Simon, Soazig

The geological evolution of the western Mediterranean exhibits complicated interactions between orogenic processes and widespread extensional tectonics. The region is located in a convergent plate margin separating Africa and Europe, and consists of marine basins – the Alboran Sea, the Algerian- Provençal Basin, the Valencia trough, the Ligurian Sea and the Tyrrhenian Sea- which formed as back-arc basins since the Oligocene. In most reconstructions, it has been stressed that back-arc extension led to drifting of continental blocks and to large-scale block rotations. The opening of the Ligurian Sea. is in fact the result of counterclockwise rotation of Corsica and Sardinia. From the point of view of seismicity, the south western Alps and northern part of the Ligurian basin are subject to frequent earthquakes of low to moderate magnitudes. However significantly destructive events are known to have occurred in the past (e.g. 1564 and 1887). Apart from these rare large events, regional studies agree in concluding that the important local microseismicity appears to be poorly focused (e.g., COURBOULEX et alii, 2007) and that, if some tectonic lines are documented onland (COURBOULEX et alii, 2001), the active structures at sea remain unknown. It is therefore an essential prerequisite to gain better insight into the deep seismogenic structures along the North Ligurian margin and even farther offshore, in the identified oceanic domain. The fact that some of these structures can undergo ruptures of Mw~6.5, such as the 1887 event (BAKUN & SCOTTI, 2006), suggests that, at least to some extent, instrumental insufficiencies in the detection and location of microseismicity is a limit to identify active faults that have not experienced large instrumented ruptures to date. The irregular coverage provided by regional seismic networks produces a bias in the recording of local seismicity. Permanent stations are naturally limited to land areas and fail to properly constrain seismicity offshore. Taking into consideration the peculiarities of regional dynamics (low strain rates, rare large events and a regular seismic activity limited to small events with M < 3-4), even onshore seismicity is insufficiently covered by permanent networks and requires dense temporary instrumenting by mobile stations. Considering the potential threat of strong offshore earthquakes, it is of first importance to characterize faults that are prone to rupture in order to quantify associated seismic and tsunami hazards. Assuming some weak seismicity exists along these faults and remains undetected by onland networks, some marine stations are necessary to address instrumental remoteness and help delineate active structures. Moreover, since the velocity models used for locations are obtained by inverting seismic data and the reliability of their locations depend, in turn, from the quality of the velocity model used for their hypocentral parameters, the constraints on the seismic path provided by a more dense seismic network may contribute to a more accurate reference model. In this study, we profited from the recent developments in sea bottom seismic instrumentation to deploy OBSs above the zones of the North Ligurian to perform seismic shots and obtain the distribution of seismic velocities with 3D active tomography. We also took the opportunity of the long term (6 months) OBSs reduced array to decrease both the detection threshold and recording distances so as to obtain more complete catalogs and better localisations.

The North-Ligurian rifted margin is singular in that it lies immediately next to the Alpine orogenic arc. It is furthermore seismically active and can experience destructive earthquakes such as in 1887 in the region of Imperia—an event that resulted in a tsunami and more than 600 casualties in spite of a coastal area that was much less densely populated than today. Out of such rare large events, the area undergoes a limited and diffuse seismic activity that can remain undetected and is generally poorly located. This results in a poor knowledge of active structures, especially at sea. Such knowledge is however required towards a quantification of the seismic hazard along the French Riviera and the Ligurian region. To this end, the GROSMarin project was undertaken with a dual objective: (1) to characterize the North-Ligurian margin from a structural standpoint—mode and degree of crustal stretching prior to oceanic accretion, segmentation along strike, subsequent evolution in an orogenic context— and (2) to identify zones of active crustal deformation at sea that are likely to generate earthquakes. The programme is a collaborative work between GeoAzur and Dip.Te.Ris (University of Genova), with some support from INGV, IFM-GEOMAR and IFREMER. It took place from April to October 2008 and consisted in the deployment of 21 ocean-bottom seismometers (OBS) on a grid spanning 50 km along strike and 25 km across, located between Nice, France, and Imperia, Italy, and ranging from mid-slope to the deep basin. This array was extended on land by the permanent stations of the French and Italian regional networks, temporarily densified by 13 portable stations. These instruments recorded the shots of a marine seismic source towed from R/V l’Atalante and were left for more than 5 months for passive surveying. The active part of the programme aims at characterizing the main structures of the margin through crustal 3D tomography; the objective of the passive part is to decrease the detection threshold of marine microseismicity and to reach a precise location of events in order to map active faults. Some of the sea and land instruments were fitted with broadband sensors to allow for teleseismic imaging of deep lithospheric discontinuities. We present the preliminary results of this experiment—in particular a first 3D tomographic model obtained from 31.500 travel times derived from our recording of active seismic shots by the OBS’s. Passive data analysis is being under progress and first relocations have been obtained. These results give an insight into the variability of the crustal structure, both along and across strike.

The deep structure of the North Ligurian margin and its contiguous Ligurian basin as well as the seismicity recorded in these zones are neither well understood nor precisely constrained. In order to better address these questions, there is a need for offshore instrumenting, which was realised for a duration of nearly 6 months during the GROSMarin (Grand Réseau d’Observation Sous-Marin) experiment. An array of 21 ocean bottom seismometers was deployed over the most active area of the margin and was complemented on land by mobile seismological stations that densified existing permanent networks. We also realised the acquisition of deep refraction seismic shots at sea in order to get a 3D distribution of velocities along the margin through travel time tomography. We present here a preliminary analysis of the seismicity recorded during this experiment and a tomographic model of the margin structures obtained using data from the offshore network only. Our results support the existence of a high velocity zone at the base of a domain interpreted as transitional between continental and oceanic ones, on the northern part of the deep basin. A very similar pattern is observed across the neighbouring margin of the Gulf of Lions and is most likely related to serpentinisation of the underlying mantle during late rifting and continental break-up. North of this transition zone, we observe the basinward crustal thinning of the continental crust beneath the margin that seemingly narrows eastward. To the south, our results hint at transition to the oceanic domain. In contrast, our velocity distribution does not reveal a transition along strike between transitional and oceanic domains, as previous works suggest. Some microseismic activity was recorded throughout the duration of the experiment, on land and at sea. The number of detected events and precision of location were both improved by our considering French and Italian permanent networks. The detection capabilities of our dense network still need to be fully exploited