Assessment of Climate-Driven Trends in Freeze-Up and Midwinter Breakups of Canadian River Ice Using a Multimodel Ensemble Framework

Abstract Midwinter breakup (MWB) events have become increasingly common in Canadian rivers, with potentially catastrophic damage linked to flooding caused by the early breakup of river ice followed by the reformation of a new ice cover. The prediction of these events is heavily reliant on the conditions of the river at freeze-up, posing challenges in modeling these events against climate change scenarios. This study proposes an ensemble modeling framework to predict the conditions of a river ice cover at the initiation of freeze, followed by forecasting the occurrence and timing of an MWB event based solely on climatic data. This framework is tested and validated against data from 52 rivers in Canada, producing a comparable level of accuracy to models producing predictions based on a combination of river and climatic data. The framework is then applied to a series of climate change data from varying emission scenarios to analyze the trends in the occurrence and timing of these events for the years 2015–2100. Consistent trends are demonstrated in the late occurrence of both freeze and MWBs compared to current averages, while the frequency of MWB events shows either no trend or a decreasing trend depending on the severity of the emissions considered. The length of time between freeze and MWB occurrence is also shown to be increasing nationally, suggesting a trend toward the late occurrence of freeze, and the early occurrence of spring breakup. The framework itself is versatile and transferable, with integration of further river ice events being a future proofed option for further study.