Though mitigation measures and research have increased over the last few decades, ice jams and associated flooding continue to be one of the most underestimated disasters in many northern countries. Operational ice jam flood forecasting systems are becoming one of the more prominent tools used in mitigating ice-related flood risk within Canada. Several forecasting systems have been adopted across the country and forecasters are constantly looking to improve the accuracy and consistency of their systems. The Lower Red River in Manitoba has been the subject in discussion of many ice jam related studies, and a data-driven ice-jam hazard forecasting system is currently in use at this site. This system differs from hydrologic model driven forecasting systems used for other ice jam prone rivers across Canada. This study focuses on identifying the methodology of the data driven ice jam flood forecasting system, along with the methodology of the forecasting procedures. Furthermore, the effectiveness of the data driven forecasting system is measured and assessed for the Lower Red River's 2020 breakup season.

Flooding is one of the most frequent and most costly natural disasters that occur throughout Canada, and although there is ongoing work to update and improve flood hazard assessments and mapping of high flood risk rivers throughout the country, most studies only delve into open water flooding. However, many rivers in Canada experience higher peak water levels due to ice jamming, resulting in severe flooding of surrounding areas. Hence, there is an urgency to expand current flood hazard assessments to include ice jam flooding for better flood management practices. One area that is often plagued with ice jam flooding is the lowest reach of Manitoba’s Red River. The Lower Red River is a low-lying river with a terminus inland delta where water levels are governed by Lake Winnipeg. Ice jam floods often divert water into the lower Red River’s floodplain that is continually being encroached by development. RIVICE, Environment Canada’s one-dimensional ice hydraulic model, was set up within a Monte Carlo framework to simulate an envelope of backwater level profiles that result from ice jams within the study site. Non-exceedance probability profiles were created from the envelope of backwater level profiles to assess ice jam flood hazard.