Huff quartile classification of rainfall intensity profiles ('storm patterns'): A modified approach employing an intensity threshold

Abstract

Many studies require a description of the temporal structure of rainfall events, which can influence water partitioning at the landsurface, and hence affect processes such as soil erosion and flash flooding (Yuan et al., 2021). A common approach is to employ the classification of events using 'Huff quartiles' (Huff, 1967). This identifies events as 'first quartile', 'second quartile', etc., depending on which dimensionless fraction of the event duration received the largest rainfall depth. Among issues attending this approach is that rainfall intensity is effectively disregarded, since quartiles with higher intensity but greater intermittency can carry less weight that those with more continuous but lower-intensity rain that yields a larger total depth. The wettest quartile can only be said to have the largest equivalent rainfall rate; this may bear little relation to actual rainfall intensities, especially in long rainfall events in which the duration of a quartile may extend from many hours to a day or more. This reduces the relevance of the quartile classification for studies of infiltration, soil loss, or related landsurface processes.This study explores a modification to the quartile approach, which involves tallying for each quartile only rainfall amounts delivered above a threshold intensity selected to be relevant to water partitioning, soil erosion, etc. Using high temporal resolution pluviography data from two Australian field stations, it is shown that this approach can result in the re-classification of 25–30% of rainfall events, and for some events this involves the maximum possible change in quartile classification by 3 steps - such that a first-quartile event becomes a fourth-quartile event, or the reverse. It is suggested that the use of an intensity threshold as explored here, that relies on actual intensities and not quartile equivalent rainfall rates, may result in a quartile classification of rainfall events that has greater relevance to the study of landsurface processes than does the original Huff system.