LAUSR

Abstract River ice processes are challenging to study due to the risks and costs associated with data collection in harsh winter conditions. Numerical models can provide valuable insight into a river's current ice regime where data is sparse. At present, most existing one-dimensional (1D) ice process models are based on an implicit finite difference solution to the Saint-Venant equations. As a result, highly dynamic events such as rapid ice jam formation or sudden ice jam release are difficult to model due to numerical instabilities that can arise if the flow approaches supercritical. This paper presents recent developments to the University of Alberta's public-domain river ice process model, River1D. The model was reformulated to accommodate natural channel geometry. This new natural channel geometry version of the model was then enhanced to include water supercooling, frazil accretion, frazil re-entrainment, anchor ice formation and release, border ice formation, under-cover transport of frazil, and ice cover formation based on leading edge stability criteria. The model was applied to the Susitna River, Alaska. The model results agreed favourably with observations of water levels, flows, water temperatures, surface ice concentrations, border ice widths, ice cover progression rates, and ice thicknesses.