Hyper-Resolution Simulation of Field-Scale Physical Processes in Watershed-Scale Integrated Hydrologic Models.

Historically, upscaling methodologies to represent local-scale physical flow and transport processes has been needed in order to incorporate their effects at the watershed scale in hydrologic models. While such an approach has been shown to work, upscaling often involves the development of empirical (i.e., parameterized) relations that are governed by intractable parameters that cannot be easily measured. This leads to the extensive calibration of the resultant models to tune parameter values, which are therefore constrained to the individual study areas from which observations are made. As efficient and robust numerical methods continue to evolve, and high performance computing (HPC) resources become more accessible to the hydrologic community, it is becoming increasing achievable to directly incorporate local-scale process detail into physics-based watershed scale models. In this work, we present an integrated subsurface/surface flow and transport model, based on the HydroGeoSphere (HGS) platform, for a highly characterized agricultural watershed in Ontario, Canada. In the HGS model, physical features that are typically simplified or upscaled in order to facilitate larger-scale hydrologic modeling (such as macropores, tile drains, and soil horizonation) are represented with a high level of numerical detail .The HGS model is composed of over one million finite element mesh nodes, 27 thousand discretely-defined line elements representing individual tile drains, and eight soil layers of which four are considered macroporous. Results produced by the hyper-resolution HGS model show that with a minimal level of physical parameter tuning during model calibration, simulated watertable depths, stream flow rates, and soil moisture levels across the 5 km² experimental watershed match closely with observed data. This work supports the concept that hydrologic modeling at large scales while still representing local-scale processes is tractable and more defensible by taking advantage of available HPC and modern, physics-based simulation tools.