Uncertainty analysis of the rate of change of quantile due to global warming using uncertainty analysis of non-stationary frequency model of peak-over-threshold series

Abstract

<jats:p>Abstract. Several methods have been proposed to analyze the frequency of non-stationary anomalies. The applicability of the non-stationary frequency analysis has been mainly evaluated based on the agreement between the time series data and the applied probability distribution. However, since the parameters of the estimated probability distribution contain a lot of uncertainty, the uncertainty in the correspondence between samples and probability distribution is inevitably large. In this study, an extreme rainfall frequency analysis is performed that fits the Peak-over-threshold series to the covariate-based non-stationary Generalized Pareto distribution. By quantitatively evaluating the uncertainty of daily rainfall quantile estimates at Busan and Seoul sites of the Korea Meteorological Administration using the Bayesian approach, we tried to evaluate the applicability of the non-stationary frequency analysis with a focus on uncertainty. From the point of view of the agreement between the time series data and the applied probability distribution, the non-stationary model was found to be slightly better. When comparing the performance of the stationary and non-stationary model from the uncertainty point of view, the uncertainty of the non-stationary model was greater than that of the stationary model since the non-stationary model included variability arising from covariates. However, it was found that if the appropriate covariate corresponding to the quantile was selected (that is, if the variability of the covariate was eliminated), the reliability of the non-stationary model could be higher than that of the stationary model. Given the covariate, it was confirmed that the uncertainty reduction in quantile estimates for the increase in sample size is more pronounced in the non-stationary model. In addition, how to use the dew point-based non-stationary frequency analysis when integrating information on global temperature rise is described. Finally, it is proposed how to quantify the uncertainty of the rate of change in the future quantile due to global warming using the rainfall quantile ensemble obtained in the uncertainty analysis process. </jats:p>