Inversion of Hydraulic Conductivity from Induced Polarisation, Part A: Methodology and Verification

Summary The estimation of hydraulic conductivity (K) from geophysical measurements is a key research topic in hydrogeophysics. In the laboratory, the Induced Polarisation (IP) method has a long story of successful K estimation, with two main approaches: i) the approach based on the link between K and imaginary conductivity σ" and ii) the approach based on the link between K and relaxation time τ. In this work, we study more in detail the estimation of hydraulic conductivity from TDIP data from the inversion point of view, re-parameterising the model space directly in terms of hydraulic properties, using both σ" and τ approaches for defining model parameters. We tested parameter correlations with the Markov chain Monte Carlo method, and performed linearized 2D inversions of synthetic data on an aquifer analog. Furthermore, we evaluated the quality of K estimation on borehole field data, against K estimates from grain size analyses and slug tests. The proposed re-parameterisation is well suited for inversion, because weak correlations exist among parameters. More important, the direct inversions of field data retrieve very good K estimates. We believe that this new inversion approach might open the way for reliable, cost-effective geophysical estimation of hydraulic conductivity in the field.