Identification of the non-stationarity of extreme precipitation events and correlations with large-scale ocean-atmospheric circulation patterns: A case study in the Wei River Basin, China

The investigation of extreme precipitation events in terms of variation characteristics, stationarity, and their underlying causes is of great significance to better understand the regional response of the precipitation variability to global climate change. In this study, the Wei River Basin (WRB), a typical ecoenvironmentally vulnerable region of the Loess Plateau in China was selected as the study region. A set of precipitation indices was adopted to study the changing patterns of precipitation extremes and the stationarity of extreme precipitation events. Furthermore, the correlations between the Pacific Decadal Oscillation (PDO)/El Nitio-Southern Oscillation (ENSO) events and precipitation extremes were explored using the cross wavelet technique. The results indicate that: (1) extreme precipitation events in the WRB are characterized by a significant decrease of consecutive wet days (CWD) at the 95% confidence level; (2) compared with annual precipitation, daily precipitation extremes are much more sensitive to changing environments, and the assumption of stationarity of extreme precipitation in the WRB is invalid, especially in the upstream, thereby introducing large uncertainty to the design and management of water conservancy engineering; (3) both PDO and ENSO events have a strong influence on precipitation extremes in the WRB. These findings highlight the importance of examining the validity of the stationarity assumption in extreme hydrological frequency analysis, which has great implications for the prediction of extreme hydrological events. (C) 2017 Elsevier B.V. All rights reserved.