Performance of frequency-domain reflectometer, capacitance, and psuedo-transit time-based soil water content probes in four coarse-textured soils

Frequency-domain reflectometry (FDR), capacitance (C), and psuedo-transit time-based (PTT) soil water content (SWC) probes have gained a broad acceptance for measurement of volumetric SWC (χ _v) in agricultural fields and other applications. The χ_v of typical coarse-textured soils can range from as low as 0.02 m³ m^-3 at permanent wilting point to 0.1 m³ m^-3 at field capacity. Thus, a small change in χv, would result in a change of a large portion of the total available water. Therefore, the SWC probes used to estimate χ _v in coarse-textured soils must be accurate and sensitive enough to detect small changes in χ_v. Few data that analyze the performance of these probes in coarse-textured soils (>90% sand) exist. The objectives of this research were to: (i) evaluate the validity of the calibration curves provided by the manufacturers for three SWC probes (CS615-FDR®, TRIME-PTT® and PRI-C®) for estimating χ_v in four coarse-textured soils, and (ii) identify whether the performance of the probes is sensitive to soil type. Soil drying cycle experiments were conducted using three fine-sandy soils (Astatula, Ellzey, and Arredondo) and a fine sandy-loam soil (Dothan). Measurements began at near-saturation and continued until the χ_v in each soil was near wilting point. The probe-estimated χv, data were statistically compared with measurements of χ_v by the gravimetric method (χ _v-grv). Significant variations were observed in probe performance between soils. The CS615-FDR probe provided the most accurate χ_v estimates. Performance of the TRIME-PTT probe was lower than that of the CS615-FDR in all soils and its estimates were significantly different than the measured χ_v-grv, values in all soils. The PRI-C probe provided the least accurate estimates overall and the root mean squared error (RMSE) ranged from 0.04 m³ M^-3 in ElIzey soil to 0.05 M³ M^-3 in other soils. Its estimates were significantly larger than the measured χ_v-grv, in all soils, except Dothan. The CS615-FDR and TRIME-PTT probes produced their least accurate estimates in Dothan soil, which had the largest clay content (22.7%) among all soils. The variations in the performance of the probes between the soils suggest, in part, that the FDR and PTT-based methods are not completely independent of the soil type and are negatively affected by the clay content of the soil. All probes should be calibrated for an individual soil under consideration to obtain accurate 0v estimations in coarse-textured soils. The calibration curves developed in this study should provide useful technical information to users when working with similar SWC probes and soils that have similar physical characteristics to those used in this research.