Comparative testing of numerical models of river ice jams

Abstract

Ice processes in general, and ice jams in particular, play a dominant role in the hydrologic regime of Canadian rivers, often causing extreme floods and affecting the life cycle of many aquatic, terrestrial, and avian species. Various numerical models have been developed to help simulate the formation and consequences of these very dynamic and often destructive jam events. To test and compare the performance of existing models, a series of three tests have been devised and coordinated by a task force appointed by the Committee on River Ice Processes and the Environment (CRIPE). The results indicate that the overall performance of the models is good when calibration data are available, but considerably more varied when the models are applied in uncalibrated or “blind” mode. However, the diversity of results in the blind mode may be attributed more to the variations in perceptions of the physical processes by the users, as compared to the mechanics of the numerical computations.