We explore the impact of roughness in crack walls on the P wave modulus dispersion and attenuation caused by squirt flow. For that, we numerically simulate oscillatory relaxation tests on models having interconnected cracks with both simple and intricate aperture distributions. Their viscoelastic responses are compared with those of models containing planar cracks but having the same hydraulic aperture as the rough wall cracks. In the absence of contact areas between crack walls, we found that three apertures affect the P wave modulus dispersion and attenuation: the arithmetic mean, minimum aperture, and hydraulic aperture. We show that the arithmetic mean of the crack apertures controls the effective P wave modulus at the low-and high-frequency limits, thus representing the mechanical aperture. The minimum aperture of the cracks tends to dominate the energy dissipation process and, consequently, the characteristic frequency. An increase in the confining pressure is emulated by uniformly reducing the crack apertures, which allows for the occurrence of contact areas. The contact area density and distribution play a dominant role in the stiffness of the model, and in this scenario, the arithmetic mean is not representative of the mechanical aperture. On the other hand, for a low percentage of minimum aperture or in the presence of contact areas, the hydraulic aperture tends to control the characteristic frequency. Analyzing the local energy dissipation, we can more specifically visualize that a different aperture controls the energy dissipation process at each frequency, which means that a frequency-dependent hydraulic aperture might describe the squirt flow process in cracks with rough walls.
Key Points:• We solve the quasi-static linearised Navier-Stokes equations coupled to elasticity equations • Seismic attenuation due to squirt-flow is strongly affected by the roughness of the crack walls • The minimum and the hydraulic apertures significantly affect the energy dissipation process presented a comparison between numerical results and an analytical model for squirt flow. In general, accepted analytical models should reproduce the equations of Gassmann (1951) in the low-frequency limit (Chapman et al., 2002). The reason is that at the relaxed state for undrained boundary conditions (low-frequency limit), the time of a half period of a passing wave allows for fluid pressure to equilibrate through FPD. At the unrelaxed state (high-frequency limit), the fluid pressure has no time to equilibrate during a half period of a passing wave and the elastic properties of the saturated material are predicted by the formulation of Mavko and Jizba (1991), which assumes that no FPD occurs during the passage of the wave. At intermediate frequencies, FPD occurs inside the cracks during the passage of the wave and part of its energy is dissipated. Nevertheless, all analytical solutions assume smooth walls for the cracks despite the fact that crack walls in rocks have been observed to present complex profiles including wall roughnes...