The Seasonal, Latitude and Temperature Dependence of Homogeneous Ice Nucleation

A long-standing scientific question producing great uncertainty in the climatic role of cirrus clouds is whether they are formed through homo- or heterogeneous ice nucleation (henceforth hom and het). Since ice particle concentrations in hom cirrus typically range from 500 to 50,000 liter-1 and in het cirrus typically range from 0.5 to 500 liter-1, the radiative properties of the two cloud types differ markedly.

We provide a means of answering this question using CALIPSO (Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation) satellite retrievals of the effective absorption optical depth (τabs) ratio, or βeff, based on Imaging Infrared Radiometer (IIR) channels at 12.05 μm and 10.6 μm. A new physical understanding of βeff reveals that it is a relative measure of the concentration of small ice crystals (length D < 60 μm) in an ice particle size distribution (PSD) and that it is also tightly related to N/IWC where N = ice particle number concentration and IWC = cloud ice water content. Combining this understanding and βeff retrievals with co-located CALIOP (Cloud-Aerosol Lidar with Orthogonal Polarization; aboard CALIPSO) measurements of geometric cloud thickness Δz_cloud, the effective diameter De-temperature (T) relationship of Heymsfield et al. (2014, JAMC) and the IIR τabs(12 μm) retrievals yields estimates of N for semi-transparent cirrus clouds, based on these relationships: τabs(12.05 μm) ≈ 3 IWP/(2 ρi De), where IWP = ice water path and ρi is the bulk density of ice; cirrus layer mean IWC = IWP/Δz_cloud, and N = (N/IWC) IWC where N/IWC is derived from βeff. Frequency distributions of IWC and their T-dependence based on this method, from 186-210 K and from 30S to 30N over oceans, are shown to agree favorably with those obtained from CALIOP and other instruments under similar conditions. Since these N retrievals are sensitive to the smallest ice crystals having the highest concentrations, they provide realistic estimates of N and thus can be used to differentiate between hom and het conditions (using 500 liter-1 as the het-hom transition marker).

Qualitatively, our results for semi-transparent single-layer cirrus clouds (meaning the lidar aboard CALIPSO can detect cloud base where T < -38°C, and that cloud emissivity at 12 μm ranges from 0.1 to 0.7) show that the occurrence of hom cirrus depends on latitude, season, orography and whether the cirrus are over land or ocean. Hom cirrus dominates in the Polar Regions during winter when cirrus cloud coverage is maximum, resulting in strong net radiative cirrus forcing. Hom cirrus continues to prevail over Antarctica during spring and fall. This is consistent with predicted seasonally-dependent concentrations of mineral dust in these regions. These hom cirrus may exert a strong greenhouse effect on polar surface temperatures as GCM studies have found, perhaps raising them several degrees relative to het cirrus conditions. The observation that het conditions manifest during polar summer (perhaps due to increased mineral dust concentrations) indicates that a delicate seasonal balance may exist between hom and het conditions. Since the Polar Regions are most sensitive to the effects of climate change and are where a climate tipping point is most likely to manifest, it is important to understand all the factors determining polar surface temperatures.

N was consistently higher over land than over the oceans. Over land, hom cirrus were more common in the N. Hemisphere between 30N-60N during winter. Finally, hom cirrus were always common over Patagonia, apparently due to orographic forcing with low mineral dust concentrations.

To effectively cool the planet, cirrus cloud climate intervention requires hom cirrus with adequate cloud coverage in the high latitudes during fall-winter (i.e. at sun angles that are low or absent). To a large degree, these requirements appear to be satisfied based on these CALIPSO retrievals.