From hydraulic root architecture models to macroscopic representations of root hydraulics in soil water flow and land surface models

Abstract. Root water uptake is an important process in the terrestrial water cycle. How this process depends on soil water content, root distributions, and root properties is a soil–root hydraulic problem. We compare different approaches to implement root hydraulics in macroscopic soil water flow and land surface models. By upscaling a three-dimensional hydraulic root architecture model, we derived an exact macroscopic root hydraulic model. The macroscopic model uses the following three characteristics: the root system conductance, Krs, the standard uptake fraction, SUF, which represents the uptake from a soil profile with a uniform hydraulic head, and a compensatory matrix that describes the redistribution of water uptake in a non-uniform hydraulic head profile. The two characteristics, Krs and SUF, are sufficient to describe the total uptake as a function of the collar and soil water potential, and water uptake redistribution does not depend on the total uptake or collar water potential. We compared the exact model with two hydraulic root models that make a priori simplifications of the hydraulic root architecture, i.e., the parallel and big root model. The parallel root model uses only two characteristics, Krs and SUF, which can be calculated directly following a bottom-up approach from the 3D hydraulic root architecture. The big root model uses more parameters than the parallel root model, but these parameters cannot be obtained straightforwardly with a bottom-up approach. The big root model was parameterized using a top-down approach, i.e., directly from root segment hydraulic properties, assuming a priori a single big root architecture. This simplification of the hydraulic root architecture led to less accurate descriptions of root water uptake than by the parallel root model. To compute root water uptake in macroscopic soil water flow and land surface models, we recommend the use of the parallel root model with Krs and SUF computed in a bottom-up approach from a known 3D root hydraulic architecture.