Introduction
Subduction is very important to plate tectonics. Wadati-Benioff zones (e.g. Isacks & Barazangi 1977, Giardini & Woodhouse 1984, Jarrard 1986) and the distribution of tomographic anomalies (e.g. Bijwaard 1999, Fukao et al. 2001) provide a snapshot of present-day conditions, suggesting that the cold solid lithosphere sinks into the convective, fluid-like mantle with different dips and shapes. The long-term evolution of subduction is still uncertain and difficult to unravel due to the transient character of the process. Hence, to better understand the dynamics of subduction, seismic data have to be integrated with other constraints; these include indirect observables (geological, petrological, geochemical and structural studies of the trench – back-arc system) and numerical and laboratory models. Modeling, in fact, is the only available tool that can provide a dynamic and evolutionary picture of the slab. The sinking of a dense material in a Stokes fluid may catch the essence of the subduction process, offering a possible explanation for a wide range of natural observations (e.g., Kincaid & Olson 1987, Griffiths et al. 1995, Zhong & Gurnis 1997, Buttles & Olson 1998, Schmeling et al. 1999, Kincaid & Griffiths 2003, Schellart 2005, Husson 2006, Capitanio et al. 2007, Becker & Faccenna 2008, Royden & Husson 2008).
Here, we provide an overview of what can be learned about subduction with laboratory models. Because no single, comprehensive model currently exists that can simulate all scales of subduction, we focus our attention on a specific area separating different temporal aspects of the process; these include the early, intermediate and long-term evolutionary stages. For this purpose, the Central Mediterranean (Figure 1) offers an ideal regional test site. This area is characterized by a very complex structural setting. It contains a northwest dipping subduction zone with an arcuate shape, which produced the Apennine chain, and an extensional domain, the Tyrrhenian Sea, which can be thought of as the recent back-arc basin of the system (e.g., Malinverno & Ryan 1986, Royden 1993, Faccenna et al. 1996). In this geodynamical setting, which has been shaped by the interplay between the Eurasian and the African plates, the net convergence of the incoming African plate at the trench has always been very low (Dewey et al. 1989, van der Voo 1993, Ward 1994, Silver et al. 1998), providing a rare opportunity to preserve the clear geological remnants of past subduction on the surface. These signatures offer the possibility to constrain the evolution of the subduction process, allowing us to study the interaction between the subduction process and back-arc opening. Presently, in addition to the laboratory models that will be described in this paper, two additional approaches have been adopted to describe the main characteristics of the Central Mediterranean subduction. The first approach combines the seismic velocity structure of the subducted lithosphere with the regional tectonic history (e.g., Wortel & Spakman 2000, Faccenna et al. 2001a, Faccenna et al. 2004). The second approach uses numerical models to investigate the stress field at the surface (e.g., Bassi & Sabadini 1994, Bassi et al. 1997, Negredo et al. 1997) and at depth (e.g., Giunchi et al. 1994, Marotta & Sabadini 1995, Giunchi et al. 1996, Carminati et al. 1998) including analysis of the mantle circulation pattern (e.g. Ismail-Zadeh et al. 2009).
In this paper, we will describe how the geodynamics of the Central Mediterranean can be efficiently illustrated in a tank, using scaled experimental models. First, we briefly describe the geological context that characterizes the Central Mediterranean area. Subsequently, after an initial technical description, we describe the laboratory modeling illustrating different phases of the subduction process. Finally, the results are placed in the context of the global picture for the Central Mediterranean. The questions addressed by this work are the following:
- Which ingredients enhanced the initiation of the subduction process?
- Which dynamic mechanism produced trench migration rates up to 6 cm/yr, with convergence rates of only a few mm/yr?
- Why did the process occur episodically, causing the opening of distinct basins (Liguro-Provençal and Tyrrhenian, along with the Vavilov and Marsili)?
- Can we reconcile the seismological images (tomography and seismic anisotropy) of the deep mantle with the available geological record?
Answers to these questions will not only improve the comprehension of the history of the Central Mediterranean but also shed light on the general evolution of the subduction process.