Discrete Fracture Network geophysical imaging from Electrical Resistivity Methods

Fractures are fundamental discontinuities to understand the exchange and the dynamic of water, matter and energy between the different parts of the critical zone. But except from boreholes and direct observation at rocks outcrops, imaging fracture network from the surface remain a challenging task in geosciences. Classical geophysical methods and their associated imaging methods are in a large extend adapted to retrieve the properties and (part of) the heterogeneities of the rock matrix and not of the discontinuities.

The aim of the study is to present a stochastic imaging method based on the direct simulation of Electrical Resistivity Tomography (ERT) dataset from Discrete Fracture Network (DFN). ERT is one of most used geophysical method in critical zone studies. And one can find numerous examples in literature of interpretation of ERT images as fracture zone, even if the inversion do not have the resolution to retrieve the fracture geometry or properties.

First the direct approach is described (from Roubinet et al., 2014) to simulate ERT dataset. Then from simple cases such one vertical or horizontal fracture, the inversion is detailed with emphasis on the convergence rate, resolution and depth of investigation. Further step towards the inversion of realistic DFN are discussed: how prior information can be used to enhance the convergence rate and how the non-unicity of the solutions can be retrieved from the stochastic inversion scheme. The inversion of DFN from ERT dataset is a first step towards coupled inversion such as pumping tests or tracers tests and ERT.