Step 1: Elastic Inversion
During the pre-stack seismic inversion, a global objective function is minimized in order to compute an optimal model for P- and S-impedance, which best explains the input angle stacks and is consistent with the geological knowledge introduced through a priori information (Tonellot, 2001). Initial low frequency, models of P- and S-impedance were created based on seismic velocity and well data. These models exclude any reservoir signature and mix the lateral variations from the velocity field and vertical resolution from wells logs filtered at 5Hz.
Final results were then extracted along well trajectories and compared to actual well log data in order to ensure a good representation of the earth model. After 25 iterations, the seismic inversion results represent very well both the earth model from the well log data and the actual partial angle stacks.
Step 2: Petrophysics
The multi-attribute rotation scheme (MARS) (Alvarez et al. 2015), was used to estimate rock properties and facies volumes from well logs and seismic inversion attributes. This workflow uses a numerical solution to estimate a transform to predict petrophysical properties from elastic attributes. The transform is computed from well log-derived elastic attributes and petrophysical properties, and posteriorly applied to seismically-derived elastic attributes. MARS estimates a new attribute, in the direction of maximum change of a target property in an n-dimensional Euclidean space formed by n attributes. The method sequentially searches for the maximum correlation between the target property and all the possible attributes that can be estimated via an axis rotation of the basis that forms the aforementioned space.
The rock properties that were estimated using the MARS approach were total porosity (PhiT), volume of clay (Vclay) and the hybrid petrophysical property water saturation plus volume of clay (Sw+Vclay). This last property, which can take values between zero and two, was used as input to build a litho-fluid facies volume based on the cut-off values. Three litho-fluid facies were defined. The green facies denotes zones where clean hydrocarbon bearing sands with thickness above seismic resolution are expected to be found, including both commercial and residual saturation given the inability of the elastic measurements to distinguish between these two. The blue facies represents clean wet sand or shaly hydrocarbon bearing sand or thin hydrocarbon bearing sand that cannot be resolved at seismic resolution. These three configurations of rock and fluid present a high degree of overlap in the elastic domain, in consequence cannot be separated using seismic data. The last litho-fluid facies (brown) represents the background trend that are composed of shales or thin wet sand. In this way 2D section of total porosity, clay content and litho-fluid facies along the 2D seismic line could be generated.
