What does return period mean on an IDF curve?
In short: A T-year return period corresponds to an annual exceedance probability of 1/T under the fitted model. A 100-year rainfall intensity therefore has about a 1% modeled chance of being exceeded in any one year. It is not an event scheduled to occur exactly once every 100 years.
Return period and non-exceedance probability
For frequency analysis, the tool evaluates the fitted inverse cumulative distribution at the non-exceedance probability p = 1 − 1/T. Each duration has its own maximum series and fitted quantile, so a return-period line connects estimates across storm durations.
Why record length matters
A return-period estimate can be much longer than the available rainfall record, but it then relies heavily on extrapolation. The tool warns when a selected return period is greater than twice the record length. It still calculates the value and annotates exports, so the analyst must decide whether the uncertainty is acceptable for the intended design use.
The EPP method is more restricted: because it reads ranks directly from the observed sample, it cannot estimate a return period beyond approximately n + 1 years for a sample of n maxima.
See the return-period tutorial and frequency-analysis formulas.
Calculate standard or custom return periods with the IDF Curve Tool.