How can X-ray computed tomography help improve the performances of a preferential water flow model in soils?

Abstract

Preferential flow pathways in soil macropores allow fast, gravity-driven water transfer that can bypass most of the soil matrix. Still, some water, and the contaminants it contains, transfer from the macropores to the matrix. This water exchange is accounted for in 1D compartment models by coupling preferential and matrix flow with simplified, physics-based equations. Still, some parameters of these exchange terms need to be calibrated, often with unrealistic values. We propose a new macropore-matrix exchange term that considers the horizontal distribution of water in the soil matrix and overcomes this shortcoming. This term was defined as the product of a macropore-matrix specific wetted interfacial area and the water flux density from macropores. The former was estimated using X-ray Computed Tomography images. The latter was obtained by solving the Richards’ equation in a horizontal representation of the soil matrix. A water transfer model equipped with this exchange term performed better, when compared to experimental data, than a version using an exchange term based on the average pressure head in the soil matrix. It significantly improved the modeled temporal evolution of drained and stored water in the soil. Water exchange was maximum after the start of a rainfall and decreased towards zero before the soil matrix was saturated. This new pseudo-2D exchange term opens up the possibility of improving the modeling of water and contaminant retention at the macropore-matrix interface and of using values of the transfer-term parameters determined experimentally or calculated by another model.