Forschungsprojekt ::

SPP 2089. Rhizosphere Spatiotemporal Organisation – a Key to Rhizosphere Functions. Teilprojekt P25: Mehrskalenmodellierung mit veränderlicher Mikrostruktur: Ein Ansatz zur Emergenz in der Rhizosphäre mit effektiven Bodenfunktionen

Projektbeschreibung

The self-organization of soil aggregates by various attracting forces influenced by geochemistry, and microbiology is studied by a novel, comprehensive model on the rhizosphere scale that explicitly represents the pore structure. To take into account specific properties of the rhizosphere the functional range of existing models for aggregation or carbon turnover has to be extended and, e.g. an explicit phase of mucilage, root dynamics or C/N specific microbial growth included. The project aims at the development of a mechanistic modeling approach that allows for dynamic structural reorganization of the rhizosphere at the single root scale and couples this evolving microscale model to the root system scale including the inference of soil functions. The SPP collaborations help to formulate the process mechanisms, and the CT scans reveal the pore structure in different settings. We model transformation processes in the rhizosphere and describe them in a rigorous way including couplings to geochemistry, microbiology, and finally to soil functions. Thus we gain insights on the development of self-organization in the rhizosphere in connection with spatiotemporal patterns of nutrients, water and biomass. The persistence of soil structure formation is studied as a function of the presence and development of gluing agents and soil carbon dynamics.
Due to its high complexity, the rhizosphere model is not amenable to large scale computations. The homogenization technique enables us to incorporate information from the rhizosphere scale on the macroscale. We advance upscaling techniques to dynamically evolving microstructures taking the spatiotemporal evolution of the rhizosphere into account. The derivation of effective parameters (e.g. diffusivity, permeability) can be an important input in existing root-water uptake models. We need a numerical approximation of the derived mechanistic model and its integration in an efficient software tool using massively parallel architecture.