Joint Probability of Extreme Streamflow and Its Day of Occurrence

Abstract

The magnitudes of peak streamflows and their return periods are important considerations in infrastructure design and hazard mitigation. Periods of high floods are usually associated with seasons and infrastructures and mitigation plans are designed accordingly. Floods, however, are also possible, even though at a much lower frequency, during low-flow seasons. Knowing the actual probability of an extreme event, given the day of the occurrence, can help in the implementation of nonstructural measures for minimizing flood damage, guide the design of more efficient mitigation plans, and optimize resources. Flood risks derived from annual maxima does not give any indication of the actual probability of experiencing a certain flow event in a given season, month, or day. In this note, a simple method of computing the nonexceedance probability of different flows of interest for any day of the year is presented. For example, the nonexceedance probability for a flow of 20,000 m3/s in the upper reach of the Lower Mississippi River Basin, obtained from water year maxima, is 0.02, but when derived from daily flows, the nonexceedance probability of experiencing such a flow in April oscillates between 0.42 and 0.56. Hence, while a frequency duration curve derived from annual maxima is useful for structural considerations, using the complete daily time series for the computation of the daily probability of flooding provides auxiliary information that can support nonstructural flood mitigation planning.