Extracting the Event Characteristics of Rainfall

Abstract

A method is proposed and demonstrated for extracting the spatial and temporal characteristics of individual rainfall events using hourly rainfall measurements at an array of rain gauges. The method is based on identifying the probabilities of overlapping rainfall events at the rain gauge locations. The method is demonstrated in a 47,000 km2 area covered by 150 rain gauges with 17 years of unbroken hourly data. The results show that the length scales of individual events in the wet season are in the range of 1.2–18.3 km depending on the event magnitude, and dry season events with corresponding magnitudes have length scales in the range of 1.5–35.9 km. The timescales of wet and dry season events generally increase with event magnitude in both seasons. The shapes and locations of cluster areas that tend to have overlapping rainfall events are substantially different between seasons, being smaller and more dispersed in the wet season and larger and lesser in number in the dry season.