Probability analysis of historical ice jam flood data for a complex reach: a case study

Abstract

Two obstacles are encountered in the probability analysis of ice jam floods: the first is the general lack of long-term hydrometric records in cold regions; the second, which compounds the first, is that the few records that are available for ice jam floods are difficult to transpose to even nearby sites. This means there must be heavy reliance placed on local historical data.A good illustration of the effort required and the rewards obtained in collecting and analyzing such data is provided by the situation at Hay River, N.W.T. The flood population at this site is completely dominated by ice jam floods for which there is no standard hydrometric data.This paper describes the approach used in the collection and analysis of the historical data at this site to define the flood probability distribution. A comparison of the resulting probability distribution with the previously designated flood zone, which was defined simply on the basis of the historical high water level, reveals the risk involved in designation based on the latter approach and demonstrates that even in the complex and apparently data-limited situation at Hay River, the systematic collection, collation, and probability analysis of historical data is well worthwhile. Key words: ice jams, historical flood analysis, probability analysis, hydrology, river floods, river ice.