A modified regional L-moment method for regional extreme precipitation frequency analysis in the Songliao River Basin of China.

The regional frequency analysis (RFA) is a widely used method in analyzing the changes of extreme precipitation (EP). The uncertainties in the identification of homogeneous subregions and the selection of optimal regional frequency distributions can largely influence the results of the RFA. In this study, the fuzzy c-means method combined with the extended Xie-Benn index (FCXB) is used to help determine the optimal division of subregions. In addition, we introduce a new comprehensive index (CI) to overcome the shortcomings of present measures and reduce the uncertainty in regional frequency distribution selection. The changes of EP at 93 meteorological stations in the Songliao River Basin (SRB) during 1960 to 2016 is analyzed. The main results show that: (1) FCXB can effectively identify the optimum number of homogeneous subregions automatically, and the corresponding subregion division is proven to be reasonable and reliable; (2) compared with the single goodness-of-fit measure, the developed CI can reduce the uncertainties in distribution selection and determine the optimal regional distribution in a reliable way; (3) the estimated EP under different return periods both decrease from the south to the north of the SRB, which indicates the risk of high-intensity EP events in the southern SRB is relatively higher. These findings can provide technical support for local policymakers to formulate effective measures to lessen the damages of the EP on ecosystem and society. • FCXB can give a reliable division of homogeneous sub-region. • The CI can select the optimal regional distributions in a reliable way. • The risk of high-intensity EP events in southern SRB is higher than northern SRB. [ABSTRACT FROM AUTHOR]