Given an imperfect forecast of the total runoff volume for a season, it is useful to determine the distribution of the forecasted runoff volume in each subperiod. A method is developed for deriving the joint distribution of the subperiod flows and the total seasonal forecast. Historical subperiod flows and the corresponding total seasonal forecast are transformed to the Gaussian domain via three‐parameter lognormal transformations. The transformed subperiod flows and total seasonal forecast are then modeled as multivariate normal, from which the conditional distribution of the runoff volume in each subperiod, given the total seasonal forecast, is obtaine