Deep electrical resistivity structure of the northern Gibraltar Arc (western Mediterranean): evidence of lithospheric slab break‐off

Terra Nova, 23, 179–186, 2011


Introduction
The convergence of the African and Iberian plates generated the Gibraltar Arc (Rif and Betic Cordilleras, Fig. 1) from the Late Cretaceous onwards (Garcı´a-Duen˜as et al., 1992;Azan˜o´n et al., 2002). Different geodynamic models have been proposed to explain the lithospheric structure of this arcshaped belt and the opening of the Alboran Basin based on Bouguer anomalies, heat flow, earthquake locations, seismic refraction, seismic tomography, geoid anomalies and elevation data (e.g. Morales et al., 1997;Ferna`ndez et al., 1998). The opening of the Alboran Basin has been suggested to be a consequence of convective removal of the thickened lithospheric root that caused uplift and extension (Platt and Vissers, 1989;Platt et al., 1998), lithospheric delamination caused by gravitational collapse of the thickened lithosphere (Seber et al., 1996;Mezcua and Rueda, 1997;Calvert et al., 2000), westwards to southwards rollback of an oceanic slab that generated back-arc extension (Royden, 1993;Lonergan and White, 1997;Gutscher et al., 2002), south-eastwards rollback of an oceanic slab attached to the African plate (Doglioni et al., 1997(Doglioni et al., , 1999a, south-eastward delamination of the subcrustal lithospheric slab (Docherty and Banda, 1995) or a vertical brokenoff piece of a previously subducting lithospheric slab (Zeck, 1996(Zeck, , 1997. The magnetotelluric method has been proved to be a useful technique for imaging the lithospheric resistivity structure beneath plate boundaries, providing constraints for geodynamic models. Continent-continent collision areas have been the focus of many studies whether or not they are active (Ledo et al., 2000;Unsworth, 2010). Continent-ocean collision areas have also been imaged clearly, depicting the subducting oceanic slab (Wannamaker et al., 1989;Brasse and Eydam, 2008;Brasse et al., 2009). Some magnetotelluric surveys have been carried out in the central Betics assuming 2-D structures (Pous et al., 1999;Pedrera et al., 2009), but this assumption can lead to incorrect interpretations in complex geological areas with 3-D structures (Garcı´a et al., 1999;Ledo, 2005). Martı´et al. (2009a) presented a 3-D resistivity model of the Central Betic Crust. To image the lithospheric structure of the Betics, we extended the study area to the whole Cordillera and included long-period data up to 20 000 s. Long-period data allow for a deeper investigation depth, which is crucial to characterize the lithosphere-asthenosphere boundary (LAB) and the electrical resistivity distribution in the lithospheric mantle and lower crust.
The Betic Cordillera (Figs 2a and S1) is divided into the External Zones and the Internal Zones (Azan˜o´n et al., 2002). The External Zones include carbonate rocks from the South Iberian palaeomargin as well as detritic rocks from the Flysch Trough Complex, with ages ranging from Mesozoic to Cenozoic. The Internal Zones, also known as the Alboran Domain, include three Palaeozoic to Triassic metamorphosed nappe complexes. Post-orogenic upper Miocene to Quaternary basins and the Alboran Basin (the backarc basin of the Gibraltar Arc) lie discordantly over these units. The north-westernmost mountain front of the Betic Cordillera is the only one that remains active (Ruiz-Consta´n et al., 2009). Fullea et al. (2010 modelled the crustal thickness as 30 km beneath the External Zones and ranging beneath the Internal Zones from more than 36 km under the highest mountains to 20 km near the coastline, which matches existing deep seismic profiles (Garcı´a-Duen˜as et al., 1994). The depth of the LAB under the Betic Cordillera increases from 100 km at the eastern boundary to 170 km at the western one (Frizon de Lamotte et al., 2004;Soto et al., 2008;Fullea et al., 2010). None of these values were obtained from electrical resistivity models, so this is the first time the lithospheric ABSTRACT Uncertainties about the lithospheric structure of the Gibraltar Arc have led to the proposal of several contradictory models to explain its geodynamic setting. Herein, we present a novel 3-D model of the lithospheric electrical resistivity distribution beneath the whole Betic Cordillera obtained by inverting both broad-band and long-period magnetotelluric data. The lithosphere-asthenosphere boundary under SW Iberia is shown to be deeper than under the Alboran Basin. In addition, the sensitivity tests confirm the presence of a N-S oriented low-resistivity anomaly at lithospheric mantle depths east of the 4°W meridian. It coincides with an area of low velocities without earthquake hypocenters and is interpreted as asthenospheric material intruded by the lateral lithospheric tearing and breaking-off of the E-directed subducting Ligurian slab under the Alboran Domain. This scenario suggests that the opening of the Alboran Basin is related to the westward rollback of this E-directed subducting slab.

Magnetotelluric data
The magnetotelluric method uses natural electromagnetic fields to characterize the structure of the subsurface. It is a valuable technique for imaging the lithosphere and the geometry of the LAB (Jones, 1999). Its investigation depth depends on both the recording period and the resistivity of the Earth.
The dataset we present consists of 100 magnetotelluric sites located across the Betic Cordillera (Figs 1 and S1); 41 of them include long-period data. The time series were processed using robust algorithms (Egbert and Booker, 1986;Chave and Thomson, 2004) with remote reference when possible. The apparent resistivity and phase curves obtained cover periods from 0.001 to 20 000 s for some of the sites, showing medium to high quality (Fig. S2).
We obtained a dimensionality map using the WALDIM code (Martıé t al., 2009b), based on the invariant rotation parameters of the impedance tensor presented by Weaver et al. (2000), to determine whether the geoelectrical structures at different depths can be identified as 1-D, 2-D or 3-D. The results (Fig. 3) show a predominance of 3-D geoelectrical behaviour for periods longer than 10 s across the whole area. Thus, a 3-D model is the most valid approach to properly characterizing the deep crustal and lithospheric electrical structure beneath the Betic Cordillera.

Geoelectrical lithospheric structure
The geoelectrical structure of the Betic lithosphere was imaged by building a 3-D resistivity model with 38 · 50 · 33 mesh elements. The initial model used for the inversion was a homogeneous 100 XAEm block, except for the sea, which was fixed at a constant 0.3 XAEm based on the bathymetry of the Alboran Sea. The WSINV3DMT inversion code (Siripunvaraporn et al., 2005) was used to invert the off-diagonal components of the impedance tensor. The misfit between the data and the model responses has an RMS value of 5.2 when using only a 5% error for the impedance values. Figure S2 shows the misfit between the data and the model responses at each site.
The resulting model ( Fig. 2b-i) is characterized at upper to middle crustal levels by a complex pattern of resistive and conductive zones. The shallow conductive zones are likely to be related to the detrital infill of Neogene basins such as the Guadalquivir Basin (CGU), the Guadix-Baza basin (CGB) and the Granada Basin (CG). The External Zones are collectively depicted down to 10.5 km as a body of heterogeneous resistivity values due to their complex structure and the variable composition of carbonate marls, mudstones and detritic rocks. To the north, south and beneath these conductive domains, the resistive zones correspond to the igneous and metamorphic rocks of the Iberian Massif (RIM) and the metamorphic Palaeozoic to Triassic rocks of the Internal Zones (RIZ), which continue to mid-low crustal levels (Fig. 2c,d).
The bases of these two resistive bodies, located in the Moho, are not visible in the model because there is no variation in electrical resistivity between the lower crust and the upper lithospheric mantle. In the Internal Zones the model shows a low-resistivity anomaly in the upper-mid crust between depths of 4.5 and 17.5 km. This conductive body (CB1) was interpreted by Martı´et al.
(2009a) to represent basic or ultrabasic rocks containing a conducting mineral phase. It is clear from our 3-D model that it has no continuity towards the west and does not appear at these depths anywhere else in the study area.
At deeper levels, the resistivity model depicts a boundary between 110 and 160 km that marks the transition from values of 500-1000 XAEm to values as low as 10 XAEm (Fig. 2h-i). These low resistivity values could correspond to asthenospheric material (Eaton et al., 2009) and, hence, this transition is interpreted as the LAB.
In accordance with the model presented by Fullea et al. (2010), the LAB is estimated to be located at 110 km under the eastern Betics, deepening to 160 km towards the western Betics.
Above this boundary the most remarkable, yet previously undescribed, feature of the model appears. This is a N-trending conductive body (CB2) located east of the 4°W meridian and extending in depth from 30 to 62 km (Fig. 2e-f). This CB2 body, located at lithospheric mantle levels, has resistivity values ranging from 5 to 15 XAEm and is subvertical, dipping almost 90°west. Despite the mainly N-trend, at its northern limit it turns 90°west and continues for about 70 km. The sensitivity tests performed on the CB2 body show that its base reaches depths of at least 62 km, but, given the loss of resolution of the magnetotelluric method beneath a conductive body (Jones, 1999) could be located as far down as the asthenosphere

Geodynamic implications
According to lithospheric seismictomography studies performed in the Betic Cordillera (Morales et al., 1999), the location of the CB2 anomaly compares well with a zone of low seismic velocity (up to a 6% decrease).
In addition, comparison of the presented resistivity model with earthquake hypocenter locations (Fig. 5  The previously presented geodynamic models have been analysed using the constraint of the CB2 body. Convective removal or gravitational collapse of a thickened lithosphere can explain the presence of asthenospheric material at lithospheric mantle levels, but the strike of the CB2 body is at odds with both hypotheses, as it is clearly oriented N-S and these hypotheses would predict the asthenospheric material to be E-W directed. The south-eastward delamination of a subcrustal lithospheric slab presented by Docherty and Banda (1995) suggests an asthenospheric upwelling matching the shape of the CB2 body but, again, the NE-SW strike this hypothesis needs is not compatible with the geometry of the CB2 body. Thus, the only remaining options are the ones involving subduction processes.
To explain the N-S strike of the CB2 body, the subduction needs to be east-or west-directed. Although the main lithospheric subduction in the western Mediterranean is west-directed (Apennine subduction; Doglioni et al., 1999b), east-directed subduction has previously been proposed to explain the lithospheric structure of the Gibraltar Arc and the opening of the Alboran Sea (e.g. Gutscher et al., 2002;Krijgsman and Garce´s, 2004;Dı´az et al., 2010;Bokelmann et al., 2010). This hypothesis combined with the lithosphere-tearing model presented by Govers and Wortel (2005) allows us to interpret the CB2 body as asthenospheric material intruded into the lithosphere (Fig. 6). The shape and location of this asthenospheric intrusion can be correlated with a detachment of the east-directed subducting slab. A slab break-off is suitable in this setting (Govers and Wortel, 2005) and explains the recent uplifting of the whole area as suggested by Zeck (1996Zeck ( , 1997. This slab  detachment is limited to the north by an E-trending lithospheric tearing in which asthenospheric material intrudes at its W limit generating the 90°turn of the northern part of the CB2 body. The E-trending lithospheric tearing corresponds to the boundary between the Iberian plate and the eastwards-subducting Ligurian oceanic lithosphere and progressively ends westwards near the 5°W meridian (Fig. 6). The resulting LAB is outlined on the resistivity model in Fig. 5. Geophysical evidence of lithospheric slab detachments has previously been found in other subduction areas such as the Mediterranean-Carpathian region (Wortel and Spakman, 2000).

Conclusions
The deep electrical resistivity model presented in this work contributes to our understanding of the lithospheric structure of the northern Gibraltar Arc, beneath the Betic Cordillera. The existence of a low-resistivity anomaly at lithospheric mantle depths east of the 4°W meridian coincides with a lack of earthquake hypocenter locations and a previously observed low velocity zone. The geodynamic setting our model suggests is based on the magnetotelluric constraints and shows that the lithospheric structure under the Betic Cordillera and the adjoining Alboran Basin is the result of the westwards roll-back of an E-directed lithospheric subduction that ends to the north in a tearing sequence. This subduction resulted, during its latest stages, in slab break-off and detachment with asthenospheric material intruding and filling the resulting gap. This model partially agrees with the E-directed subduction proposed previously (Royden, 1993;Lonergan and White, 1997;Gutscher et al., 2002) but introduces lateral lithospheric tearing and slab break-off. The outlining of the geometry of the LAB, located at depths between 110 km (NE) and 160 km (SW), corresponds well with the geometries presented by previous works, the only exception being the previously undescribed asthenospheric intrusion.

Supporting Information
Additional Supporting Information may be found in the online version of this article: Figure S1. Geological map of the Betic Cordillera with the location of the used magnetotelluric sites. Figure S2. xy and yx apparent resistivity and phase for the data and model responses for the whole dataset. Site locations can be extracted from Figure S1.
Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author of the article.