Unraveling the amplified role of urbanization on occurrence likelihood of precipitation extremes through nonstationary model in Huaihe River Basin, China

Abstract

Under the combined influence of urbanization and climate change, the frequency and severity of precipitation extremes in the Huaihe River Basin of China (HRB) have intensified over recent decades. This underscores the importance of considering trend-induced nonstationarity in the risk analysis of rainfall extremes. This study utilized daily precipitation observations from 125 rural, suburban, and urban stations in the HRB to develop a nonstationary Generalized Extreme Value (GEV)-based model. The aim was to investigate the spatiotemporal evolution patterns of precipitation extremes (PEs) by integrating physical factors into distribution parameters as potential covariates. A risk amplification factor (RAF) was derived from comparing the recurrence levels between nonstationary and stationary GEV models to elucidate the amplified role of urbanization processes on PEs across different types of stations. Furthermore, for the stations suffering both the urbanization and climate change-induced nonstationarity, the singular impact of urbanization was isolated through the comparative analysis of RAF from ${\mathrm{GEV}}_{\mu /\sigma (Urb+Cli)}$ and ${\mathrm{GEV}}_{\mu /\sigma \left(Cli\right)}$. The study found that all PEs exhibited increasing trends, with significantly increasing trends concentrated in the northern region of the HRB. Urbanization significantly influenced the trend-induced nonstationarity of Rx1day and Rx5day series, while its impact on the R95P and R25 series was negligible in rural stations. Urbanization had the most substantial impact on Rx1day and Rx5day, with noticeable changes, while its influence on R25 was minimal. Additionally, the changes in recurrence levels for suburban and urban areas were more pronounced than those in rural areas, particularly for Rx1day, Rx5day, and R95P.