+Dynamic identification and risk analysis of compound dry-hot events considering nonstationarity

Duo to the influence of anthropogenic and climate change, the traditional drought or hot event identification indices under the stationary assumption of the probabilistic behavior of a hydrometeorological variable could be no longer valid. This study proposed a nonstationary framework for the identification of compound dry-hot extremes and risk assessment considering climate change for the warm season during the period 1901–2017 in Weihe River Basin (WRB) and Fenhe River Basin (FRB), China. The proposed framework is composed of three phases: (1) calculating nonstationary precipitation index (NSPI) and standardized temperature index (NSTI) by incorporating large-scale climate indices as covariates of fitted distribution parameters; (2) calculating nonstationary compound dry-hot index (NCDH) based on nonstationary copula models; and (3) bivariate nonstationary risk analysis of compound dry-hot events through two regional characteristics called relative affected area (RA) and mean severity of area (MSA). The proposed NCDH using climate indices as covariates was found to be superior in capturing compound dry-hot characteristics and revealing the trend of temporal and spatial changes of WRB over the period of 1901–2017. Based on the nonstationary risk assessment of regional characteristics, the occurrence risk of compound extremes in Sub-Basin 1 of WRB was less than that in the other two Sub-Basins. Considering the importance of nonstationarity during the above three modelling phases, the nonstationarity in the 1st phase modelling of average precipitation (AP) and average temperature (AT) to derive NSPI and NSTI could have a greater impact on the final results of nonstationary risk assessment of compound dry-hot events than nonstationarity in the other two phases. This nonstationary compound dry-hot index provides a new insight into compound extreme identification and risk assessment that can adapt to a changing environment.