Temporal continuity of extreme precipitation events using sub‐daily precipitation: application to floods in the Ebro basin, northeastern Spain

Abstract

ABSTRACTExtreme precipitation events have been studied widely in various regions because of their implications for hazard assessment and risk management, particularly those of a hydrologic nature. Extreme precipitation events have been commonly defined using daily precipitation series owing to a lack of long‐term data at higher temporal resolution. However, because of their potentially dramatic consequences, intense sub‐daily precipitation is of great importance for risk assessment. Here, we define a new concept of extreme precipitation event based on temporal continuity of that event through run analysis using hourly precipitation. The aim of this definition is to record intense precipitation events over a very short period as well as very persistent rainfall events that produce heavy accumulated precipitation over several days. Both types of event can present considerable hydrologic hazards and are not considered in the definition of extreme precipitation event that is based on a daily time scale. We applied this method to the 29 greatest flood events in the Ebro River basin during 2000–2014 to determine whether our definition can be used to identify events producing hydrologic risks. In addition, we compared the synoptic patterns and trends of events defined using this method and those based on a daily time scale. The results show that the new definition allowed for identification of the 29 analysed flood events (as opposed to extreme events based on a daily time scale), among which eight events were not identified. Moreover, using the new definition facilitates establishment of synoptic patterns and more detailed study of trends by discriminating events according to their duration. The results reveal how the use of daily precipitation can hide trends of shorter events, because positive trends are more dominant in extreme precipitation events at sub‐daily scale.