During the 2006 Tropical Warm Pool International Cloud Experiment conducted in the region near Darwin, Australia, the Scaled Composites Proteus aircraft executed spiral profiles and flew horizontal legs through aging cirrus, fresh anvils, and cirrus of unknown origin. Data from 27 Jan., 29 Jan., and 2 Feb., when all the microphysical probes a Cloud and Aerosol Spectrometer (CAS), a Cloud Droplet Probe (CDP), a Cloud Imaging Probe (CIP), and a Cloud Particle Imager (CPI) were working, are used to investigate whether a single parameterization can be used to characterize tropical cirrus in terms of prognostic variables used in large-scale models, to calculate the single-scattering properties (scattering phase function P11 and asymmetry parameter g) of aggregates and small ice crystals that more closely match observed ice crystals, and to quantify the influences of small ice crystals on the bulk scattering properties of tropical cirrus.
A combination of CDP (D < 50 μm), fits (50 < D < 125 µm), and CIP (D > 125 μm) distributions is used to represent ice crystal size distributions. The CDP measurements are used for small ice crystals because comparison between the CAS and CDP suggested the CAS was artificially amplifying small ice crystal concentrations by detecting remnants of shattered large ice crystals. Artifacts in CIP images are removed or corrected and then CIP measurements are used to represent large ice crystals. Because of the uncertainties in both the CPI and CIP for 50 < D < 125 μm, the incomplete gamma fitting method with the CDP (D < 50 μm) and CIP (D > 125 μm) measurements as input is used to characterize these distributions.
A new quasi-automatic habit classification scheme is developed. For all days, small quasi-spheres dominated the contributions from all ice crystal sizes (D > 0 μm, by number) for all 3 days. The areal fraction (D > 200 μm) from bullet rosettes and their aggregates was 48% and 60% for 27 and 29 Jan., respectively, but only 7% for 2 Feb, whereas the fraction of aggregates of plates was 46.2% for 2 Feb. and only 7.2% and 1% for 27 and 29 Jan., respectively. The difference in ice crystal habits sampled on the different days is likely associated with the difference between fresh anvil cirrus on 2 Feb. and aged cirrus bands on the 27 and 29 of Jan.
Because of variations in microphysical properties (i.e., number concentration, median mass dimension, and fit variables of gamma distributions) it is also shown that variables in addition to ice water content and temperature are required to represent the characteristics of cirrus with different origins in large-scale models.
Aggregates of bullet rosettes and aggregates of plates are shown to scatter more light in the lateral and backward scattering region and less light in the forward scattering region compared to their component crystals, leading to a decrease in g for aggregates. To represent the three-dimensional shape of aggregates of plates, three parameters, the aggregation index ( AI), the area ratio (AR), and the normalized projected area (An), are introduced and the single-scattering properties of aggregates of plates are shown to depend heavily on AI.
A new model (budding Bucky ball, 3B) for the shape of small ice crystals is developed based on the shapes of ice analogues grown in laboratory experiments. The 3B scatters more light in the lateral, and backward direction and less in the forward direction compared with other existing models currently used to describe small crystal shape (i.e., Gaussian random sphere and droxtal). The combination of the reduction in the forward scattering and enhancement in the lateral and backward scattering causes 11.13% and 8.74% decreases in g for the 3B compared with that for Gaussian random sphere and droxtal, respectively.
The impacts of variations in small ice crystal shapes and concentrations on bulk scattering properties of tropical cirrus are quantified. The calculated mean asymmetry parameter g¯ for the fresh anvil (i.e., 2 Feb) is larger than that for cirrus bands of varying ages (i.e., 27 and 29 Jan.) for -60 < T < -45°C and -45 < T < -30°C where the fractional contributions of small ice crystals to total cross sectional area are small. The impact using different models for small ice crystals on g¯ is largest at lower temperatures (T < -60°C). The impact of enhanced number concentrations of small ice crystals on the bulk scattering properties depends on the assumed shapes of small ice crystals, which is largest (smallest) in the temperature ranges of -45 < T < -30 T (T < -60°C) where the CAS/CDP ratio of N of small ice crystals is maximum (minimum).