Probability distributions for peak stage on rivers affected by ice jams

Evaluation of alternatives to reduce the severity of ice-jam flooding requires the development of probability distributions for ice-affected stage at non-gauged locations along a river reach. Difficulties arise as the ice-affected peak stage is not only a function of discharge and channel hydraulic properties, as is the case for the open water period, but is also influenced by the location, extent, thickness, and roughness of ice jams. This paper presents a new method for estimating winter peak stage along a river reach by dividing the record of winter peak discharges into distinct ice-jam and no-ice-jam populations. Rating curves may then be developed for each population using gradually varied-flow analysis and existing equilibrium ice-jam theory. Stage-probability distributions for the ice jam and no-ice-jam populations may then be combined into a single stage-frequency relationship representing the entire winter period. A case study is presented, in which the method is used to develop probability distributions for the annual maximum stage along the Winooski River at Montpelier, Vt. Viewed in the context of the historical record, the method yields reasonable results.