Investigating River Ice Thickness and Analyzing Ice Breakup Flooding Events for Canadian Rivers

Ice formation in rivers within Canada and other colder climates causes changes to the river hydraulics; as the ice cover breaks, it can be challenging to predict the breakup event and potential unexpected flooding. Determining the timing of a breakup event is critical for the planning and mitigating flooding and subsequent damages. Hydrometric monitoring was performed for both the Thames River and the Humber River in Ontario, Canada, along with compiling a large dataset for different size rivers from the Canadian River Ice Database. By combining the traditional ice thickness estimation technique of Cumulative Degree-Days of Freezing, hydraulic flow properties of a river, and advanced machine learning methods, a new, more accurate formula for ice thickness estimation was developed. This new method was used in conjunction with ice stability index to forecast the timing of the ice breakup and jamming events for rivers of varying sizes. Given the recent developments in drone technology, drone photogrammetry was employed to accurately construct 3D models of the ice cover and ice-jam events. This research contributes to the application of new machine learning and remote sensing techniques in the field of river ice analysis and flood risk mitigation in an urban stream.