Numerical modeling of fluid effects on seismic properties of fractured magmatic geothermal reservoirs

Abstract. Seismic investigations of geothermal reservoirs over the last 20 years have sought to interpret the resulting tomograms and reflection images in terms of the degree of reservoir fracturing and fluid content. Since the former provides the pathways and the latter acts as the medium for transporting geothermal energy, such information is needed to evaluate the quality of the reservoir. In conventional rock physics-based interpretations, this hydro-mechanical information is approximated from seismic velocities computed at the low-frequency (field-based) and high-frequency (lab-based) limits. In this paper, we demonstrate how seismic properties of fluid-filled, fractured reservoirs can be modeled over the full frequency spectrum using a numerical simulation technique which has become popular in recent years. This technique is based on Biot's theory of poroelasticity and enables the modeling of the seismic velocity dispersion and the frequency dependent seismic attenuation due to wave-induced fluid flow. These properties are sensitive to key parameters such as the hydraulic permeability of fractures as well as the compressibility and viscosity of the pore fluids. Applying the poroelastic modeling technique to the specific case of a magmatic geothermal system under stress due to the weight of the overlying rocks requires careful parameterization of the model. This includes consideration of the diversity of rock types occurring in the magmatic system and examination of the confining-pressure dependency of each input parameter. After the evaluation of all input parameters, we use our modeling technique to determine the seismic attenuation factors and phase velocities of a rock containing a complex interconnected fracture network, whose geometry is based on a fractured geothermal reservoir in Iceland. Our results indicate that in a magmatic geothermal reservoir the overall seismic velocity structure mainly reflects the lithological heterogeneity of the system, whereas indicators for reservoir permeability and fluid content are deducible from the magnitude of seismic attenuation and the critical frequency at which the peak of attenuation and maximum velocity dispersion occur. The study demonstrates how numerical modeling provides a valuable tool to overcome interpretation ambiguity and to gain a better understanding of the hydrology of geothermal systems, which are embedded in a highly heterogeneous host medium.