Imaging spectral electrical properties of variably saturated soil columns

Accurate knowledge about the spatial distribution of soil h ydraulic parameters is highly important for modelling the movement of water and solutes in saturated and unsaturated soils on several scales. Electrical impedance tomography (EIT) is a promising tec hnique to non-invasively acquire information on hydraulic conductivity, because the measured complex conductivity is sensitive to soil pr operties that also contro l the hydraulic properties (w ater saturation, porosity, specific surface and pore/grain size distribution). Ev en more information is available if EIT is performed in a spectral manner (i.e. performing EIT measurements over a range of frequencies), which yields information on the spa tial distribution of the spectral electrical properties of the soil. Relationships between spectral electrical and hydr aulic parameters have s hown strong correlations in several studies. Our ability to obtain such distributions of spectra l electrical parameters in an EIT inversion approach is influenced by several factors, among which data accuracy, sensitivity and regularization strength and type are typically the most importa nt. Numerical studies of the reproducibility of spectral electri cal parameters in an inversion a pproach showed that an effective electrical relaxation time should be less affected by these factors than chargeability or DC- conductivity. In order to i nvestigate this finding experimentally a nd to test the actual reproducibility of spectral electrical pr operties in an imaging approach, EIT measurements were conducted on a range of packed soil columns with materials whic h were well defined in terms of their spectral electrical properties using effective spectroscopic measurements.