An inovative regional frequency analysis approach for robust extreme precipitation assessment in data-rich and climatically diverse regions

Abstract

This study addresses regional frequency analysis (RFA) uncertainties caused by difficulties in identifying homogeneous subregions and choosing the best regional frequency distributions. The study modifies Hosking and Wallis (1997)'s approach to improve regionalization, especially in regions with many gauge stations. The proposed method uses 512 Iranian gauges to identify three primary regions based on annual precipitation patterns. Examining data uniformity, regional variations, frequency distributions, and quantiles for exceptional events are crucial. L-moments are important in the analysis because they estimate distribution parameters and help evaluate heterogeneity and choose distributions. The study emphasizes the importance of considering distributional characteristics beyond the mean to ensure homogeneous clusters. The findings indicate that annual precipitation patterns in Iran are spatially heterogeneous. Despite challenges, the proposed regionalization approach finds homogeneous regions that can be represented by fitted distributions. The approach's ability to accommodate spatial intricacies and tailor analysis to specific climates is shown by disaggregated area fit assessments. Thus, the study illuminates Iran's hydrological conditions-specific RFA methodology. This improves extreme precipitation estimates and aids water resource management and strategic planning. The methodology can meet different user needs and be implemented in comparable regions worldwide.