STOCHASTIC MODELLING APPROACH TO IMPROVE ICE-JAM FLOOD RISK MANAGEMENT

Abstract

In the northern hemisphere, ice jamming can result in devastating flood events along many rivers. Understanding the physical and hydraulic processes of ice jam formation and predicting ice jam floodwater levels is a key requirement for ice jam flood management and planning. Over the years, river ice numerical modelling techniques have advanced and achieved unprecedented successes in simulating ice jam occurrences and associated impacts. However, there are still some limitations in understanding the impacts of model components on flood hazard delineation and finding a reliable modelling approach for implementing an effective ice jam flood mitigation strategy. Besides, there is a lack of a reliable modelling approach to quantify the severity of ice jam flooding along many rivers in northern communities under future climatic conditions. The main objectives of this dissertation are to address these research gaps using a stochastic modelling framework. A framework was developed to incorporate the model parameters, boundary conditions and digital elevation models (DEMs) in a global sensitivity analysis of flood hazard delineation. The global sensitivity analysis of these components shows that ice jam flood delineation is highly sensitive to DEMs and boundary conditions. The severity of ice-jam flood hazard under future climatic conditions has been quantified applying a novel modelling framework. The modelling results show that mean stage frequency distribution (SFD) is projected to be lower under the projected changes in climate in the period of 2041-2070 compared to the baseline period of 1971-2000 along the Athabasca River at Fort McMurray, Canada. Since the risk of ice jam flooding is exist both at present and in the future, a methodological framework was developed to explore appropriate mitigation measures to reduce ice jam flood risk along the Athabasca River at Fort McMurray. A total of three ice-jam flood mitigation measures (artificial breakup, sediment dredging and dike installation) were examined using a stochastic modelling framework for the potential to reduce ice-jam flood risk. The results show that, while sediment dredging may be able to reduce a certain level of expected annual damages in the town of Fort McMurray, Alberta, Canada, artificial breakup and dike system may be the most effective measure to reduce the amount of expected annual damages.