oa Integrating Rock Physics, Seismic Reservoir Characterization and Static Modeling of Carbonates; a Case Study from the UAE

An integrated workflow to generate seismically constrained reservoir models is described. Several key<br>technologies were used including carbonates rock physics estimation, seismic forward modeling and<br>comparison to surface 3D seismic, as well as probabilistic seismic inversion.<br>Using the SUN rock model as a framework, the team derived relationships between the reservoir<br>properties (like modeled facies, porosity and fluid content) and the elastic properties (Vp, Vs and<br>density). The theory of the Sun model was chosen as a key for relating reservoir properties to seismic:<br>relations were identified between Suns frame flexibility factor (describing the elasticity of the rock<br>frame determined by pore geometry), velocities, densities and porosities. Those relations were<br>compared to the information on sedimentology, diagenesis, structural position and reservoir rock<br>types. Detailed well to seismic match enabled estimating a fit-for-purpose average wavelet to be used<br>over the entire field. The fluid substitution of the well logs, log blocking, and the fluid properties made<br>it possible to model the fluid content impact on the seismic, and to better understand the impact of peg<br>-leg multiples still present in the seismic data.<br>The generation of 3D synthetic seismic based on the static model included the use of the relations<br>obtained from the rock physics model, the well log blocking and the derived seismic wavelet. The<br>match between real and modeled synthetic seismic indicates how well the parameters in the static<br>model describe the reservoir, and the relevance of the variables (rock & fluid properties, layering, and<br>wavelet) included in the forward modeling. The seismic match was improved by iteratively fine tuning<br>the different variables used to generate the synthetic seismic. The optimization process highlighted<br>which variables control the seismic response. These were subsequently used to define the stochastic<br>parameters and the uncertainties in the probabilistic seismic inversion. The inversion algorithm used<br>utilizes the constrained static model as input and can invert to any of the variables present in it.<br>Therefore it was possible to obtain probability distributions for porosity, fluid saturation and rock<br>rigidity in any location of the reservoir that match the seismic.<br>The workflow was applied to the Bab field in Abu Dhabi, UAE. The resulting static model reflects more<br>accurately the lateral variability of the rock properties while preserving vertical resolution.