Intercomparison of the performance of three different versions of Cloud Particle Imagers (CPIs)

Cloud particle size distributions (PSDs) are used to determine the fundamental microphysical and radiative properties of clouds. To measure PSDs, optical array probes (OAPs) have been used extensively over the past 40 years. Past studies have indicated inherent inaccuracy of OAPs in recording and sizing cloud particles, especially those with maximum dimensions less than 125 μm. A Cloud Particle Imager (CPI) is another probe that provides high resolution images of cloud particles but whose use to quantitatively determine PSDs has yet to be well established. Connolly et al. (2007) has suggested that CPIs may be able to accurately measure PSDs using sufficiently large averaging times provided that a calibration factor is determined. Following previous OAP calibrations and adding modifications necessary for the optical system of a CPI, the Connolly et al. (2007) method for determining CPI particle size and depth of field assumes that shadow images of transparent spherical particles are adequately described by the Fresnel-Kirchhoff diffraction. Hence errors in sizing depend on the distance (Z) between the object plane and a particle. The original Connolly calibration method, developed using data acquired with CPI version 1.0, cannot be directly applied to other CPI versions because of different optical alignments and other characteristics.

In this study, three different CPIs (versions 1.0, 1.5, and 2.0) were calibrated by moving glass calibration beads and ice analogues of known size parallel to the object plane. Measures of Z, focus, apparent maximum dimension, and background image were used to derive calibration parameters for each CPI version in two empirical equations that were then applied to in-situ CPI data to determine particle size and depth of field, and hence PSDs. Three CPIs were installed at the base of the Manchester Ice Cloud Chamber and connected to air pumps used to draw cloud through the sample volume. Warm liquid clouds at a temperature of 1-2 ^oC and ice clouds at a temperature of -5 ^oC were generated and PSDs of these clouds were determined by applying the results of the calibration.

The calibration parameters of the different CPIs will be compared and impacts of the calibration on sizing error will be shown in this presentation. Further, influences of the calibration on the performance of each CPI will be shown using cloud chamber measurements to test the reliability of different versions of CPIs in measuring PSDs.