A New Outlook on Ice Cloud through Sub-Millimetre-Wave Scattering

Scattering by atmospheric ice at sub-mm-wave frequencies is a challenge to both the cloud physics and light scattering communities owing to scattering at these frequencies being dependent on assumptions about the particle size distribution, ice crystal shape, orientation and size. Moreover, the scattering also depends on how the particle density is assumed to evolve with size. As there is as yet no prediction of a universal PSD or mass–dimension or density–dimension relationship, the modelling of ice crystals, so as to conserve the observed scattering and ice mass, is potentially problematic. In this presentation, the challenge presented by sub-mm-wave scattering is explored through the study of an ice cloud case using a new sub-mm spectral-like radiometer that was deployed on board an aircraft.

Here, we evaluate the predictive quality of applying members from an ensemble model of cirrus ice crystals to modelling observed sub-millimetre brightness temperatures. The airborne straight and level near-nadir observations used here were from a case of ice cloud, which occurred during a winter period. The airborne microwave observations were obtained using the International Submillimetre Airborne Radiometer (ISMAR) [1], as the observations collected were at near-nadir we do not as yet consider polarisation. The ISMAR instrument has five central frequencies located between 118 and 664 GHz, with a number of sub-channels situated around some of the central frequencies to obtain spectral-like observations. The frequency selected for presentation is the 664 GHz “window” channel. This channel selection reduces uncertainties in modelling the gaseous spectroscopy, thereby enabling the scattering properties of members of the ensemble model to be more directly evaluated at this frequency. This is also the frequency that is most sensitive to assumptions about the ice crystal models and microphysics.

The methodologies adopted for the calculation of the single-scattering properties of the ensemble model members at this frequency have been previously peer-reviewed and published [2, 3]. As such, this presentation concentrates on the application of these methodologies to the interpretation of the airborne ISMAR observations using a fast, state-of-the-art line-by-line radiative transfer model [4]. Moreover, state-of-the-art airborne observations of particle size distributions (PSDs) were also collected from the ice cloud case. These in-situ PSDs, as well as an often used database of in-situ PSDs collected during the SPARTICUS campaign in 2010, are applied to the two most compact and spatial hexagonal ice aggregate members of the ensemble model. A further ice aggregate model, called the Voronoi model, forming a chain of polyhedral particles, constructed to follow an observed density–dimension relationship, was also applied so as to simulate the observations.

From the in-situ PSDs, geometric optics-based power law relationships have been previously obtained between the ice water content and the bulk extinction coefficient [5]. These same geometric optics-based relationships were estimated using the area–dimension power laws predicted by the ensemble model members and the Voronoi model. The best-fit ensemble model members to the observed power laws, and the Voronoi model, were applied in order to simulate the sub-mm-wave observations. Thus, we demonstrate consistency of model application from the limit of geometric optics (i.e. typically at visible wavelengths) to the sub-mm.

In this presentation, we demonstrate a general overlap between the uncertainty in the radiative transfer simulations assuming the ensemble model members and the uncertainty in ISMAR brightness temperature observations at 664 GHz. However, portions of the straight and level runs were either simulated well with the compact aggregate model member or a three-component model, consisting of the two members of the ensemble model and the Voronoi particle, but never with one and the same model. Owing to the Voronoi model being the most spatial of all the models, this model simulated, to within the upper end of the experimental uncertainty, the ISMAR observations, but never the coldest observations at the highest sub-mm-wave frequency. However, if a different density–dimension relationship were to be adopted in the modelling of the Voronoi model that predicted higher mass values, then this should result in an improved agreement with the observations. It is as yet unclear as to which density–dimension relation is best to apply in general. These observations indicate changes in microphysics in terms of the mass–dimension profile and/or the size of the ice crystals and, therefore, represent a challenge to the global retrieval of ice cloud properties using the Ice Cloud Imager (ICI), which is due for launch around 2022. A further uncertainty is the assumed parametrised shape of the PSD. We also show in this presentation that the choice of PSD and ice crystal models are of equal importance in interpreting sub-mm-wave observations.

[1] Fox, S et al., 2017: ISMAR: an airborne submillimetre radiometer. Atmos. Meas. Tech., doi:10.5194/ amt-10-477-2017.
[2] Baran, A. J., et al., 2018: The applicability of physical optics in the millimetre and sub-millimetre spectral region. Part II: Application to a three-component model of ice cloud and its evaluation against the bulk single-scattering properties of various other aggregate models. JQSRT. 206, 68-80.
[3] Baran, A. J., Hesse E., and Sourdeval O., 2017: The applicability of physical optics in the millimetre and sub-millimetre spectral region. Part I: The ray tracing with diffraction on facets method. JQSRT. 190, 83-100.
[4] Havemann, S et al., The Havemann-Taylor Fast Radiative Transfer Code (HT-FRTC): a multipurpose code based on Principal Components, submitted to JQSRT (February 2018).
[5] Fox, S et al., 2017: ISMAR: an airborne submillimetre radiometer. Atmos. Meas. Tech., doi:10.5194/ amt-10-477-2017.