River ice jams present a serious threat to life and property. The advancement of knowledge pertaining to river ice jams is impeded by a lack of quantitative data describing their evolution. Venturing on top of an ice jam for the purpose of collecting data is prohibitively dangerous and due to the difficulties associated with the prediction on the timing and location of their occurrence, organizing data collection efforts are not logistically practical. The safe and economic alternative is to collect data on a physical model under controlled laboratory conditions.

A hydraulic model of ice jam processes and a data collection system were designed for the purpose of obtaining continuous time-series data describing water levels, ice jam thickness, flow velocities, ice cover progression, and discharge data during the formation of a model ice jam. The results of 40 tests are presented in this thesis. Two primary testing scenarios were explored: the first series of tests investigated ice jams formed under a constant supply rate of flow (carrier discharge); the second series of tests investigated ice jams formed by destabilizing an initial ice jam by a rapid increase in discharge to form a new thicker stable ice jam.

Detailed analyses of velocity data obtained under stable ice covers for eight different Canadian rivers were conducted to establish the feasibility of using unique point velocities to describe the channel average velocity. The success in finding these unique relationships for natural irregular channels extended confidence in the use of index velocity methods during the experimental investigations.

The work presented in this thesis represents the first comprehensive set of data describing the dynamic nature of ice jam evolution under a constant carrier discharge and the dynamic response of the resulting stable ice jam accumulations to a rapid increase in carrier discharge. This work also presents the first published attempt at obtain discharge estimates under an ice jam during formation. These data contribute to an improved understanding in ice jam evolution and will facilitate validation of dynamic ice jam computational models.