The existence of cloud free air masses
in the state of supersaturation with respect to ice was proven almost 60
years ago. E. Glückauf (1945) found from hygrometer data obtained
over southern England that (very high) supersaturation with respect to
ice occurs very frequently in the upper troposphere. H.Weickmann concluded
in his 1945 review paper on "Shapes and formation of atmospheric ice crystals"
(Weickmann 1945) that ice crystals in the atmosphere, i.e. cirrus clouds,
form mainly via the water phase and not as soon as ice saturation is reached.
He characterized the ice forming regions in the upper troposphere and the
(lowermost) stratosphere as regions of high ice-supersaturation but with
small absolute humidity. The insights of Glückauf and Weickmann were
obviously deemed unimportant for numerical weather prediction and climate
studies; after 50 years there is still no single weather prediction model
(to the author's knowledge) that would allow ice-supersaturation to occur.
The single exception we are aware of is one version of ECHAM (Kärcher
and Lohmann 2002; Lohmann and Kärcher 2002).
Cloud free ice-supersaturated air masses
have been detected by various types of hygrometers during several airborne
measurements campaigns like POLINAT (Ovarlez et al. 2000), SUCCESS (Jensen
et al. 1998) or SONEX (Vay et al. 2000) and in the Measurement of ozone
by Airbus in service aircraft (MOZAIC)
project (Marenco et al. 1998). It is now even
possible to use carefully calibrated and corrected RS80A radiosondes (Nagel
et al. 2001) to detect ice supersaturation (Spichtinger et al. 2003).
The realisation that there is often substantial
ice supersaturation in the upper troposphere is also in accordance with
recent results from laboratory work on homogeneous freezing nucleation of aqueous
solution droplets (currently believed to be the dominant cirrus formation
mechanism below -40°C). Koop et al. (2000) were able to show that the freezing of
aqueous solution droplets in the cold upper troposphere needs substantial
supersaturation of generally more than 40% (in order to make the solutions sufficiently
thin). The mere existence of cirrus clouds therefore implies that there
must be a lot of supersaturation in the upper troposphere.
A good marker of ISSRs is persistent condensation
trails (contrails) when the sky is otherwise free of clouds. Since the
mixing process in an aircraft exhaust plume can create very high degrees of supersaturation
even in dry ambient air, the formation of contrails does not require as
high ambient humidity as the formation of natural cirrus. Contrails can therefore
decorate the sky when no cirrus clouds are around. Contrail persistence
however requires at least ice saturation. A sky full of contrails but without cirrus therefore
shows that there must be an ISSR.
In the recent years it has become quite
evident that cloud free air in the upper troposphere is often in a thermodynamic
state of supersaturation with respect to ice. Such regions have been termed "ice-supersaturated
regions" (ISSRs, Gierens et al., 1999).