively. The data sug-gest that the superinflation of the sea-floor is generated by both the decreasedrate of spreading and an increase inmagma supply (22% during the Jara-millo and 55% during the Brunhes).The magmatic impulse was generatedby slab tearing of the Ionian subduc-tion systemanda subsequent upwellingof asthenospheric mantle around theslab edge beneath the Marsili region(Trua et al., 2003).This interpretation is confirmed by

other observations, such as the cut-offof deep seismicity on the northernsector of Sicily (Chiarabba et al.,2008) and below the Marsili area(Rosenbaum et al., 2008) and thepresence of Ocean Island Basalt(OIB) products with a slight contam-ination from subduction related fluidsof Quaternary age on the Etna andUstica volcanoes (Beccaluva et al.,1982) and the Prometeo Lava field(Trua et al., 2003). Furthermore,palaeomagnetic analysis of the Pleist-ocenic mudstone of Sant�Arcangelobasin (on-shore Calabria, Fig. 1) hasrevealed ultra-fast vertical axis rota-tion during the Jaramillo chron inter-preted as a consequence of slabtearing (Mattei et al., 2004).

Conclusions

Magnetic anomalies define the chro-nological evolution of the Marsilibasin, clarifying the relationshipbetween the spreading episode of thebasin and the building phase ofthe Marsili seamount. The genesis ofthe Marsili volcano is linked to theparticular setting of the Ionian sub-duction system during the Pleistocene,with a combination of slab tearing at

the edges of the Ionian slab andthermal constriction of the areas sur-rounding the Marsili basin.Our magnetic chronology dates the

beginning of the growth of Marsilivolcano in the Jaramillo, thus datingthe slab tearing at 1.07 Ma. Since theMatuyama, the Marsili basin openedwith an average full spreading rate of2.8–3.1 cm yr)1, followed by uniformdeceleration during the Jaramillo sub-chron. This latter marks the beginningof a phase during which slowerspreading (1.8 cm yr)1) coexisted withthe vertical growth of the volcanicedifice. Mantle upwelling along theslab tears induced a strong magmaticpulse during this phase. The slabtearing effects in the Marsili area arecorrelated with several mag-matic ⁄deformational events occurringin different locations (on shore Cala-bria, off-shore Sicily). The commonage of these events (1.07 Ma) could bethe result of geodynamic phenomenathat affected the entire southernTyrrhenian Sea. Magnetic data pro-vide a snapshot of the Marsili basinevolution when seafloor spreading wasreplaced by vertical superinflation ofthe seamount because of increasingmagma supply that began in theJaramillo subchron. The transitionwas fast, lasting about 0.3 Ma, andwas related to a general reorganiza-tion of the Tyrrhenian subductionsystem involving a deep mutation ofthe subducting slab.

Acknowledgements

We thank the Captain, officers and crew ofN ⁄R Universitatis of Conisma for theirassistance during the Prometheus cruise.We are grateful to Prama S.r.l for thefinancial support. We especially thankEnrico Bonatti, Richard J. Blakely, Ralphvon Frese and Laurent Jolivet fortheir notes and stimulating comments. Weacknowledge two referees for their scien-tific comments. Figures were prepared withGMT (Wessel and Smith, 1998).

References

Amante, C. and Eakins, B.W., 2008.ETOPO1 1 Arc-Minute Global ReliefModel: Procedures, Data Sources andAnalysis. National Geophysical DataCenter, NESDIS, NOAA, U.S. Depart-ment of Commerce, Boulder, CO.

Beccaluva, L., Rossi, P.L. and Serri, G.,1982. Neogene to Recent volcanism ofthe Southern Tyrrhenian–Sicilian area:

implications for the geodynamic evolu-tion of the Calabrian arc. Earth Evol.Sci., 3, 222–238.

Cande, S.C. and Kent, D.V., 1995. Revisedcalibration of the geomagnetic polaritytime scale for the Late Cretaceous andCenozoic. J. Geophys. Res., 100, 6093–6095.

Caratori Tontini, F., Cocchi, L. andCarmisciano, C., 2008. Potential fieldinversion for a layer with uneven thick-ness: The Tyrrhenian Sea density model.Phys. Earth Planet. Inter., 166, 105–111.

Chiarabba, C., De Gori, P. and Speranza,F., 2008. The southern Tyrrheniansubduction zone: Deep geometry,magmatism and Plio-Pleistoceneevolution. Earth Planet. Sci. Lett., 268,408–423.

Doglioni, C., 1991. A proposal of kine-matic modelling for W-dipping subduc-tions - Possible applications to theTyrrhenian - Apennines system. TerraNova., 3, 423–434.

Faggioni, O., Pinna, E., Savelli, C. andSchreider, A.A., 1995. Geomagnetismand age study of Tyrrhenian seamounts.Geophys. J. Int., 123, 915–930.

Gueguen, E., Doglioni, C. and Fernandez,M., 1998. On the post-25 Ma geody-namic evolution of the western Medi-terranean. Tectonophysics., 298, 259–269.

Kamesh Raju, K.A., Chaubey, A.K.,Amarnath, D. and Mudholkar, A., 2008.Morphotectonics of the Carlsberg Ridgebetween 62�20¢ and 66�20¢E, NorthwestIndian Ocean. Mar. Geol., 252, 120–128.

Kastens, K.A., Mascle, J., Auroux, C.A.,Bonatti, E., Broglia, C., Channell, J.,Curzi, P., Emeis, K.C., Glacon, G., Ha-segawa, S., Hieke, W., Mascle, G.,McCoy, F., McKenzie, J., Mendelson,J., Mueller, C., Rehault, J., Robertson,A., Sartori, R., Sprovieri, R. and Torii,M., 1988. ODP Leg 107 in theTyrrhenian Sea: Insights into passivemargins and back-arc basinevolution. Geol. Soc. Am. Bull., 100,1140–1156.

Kastens, K.A., Mascle, J., Auroux, C.A.,Bonatti, E., Broglia, C., Channell, J.,Curzi, P., Emeis, K.C., Glacon, G.,Hasegawa, S., Hieke, W., Mascle, G.,McCoy, F., McKenzie, J., Mendelson,J., Mueller, C., Rehault, J., Robertson,A., Sartori, R., Sprovieri, R. and Torii,M., 1990. Proceedings of the Ocean Dril-ling Program, Scientific Results Volume107. College Station, Texas, 722pp.

Larter, R.D., Vanneste, L.E., Morris, P.and Smythe, D.K., 2003. Structure andtectonic evolution of the South Sand-wich arc. In: Intra-Oceanic SubductionSystems: Tectonic and Magmatic Pro-cesses (R.D. Larter and P.T. Leat, eds).Geol. Soc. Lond. Spec. Publ., 219,255–284.

Terra Nova, Vol 21, No. 5, 369–374 L. Cocchi et al. • From spreading ridge to huge seamount in the Marsili Basin

.............................................................................................................................................................

� 2009 Blackwell Publishing Ltd 373

Ligi, M., Bonatti, E., Bortoluzzi, G.,Carrara, G., Fabretti, P., Gilod, D.,Peyes, A.A., Skolotnev, S. and Turko,N., 1999. Bouvet triple junction inthe South Atlantic: Geology andevolution. J. Geophys. Res., 102,29365–29385.

Malinverno, A. and Ryan, W.B.F., 1986.Extension in the Tyrrhenian Sea andshortening in the Apennines as a resultof arc migration driven by sinking oflithosphere. Tectonics, 5, 227–245.

Marani, M.P. and Trua, T., 2002. Thermalconstriction and slab tearing at the ori-gin of a superinflated spreading ridge:Marsili volcano (Tyrrhenian Sea).J. Geophys. Res., 107 (B9), 2188.doi:10.1029/2001JB000285.

Marani, M.P., Gamberi, F., Binns, R.A.and Parr, J.M., 1999. Results from re-cent seabed sampling of hydrothermaldeposits in the 0073outh Tyrrhenian Sea.In: Mineral Deposits: Processes to Pro-cessing (C.J. Stanley et al., eds), pp. 547–550. A.A. Balkema Publishers, Rotter-dam.

Marani, M.P., Gamberi, F. and Bonatti,E., 2004. From seafloor to deep mantle:architecture of the Tyrrhenian backarcbasin. Mem. Descr. Carta Geol. d�It., 44,195.

Mattei, M., Petrocelli, V., Lacava, D. andSciattarella, M., 2004. Geodynamicimplications of Pleistocene ultra-rapidvertical-axis rotations in the SouthernApennines, Italy. Geology, 32, 789–792.

Nicolosi, I., Speranza, F. and Chiappini,M., 2006. Ultrafast oceanic spreading ofthe Marsili Basin, southern TyrrhenianSea: Evidence from magnetic anomalyanalysis. Geology, 34, 717–720.

Paltrinieri, D., Viezzoli, C., Signanini, P.,Di Sabatino, B., D�Anna, G., Calcara,M., Bortoluzzi, G., Cocchi, L. andCaratori Tontini, F., 2006. Report onthe morphobathymetric, magnetometric,gravimetric, CTD, water and bottomsampling investigations during cruiseMRS06 aboard R ⁄v Universitatis,ISMAR-CNR Interim Report, Bologna.

Patacca, E., Sartori, R. and Scandone, P.,1990. Tyrrhenian Basin and Apenninicarcs: kinematic relations since LateTortonian times. Memorie della SocietaGeologica Italiana, 46, 425–451.

Rosenbaum, G. and Lister, G.S., 2004.Neogene and Quaternary rollbackevolution of the Tyrrhenian Sea, theApennines, and the Sicilian Maghre-bides. Tectonics, 23, TC1013.doi:10.1029/2003TC001518.

Rosenbaum, G., Gasparon, M., Lucente,F.P., Peccerillo, A. and Miller, M.S.,2008. Kinematics of slab tear faultsduring subduction segmentation andimplication for Italian magmatism.Tectonics, 27, TC2008. doi:10.1029/2007TC002143.

Royden, L., 1988. Flexural behavior of theContinental Lithosphere in ItalyConstraints imposed by gravity anddeflection data. J. Geophys. Res., 93,7747–7766.

Sartori, R., Torelli, L., Zitellini, N.,Carrara, G., Magaldi, M. and Mussoni,P., 2004. Crustal features along a W–ETyrrhenian transect from Sardinia toCampanian margins (Central Mediter-ranean). Tectonophysics, 383, 171–192.

Scarascia, A., Lozej, A. and Cassinis, R.,1994. Crustal structures of the LigurianTyrrhenian and Ionian seas and adjacent

onshore areas interpreted from wide-angle seismic profile. Bollettino diGeofisica Teorica e Applicata, 36, 4–19.

Selli, R., Lucchini, F., Rossi, P., Savelli, C.and Del Monte, M., 1977. Dati geo-logici, petrochimici, e radiometrici suivulcani centro-tirrenici. Giorn. Geol., 2,221–246.

Thomas, C., Livermore, R. and Pollitz, E.,2003. Motion of the Scotia Sea plates.Geophysical Journal International, 155,789–804.

Tivey, M.A., Rona, P.A. and Schouten, H.,1993. Reduced crustal magnetizationbeneath the active sulfide mound, TAGhydrothermal field, Mid-Atlantic Ridge.Earth Planet. Sci. Lett., 115, 101–115.

Tivey, M., Rona, P.A. and Kleinrock,M.C., 1996. Reduced crustal magnetiza-tion beneath hydrothermal mounds:TAG hydrothermal field, Mid-AtlanticRidge, 26�N. Geophys. Res. Lett., 23,3511–3514.

Trua, T., Serri, G., Marani, M., Renzulli,A. and Gamberi, F., 2002. Volcanolo-gical and petrological evolution ofMarsili Seamount (southern TyrrhenianSea). J. Volcanol. Geoth. Res., 114,441–464.

Trua, T., Serri, G. and Marani, M.P., 2003.Lateral Flow of African mantle belowthe nearby Tyrrhenian Plate: geochemi-cal evidence. Terra Nova, 15, 433–440.

Wessel, P. and Smith, W.H.F., 1998. New,improved version of the generic mappingtools released: EOS. Trans. Am. Geo-phys. Union, 79, 579.

Received 17 March 2009; revised versionaccepted 5 June 2009

From spreading ridge to huge seamount in the Marsili Basin • L. Cocchi et al. Terra Nova, Vol 21, No. 5, 369–374

.............................................................................................................................................................

374 � 2009 Blackwell Publishing Ltd