Methodology to monitor the seismic response to injected carbon dioxide

• Published in: Acta geophysica 2024-04, Vol.72 (2), p.1343-1353
• Main Authors: Alajmi Mamdoh, Gei Davide, Carcione, José M, Qadrouh, Ayman N, Santos, Juan, Ba Jing
• Format: Article
• Language: English
• Subjects: Carbon dioxide; Clay; Correlation; Flow simulation; Fluid flow; Gas viscosity; Injection; Mudstone; P waves; Permeability; Pore pressure; Pore water pressure; Porosity; Preferential flow; Pressure variations; S waves; Sandstone; Seismic activity; Seismic response; Seismograms; Seismological data; Sensitivity analysis; Shales; Simulation; Spectral methods; Viscosity; Wave velocity
• ISSN: 1895-6572; 1895-7455
• DOI: 10.1007/s11600-023-01199-x

Petrophysics and fluid-flow simulations are used to build a realistic pre- and post- CO2 injection geological model for the Utsira formation at the Sleipner field, and the Fourier pseudospectral method is employed to compute synthetic seismograms. The methodology can be used to perform a seismic sensitivity analysis for the detection of carbon dioxide. We built the model solely based on the porosity and clay content of the formations with the aid of fluid-flow and seismic simulations. The pressure map before the injection is assumed to be hydrostatic for which a reference porosity map is defined. The injection induces pore pressure variations and partial saturation, which affect the poroelastic properties and hence the associated seismic response. A proper porosity–permeability–clay content relation is one of the key factors since permeability determines the preferential flow directions and the distribution of the CO2 plume. The petrophysical model is based on a shaly sandstone (or sandy shale) to represent the caprock, Utsira Sand and embedded mudstone layers. The composite permeability (anisotropic) is analogous to the inverse electrical resistance model. Gas viscosity depends on pressure and temperature. The P- and S-wave velocities are obtained from Gassmann equation (pre-injection) and White’s mesoscopic model (post-injection), which also yields the P-wave quality factor in the case of partial (patchy) saturation. To model a realistic situation, we implement a fractal variation of the porosity and clay content, based on the von Kármán correlation function. We then compare the real and synthetic seismograms (pre-injection and post-injection) and show the effect of attenuation on the seismic data. Simulations and real data show a remarkable match.