Identifying Characteristics of Guam’s Extreme Rainfalls Prior to Climate Change Assessment

Abstract

Extreme rainfall and its consequential flooding account for a devastating amount of damage to the Pacific Islands. Having an improved understanding of extreme rainfall patterns can better inform stormwater managers about current and future flooding scenarios, so they can minimize potential damages and disruptions. In this study, the scaling invariant properties of annual maximum precipitations (AMPs) are used for describing the regional patterns of extreme rainfalls over Guam. AMPs are calculated at seven stations in Guam and exhibit distinct simple scaling behavior for two different time frames: (1) from 15 min to 45 min; and (2) from 45 min to 24 h. With these two different behaviors, the conventional estimation methods for sub-hourly durations overestimate the frequencies at a site in which breakpoints are clearly observed, while the proposed Scaling Generalized Extreme Value (GEV) method, based on the Scaling Three-NCM (S3NCM) method, provides comparable estimates. A new regional extreme rainfall analysis approach based on scaling exponents is introduced in this study. Results show distinct extreme rainfall patterns over Guam. Moreover, the numerical and graphical analyses identify that a tropical cyclone may increase daily AMPs by 3%, on average.