Overview
Flood risk estimates are inherently uncertain. Every model which estimates the e.g. 100-year flood depths in your area is a guess. It's straightforward for the modelers to determine how good that guess is---does that 7 feet of flood depth reflect a certain estimate of 7 feet, a plausible range of 5-9 feet, or a plausible range of 0-14 feet? Resampling methods reveal uncertainty in estimating 100-year streamflow or rainfall, and validation error approximates hydrodynamic modeling uncertainty in estimating resulting flood depths. Sensitivity analysis can reveal the impact of drivers which aren't modeled, such as soil moisture variability.
The Harm
Many commercial vendors actively hide their uncertainty information. This causes harm; if your actual flood risk is uncertain, so are the benefits of risk mitigation projects. Floodplain managers seeking to maximize the ROI or cost-benefit ratio will inevitably fail to do so under rational expectation if they do not have access to uncertainty information.
Pending Documentation
Pending documentation will use rainfall estimates from Atlas 14, the SFINCS open-source hydrodynamic model, and simplified HAZUS-based depth-damage modeling to characterize the expected financial damages caused by the withholding of uncertainty information, and further damages caused by failure to account for variability in soil moisture. These results will then be demonstrated with a user-friendly Jupyter notebook, with open source and readily reproducible workflow.