Formación y estructura de la cuenca del Tirreno en el contexto de retrarco del Mediterraneo Occidental

Resumen   Abstract   Índice   Conclusiones

Prada Dacasa, Manuel

2015-A
Descargar PDF  

 
Resumen

In this thesis I present a geophysical study that aims at defining the structure and petrological nature of the main geological domains in the Central Tyrrhenian basin, and to investigate the mechanisms involved in their formation.

The Tyrrhenian basin is a Neogene back-arc basin located in the Central Mediterranean region. Its formation is fundamentally related to the E-SE slab-rollback and retreat of the Apennines-Calabrian subduction system during the last ~11-12 Ma (Serravallian-Tortonian). The consequent back-arc extension led to the formation of triangular-shaped basin with a peculiar crustal structure where:

i) Extension increases from North to South, so rifting evolution can be inferred from along-axis crustal structure variations 

And ii) the original deformation is perfectly preserved and clearly observed in the bathymetry due to the lack of sedimentation

All these features turn this back-arc basin into an ideal natural laboratory to study lithospheric extension processes, and therefore, the aim of many geophysical studies since the late 70s [Trincardi and Zitellini, 1987]. However, despite of the many geophysical experiments conducted in the area (mostly seismic reflection lines), there is no reliable information concerning the petrological nature of the main geological domains in the area, and hence, formation processes cannot be properly constrained. 

As mentioned above, in this thesis I try to shade some light to this issue by presenting modern geophysical data acquired in the Tyrrhenian during the MEDOC (2010) survey within the framework of the MEDOC project, which was designed to improve our understanding of the origin and evolution of rifted margins. 

The method applied in the present work focuses on the analysis, processing, modeling, and interpretation of coincident Wide-Angle Seismic (WAS), Multichannel Seismic (MCS), and gravity data corresponding to the two longest transects acquired during the MEDOC survey in the Central Tyrrhenian basin, the southern Line GH/MEDOC-6 (~450 km) (40ºN) and the northern Line EF/MEDOC-4 (~400 km) (40.5ºN). Both lines run across the Central Tyrrhenian basin from Sardinia in the East to the Campania margin in the West.

The processing of MCS data provides the tectonic structure and geometry of the sedimentary basins, whereas the modeling of WAS data from travel-time tomography provides 2D seismic velocity models from which the velocity distribution of the crust and uppermost mantle, and the geometry of the crust-mantle boundary are inferred. The WAS models are then converted to density models using existing empirical relationships for different lithologies in order to test which of the different hypothesis concerning the petrological nature of the basement (e.g. continental/oceanic crust or exhumed mantle) explain better the observed gravity anomaly. The results obtained from the MCS, WAS and gravity data of each line, have been complemented with a thorough analysis of the velocities of the WAS-derived models, which consisted of comparing the depth-velocity structure of the models with existing 1D velocity-depth references of continental and oceanic crust, and exhumed mantle [White et al., 1992; Christensen and Mooney, 1995; Dean et al., 2001; Sallarès et al., 2013a]. All these results, together with the integration of geological data from rock sampling of the seabed [Colantoni et al., 1981; Kastens and Mascle, 1990] reveals the existence of three geological domains in the Central Tyrrhenian, that is: continental crust, magmatic crust, and exhumed mantle.

Along both lines, continental crust is observed beneath the island of Sardinia, and beneath the Sardinia and Campania margins. In Sardinia the crust is ~25 km thick, whereas in the margins the crust thins toward the center of the basin from 25-22 km to 10-13 km thick, implying a stretching factor ß > 2. Only in the northern Line EF/MEDOC-4, continental crust is observed in between both continental margins, in the so-called Sechi segment, being ~8 km thick (ß > 3).

The magmatic crust is observed beneath the Cornaglia Terrace and the new-defined Campania Terrace in both lines. This type of crust is characterize by a seismic velocity structure that closely resembles to an oceanic crust (e.g. Atlantic oceanic crust) [White et al., 1990], but with lower velocity gradient in the L3 region. Based on observations made in other back-arc basins of the world, such as in the western Pacific [e.g. Martinez et al., 2007], it is suggested in this thesis that this type of crust could be formed by back-arc rifting and/or back-arc spreading processes near the volcanic arc, so that its formation would be influenced by two types of magmas: 1) those produced by pressure-release melting or related with the passive decompression of the underlying mantle, and 2) those produced by hydrous flux melting or derived from the subducted slab.

The exhumed mantle domain is observed beneath the V-shaped Magnaghi and Vavilov basins. The basement beneath these basins is characterized by the lack of Moho reflections in both WAS and MCS data and by a vertical velocity structure similar to other regions where basement is made of serpentinized mantle [Dean et al., 200; Sallarès et al., 2013a]. Additionally, this domain hosts large basaltic structures (i.e. Magnaghi and Vavilov seamounts, and D’Ancona and Gortani Ridges) that are intersected by Line GH/MEDOC-6, and hence imaged in depth by the velocity model.

The comparison between the results of Line EF/MEDOC-4 (northern line) with those of the Line GH/MEDOC-6 (southern line) reveals that the velocity and tectonic structure of the three geological domains differ in some regions from north to south. These differences are most likely attributed to the southward increase of extension that characterizes the Tyrrhenian basin [e.g. Faccena et al., 2001; Sartori et al., 2004].

Finally, to explain the mechanism involved in the formation of these domains, I examine the modes of back-arc basin formation proposed to explain the formation of the western Pacific basins [Martinez et al., 2007; Dunn and Martinez, 2011], as well as the causes that may have led to mantle exhumation [Pérez-Gussinyé, et al., 2006]. In summary, the proposed conceptual model is based on a slab rollback and depleted mantle setting, in which production of extension-related melting is limited, thus, crustal accretion is attributed to hydrous flux melting. The model presents 5 stages of opening that includes: (I) a normal subduction scenario followed by (II) development of back-arc rift, (III) initiation of back-arc spreading, (IV) mantle exhumation, and finally (V) emplacement of large volcanic edifices in the central parts of the basin.

In summary, the basement configuration inferred from modern geophysical data in this thesis led to a completely new definition of geological domains in the Central Tyrrhenian. According to the presented distribution of the basement, rifting in the Central Tyrrhenian basin would have started with continental crust extension, continued with back-arc spreading leading to generation of magmatic back-arc crust, and followed by mantle exhumation intruded by later magmatic episodes. The interpretation of these results differ from current conceptual models of the formation of rifting systems involving mantle exhumation and indicate that the response of the continental lithosphere to extension processes may be more complex than previously assumed.

 

Reference in the summary:

Christensen, N., and W. Mooney (1995), Seismic velocity structure and composition of the continental crust: a global view, J. Geophys. Res, 100 (B7), doi:10.1029/95JB00259.

Colantoni P., A. Fabbri, P. Gallignani, R. Sartori, and J.P. Rehault (1981) Carta Litologica e Stratigrafica dei Mari Italiani, scala 1/1.500.000, Litografia Artistica Cartografica, Firenze, Italy.

Dean, S.M., T.A. Minshull, R.B. Whitmarsh, and K.E. Louden (2000), Deep structure of the ocean-continent transition in the southern Iberia abyssal plain from seismic refraction profiles: the IAM-9 transect at 40º20’N, J. Geophys. Res. B: Solid Earth, 105 (B3), 5859-5885.

Dunn, R.A., and F. Martinez (2011), Contrasting crustal production and rapid mantle transitions beneath back-arc ridges, Nature, 469, 198-202, doi: 10.1038/nature09690.

Faccena, C., T.W. Becker, F. P. Lucente, L. Jolivet, and F. Rossetti (2001), Hystory of subduction and back-arc extension in the Central Mediterranean, Gophys. J. Int., 145, 809-820.

Kastens, K., and J. Mascle (1990), The geological evolution of the Tyrrhenian Sea: an introduction to the scientific results of ODP LEG 107, In Kastens, K.A., Mascle, J., et al. (Eds.), Proceedings of the Ocean Drilling Program, Scientific Results 107, pp. 3-26, doi:10.2973/odp.proc.sr.107.187.1990

Martinez, F., K. Okino, Y. Ohara, A.-L. Reysenbach, and S.K. Goffredi (2007), Back-arc basins, Oceanography, 20 (No.1), 116-127.Pérez-Gussinyé, et al., 2006

Sallarès, V., S. Martinez-Loriente, M. Prada, E. Gràcia, C. Ranero, M.-A. Gutscher, R. Bartolomé, A. Gailler, J.J. Dañobeitia, and N. Zitellini (2013a), Seismic evidence of exhumed mantle rock basement at the Gorringe Bank and the adjacent Horseshoe and Tagus abyssal plains (SW Iberia), Earth Planet. Sci. Let., 365, 120-131, doi: 10.1016/j.epsl.2013.01.021.

Sartori, R., L.Torelli, N. Zitellini, G. Carrara, M. Matteo, P. Mussoni (2004), Crustal features along a W–E Tyrrhenian transect from Sardinia to Campania margins (Central Mediterranean), Tectonophysics, 383, 3–4, 171-192, doi: 10.1016/j.tecto.2004.02.008.

White, R.S., D. McKenzie, and R.K. O’Nions (1992), Oceanic crustal thickness from seismic measurements and rare earth element inversions. J. Geophys. Res. 97, 19683–19715.  


 
Abstract

In this thesis I present a geophysical study that aims to define the structure and petrological nature of the main geological domains in the Central Tyrrhenian basin, and to investigate the mechanisms involved in their formation.

The geophysical data used in this thesis was acquired during the MEDOC (2010) survey within the framework of the MEDOC project, which was designed to improve our understanding of the origin and evolution of rifted margins. The MEDOC survey was focused on the Tyrrhenian back-arc basin, an extraordinary natural laboratory to investigate the structure and evolution of back-arc basins, and hence rifted margins, since 1) extension evolves from north to south [Kastens and Mascle, 1990], 2) the original structure is preserved, and 3) it is small enough to be investigated with a single marine survey.

The present work is based on the analysis, processing, modeling, and interpretation of coincident Wide-Angle Seismic (WAS), Multichannel Seismic (MCS), and gravity data corresponding to the two longest transects acquired during the MEDOC survey in the Central Tyrrhenian basin, the southern Line GH/MEDOC-6 (~450 km) and the northern Line EF/MEDOC-4 (~400 km). Both lines run across the Central Tyrrhenian basin from Sardinia to the Campania margin.

The processing of MCS data provides the tectonic structure and geometry of the sedimentary basins, whereas the modeling of WAS data from travel-time tomography provides 2D seismic velocity models from which the velocity distribution of the crust and uppermost mantle, and the geometry of the crust-mantle boundary are inferred. The WAS models are then converted to density models using existing empirical relationships for different lithologies in order to test which of the different hypothesis concerning the petrological nature of the basement (e.g. continental/oceanic crust or exhumed mantle) explain better the observed gravity data. The results obtained from the MCS, WAS and gravity data of each line, have been complemented with a thorough analysis of the velocities of the WAS-derived models, which consisted of comparing the depth-velocity structure of the models with existing 1D velocity-depth references of continental and oceanic crust, and exhumed mantle [White et al., 1992; Christensen and Mooney, 1995; Dean et al., 2001; Sallarès et al., 2013a]. All these results, together with the integration of geological data from rock sampling of the seabed [Colantoni et al., ii 1981; Kastens and Mascle, 1990] reveals the existence of three geological domains in the Central Tyrrhenian, that is: continental crust, magmatic crust, and exhumed mantle.

Along both lines, continental crust is observed beneath the island of Sardinia, and beneath the Sardinia and Campania margins. In Sardinia the crust is ~25 km thick, whereas in the margins the crust thins toward the center of the basin from 25-22 km to 10-13 km thick, implying a stretching factor ß > 2. Only in the northern Line EF/MEDOC-4, continental crust is observed in between both continental margins, in the so-called Sechi segment, being ~8 km thick (ß > 3).

The magmatic crust is observed beneath the Cornaglia Terrace and the newdefined Campania Terrace in both lines. This type of crust is characterize by a seismic velocity structure that closely resembles to an oceanic crust (e.g. Atlantic oceanic crust) [White et al., 1990], but with lower velocity gradient in the L3 region. Based on observations made in other back-arc basins of the world, such as in the western Pacific [e.g. Martinez et al., 2007], it is suggested in this thesis that this type of crust could be formed by back-arc rifting and/or back-arc spreading processes near the volcanic arc, so that its formation would be influenced by two types of magmas: 1) those produced by pressure-release melting or related with the passive decompression of the underlying mantle, and 2) those produced by hydrous flux melting or derived from the subducted slab. The exhumed mantle domain is observed beneath the V-shaped Magnaghi and Vavilov basins. The basement beneath these basins is characterized by the lack of Moho reflections in both WAS and MCS data and by a vertical velocity structure similar to other regions where basement is made of serpentinized mantle [Dean et al., 200; Sallarès et al., 2013a]. Additionally, this domain hosts large basaltic structures (i.e. Magnaghi and Vavilov seamounts, and D’Ancona and Gortani Ridges) that are intersected by Line GH/MEDOC-6, and hence imaged in depth by the velocity model.

The comparison between the results of Line EF/MEDOC-4 (northern line) with those of the Line GH/MEDOC-6 (southern line) reveals that the velocity and tectonic structure of the three geological domains differ in some regions from north to south. These differences are most likely attributed to the southward increase of extension that characterizes the Tyrrhenian basin [e.g. Faccena et al., 2001; Sartori et al., 2004].

Finally, to explain the mechanism involved in the formation of these domains, I examine the modes of back-arc basin formation proposed to explain the formation of the western Pacific basins [Martinez et al., 2007; Dunn and Martinez, 2011], as well as the causes that may have led to mantle exhumation [Pérez-Gussinyé, et al., 2006]. In summary, the proposed conceptual model is based on a slab rollback and depleted mantle setting, in which production of extension-related melting is limited, thus, crustal accretion is attributed to hydrous flux melting. The model presents 5 stages of opening that includes: (I) a normal subduction scenario followed by (II) development of backarc rift, (III) initiation of back-arc spreading, (IV) mantle exhumation, and finally (V) emplacement of large volcanic edifices in the central parts of the basin. In summary, the basement configuration presented in this thesis led to a completely new definition of geological domains in the Central Tyrrhenian. According to the presented distribution of the basement, rifting in the Central Tyrrhenian basin would have started with continental crust extension, continued with back-arc spreading leading to generation of magmatic back-arc crust, and followed by mantle exhumation intruded by later magmatic episodes. The interpretation of these results differ from current conceptual models of the formation of rifting systems involving mantle exhumation and indicate that the response of the continental lithosphere to extension processes may be more complex than previously assumed.

 

 
Índice

Part I: Introduction 1

Chapter 1: General Introduction 3

1.1 Background and motivations         3

1.2 Objectives 7

1.3 Basic concepts 9

1.3.1 Plate tectonics 9

1.3.1.1 Extension and continental rifting at divergent boundaries                11

1.3.1.2 Subduction at convergent boundaries         14

1.3.1.2.1 Slab rollback         16

1.3.1.2.2 Back-arc basins        17

1.3.2 Nature of the basement        21

1.3.2.1 Continental crust        22

1.3.2.2 Oceanic crust        27

1.3.2.3 Exhumed mantle         30

Chapter 2: Geological & Geodynamics setting         35

2.1 The Western Mediterranean back-arc basins         35

2.2 The Tyrrhenian back-arc basin        38

2.2.1 Geodynamic evolution of the Tyrrhenian basin region         39

2.2.2 Tectonic setting of the Tyrrhenian basin         40

2.2.3 Previous work on the rock-type distribution of the Tyrrhenian basin        46

2.2.4 Neogene-Quaternaty Volcanism         54

Part II: Methodology       59

Chapter 3: Acquisition & Modeling of Wide-Angle Seismics         61

3.1 Introduction        61

3.2 Acquisition system         61

3.3 Wide-Angle seismic data processing and seismic phase identification         65

3.4 Wide-Angle seismic modeling         67

3.4.1 Layer-stripping strategy         72

3.4.2 Uncertainty analysis        73

Chapter 4: Gravity        77

4.1 Introduction        77

4.2 Data acquisition and corrections        77

4.3 Empirical relationships between compressional-wave 

velocity (Vp) and density (?)         79

4.4 Gravity modeling        82

Chapter 5: Acquisition, Analysis & Processing of Multichannel Seismics         85

5.1 Introduction        85

5.2 Acquisition system and geometry of the experiment         85

5.3 Data processing        89

Part III: Results & Discussion     105

Chapter 6: Line GH/MEDOC-6       107

6.1 Data analysis       107

6.1.1 Wide-Angle Seismic Data       108

6.2 Results       109

6.2.1 Velocity structure       109

6.2.1.1 Uncertainty analysis       112

6.2.2 Tectonic structure       113

6.2.3 Automatized interpretation of basement affinity based on

P-wave depth-velocity profiles       123

6.2.4 Density structure       126

6.3 Characterization of geological domains along Line GH/MEDOC-6       128

6.3.1 The continental margins of Sardinia (Domain 1) and 

Campania (Domain 5)       128

6.3.2 Magmatic crust in the Cornaglia Terrace (Domain 2) and 

the Campania Terrace (Domain 4)       129

6.3.3 Exhumed mantle-rock basement at the Magnaghi and

Vavilov basins (Domain 3)       130

6.3.4 Volcanic intrusions in mantle-rock basement       132

Chapter 7: Line EF/MEDOC-4       135

7.1 Data analysis       135

7.1.1 Wide-Angle Seismic Data       135

7.2 Results       138

7.2.1 Velocity structure       138

7.2.1.1 Uncertainty analysis       141

7.2.2 Tectonic structure       142

7.2.3 Automatized interpretation of the basement affinity based on

P-wave depth-velocity profiles       151

7.2.4 Density structure       154

7.3 Characterization of geological domains along Line EF/MEDOC 4       156

7.3.1 Contiental crust in the Sardinia and Campania margins, and

the Sechi segment (Domains 1a, 1b, and 5)       156

7.3.2 Magmatic back-arc crust in the Cornaglia and Campania Terraces

(Domains 2a and 4), and in the Farfalla segment (Domain 2b)       157

7.3.3 Possible presence of exhumed mantle rocks at the northernmost

Vavilov basin (Domain 3)       158

Chapter 8: Geological domains of the Central Tyrrhenian: Geodynamic implications      159

8.1 Distribution of the geological domains in the Central Tyrrhenian basin       159

8.1.1 Continental crust       160

8.1.2 Magmatic back-arc crust       164

8.1.3 Exhumed mantle       166

8.2 Implications of the new geological domains defined in the Central Tyrrhenian

basin       167

8.3 Conceptual model of the basement formation in the Central Tyrrhenian 

basin       169

Part IV: Conclusions     175

Chapter 9: Conclusions       177

9.1 Geophysical cross-sections       177

9.2 Basement formation model       180

Chapter 10: Outlook       183

10.1 Geophysical transects       183

10.2 Drilling and dredging       185

10.3 Numerical Modeling       186

Part V: References     189

ANNEX       207

A.1 Wide-Angle Seismic record sections       209

A.1.1. Line GH       209

A.1.2 Line EF       240

A.2 Multichannel Seismic profiles       273

A.2.1 MEDOC 6     273

A.2.2 MEDOC 4       275

A.2.2.1 MEDOC 4 Processing evolution       277

 

 
Conclusiones

9.1 Geophysical cross-sections
The results obtained from WAS and gravity modeling, together with observations based on the MCS profiles MEDOC-4 and 6 reveal three different basement affinities in the Central Tyrrhenian basin, that is, continental crust, magmatic back-arc crust and exhumed mantle. In addition, combination of these results with morpho-tectonic observations from bathymetry, and geological information from rock sampling and drilling [Colantoni et al., 1981; Kastens and Mascle, 1990] has allowed to interpret the plan view distribution of these domains (Figure 8.5). This section presents the main conclusions concerning the distribution of continental crust, magmatic backarc crust, and exhumed mantle in the Central Tyrrhenian basin.
 
Continental crust
In Line GH this type of crust is observed in Domain 1 and 5 (Figure 8.1b and 8.4b). Domain 1 is formed by the continental crust of Sardinia and its margin, which is characterized by landward and seaward normal faults and crustal thinning from ~22 km to ~13 km-thick. The conjugate margin of Campania (Domain 5) might represent the westernmost part of the Italian continental crust.
Similarly, in Line EF continental crust is also observed beneath the Island and margin of Sardinia in Domain 1a, characterized by a significant crustal thinning from ~25 to ~10 km-thick. However, the northern and southern geophysical segments of the Sardinia margin display differences in crustal velocity structure (Figure 8.3a) that might be attributed to differences in the petrological composition (e.g. granitic, basaltic, metamorphic). To the east, the continental Domain 1b is identified along the Sechi segment as a highly stretched portion of continental crust (ß>3) (Figure 8.1a and 8.4a).
This domain appears to belong to a major continental structure that extends northward and southward but does not reach Line GH (Figure 8.5). Finally, in Line EF the conjugated Campania margin (Domain 5) is also intersected and interpreted to represent the westernmost extension of the Italian continental landmass (Figure 8.5).
Continental crust is extended toward the Northern Tyrrhenian (Figure 8.5), where it is believed that is partially intruded by magmas at lower crustal levels [Moeller et al., 2013].
 
Magmatic back-arc crust
In Line GH it is represented by the Cornaglia Terrace (Domain 2) and the Campania Terrace (Domain 4) (Figures 8.1b and 8.4b). The transition between continental and magmatic crust is only observed by a marked, abrupt increase of velocities (> 7 km/s) in the lower crust (at ~125 km in Figure 8.1b). Large-scale normal faulting has affected the Cornaglia Terrace (Figure 6.5a), whereas in the Campania Terrace, normal faulting is less important (Figures 6.5b). The crustal structure in both regions is represented by velocities slightly lower than those found in a 0-7 Ma-oldoceanic crust [White et al., 1992]. This has been discussed and finally interpreted as indicative of back-arc spreading close to the active volcanic arc [Martinez et al., 2007].
In Line EF, this type of crust is encountered in Domains 2a, 2b, and 4 (Figure 8.1a). Domain 2a is represented by the northern region of the Cornaglia Terrace showing lower crustal velocities >7 km/s and a rather homogeneous velocity (Figure 8.6a) and tectonic (Figure 7.4) structure compared to that of Line GH in the south (Domain 2) (Figure 6.5a). Eastward of Line EF, lower crustal velocities > 7 km/s are also observed in Domains 2b and 4 in the Farfalla segment and the northern Campania Terrace, respectively (Figures 8.1a and 8.4a). The Farfalla segment (Domain 2b) does not present any counterpart in Line GH (Figure 8.1 and 8.5), and it is interpreted as a continental crust affected by magmatic underplaiting or either serpentinized mantle rocks at the base of the crust [e.g. Chamot-Rooke et al., 1999; Dean et al., 2000; Funk et al., 2004]. In contrast, the Campania Terrace is intersected by Line GH in the south (Figure 8.1, 8.4, and 8.5), displaying crustal velocity, tectonic, and bathymetric variations from north to south (Figures 6.4, 7.4, and 8.1, 8.6b). These differences suggest that the northern terrace might correspond to a stretched continental crust heavily intruded by rift-related magmas, while the southern terrace might correspond to a magmatic back-arc crust, similar to that of the Cornglia Terrace (Domain 2 and 2a). However, drilling in the Cornaglia and Campania Terraces has only evidenced the presence of isolated continental blocks (e.g. Flavio de Gioia seamount in Figure 8.1b, 8.4b, and 8.5) [Colantoni et al., 1981]. Hence, further rock sampling and geochemical information of the basement in these two terraces is necessary to better constrain its petrological nature.
 
Exhumed mantle
In Line GH, the presence of a basement made of exhumed mantle rocks is evidenced in Domain 3, which intersects both V-shaped Magnaghi and Vavilov basins (Figure 8.1b, 8.4, and 8.5). The main characteristics of this domain are the lack of Moho reflections in both WAS and MCS data and a vertical velocity structure characterized by a strong velocity gradient in the topmost 3-4 km similarly to other regions where basement is interpreted to be made of serpentinized mantle [Reid, 1994; Dean et al., 2000; Van Anvendonk et al., 2006; Sallarès et al., 2013a].
In Line EF, the presence of a mantle rock basement is less clear, since the velocity gradient is lower than the velocity-depth references of exhumed mantle, and hence fits better with that of oceanic crust (Figure 8.7). However, the lack of Moho reflections in the MCS data and in most of the WAS receivers of this region supports
the exhumation of mantle (Figures 7.4, 7.5, and A.1.2.25). Thus, I tentatively suggest that the lower velocity gradient of the exhumed mantle basement might be attributed to the influence of basaltic intrusions like sills and dicks. Nevertheless, more petrological information of the basement is needed to confirm this interpretation.
Additionally, this domain hosts large basaltic edifices related to oceanic crustal accretion (i.e. Sites 655 and 651 in Figures 8.5) [Beccaluva et al., 1990] and to interplate volcanism (i.e. Magnaghi and Vavilov seamounts in Figures 8.4 and 8.5) [Robin et al., 1978; Savelli, 2002]. Some of these structures are intersected by Line GH and
imaged in depth showing that they are local and hence not representative of the entire basement nature (Figure 6.10).
The overall differences in the seismic velocity structure and the tectonic signature between domains of the northern Line EF and the southern Line GH are most likely attributed to the southward increase of extension that characterizes the Tyrrhenian basin [e.g. Faccena et al., 2001; Sartori et al., 2004].
The analysis of these new geophysical data has led to a radically new definition of geological domains in the Central Tyrrhenian. According to the presented distribution of the basement, rifting in the Central Tyrrhenian basin would have started with continental crust extension, continued with back-arc spreading leading to
generation of magmatic back-arc crust, followed by mantle exhumation and subsequent magmatic intrusions.
 
9.2 Basement formation model
To explain the formation of the basement in the Central Tyrrhenian basin I propose a new conceptual model for the opening of the basin (Figure 8.9). This model is based on those previously proposed to explain back-arc extension and formation of the Western Pacific basins in slab rollback settings [Martinez et al., 2007; Dunn and
Martinez, 2011]. 
The most relevant observation concerning the evolution of the basin is the presence of a wide exhumed mantle band in the Central Tyrrhenian despite the fast spreading rates (5-6 cm/yr) and the apparently “normal” mantle temperatures at the time of extension. This makes that the most likely cause for the lack of  ubstantial magmatism at the spreading, and hence formation of typical oceanic crust, is the presence of a depleted mantle source.
In summary, the presented model is based on 5 stages of opening in which a depleted mantle setting constrains the production of pressure-release magmas, and thus, back-arc crustal accretion is mainly explained by the influence of subduction-related melts (i.e. hydrous flux melting in Figure 8.9). The main features of these stages are briefly exposed below:
I. The first stage consists of a normal subduction scenario, in which slab-derived melting originates arc volcanism (Figure 8.9a). The mantle beneath the overriding plate is assumed to be depleted in composition. This stage may likely illustrate the latest phases of opening of the Liguro-Provençal basin at ~20-15 Ma [Gattacceca et al., 2007].
II. In the second stage back-arc rift forms as a result of the slab rollback (Figure 8.9b). In this stage, extension causes the passive decompression of the underlying asthenospheric mantle, which in turn results in the production of pressure-release melts. However, since the mantle is assumed to be depleted, this production is rather limited. In contrast, since rifting occurs near the volcanic arc a large amount of hydrous flux melting is accreted in the lower crust forming thereby a new magmatic crust. The composition and crustal structure of this new magmatic crust would differ from those of normal oceanic crust [Martinez et al., 2007; Dunn and Martinez, 2011]. Stage II, then, illustrates the formation of both magmatic crusts interpreted beneath the Cornaglia and the Campania Terrace (Domains 2, 2a, and 4) (grey crust in Figure 8.9c).
III. In the third stage back-arc spreading is initiated (Figure 8.9c). However, as hydrous flux melting migrates with subduction front trenchward, magma supply in the spreading center diminish considerably because the depleted mantle source is not able to produce large amounts of melting by itself. Consequently, no oceanic crust is generated, but the limited amount of melt is emplaced as intrusive bodies forming sills and dikes, as suggested by the ODP site 651 (Figure 8.5) [Beccaluva et al., 1990], and isolated oceanic ridges like the Gortani Ridge [4.3 Ma] (ODP 655 in Figure 8.5) [Beccaluva et al., 1990; Bertrand et al., 1990].
IV. In the fourth stage (Figure 8.9d) extension leads to the exhumation of mantle rocks, while the little amount of pressure-release melting continues forming isolated oceanic ridges. This stage would illustrate the formation of the basement that floors the Magnaghi and Vavilov basins (Figure 8.5). V. Finally, the last stage illustrates the emplacement of large volcanic edifices in the central part of the basin, as a result of the sporadic magmatism dated at 3-0.1 Ma (i.e. posterior to the exhumation of the mantle) and attributed to an intermittent episode of intra-plate magmatism (e.g. Vavilov seamount in Figure 8.5) (Figure 8.9e) [Robin et al., 1978; Savelli, 2002].