Modeling root growth in heterogeneous soils

Javaux, Mathieu [UCL] Schnepf, Andrea [Agrosphere, Forschungszentrum, Juelich, Germany] Huber, Katrin [Agrosphere, Forschungszentrum, Juelich, Germany] Mrandage, Shehan [Agrosphere, Forschungszentrum, Juelich, Germany] Landl, Magdalena [Agrosphere, Forschungszentrum, Juelich, Germany] Vanderborght, Jan [Agrosphere, Forschungszentrum, Juelich, Germany] Vereecken, Harry [Agrosphere, Forschungszentrum, Juelich, Germany] et al.[show all]

Heterogeneity is a key characteristic of soils: pedological horizons, bio-pores, soil aggregates, etc. generate heterogeneous distribution of soil hydro-mechanical properties. Root development is affected by this spatial variability, and root growth, architecture, uptake and eventually yields are possibly impacted by soil heterogeneity. Yet, most of hydrological models, and in particular soil-plant models neglect heterogeneity to a certain point, for sake of simplicity. In this presentation, a detailed three dimensional root model is presented, which allows one to explicitly represent the interactions between root and soil for growth and water uptake based on bio-physical laws. Root development functions are parameterized based on root age and order. Root growth is affected by a soil strength tensor, which defines a preferential growth direction. This model is used to investigate how biopores (density and inclination) affect growth and root water uptake. Qualitative comparison is made to literature data. A sensitivity analysis reveals that root distribution and plant-water relationships are affected by the presence of macropores. Another application of such models is to assess how soil root sampling schemes in the field potentially bring information of root architectural traits. Virtual experiments were conducted to simulate root distributions in soils, and classical root sampling schemes were simulated. Inverse modeling with simpler architectural models to retrieve specific root traits demonstrates the sensitivity of current models to root specific traits and suggests that sampling schemes should be adapted as a function of root system architecture types.