No abstract

Despite numerous applications of time‐domain reflectometry (TDR), serious difficulties remain in estimating accurate soil water contents under field conditions, especially in fine‐textured soils. We developed a physically based calibration model to predict the frequency‐ and temperature‐dependent complex dielectric response of soils. The model was used to predict frequency‐dependent attenuation and a single “effective” frequency approximation of apparent permittivity of the soil. Effective frequency was predicted to decline from 450 to 160 MHz as water contents increased from air dry to saturation. Predicted frequency decline was small for an input bandwidth of 130 MHz, reflecting that modeled polarization mechanisms associated with relaxation frequencies above 100 MHz were responsible for most of the frequency‐dependent attenuation. For specific surface areas ranging from 150 to 300 m2 g−1, simulations indicate that ignoring dielectric and conductive losses or the associated decline in effective frequency results in a 5 to 22% underestimation of the apparent permittivity. Both the power‐law and de Loor–Dobson mixing models gave a reasonable approximation to the measured apparent permittivity for a silty clay loam (34% clay) across the entire water content range. Moreover, the models were able to describe the behavior of apparent permittivity in response to temperature for two soils with contrasting bulk electrical conductivity contributions to losses. These results demonstrate that loss mechanisms and declines in effective frequency need to be considered to accurately predict the soil water content of fine‐textured soils.

No 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.

No abstract