Quantitative imaging of spectral electrical properties of variably saturated soil columns

Abstract

Measurements of the effective complex electrical conductivity in the mHz to kHz frequency range using Spectral Induced Polarization (SIP) have shown promise for the characterization of soil physical and biogeochemical properties and processes. In principle, the spatial distribution of the complex electrical conductivity can be determined using Electrical Impedance Tomography (EIT). However, little is known about the accuracy of EIT imaging results, which is known to be affected by the need for a consistent treatment of variable data quality for different frequencies in the inversion process. Therefore, we aim to evaluate the ability of EIT to accurately determine the distribution of spectral electrical conductivity in the mHz to kHz frequency range. For this, we used artificially packed lysimeter columns with known structure and spectral electrical properties. EIT imaging of these columns was performed during drainage in order to investigate data quality and the accuracy of soil spectral electrical property determination for different saturation and contact impedances of the electrodes. Spectral imaging results were analyzed using a Debye-decomposition approach and compared with effective SIP measurements on independent samples of the same material. Overall, there was a good correspondence between imaged and independently measured complex electrical conductivity. Furthermore, it was found that images of peak relaxation time distribution were much less affected by spatial smoothing than images of the normalized chargeability. It was concluded that state-of-the-art data processing and inversion procedures in combination with advanced EIT equipment can be used to accurately image the spatial distribution of spectral electrical properties.