Flood Frequency Hydrology: Navigating Uncertainty in Flood Design

Flood frequency analysis is a cornerstone of hydrologic studies, providing a probabilistic framework to relate the magnitude of extreme events to their frequency of occurrence. This methodology is critical for designing flood-related infrastructure, conducting economic evaluations of flood control projects, and delineating floodplains. However, its utility depends heavily on data quality, model selection, and parameter estimation, each of which introduces uncertainties that become especially significant for rare events.

This presentation will address key sources of uncertainty, including model choice, parameter inference methods, and sample size limitations. Strategies for incorporating these uncertainties into engineering practice are discussed, with an emphasis on probabilistic representations and innovative design approaches. An exceptional flood, a "black swan" event, is used to illustrate the paradox of increased uncertainty despite improved information. This case underscores the importance of expanding flood analyses through historical records, regionalization, and causal modeling, particularly in the context of a changing climate.

The presentation will be designed to foster cross-discipline exchange in the quantification of uncertainty in Earth Sciences.