Search

Process-based models could be used to predict the soil saturated extract salinity from extractable ion contents. However, a rigorous validation of such models for this purpose had not been carried out. A process-based model to predict the main inorganic ion composition, electrical conductivity, and pH of the saturated paste extract from extractable ion contents was developed step by step. The model development started from the principle of matter conservation in the soil solution as it concentrates from the 1:5 to the saturated paste extract. The need to include new hypotheses in the model was studied through calculation and analysis of standardised differences between measurements and model predictions. Therefore, best estimates of saturation paste extract properties occurred after taking into account the following: (i) free equilibration of the soil solution with the minerals calcite and gypsum under CO2 partial pressure of the saturated paste; (ii) further equilibration of the soil solution with the exchange complex; and (iii) determination of salt contents within the diffuse double layer (DDL) of the soil colloids, i.e. the anion exclusion. The last was necessary because the extracts where the determination of soil extractable anions (1:5 extracts), cations, and cation exchange capacity (CEC) were carried out were separated from their suspensions through centrifugation. Therefore, the reliable prediction of soil saturated paste extract salinity demands data on soil extractable ion contents, CEC, and the quantification of salts within the DDL. © 2012 CSIRO.

Affordable, commercial dielectric sensors of the frequency domain reflectometry (FDR) and capacitance-conductance (CC) types estimate the dielectric permittivity (εb) and electrical conductivity (σb) of bulk soil. In this work, an equation was obtained to estimate the pore-water electrical conductivity (σp), which is closely related to the soil salinity in contact with plant roots, from εb and σb data, by combining the simplified dielectric mixing (SDM) model that relates εb to the soil volumetric water content (θ), with the Rhoades equation that relates θ and σb to σp. This equation was calibrated with measurements of εb and σb obtained with the Delta-T WET (FDR) and the Decagon 5TE (CC) sensors, in 20 pots filled with a clay loam soil and arranged as combinations of four levels of soil moisture with five levels of soil salinity. The calibrations were performed against reference θ and σp values. The σp was calculated with the chemical equilibrium model SALSOLCHEMEC and used as a more reliable reference than the electrical conductivity of the soil wetting water. For both sensors, the SDM model on the one hand, and the Rhoades equation on the other, provided the most accurate estimations using the least number of parameters regarding their respective alternatives, i.e. the third-order polynomial and the Hilhorst equation. The combined equation for estimation of σp subsequently provided root mean square deviations of 3.1 (WET) and 4.1 (5TE) dSm-1, which decreased to 1.5 and 2.6dSm-1 for θ >0.22m3 m-3, and σb 0.22m3 m-3 and σb