Ice jam behaviour has significant influence on the local ecology of inland deltas. The occurrence and/or absence of ice jams and the subsequent flooding can water or dry riverine delta floodplains and affect their ecological integrity. To better understand and quantify ice jams, historical and future ice jam behaviour is investigated and compared in this thesis from different perspectives. Modelling is a widely used method in the ice jam research field. A literature review on ice jam modelling research shows that previous studies made great advances in ice jam theory, however some research gaps are identified. Due to data limitations, ice jams in remote and data sparse areas are less understood. Similarly, understanding the impacts of climate change on future ice jam behaviour is still limited. To explore ice regimes in data sparse areas, an inland delta (the Slave River Delta) that lacks long term ice jam monitoring data was chosen as the study site. Input factor determination is a prerequisite for setting up numerical simulation of an ice jam. The inflowing ice volume, an input factor of the hydraulic model RIVICE, is determined from using remote sensing datasets as well as a simple ice thickness calculation method (the Cumulative Degree-Day of Freezing method). A calculation algorithm of ice volume estimation is proposed. After calibrating the RIVICE model for the Slave River Delta (SRD), two sensitivity analysis (SA) methods were implemented to evaluate the RIVICE model and to identify the sensitive parameters/boundary conditions. Ice thickness is identified as a sensitive input parameter of RIVICE, for which an ice thickness calculation framework (IceThick-RS) is developed. IceThick-RS, incorporating polarimetric parameters of C-band RADARSAT-2 images into the ice thickness calculation, is used to calculate river ice thicknesses at Fort Smith along the Slave River. The results calculated by using the IceThick-RS show better consistency with the gauged data than the Cumulative Degree-Day of Freezing method. A novel ice jam behaviour projection framework is developed to investigate climatic effects on ice jams. Future ice jam behaviour was quantified. The projected climatic dataset, CanRCM4-WFDEI-GEM-CaPA, was employed to drive the hydrological and hydraulic models to simulate the effects of future hydro-climatic conditions on ice jam behaviour. Trends of later ice jam initiation date and decreasing possibility of ice jamming were projected, which may exacerbate the drying trend in the SRD. On the other hand, higher extreme backwater elevations induced by ice jams in the SRD were also projected, which may increase the future ice jam flooding possibility. The inflowing ice volume estimation method, ice thickness calculation framework, and future ice jam behaviour projection framework are three important methodological contributions of this research.
Ice jam, numerical modelling, climate change, the Slave River, parameter sensitivity analysis, remote sensing
Doctor of Philosophy (Ph.D.)
School of Environment and Sustainability
Environment and Sustainability