Characterising faults in geothermal fields using surface waves: a numerical study

Subsurface characterisation of geothermal fields is important for the expansion of geothermal energy as a low-carbon resource. Faults and fractures provide secondary permeability, thus, their characteristics are crucial parameters in deep geothermal fields. Analysis of ambient seismic noise provides a relatively cheap and widely accessible method for constraining faults and fractures in geothermal settings.

 

Three-component (3C) beamforming is an array-based method that extracts the polarizations, azimuths, and phase velocities of coherent waves as a function of frequency from ambient seismic noise, offering a comprehensive understanding of the seismic wavefield. 3C beamforming can be used to determine surface wave velocities as a function of depth and the direction of propagation of waves. It is assumed that anisotropic velocities relate to the presence of faults, giving an indication of the maximum depth of the permeability essential for fluid circulation and heat flow throughout a geothermal field. Previous results suggests that some structures have a stronger effect on surface wave velocities than others. Numerical models are essential to study these relationships in more detail.

 

Here we present a numerical simulation of wave propagation through a model of the subsurface, with anisotropy depicted as faults. This is employed by a rotated staggered grid (RSG) finite-difference (FD) scheme. We model a homogeneous half-space with a fault-like structure (40 m width), changing fault parameters, such as depth, width, velocities and internal conditions of the fault (“fill”). We generate surface waves from a single source as well as multiple sources emulating an ambient noise wavefield.

 

We then use 3C beamforming on the synthetic data to characterise the modelled wavefield and observe the types of waves present. The polarisation and beam power of the synthetic data denote the composition of the synthetic wavefield and what percentage are retro- and prograde Rayleigh waves and Love waves. To investigate the strength of anisotropy introduced by a single fault we propagate surface waves across the fault in different directions, estimating velocities from array recordings using the beamformer. We are further able to assess the sensitivity of Rayleigh waves towards anisotropy at depth by considering Rayleigh waves at different frequencies sampling different depths.