Stéphane Ruy

Predicting preferential water flow in soils by traveling-dispersive waves

Abstract Rapid preferential drainage or by-pass flow of water and pollutants occurs in soil macropores such as burrows and channels formed by earthworm activity in soils. We show that preferential flow through these non-capillary pores can be described by a traveling-dispersive wave. This wave is the solution of a non-linear convective–dispersive equation (KDW model) that depends on three transport parameters: two are related to a convective celerity and the other one is a dispersive coefficient. We show that the flux–mobile water content relation is hysteretic and that it can be described by a non-linear function of the mobile water content and its first time derivative. By combining the latter relation with the continuity equation we derive the KDW model. This model can be viewed as a second-order correction of the purely convective kinematic wave model. The dispersive term incorporates the large-scale effects of dissipative forces without resolving the small-scale conservation equations in detail. We further present numerical solutions for the signaling problem and a direct method for estimating model parameters. The model is validated with data obtained from laboratory infiltration experiments on soil columns. The experiments were carried out in repacked soil columns inoculated with Allolobophora chlorotica earthworms. Varying rainfall intensities were applied at the top surface of the columns with a rainfall simulator. Both the mean of mobile water content within the columns and the drainage hydrograph at the bottom were recorded in time. The parameters of the model were estimated from the experimental flux–mobile water content relation. A very good agreement was found between model prediction and experimental data.

Assessment of earthworm contribution to soil hydrology: a laboratory method to measure water diffusion through burrow walls

The capacity for water diffusion in burrow walls (i.e. the coefficient of sorptivity) either burrowed by Lumbricus terrestris (T-Worm) or artificially created (T-Artificial) was studied through an experimental design in a 2D terrarium. In addition, the soil density of earthworm casts, burrow walls (0–3 mm around the burrow) and the surrounding soil (>3 mm) were measured using the method of petroleum immersion. This study demonstrated that the quantity of water which transits through burrows of L. terrestris in the soil matrix was lower than that transited through soil fractures, due to a reduction of soil porosity in burrow walls (compaction: cast > worm’s burrow walls > surrounding soil > artificial burrow walls). Earthworm behaviour, in particular burrow reuse with associated cast pressing on walls, could explain the larger burrow wall compaction in earthworm burrows. If water diffusion was lower through the compacted burrows, burrow reuse by the worms makes them more stable (worms would maintain the structure over years) than unused burrows. The present experimental design could be used to test and measure the specific differences between earthworm species in their contributions to water diffusion. Probably, these contributions depend on the presumed related-species behaviours which would determine the degree of burrow wall compaction.

Unsaturated hydraulic conductivity of structured soils from a kinematic wave approach

Measuring soil hydraulic conductivity in the immediate vicinity of soil water saturation remains challenging, particularly for structured agricultural soils. As a consequence, it is not possible to accurately predict infiltration and drainage rates near saturation with the Darcy-Richards model. Since these rates are accurately predicted by the kinematic wave approach near saturation in structured soils, we assume that the flux-water content relationship could be used to assess hydraulic conductivity within this range of soils and moisture content. Two types of experiments were performed on a structured loamy clay soil in the laboratory: (i) infiltration experiments on samples at laboratory capacity, in order to apply the kinematic wave theory, and (ii) evaporation experiments on saturated samples, in order to assess hydraulic conductivity using Winds method. Estimating the relationship between the flux of drained water and macropore moisture content according to the kinematic wave theory requires the knowledge of two parameters: a macropore-flux distribution index a and a conductance term b. These parameters were fitted using both a current and a new method. Currently parameters are estimated from the falling limb of the drainage hydrograph whereas the new method uses from the evolution of the drained water flux versus macropore saturation after the end of rainfall. Results have shown that the new method is more reliable than the current one. The estimation of kinematic wave parameters is not influenced by rainfall intensity but depends on initial water content and temporal evolution of the soil sample. Hydraulic conductivity data assessed by both the kinematic wave theory and Darcys theory seem consistent at values close to “laboratory capacity”, defined in this work as volumetric water content after one-day draining. Thus, the kinematic wave approach could be a reliable tool to assess the hydraulic conductivity of macroporeus soils near saturation.

Numerical modelling of water infiltration into the three components of porosity of a vertisol from Guadeloupe

We present a mechanistic model of soil deformation and water infiltration into a Vertisol of Guadeloupe (French West Indies), which accounts for the three components of porosity of this soil (matric, structural and macro-cracks). Time and space scales are respectively of several hours and the prism delimited by macro-cracks. The model accounts for water movements from the structural porosity and from the macro-cracks into the matric porosity. It simulates the non-equidimensional deformations of the prism resulting from water storage in the matric porosity. Inputs of the model, measured in the laboratory, are the shrinkage curve, the retention curve and the hydraulic conductivity of the matric porosity. The anisotropy ratio of the soil deformation was measured in situ. Experiments were conducted in situ to provide structural parameters and data to fit and test the model. It is possible to find a unique set of parameters for each experiment. However, parameters significantly differ from one experiment to another. The model shows that the structural water flow regulates the partition of water infiltrating within the prism and of water flowing in the macro-cracks. The model does not accurately predict water infiltration because of a poor modelling of water flow in the structural porosity.