Three-dimensional numerical model validation: Issues and directions

Recently, numerical models have been used more frequently in hydraulic design projects. This is particularly true in the design of fish passage facilities. Advances in numerical modeling methodologies and the ability to collect high-resolution position data of salmonids within a reservoir have led to interesting issues of numerical model calibration. Presently, more emphasis is being placed on higher-order hydrodynamics, such as accelerations, turbulence and energy dissipation. Therefore, it has become subsequently more important to understand the ability of a numerical model to predict such quantities accurately. The purpose of the paper is to describe the detailed approach taken to calibrate a fully three-dimensional numerical model of two main stem Columbia and Snake River dams, namely Rocky Reach Dam and Lower Granite Lock and Dam. The paper presents detailed calibration methods that show validation of the numerical model to field and physical model velocities; field measurements of velocity fluctuations due to unsteadiness as well as turbulence; and the effect of surface roughness. Additionally, data will be presented showing the impact of scaling factors and limitations of spatial boundaries from physical models used to provide validation data. The paper concludes by discussing the need for accepted standards for numerical model validation. Copyright ASCE 2004.