The Swedish guidelines for design flood determination are examined with emphasis on sensitivity analyses of modelling aspects. The research is mainly based on conceptual hydrological modelling with the HBV model. An automatic calibration methodology is presented and discussed. The influence of model calibration, model structure and climatic variability on the design flood simulation as well as the preset meteorological and hydrological conditions suggested in the guidelines are studied. The thesis also includes a discussion of the annual exceedance probability of the Swedish design floods by comparisons with observed extreme floods and frequency analyses.

The design precipitation and snowpack magnitude are the most important rescriptions in the Swedish guidelines for design flood determination. Consequently, the regional design precipitation sequence and all suggested correction factors to the design rainfall sequence (area, altitude and season) are crucial. In application, the most important factors are: choice of  hydrological model, model calibration and selected climate period. These factors can lead to an uncertainty in the order of plus/minus 20 % on the design flood peak in catchments with good quality data and good model performance. In catchments with poor data quality, the uncertainty in design flood simulation will be larger. The corresponding uncertainty on reservoir water stage development will depend on local conditions.