GNGTS 2018 - 37° Convegno Nazionale

570 GNGTS 2018 S essione 3.1 identified. Moreover, due to the shallow water conditions of the area, the potential transfer of gas from sediment to the water column and then into the atmosphere is supposed to be rapid. Therefore, the importance of understanding the mechanisms of natural gas seepage has direct climatological implications, since CH 4 is the 2nd most significant long lived greenhouse gas. Data analysis. The seismic data analyzed in this work were collected in the NorthernAdriatic Sea within the framework of two projects, i.e., STENAP “Stratigraphic and Tectonic Evolution of the Northern Adriatic Sea” and GANDI “GAs emissions in the Northern ADriatIc Sea”. Two perpendicular seismic lines, STENAP 08 and GANDI 09, were chosen for the analysis (Fig.1). All the well log data used in this work come from the ViDEPI database, designed to make all the documents concerning Italian oil exploration easily accessible. Amongst the 18 composite logs available in the study area, 5 wells, crossing the selected seismic lines or located nearby, have been analysed. However, not every well includes all the fundamental logs, so that relationships between logs had to be evaluated, in order to extrapolate the quantities needed for the analysis. Log Processing. Log data were digitized, edited and resampled to be consistent with the seismic frequency content. Experimental relationships between each pairs of signals were investigated via cross-correlation, fit and χ 2 tests. The basic idea was to seek for a common trend in the log properties of each well and of the different wells, which could be representative of rock formations. This has then been used to find the best empirical relationship between the available and missing logs. Density profiles were reconstructed for each borehole of interest with an iterative procedure of comparison between synthetic and real signals. The procedure was guided by a geological interpretation of the lithostratigraphic column of the boreholes, and based on the geological information taken from the technical drilling reports. Density profiles were then used, together with sonic logs, to calculate acoustic impedance at well-location and to perform acoustic inversion. Seismic Data Processing. Afirst processing flow has been applied to the two chosen seismic profiles (STENAP 08 and GANDI 09) to better image the subsurface. Data were strongly affected by multiples, especially water-bottom reverberations, so that the whole procedure was focused on removing them. Apart from filtering out low frequency noise, pre-stack processing consisted in deconvolution in τ-p domain; correction for spherical divergence; a trimmed mean dip filter (TMDDF); migration and predictive deconvolution. TMDDF removes high amplitude random noise and locally weak coherent events without eliminating useful information by too severe lateral filtering. Data were then migrated in time, in common-offset, with a double-square root operator and of aperture. Velocity fields were built using the sonic log for the shallower part and through a semblance analysis for the deeper part. In post-stack, a F-X deconvolution and a time-variant filter were applied. This processing flow resulted in seismic sections where the geometry of the seismic event has been well displayed, crucial for geological interpretation. A second processing flow used specific algorithms to improve the S/N ratio and to remove multiples without affecting the relative amplitude information. This implies that specific processes, such as deconvolution or migration, which are commonly used and are very effective to enhance data quality, were not applied because influencing the original amplitude of the signal. Instead, the focus of the ‘true amplitude’ flow is the surface-related multiple elimination (SRME), an adaptive amplitude-preserving algorithm which does not assume any model of the subsurface nor source signature. The resulted processed lines are the ones used in the correlation, inversion and gas content quantification. Post-stack stratigraphic inversion. Stratigraphic acoustic inversion combines seismic and log data to produce P-impedance (IP) sections. IP are crucial in costraining every kind of reservoir model and they are used in this work as input data in the porosity estimation with EMT (Effective Medium Theory). The coupling between data acquired at different scales of resolution is very powerful because it gives back a calibrated information on a broad-band frequency. The inversion performed here is based on a geological ’a priori’model, which reduces the space of solutions. The whole procedure can be summarized in three steps: well-seismic