The concept of "normalized" distribution to describe raindrop spectra: A tool for hydrometeor remote sensing

The concept of normalization of DSD developed in this paper is founded upon two reference variables, the liquid water content LWC and the mean volume diameter D-m. We show mathematically that it is appropriate to normalize by N-0* proportional to LWC/D-m(4) with respect to particle concentration and by D-m with respect to drop diameter. The major point of our approach is t hat it is totally free of any assumption about the shape of the DSD. This n ew normalization has been applied to the ensemble of the airborne microphys ical data of TOGA-COARE (21 flights of the NCAR Electra). The classificatio n of the TOGA-COARE raindrop spectra into 4 categories (one stratiform, and three convective [0-10mm/h, 10-30mm/h, 30-100mm/h]) allowed us to identify the following features: (i) There is a distinct behavior of N-0* between s tratiform and convective rains; a typical "stratiform" value is 3x10(6)m(-4 ), while it is 2x10(7)m(-4) in convective.(ii) Between the convective rain categories, there a clear trend for D-m to increase as the rainfall rate. ( iii) The "average" normalized shape of the DSD is remarkably stable between the four rain categories. This normalized shape departs from the exponenti al, but also from all the analytical shape considered up to now (as Gamma, log-normal, modified Gamma). (C) 2000 Elsevier Science Ltd. All rights rese rved.