Soil Permittivity Response to Bulk Electrical Conductivity for Selected Soil Water Sensors

Abstract

Bulk electrical conductivity (σa) can dominate the low frequency dielectric loss spectrum in soils, causing changes in the permittivity and errors in estimated water content. We examined the dependence of measured apparent permittivity (Ka) on σa in contrasting soils using time‐domain reflectometry (TDR), a digital time‐domain transmission (TDT) sensor, and a capacitance sensor (5TE) during near saturated solute displacement experiments. Sensors were installed in columns packed with fine sand or a clay loam soil. Displacement experiments were completed by first equilibrating columns with 0.25 dS m−1 CaCl2, introducing a step pulse of ∼4.7 dS m−1 CaCl2 and, after equilibration, displacing the resident solution with 0.25 dS m−1 CaCl2. Using TDR, measured Ka increased with increasing σa; however, the slope of this response averaged 3.47 m dS−1 for clay loam compared with 0.19 m dS−1 for sand. The large response in the clay loam was attributed to relaxation losses that narrowed the effective bandwidth from 821 to an estimated 164 MHz. In contrast, the effective frequency in sand averaged 515 MHz. Permittivity measured using the TDT probe exhibited little or no sensitivity to σa (<0.32 m dS−1) in both media. Measured Ka using the 5TE probe declined with increasing σa up to 1 to 1.8 dS m−1 and then increased thereafter with net negative responses for sand (ΔKa/Δσa = −3.1 m dS−1) and net positive responses for the clay loam (ΔKa/Δσa = 2.9 m dS−1). Consideration of the Ka–σa response is required for accurate soil water content estimation (±0.03 m3 m−3) in the presence of solution EC variations using TDR in fine‐textured soils or the 5TE sensor in all media. Large differences in the sampling volumes between 5TE‐measured bulk EC and permittivity confounded the Ka–σa response in the presence of a concentration gradient.