Event-based rainfall dynamics and storm frequency in arid and semi-arid climates: An innovative standardized approach for rainstorm analysis

• A multi-phase analysis integrates MIET, clustering, and GIS to investigate storms. • Optimal MIET (6–9 hr) captures storm Features; ANE outperforms CV in arid climates. • GAM, P3, and WEIB are robust for rainstorm frequency across nine diverse climates. • The novel SAI identifies storm severity and retrieves design storm boundaries. • Coastal areas as highly vulnerable to extreme rainstorms and flash flood hazards. Precipitation, a fundamental component of hydrological processes, is characterized by sporadic intense storms in arid and semi-arid regions and often triggers severe flash floods that threaten infrastructure, ecosystems, and human safety. This study explores the dynamics of event-based rainfall characteristics in dry climates, employing a detailed analysis of Minimum Inter-Event Time (MIET) and frequency investigation. The utilized hourly rainfall datasets encompass all available pixels at a 0.1-degree resolution across the United Arab Emirates (UAE) over 21 years. Cluster analysis highlights the diverse climatic profiles within the UAE, resulting in nine distinct regions. The study examines the application of three statistical methods for MIET estimation and conducts spatiotemporal analyses, revealing the influence of MIET variation on storm patterns. The findings demonstrate that the optimal MIET values range from 6 to 9 h across the nine zones. The Average Annual Number of Events (AANE) approach aligned closely with storm volume examination in determining appropriate MIET values. In contrast, the Coefficient of Variation (CV) method proved statistically inefficient in dry climates, underscoring the non-exponential nature of inter-event intervals. Event-based rainfall analysis indicates an increased storm frequency, exceeding 22 events per year along the northern coast and the northeastern mountainous lands, while prolonged dry conditions lasting over 23 days are prevalent in the interior. Temporally, the Mann-Kendall test revealed no significant monotonic trends in rainstorm datasets across most of the UAE at a 95 % confidence interval. However, the northeastern regions exhibited upward trends, indicating a heightened flood risk. The comprehensive frequency analysis, incorporating ten probability distributions and five statistical indices, shows the robust performance of Gamma, Pearson type III, and Weibull distributions in capturing storm features. Subsequently, these optimal models were integrated into the Standardized Attribute Index (SAI) analysis, providing a novel perspective for rainstorm classification and evaluation while effectively identifying the lower boundaries of design storms. Coastal regions are highly vulnerable to extreme events and flash flood hazards, with elevated near-normal upper limits. Ultimately, this study contributes to a refined understanding of rainstorm behaviors in arid and semiarid climates, offering critical insights for hydrological modelling and water resource management.