Statistical Characteristics of Extreme Precipitation Events Over Tropical Oceans and Land Using Five Years of IMERG Data

Abstract

This study provides a statistical analysis of tropical extreme precipitation events (EPEs) using fiveyears (2006-2010) of IMERG V06 Final Run data. EPE life cycles and spatial structures aretracked with a Lagrangian algorithm that allows spatial and temporal gaps and captures both lightand heavy/extreme precipitation. Across the tropics, we identify 6.84 million EPEs (~3.99 millionoceanic and ~2.85 million land) and examine their frequency and key characteristics, includingduration, intensity, rainfall, and areal coverage, over tropical land and oceans. Geographic andseasonal patterns follow major convective regions, such as the ITCZ, with event counts migrating meridionally alongside the seasonal progression of tropical rainfall. Local event counts are dominated by 1–6-h events everywhere, while 6–12-h events make up a higher percentage overland. Diurnal cycles differ sharply: land EPEs peak in the afternoon, whereas oceanic eventsexhibit weaker morning and late-evening peaks. Substantial land-ocean contrasts also emerge inEPE properties. Oceanic 6–24-h events exhibit larger median and interquartile ranges of arealcoverage, extreme rainfall, and total rainfall than their land counterparts, whereas for 1–6-h eventsthe opposite holds. Relationships between areal coverage and total rainfall are broadly linear across several orders of magnitude, with upper bounds consistently larger over the ocean. Instantaneous attributes show similar linearity but greater scatter and comparable upper bounds across land and ocean, highlighting the role of duration in controlling rainfall. These results suggest that EPE characteristics are shaped by local environments and may respond differently to climate change over land and ocean.