Investigating the variation and non-stationarity in precipitation extremes based on the concept of event-based extreme precipitation

•The concept of event-based extreme precipitation (EEP) was proposed.•Spatio-temporal changes and possible non-stationarity of EEP events were analyzed.•The proposed EEP concept was more appropriate in the extreme precipitation analysis.

Extreme precipitations (EP) could induce a series of social, environmental and ecological problems. Traditional EP analysis usually investigated the characteristics based on a fixed time scale and therefore ignored the continuity of EP occurrence. As a result, a comprehensive assessment on the influence and consequence of the EP occurring during consecutive time periods were largely eliminated. On the other hand, the characteristics of EP, including variables such as frequency, intensity and extreme volume, were commonly defined without sufficient consideration of the local tolerance capacity (which can be represented by a threshold level of EP) and therefore would sometimes be inappropriate. In this study, we proposed a concept of event-based extreme precipitation (EEP) by considering the continuity of EP and defined the statistical variables for the characteristics of an EEP event by taking account of local tolerance capacity. An EEP was identified as a collection of precipitation data over the consecutive time period in which all the precipitation amounts are above the pre-defined threshold, and EEP events are separated by at least one time step (e.g., day or hour) with precipitation amount below the threshold. As a case study which in fact motivated our proposal, we investigated the changes and variations of EEP with the consideration of potential non-stationarity in the Hanjiang River Basin of China (HJRB) during the time period of 1960-2013. Results showed that the concept of EEP, which could reflect the impact of continuity of EP occurrence and mirror the differences of local tolerance capacity, was more appropriate than the traditional method in the EP analysis.