Investigating electrical polarization signatures of sugar beet and maize: A field study using spectral electrical impedance tomography

Despite their vital role for agricultural management practices and plant breeding experiments, it is still challenging to characterize plant roots non-invasively in their natural environment. A promising new method for plant root characterization is the spectral electrical impedance tomography (sEIT) method, which is able to image the conductive and polarizable subsurface properties with high spatio-temporal resolution. Electrical polarization signatures have been shown to be sensitive to root structure and activity, although superimposed soil signatures complicate the interpretation. Recent studies have demonstrated that impedance measurements can be used to estimate root traits under laboratory conditions, especially in hydroponic experiments. However, field studies using sEIT on plant-root systems are still scarce.

In this study we present a field dataset of multi-frequency sEIT measurements on sugar beet and maize. Three different growth stages were measured during a whole growing season. We performed complex resistivity inversions for each measurement frequency, and subsequently analyzed the spatially resolved spectral response using a Debye decomposition analysis. Characteristic relaxation times, extracted from the spectral analysis, serve as proxies indicating the length scales of the observed polarization processes. We find that the physiologically different plant root systems cause distinct polarization responses in the low-frequency range. While both root systems exhibit an increasing polarization response towards higher frequencies, sugar beet develops an additional low-frequency polarization peak near 10 Hz later in the season, corrseponding with increasing size of the sugar beets. We attribute this peak to the polarization of root structures associated with the macroscopic dimensions of the beet roots, and demonstrate this link through the correlation of the retrieved mean relaxation time at the sugar beet positions with the square of the respective maximum beet diameter. Additionally, we evaluate the intrinsic spectral form of the polarization signatures extracted from the maize root area, and find a moderate correlation with the fresh biomass.

In conclusion, our results highlight that sEIT can be used in the field for plant root trait estimations, but structurally differing plants require different analysis procedures to extract root information. Additionally, environmental factors, like a varying soil composition or soil water content, have a strong influence on the measured impedance signal, and can make precise root trait estimation difficult.