The capability of derived probability distribution function

One of the concerns in Iran is the estimation of peak flows with different return periods in watersheds having insufficient or no hydrometric data. In this paper the ability of using combined probability of daily and five days before rainfall for deriving peak flow probability distribution function was investigated at Taleghanrood-Glinack watershed. For this purpose the 144 observed peak flows and their daily and five days antecedent rainfalls were cted and the combined probability of both of them calculated and realized equal with the accidence probability of those similar peak flows. The peak flows and their probabilities plotted and shifted to the condition that the frequency of mean peak flows equal to 50% and the mathematic equation of envelope curve of recession branch of stochastic curve was created. Based on the prepared equation the peak flows with different return periods were estimated. Then, based on 30 annual series of peak flows which were independent of former used data, fitted with seven general probability distribution curves and the peak flows with different return periods estimated and compared with synthetic quantities of peak flows. The result shows that the distribution function that extracted based on envelope curve of frequency of peak flow without corresponding to mean has the best compatibility with probability distribution of Log Normal two parameters. Contemporary the application of cumulative combined probability curve causes large error to estimate of peak flows with different return periods.