Investigating raindrop size distributions in the (L‐)skewness–(L‐)kurtosis plane

Abstract

Skewness–kurtosis (β3−β4) and L‐skewness–L‐kurtosis (τ3−τ4) planes are proposed here as diagnostic tools to guide the identification of drop size distributions (DSDs) of rainfall at the ground. Firstly, we have determined β3−β4 and τ3−τ4 domains of 13 distribution families, namely normal, exponential, gamma, truncated gamma, log‐normal, truncated log‐normal, Weibull, hyperbolic, generalized hyperbolic, log‐logistic, skewed Laplace, Johnson SB and Johnson SU. These include those most used to represent DSDs and, in general, particle size distributions (PSDs). Secondly, we have considered 1 min and 2 min disdrometric data, collected at six sites in the United States, and reported the empirical couples (β3,β4) and (τ3,τ4) in the moment diagrams. The location uncertainty of the empirical couples in the diagrams, mostly due to the occurrence of sampling errors, has been thoroughly investigated. The variability of the empirical DSD couples (β3,β4) and (τ3,τ4) is well described by truncated gamma, truncated log‐normal and Johnson SB over the other considered distributions. However, a Monte Carlo analysis has shown that the Johnson SB is the most adequate distribution in describing the drop size variability, being characterized by the lowest level of uncertainty.