Flood events in Europe are characterised by different generating mechanisms determining large varieties of peaks, volumes and hydrographs. Understanding such mechanisms is crucial not only for deterministic or stochastic modelling of floods, but also for practical purposes such as hydrological design . In this study, the driving mechanisms of floods are analysed and the associated catchment and atmospheric attributes controlling these flood types are identified through a classification and regression tree approach. In addition, the relevance of flood types for estimations of flood probabilities is analysed using type-based flood statistics. It is shown which flood types dominate the more frequent floods and which flood types are most relevant for extreme flood events. Ordinary and extraordinary floods are identified by a Likelihood-Ratio test and tested for a significant difference in the frequency distribution of flood types. Our results show that the flood types vary regionally in Europe with distinct clusters. In the Alpine region, heavy rainfall floods are responsible for the most extreme flood events, while in the northern parts of Europe flood events caused by snowmelt lead to the largest peaks. This is reflected in the flood statistics by the type-specific distributions, which have a different tail heaviness. The results of this study demonstrate that flood types are a main explanatory factor when detecting regional patterns of statistical attributes of flood events.