One of the main obstacles to the routine application of physically-based hydrological models is the lack of data with which to initialize, drive, and evaluate model simulations. Consequently, the modeler encounters a large number of degrees of freedom when trying to match observations with simulations, and it is unclear whether the right results are achieved for the wrong reasons.
The BOREAS large-scale field experiment, conducted from 1993 - 1996 in Central Canada, has provided excellent opportunities to improve our understanding of hydrological processes over a range of scales. The intensity, scope, and breadth of the observations collected, limits the number of degrees of freedom a modeler has available to match observations with simulations.
As part of a study of energy and moisture fluxes in the boreal forest, BOREAS data were used to analyze and improve the formulation of the Distributed Hydrology-Soil-Vegetation Model (DHSVM). Initial application of the model showed that when the choice of model parameters was constrained by BOREAS field measurements, DHSVM tended to melt the snow too early under the forest canopy. This problem was traced to the manner in which partial canopy closure was handled in canopy radiative transfer computations. In addition the model exhibited a bias in the diurnal cycle of the sensible heat flux, which was traced to problems with the ground heat flux algorithm. Improvements in the calculation of radiative transfer of shortwave through the canopy, and a new soil heat flux algorithm, led to substantial improvements in model results