GNGTS 2019 - Atti del 38° Convegno Nazionale

736 GNGTS 2019 S essione 3.3 recorded and hence full elastic data are available. Since the receiver systems lay on the sea floor, they are much less affected by obstacles on the sea surfaces as FPSO (Floating Production Storage and Offloading), platforms, etc. Moreover, a great advantage especially for 4D applications is that node positions are well determined and can be reproduced much better than the towed-streamers acquisition, typically affected by feathering. Another side product of an OBN survey is the possibility to exploit the down-going wavefield (i.e., the receiver ghost field) for the so called “mirror” imaging (Grion et al. , 2007; Hanafy et al. , 2015), resulting in a wider illumination area with respect to the migration of primaries only. In the market, different kind of nodes suitable for deep water can be found and they differ from each other mainly for the shape (rectangular or circular), size (in the order of tens of centimeters), weight (10 to 50 Kg in sea water) and operating life (45 to 180 days). Despite these differences, all the nodes are characterized by 4 sensors: 1 hydrophone and 3 geophones. The OBN is set in continuous recording: the node starts to record when it is turned on, usually just before the splash in the water, and it could continue until it is recovered or it runs out of battery. Differently from the ocean bottom cable, the deployment is conducted by means of ROV (Remote Operated Vehicle) that positions each node on the sea floor independently from each other. Hence, operations on sea surface are reduced to the receiver vessel and, once the nodes are deployed, to the source vessel only. Once all the nodes are recovered and data downloaded, the deliverable seismic traces are obtained as segmentation of the unique and long trace blindly and continuously recorded by the node for the whole duration of the survey. As it is, the recording length can be considered not anymore an acquisition parameter but the first processing parameter. In order to perform a good segmentation, the GPS clock timing synchronization of the node and of the shot is of paramount importance. With this regards, the nodes can be equipped with one of the two more common devices, that differ from each other for the clock drift effect varying with time: atomic clock and oven oscillator system. Since each node is an autonomous system without real-time connection to the surface, one of the main drawback of OBN is that the data QC during acquisition is not possible and so any eventual failure of the system is known only once the survey is completed. Survey design. In order to solve the aforementioned geophysical challenges in the area of interest, the optimal and feasible acquisition parameters were defined by means of ray tracing and both acoustic and elastic modelling. These parameters consist mainly in the definition of maximum offset (and hence the source area), the migration halo (that determines the node area) as well as the nodes and shots spacing. Different synthetic datasets were generated (with different streamer and OBN layouts) and then pre-stack depth migrated to evaluate which one provides the better imaging in the subsalt areas and to assess the benefits of an OBN with respect to conventional towed-streamers acquisition. These acquisition parameters were defined not only to improve the poor subsalt imaging, but also with regards of full-waveform inversion, as well as keeping lower as much as possible the effects of contamination of converted waves. As far as the nodes, and hence also the shots, mutual position is concerned, a triangular grid, and hence a hexagonal bin, is adopted (Fig. 1). In this acquisition geometry, in order to minimize the amount of nodes, the spacing of these devices is quite large but such to still guarantee a reasonable Signal-to-Noise ratio, while the shot points are quite dense in order to guarantee a good resolution. In order to sample a signal that is varying in space, Petersen and Middleton (1962) demonstrated that for the same maximum wavenumber, a triangular grid better optimizes the distribution of the samples with a reduction of 13.4% with respect to the square one. In this kind of study, an analysis of the sea floor is of paramount importance in order to understand where the nodes effectively will lay. In the area of interest, the sea floor is characterized by a general very gentle slope (less than 2°) and by furrows (elongated depressions, few meters deep with a gentle slope up to 4°). As it is, there are no criticalities with regards to the node deployment, except over the salt bodies, where the sea floor is much more corrugated and characterized by the presence of pockmarks: large (up to 350 m) isolated or coalescent