Factors affecting choice of distribution for flood series

Rival candidate distributions for modelling flood series may give markedly different magnitude-return period (Q-T) relationships, particularly at high T, A priori theoretical arguments, empirical criteria for distribution choice and previously used methods of discrimination between candidate distributions are discussed. The effects which method of parameter estimation, treatment of outliers, inclusion of large historical flood values, data transformations and causal composition of the flood population may have on choice of distribution are also considered.