Destructive breakout floods that are generated by the breaching of embankments need to be assessed in the risk evaluation of potential hazards to prepare emergency action plans. The current standard practice to assess the breaching of embankment dam is with statistically derived parametric equations based on historic dam failures. However, the prediction uncertainties of such equations are widely recognised to be significant. These equations exclude the consideration of essential factors such as material properties, erodibility, configuration and construction. This study focuses on investigating and presenting the available statistically derived breach equations and comparing them individually and compares the accuracy of predicting the breach parameters of earthfill and rockfill embankments. Additionally, two simplified physically-based breach models were investigated how they go about modelling overtopping breach formations. Data from two field experiments with different configuration subjected to overtopping failure is used as input to simulate overtopping failure with the simplified physically-based breach models. The results are analysed and the models compared. Sensitivity analysis was conducted for both models to detect both sensitivity to change of three parameters as well as instabilities. Furthermore, two preliminary physical models of a rockfill dam were conducted in the hydraulic laboratory at the Norwegian University of Science and Technology. The process of conducting such tests is illustrated and limitations discussed as well as the breach process analysed.