Search

Merritt, B.C. has a history of suffering damages from ice-jam and open-water floods. The most damaging flood on record happened in the city in 1991 due to an ice jam on the Coldwater River. A numerical ice-jam model is calibrated using the field data collected during the 1991 flood event, and the resulting model coefficient values are shown to be comparable to what has been found in previous applications. With the 'calibrated' input, the model is then applied to different dyking scenarios as a means of protecting the city from future ice-jam events. It is shown that the common approach of placing dykes at the river's edge is not a satisfactory solution because it can aggravate winter floods by enhancing peak water levels, and augmenting ice-jam surge severity. This difficulty can be addressed by setting the dykes back for suitable distances, thus providing for adequate ice storage, and reducing the length of ice jams.

Following a brief account of the development of ice jam stability theories since the 1960s, inconsistencies in their interpretation for practical applications are identified. In particular, the widely used 'bell-shaped curve' is often applied under conditions that do not satisfy the assumptions made in order to derive it. Using the results of later theoretical and numerical developments, it is shown that such applications can lead to significant errors. Analytical predictions of the equilibrium thickness of ice jams should be confined to equilibrium reaches. A simple rule is formulated to resolve the ambiguity of the bell-shaped curve when used to determine ice jam thickness. Use of numerical solutions to the differential equations of stability is recommended for predictions of nonequilibrium ice jam characteristics.