Flood-loss modelling for assessing impacts of flood-frequency adjustment, Middle Mississippi River, USA

This study modelled flood losses (economic damages) along the Middle Mississippi River (MMR) (1) using current US government estimates of flow frequencies and (2) using frequencies based on the original, unaltered discharge measurements. The official flood frequencies were quantified in the Upper Mississippi River System Flow Frequency Study (UMRSFFS), but as a last step in that study, early discharges along the MMR were reduced by up to 54% to reflect a purported bias in early measurements. Subsequently, early discharge measurements were rigorously tested, and no such bias was found. Here, flood damages were quantified using a combination of one-dimensional hydraulic modelling and flood-loss modelling. For all recurrence intervals, damages were much less using the UMRSFFS flow frequencies compared with the frequencies based on the original discharge measurements, with differences ranging up to 79% (100-year event) and $2.9bn (200-year event). Annualized losses in the study area based on the UMRSFFS frequencies were just $41.6m versus $125.6m using the raw frequencies (an underestimation of 67%). These totals do not include flood losses elsewhere along the MMR, including in metropolitan St Louis. In summary, a seemingly small methodological adjustment in this case, a single hidden adjustment, not documented anywhere within the UMRSFFS can have dramatic societal impacts in terms of underestimation of flood probabilities and flood risk. Copyright (C) 2012 John Wiley & Sons, Ltd.