Development of an ice jam database and prediction tool for the Lower Red River

Abstract The Lower Red River in Manitoba regularly experiences springtime ice jam flooding, with the most severe events occurring between Lockport and Netley Lake. A database of ice jam events was developed through newspaper archives and historical stage data. Each event was given a severity rating from 1-5, based on the resulting ice jam flood. This facilitated an investigation of ice jam timing and frequency on this section of the Lower Red River and the development of a threshold-based ice jam prediction tool. Out of 54 ice jam events from 1962-2017, all ice jam events occurred when the peak spring flow exceeded 1000 cms and all severe events (severity 3+) occurred when peak spring flows exceeded 1500 cms. The threshold ice jam prediction model was developed using five meteorological and hydrometric parameters including accumulated degree day of thaw, accumulated degree day of freezing, freeze-up water level, snow on ground, and rain equivalence. The model was able to differentiate all severe event years from non-event years with only one false positive result. By gaining a better understanding of ice jamming in the area, this research provides an accessible prediction method that can help guide decisions related to the risk and severity of spring ice jamming.