Probability‐interpretation of the observed return‐periods of floods

The purpose of this article is to show by numerical examples how the problem of flood‐flows can be treated from the view‐point of probability without making any artificial or arbitrary assumptions whatever. It is customary in hydrology to calculate either the observed recurrence or the observed exceedance‐intervals of floods. In a recent article [see 3 of “References” at end of paper] we have shown that there is only one theoretical value which corresponds to both. We call it the return‐period. It is a statistical function as is the distribution or the cumulative frequency. Therefore these calculations are related to the distribution of flows. We have reproduced this distribution by the theory of largest values. As the flood‐flows are the largest flows within N = 365 daily observations, this method assumes that N = 365 is a large number and that the discharges can be considered as unlimited statistical variables.