Raindrop Size Distribution variability from high resolution disdrometer networks

The characteristics of the raindrop size distribution (DSD) have been widely studied since Marshall and Palmer (1948) introduced specific version of exponential distribution for the observed size spectra, based on measurements of raindrops records on dyed filter papers. Across the decades, interest in measuring and studying rain DSD has grown due to applications in cloud physics studies, in calibration of space-borne and ground-based microwave active precipitation sensors and in soil science and agriculture. The study of DSD and of the processes that determine it, are always been challenging from both theoretical and experimental point of view. Moreover, the study of DSD in natural rain is hindered by the difficulties (logistic and economic) in the management of dense disdrometer networks. Based on the unprecedented datasets available, this Thesis aims to contribute in characterizing, from a microphysical point of view, the precipitation structure and the processes that generate it. In particular, the vertical and horizontal DSD variability is analyzed, starting from the study of collisional break-up mechanism in natural rain. The signature of collisional break-up, first evidenced in a particular shape of Doppler power spectrum of a microwave disdrometer, is then searched and characterized in DSD spectrum, assessing its variability with altitude. The horizontal variability of DSD is studied both analyzing the occurrence of equilibrium DSD among the different datasets available and evaluating the correlation of integral and non-integral DSD parameters at small scale. In the first part of the Thesis, an overview on past and recent studies on different aspects of DSD is given. The main mechanisms that govern the rain development are firstly summarized, then the DSD parameterization and the DSD variability in natural rain are discussed. Finally, the description of the characteristics of instruments and of the field campaigns considered in this work are presented. The vertical variability of DSD has been studied thanks to the development of specific algorithms able to detect and characterize both the collisional break-up and the equilibrium DSD. I analyzed the signature of collisional break-up both on the Pludix Doppler power spectrum and on DSD spectrum. The analysis is carried out developing two algorithms that detect the collisional break-up as well as estimate the break-up diameter as function of altitude. The results show a decrease of break-up diameter with altitude, due to the reduction of air density, that plays a critical role in the energetic balance of the collision between two raindrops. The analysis also indicates that, regardless the altitude, the collisional break-up occurs if the kinetic energy of the collision exceeds 12.2 μJ. The results, together with the detailed analysis of some case study at high altitude (over the Tibetan Plateau), show also that the dominance of the break-up process is required to reach the equilibr ium DSD. The study of the DSD variability was deepened focusing the analysis on the 2DVD DSD properties to evaluate the occurrence of equilibrium DSD in natural rain. Another algorithm, based on 2DVD characteristics, is set up to automatically detect the equilibrium DSD by using the great amount of high quality disdrometric data available from the datasets of Ground Validation program of NASAGlobal Precipitation Measurement mission. The results shows a good agreement between the experimental equilibrium DSD and the equilibrium DSD obtained by theoretical models. The analysis shows also that the equilibrium DSD is mainly reached during convective rain and its dependence on season and latitude (no equilibrium DSD is observed at high latitude - 60°N). The occurrence of equilibrium DSD is a rare event in natural rain (maximum 8% of selected minutes), while an increase is observed if transition situations are considered. The results are also analyzed to estimate the goodness of fitting the equilibrium DSD by a three parameter gamma distribution, that is widely used to parameterize the DSD. The low correlation between the experimental DSDs and the gamma distribution evidences that the gamma is not the best parametric form to fit the experimental equilibrium DSD. The behavior of the rain and DSD parameters is studied as function of break-up occurrence and shows that they can be considered an additional indicators to screen out the situations that are not expected to reach the equilibrium DSD. The data collected from two high-resolution disdrometric dataset are used to study the horizonta l DSD spatial variability at small scale. The size of the measuring fields are different but comparable with a ground radar pixel or satellite footprint and this makes the analysis of the particular interest . The rainfall rate and other DSD parameters are analyzed using a three parameter exponential function to estimate their correlation at small scale. The estimated correlation distance shows that the most of the rain and DSD parameters are correlated within a radar pixel or satellite footprint (generally, the integral DSD parameters – rainfall rate, radar reflectivity, liquid water content, etc. – are less correlated than the non integral DSD parameters – maximum diameter, mean mass diameter, etc.). The root mean square error evidences a very good fit of the function used with respect the experimental data, indicating a good reliability of data. The results presented in this Thesis, first, increase the knowledge of break-up phenomenon and its effect on the DSD up to reach the equilibrium DSD, and can be used to improve the parameterizat ion form for break-up and equilibrium DSD occurrence and the modeling of cloud and precipitat ion mechanisms. Secondly, they give reliable indications about the spatial variability of the structure of precipitation within a radar pixel and/or a satellite footprint, with an immediate application to the interpretation of remote sensing measurements to improve precipitation retrieval from radar/satellite measurements, especially after the launch of Dual-frequency Polarization Radar in the frame of Global Precipitation Measurement mission. The results obtained in this Thesis lead to the study of many other aspects that can be investigated to better characterize the precipitation. The time evolution of the precipitation with particular emphasis to the time necessary to the break-up to modify the DSD to reach equilibrium DSD can be investigated by using the algorithms proposed here. A new parameterization of DSD affected by break-up and of equilibrium DSD is necessary to improve the remote sensing of precipitation. Finally, a deeper study of DSD spatial variability is needed to have more information about rain structures at small/medium spatial scales, by different techniques and datasets in different season/location.