Nonstationary frequency analysis of the recent extreme precipitation events in the United States

The intensification of the hydrologic cycle due to climate change is likely to influence the extreme precipitation characteristics (i.e., intensity, duration and frequency). These precipitation characteristics are integrated to construct Intensity-Duration-Frequency (IDF) curves that are widely used to design civil infrastructure systems. These IDF curves are typically derived based on the stationary assumption, however, the frequency and intensity of extreme precipitation events likely to become nonstationary as a consequence of climate change. During the past decades, unusual extreme precipitation events with more than thousand-year return periods were recorded in the United States. This study investigates the nonstationary nature of the most recent extreme precipitation events occurred over different durations (1-, 3- and 5-days) by incorporating time-varying covariates, such as time, maximum temperature, mean temperature, and the El Nino Southern Oscillation cycle (ENSO). The nonstationary frequency analysis for these extreme events was conducted using nonstationary Generalized Extreme Value distribution by incorporating the time-varying covariates. It was observed that most of the temporal evolution of extreme precipitation events follow the nonstationary pattern, which may be due to the increase in the magnitude of recent extreme precipitation events, especially during hurricane events. Different combination of covariates can potentially influence the nonstationary frequency analysis, and the type of covariate may differ when the accumulated period of extreme precipitation event increased. Based on the Nonstationary Extreme Value Analysis, the return periods associated with extreme precipitation events significantly reduced compared to the stationary approach.