Evaporation Study for Real Soils Based on HYPROP Hydraulic Functions and Micro-CT-Measured Pore-Size Distribution

Abstract

Evaporation—a key process for water exchange between soil and atmosphere—is controlled by convective and diffusive surface fluxes that determine the functional time dependence of the evaporation rate (). Recent studies demonstrated that only a pore-scale surface flux model can capture the correct () curve. These studies also showed that a realistic estimate of the hydraulically connected region (HCR) of the pore-size distribution (PSD) is crucial for coupling surface flux to internal water flux. Since previous studies were often based on natural sands and glass beads, the main focus of our study was to test these conclusions for real soils. Therefore, we investigated the evaporation process within undisturbed soil columns of a sandy soil and loamy sand and measured the hydraulic functions via HYPROP experiments (a system to measure hydraulic properties using the evaporation method). Based on the isolated pore evaporation (IPE) model using a discretized form of the PSD, we developed a continuous IPE model and applied it to our experiments. Because the PSD plays a central role in the IPE model, we determined the PSD of the loamy sand soil via X-ray microtomography (μCT) for pores >19 μm. The consistency of the experimental data, i.e., (i) the retention curve for deriving the HCR of the pore size distribution, (ii) the unsaturated hydraulic conductivity for calculating the characteristic lengths of the evaporation process, and (iii) the high accuracy of the mass loss data strongly support the HYPROP method for this kind of complex evaporation experiment. The continuous IPE model describes the characteristic Stage 1 behavior well (functional form of the evaporation rate and length of Stage 1) for both soil types if a realistic HCR estimate is used that (i) is derived from a characteristic length analysis estimating the lower boundary of the HCR and (ii) the upper range of the HCR is based on the true PSD derived from μCT data.