Evaluation of a Dielectric Sensor for Measurement of Soil-Water Electrical Conductivity

The salinity status of soil can be obtained through the measurement of the pore-water electrical conductivity sigma(p). In the present study the WET sensor's capability in predicting sigma(p) through the simultaneous measurements of the soil bulk electrical conductivity (sigma(b)) and the soil dielectric constant (K) is investigated. The estimation of sigma(p) is based on a model by Hilhorst, which relates sigma(p) to the two other quantities K and sigma(b) and an extra fitting parameter K-0 which is incorporated in the software of the device. The study involved experimental measurements in the laboratory using four different soils with a large range in soil texture and volumetric water content theta. In each soil type four different electrical conductivity aqueous solutions were used. The results exhibited a rather strong linear relationship between K and sigma(b). The slope of this linear relationship appeared to depend on both the soil type and the electrical conductivity of the pore water. The value of K-0 seems to be soil specific and increases when the salinity level increases. The sigma(p) prediction according to the linear model of Hilhorst deviates for almost all soils tested, except for the case of the sandy soil. Besides the linear model of Hilhorst, the Malicki and Walczak linear model was also tested for all soils under investigation and the model of Munoz-Carpena was tested for sand. Malicki and Walczak model performance at predicting sigma(p) values was approximately the same as the Hilhorst model for sand and slightly better for sandy loam. In general, one could argue that the linear models could predict sigma(p) with some accuracy for the cases of coarse porous media.