Lateral Discrimination of Sand Packets with 3D-Seismic-Derived Acoustic Impedances, La Itala, Golfo San Jorge, Argentina

Abstract

Abstract The depositional environment in the west section of the Golfo San Jorge Basin is characterized by fluvial sand-shale sequences in which the reservoirs rarely have thickness that exceeds 10m, while average thickness is 4m. The acquired seismic has a tuning thickness of 17m, so that the wavelet effect and thin bed interferences make the seismic derived attribute analysis not ideal for lateral and vertical delineation of sandstone complexes. The seismic limitations of only describing interfaces without discriminating lithologies, is overcome inverting 3D seismic traces into pseudo logs of Acoustic Impedances by applying a constraint sparse spike algorithm. Both well data (lithology and compaction) and horizons (structural framework) help to constrain the process, thus the reflectivity series found meets the criteria of minimizing the residual between the synthetic model and the real 3D seismic trace. An Acoustic Impedance volume is derived from the reflectivity series, (the frequency spectrum is limited to the seismic bandwith lacking the low frequencies corresponding to compaction) and is called bandlimited. Impedance logs from key wells were generated, filtered to a range between 0-7 Hz and then interpolated (honoring both the stratigraphy and the structural framework) to generate a low frequency model of impedances. By merging the low frequency model with the bandlimited volume a total Acoustic impedance volume is obtained. Well data analysis describes that Acoustic impedances not only at well frequency but also at seismic frequency are visualized as maximum for sandstone packets. The acoustic impedances derived from the inversion process are sensitive to the presence of a stack of several sand units and hence can be used to map lateral variations of petrophysical reservoir properties. We have attempted to quantify the petrophysical properties of such sand sequences using extensive well control. To do so we first low-pass filtered the well logs to exhibit vertical variations consistent with those of the 3-D seismic data. The low-pass filtered logs were further analyzed via lithology-based cross-plots to ascertain their effective petrophysical properties. This procedure successfully provided us a quantitative model to predict vertical and lateral variations of petrophysical properties from seismic-derived acoustic impedances. We have also developed a similar procedure based on attributes extracted from the reflectivity data generated as by-product of the inversion process. This effectively allowed us to delineate the lateral and vertical petrophysical behavior of useful sand units for subsequent in-fill and step-out development operations. We have successfully tested this methodology on a number of new wells in the area.