Abstract. Distributed and continuous catchment models are used to simulate water and energy balance and fluxes across varied topography and landscape. The landscape is discretized into computational plan elements at resolutions of 10 1 -10 3 m, and soil moisture is the hydrologic state variable. At the local scale, the vertical soil moisture dynamics link hydrologic fluxes and provide continuity in time. In catchment models these local-scale processes are modeled using 1-D soil columns that are discretized into layers that are usually 10 −3 -10 −1 m in thickness. This creates a mismatch between the horizontal and vertical scales. For applications across large domains and in ensemble mode, this treatment can be a limiting factor due to its high computational demand. This study compares continuous multi-year simulations of soil moisture at the local scale using (i) a 1-pixel version of a distributed catchment hydrologic model and (ii) a benchmark detailed soil water physics solver. The distributed model uses a single soil layer with a novel dualpore structure and employs linear parameterization of infiltration and some other fluxes. The detailed solver uses multiple soil layers and employs nonlinear soil physics relations to model flow in unsaturated soils. Using two sites with different climates (semiarid and sub-humid), it is shown that the efficient parameterization in the distributed model captures the essential dynamics of the detailed solver.

Basin response and hydrologic fluxes are functions of hydrologic states, most notably of soil moisture. However, characterization of hillslope-scale soil moisture is challenging since it is both spatially heterogeneous and dynamic. This paper introduces an entropy-based and discretization-invariant dimensionless index of hydrologic complexity H that measures the distance of a given distribution of soil moisture from a Dirac delta (most organization) and a uniform distribution (widest distribution). Applying the distributed hydrologic model MOBIDIC to seven test basins with areas ranging 10 0 210 3 km 2 and representing semiarid and temperate climates, H is shown to capture distributional characteristics of soil moisture fields. It can also track the temporal evolution of the distributional features. Furthermore, this paper explores how basin attributes affect the characteristic H, and how H can be used to explain interbasin variability in hydrologic response. Relationships are found only by grouping basins with the same climate or size. For the semiarid basins, H scales with catchment area, topographic wetness, infiltration ratio, and base flow index; while H is inversely related to relief ratio.

Abstract. Distributed and continuous catchment models are used to simulate water and energy balance and fluxes across varied topography and landscape. The landscape is discretized into plan computational elements at resolutions of 101–103 m, and soil moisture is the hydrologic state variable. At the local scale, the vertical soil moisture dynamics link hydrologic fluxes and provide continuity in time. In catchment models these local scale processes are modeled using one-dimensional soil columns that are discretized into layers that are usually 10−3–10−1 m in thickness. This creates a mismatch between the horizontal and vertical scales. For applications across large domains and in ensemble mode, this treatment can be a limiting factor due to its high computational demand. This study compares continuous multi-year simulations of soil moisture at the local scale using (i) a 1-D version of a distributed catchment hydrologic model; and (ii) a benchmark detailed soil water physics solver. The distributed model uses a single soil layer with a novel dual-pore structure, and employs linear parameterization of infiltration and some other fluxes. The detailed solver uses multiple soil layers and employs nonlinear soil physics relations to model flow in unsaturated soils. Using two sites with different climates (semiarid and sub-humid), it is shown that the efficient parameterization in the distributed model captures the essential dynamics of the detailed solver.