Towards a Generic Method to Parameterize Hydrologic Land Surface Schemes

One of the challenges in large scale physically based and semi-distributed hydrologic land surface schemes (HLSSs) is how to parameterize, in a computationally efficient manner, the spatial variability of the landscape and other physiographic characteristics of the catchment. In hydrological models, hydrological similarity is commonly used to identify and group units that have a locally uniform hydrological response to meteorological input. The grouping is designed to balance an accurate representation of the spatial heterogeneity of a catchment with realistic computational demand. The basis of hydrological similarity in HLSSs is dependent on coupled horizontal and vertical processes. A generic method of parameterization is developed and tested whereby classifications based on vegetation are used for parameters related to the vertical land surface processes and classifications based on ecologically distinctive areas are used for parameters related to the horizontal hydrological processes and the vertical sub-surface processes. The ecologically distinctive areas are based on geological, landform, soil, vegetative, climate, wildlife, water, and human factors. The approach: 1) maintains a balance between an efficient parameterization and a physically-based model, 2) reduces the need for model calibration, and 3) allows for the transfer of parameters to other basins. Application of this method to three sub-basins of the South Saskatchewan River Basin (SSRB) in the Rocky Mountains shows that this approach provides consistent and satisfactory stream flow predictions with minimal calibration, and making optimal use of easily obtainable landscape information.