Three different methods are compared to estimate the instantaneous peak flow (IPF) from the corresponding maximum daily flow (MDF), as the daily data are more often available at gauges of interest and often with longer recording periods. In the first approach, simple linear regression is applied to calculate IPF from MDF values using probability weighted moments and quantile values. In the second method, the use of stepwise multiple linear regression analysis allows to identify the most important catchment descriptors of the study basin. The resulting equation can be applied to transfer MDF into IPF. With the third method, the temporal scaling properties of annual maximum flow series are investigated based on the hypothesis of piece wise simple scaling combined with the generalized extreme value distribution. The scaling formulas developed from three 15 min stations in the Aller-Leine river basin of Germany are transferred to all daily stations to estimate the IPF. The method based on stepwise multiple linear regression gives the best results compared with the other two methods. The simple regression method is the easiest to apply given sufficient peak flow data, while the scaling method is the most efficient method with regard to data use.