Improved Representations of Ice Microphysical Properties Derived from In-Situ Aircraft Observations during ARM Field Campaigns

Poster PDF

Description

To improve representations of cloud processes in models, a database on the dependence of ice cloud properties on environmental factors has been developed using ARM field campaign data. Error analysis is included in the database in order to determine whether differences between cloud properties sampled in different conditions exist and whether variations between observed and modeled fields are statistically significant. Here the techniques used to determine uncertainties in observed microphysical properties are shown, concentrating on statistical, observational and processing sources of error. These uncertainties are taken into account when comparing derived properties from cloud probe measurements made in stratiform regions behind a mesoscale convective system sampled on 20 May 2011 during the Mid-latitude Continental Convective Clouds Experiment against fields simulated using the Weather Research and Forecasting model with 3 different spectral bin microphysics schemes. Comparisons between the intercept (N0), slope (λ) and shape (μ) of observed and simulated size distributions (SDs) are quantified using ellipsoids in (N0, λ, μ) phase space to represent volumes of equally realizable solutions. Large differences in SDs are found among the 3 bin schemes and between the simulation and observation. Various microphysical process rates (e.g., nucleation, diffusion, aggregation, melting) are output and compared for different schemes. Assumptions about particle properties (mass/terminal velocity-dimensional relations, etc.) and representations of processes in different bin schemes are investigated to explain differences between models and observations. Knowledge of single-scattering properties is also required to determine cloud radiative effects. Although it is known that scattering properties of non-spherical ice crystals differ from those of spheres, the impacts of those differences on parameterizations for models and retrievals are unknown. Using several idealized models to represent small crystal shape (droxtals, Gaussian random spheres, budding Bucky balls, Chebyshev particles) and exact (discrete dipole approximation) and approximate (improved geometric optics method (GOM) and conventional GOM) numerical methods, single-scattering properties are computed at non-absorbing and absorbing wavelengths. Impacts of varying crystal shape, size, area ratio and absorption on the single-scattering properties and radiative forcing are discussed.