Different methods of flood prediction

The main methods of estimating (predicting) floods are:

1. Envelope curves: the maximum observed floods of many catchments are plotted against area for a region, and the upper envelope is used for a quick estimate.
2. Empirical formulae: regional relations such as Dickens, Ryves, Inglis, Fuller and the WECS/DHM equations of Nepal, $Q=CA^n$, where A is the catchment area.
3. Rational method: $Q_p=CiA/360$ (A in ha, i in mm/h), for small catchments.
4. Unit hydrograph method: the design storm's rainfall excess is applied to a unit hydrograph (observed or synthetic) to get the flood hydrograph.
5. Flood frequency studies (statistical methods): a probability distribution (Gumbel, log-Pearson III, log-normal) is fitted to the annual flood series, and the flood of a given return period is extrapolated. Both analytical and graphical (plotting position) approaches are used.
6. Watershed models and routing: rainfall-runoff models (HEC-HMS etc.) for gauged and ungauged catchments, and regional flood frequency analysis for ungauged sites.