Analysis of cirrus particle size distributions from three in situ flight campaigns: Applications to cirrus microphysics and parameterization, remote sensing, and radar forward model simulators.

摘要

A database of cirrus particle size distributions (PSDs), with concomitant meteorological variables, has been constructed using data collected with the Two-dimensional Stereo (2D-S) probe. Parametric functions are fit to each measured PSD. Full statistical descriptions are given for unimodal fit parameters. Three statistical tests were developed in order to determine the utility of bimodal fits and the efficacy of unimodal fits, and an investigation into the relationship between the parameterized PSDs and several meteorological variables was made.;Next, a parameterization of a "universal" cirrus PSD is given. This parameterization constitutes an improvement on earlier works due both to the size of the dataset and to updated instrumentation. Despite earlier works that predicted a gamma-distribution tail to the universal ice PSD, it is shown here that the tail is best described by an inverse gamma distribution. A method for predicting any PSD given the universal shape and two independent remote sensing measurements is demonstrated.;The constructed PSD database is then used to address a straightforward question: how similar are the statistics of PSD datasets collected using the recently developed 2D-S probe to cirrus PSD datasets collected using older Particle Measuring Systems (PMS) 2D Cloud (2DC) and 2D Precipitation (2DP) probes? It is seen, given the same cloud field and given the same assumptions concerning ice crystal cross-sectional area, density, and radar cross section, that the parameterized 2D-S and the parameterized 2DC predict similar distributions of inferred shortwave extinction coefficient, ice water content, and 94 GHz radar reflectivity. However, the parameterized 2DC predicts a statistically significant higher number of total ice crystals and a larger ratio of small ice crystals to large ice crystals.;Finally, the beginnings of two works in their early stages are presented. First, the probability structure of the parameterized PSDs is considered in light of application to the Bayesian inference of cirrus microphysical properties from remote sensing measurements. Then, the collection of measured PSDs, along with a forward model for radar reflectivity, is used to investigate uncertainty in computations of radar reflectivity from modeled moments of cirrus cloud PSDs.