Usually the flood frequency analysis uses the most significantly flood event in a year. In a statistical population of 100 years are exactly 100 flood events with different height or discharge. Each event isn't really a flood event. In a dry year the most significantly flood has an mean discharge.
Very often has a year two ore more sidnificantly flood events, but they were eliminated.
The idea is to count ALL flood-events, irrelevant the time when they occure. This might be have effect on the flood frequency with return periods between 2 and 10 years. This needs to determine a limit value and how to work with more than one peak during an event.
My question to peers is: