Fig. 1
Meteorological situation at 00 UTC, 16 September 2000 from Bracknell UK MetOffice Analysis for surface pressure including fronts. This analysis was kindly provided by UK Met Office under open Government Licence (see also http://www.nationalarchives.gov.uk/doc/open-government-licence/).
Fig. 2
Backward trajectory calculations with LAGRANTO (for 42 h). The colour indicate altitude in pressure coordinates. Note, that the air masses are quite collocated in the vertical direction after passing about longitude λ≈0°, which corresponds roughly to a time of t=12–14 h relative to the starting point. The upward motion during the last 6 h correspond to a median value of med(w)~0.03 m s−1.
Fig. 3
Backward trajectory calculations for investigating the air streams, started at t=0 h (corresponding to 06 UTC, 16 September 2000) at ECMWF model levels 31–34. On the top panel, the altitude in pressure coordinates is shown, whereas on the bottom panel the temperature evolution is displayed.
Fig. 4
Vertical profiles of temperature, relative humidity over ice, potential temperature and horizontal wind, respectively, from high-resolution radiosonde data (Δz~50 m) and ECMWF operational analyses are shown. Additionally, the upper and lower boundary of the potentially unstable layer are indicated in the profiles, as derived from equivalent potential temperature.
Fig. 5
Determination of unstable layer using ‘vertical velocity’ as derived from radiosonde data and equivalent potential temperature. On the left panel the vertical velocity perturbation relative to the mean value for radiosonde ascent (<w>=5.15 m s−1) is shown. The middle panel represents a zoom of the left panel into the interesting altitude range of 7≤z≤12 km. For comparison, the equivalent potential temperature profiles (radiosonde/ECMWF) are shown in the right panel.
Fig. 6
AVHRR image (channel 5: λ=11.5–12.5µ m) at 05:31 UTC, September 16, 2000, showing thick cirrus clouds over Germany. This satellite image was kindly provided by NERC Satellite Receiving Station, Dundee University, Scotland (http://www.sat.dundee.ac.uk/).
Fig. 7
Initial profiles of temperature, relative humidity over ice and horizontal wind, respectively, for realistic simulations with the EULAG model.
Fig. 8
Initial profiles of potential and equivalent potential temperature for realistic simulations with the EULAG model. Top: Whole vertical profile, bottom: Zoom into potentially unstable layer.
Fig. 9
Time evolution of convective cells inside the ISSR in the x-z-plane (horizontal extension in kilometres vs. altitude in kilometres) – early stage of cirrus cloud (t=75, 90, 105, 120, 135, 150, 165, 180, 195, 210 min). Colour bar: relative humidity over ice, black lines: isolines of ice water content (increment: ΔIWC=5 mg m−3).
Fig. 10
Time evolution of convective cells and the layer cloud in the x-z-plane (horizontal extension in kilometres vs. altitude in kilometres) at a later stage (t=270/330/360 min). Colour bar: relative humidity over ice, black lines: isolines of ice water content (increment: ΔIWC=5 mg m−3).
Fig. 11
Time evolution of a tracked cell during the time interval 90 ≤t≤130 min (Δt=10 min). The panel shows the time evolution of vertical velocity perturbations, relative humidity over ice, ice crystal number concentrations and ice water content, respectively. Left: 2D structure of the cell. Right: Vertical profiles at the centre of the cell.
Fig. 12
Timescales and timescale ratios for the tracked cell at simulation time t=130 min. Top row: Timescales vs. ice mass mixing ratio for relative humidity timescales (left panel) and ice mass concentration timescales (right panel). Bottom row: Timescale ratios vs. vertical velocity for relative humidity timescales (left panel) and ice mass concentration timescales (right panel).
Fig. 13
Timescales for the whole simulation time. Top row: RHi timescale ratios, left: ratios ; right: . Bottom row: q c timescale ratios, left: ratios , right: .
Fig. 14
Joint two-dimensional probability density for relative humidity over ice inside cirrus clouds. The frequency of occurrence for events of certain relative humidity over ice and ice water content is shown for events with ice crystal number concentration larger than 1L−1. Thermodynamic equilibrium (RHi=100%) is indicated by a black line.
Table 1. Probability (in percent) for cloudy air (Ni ≥1 L−1) of a certain state of (super/sub)-saturation and ice water content, respectively
RHi≤100%36.477.360.38RHi~100%11.492.930.18RHi≥100%52.0414.231.68RHi≥120%20.085.680.90RHi≥140%5.241.550.22
[i] This table represents integrated values of the two-dimensional probability density, as shown in Fig. 14.
Fig. 15
Probability density of relative humidity over ice in cloudy air (red), clear air (green) and all data (blue) for the time intervals 60≤t≤210 min (top panel, active convection) and 210≤t≤400 min (bottom panel, decaying convection). For the later time period, the distributions are fitted by an exponential distribution of the form p(RHi)=a·exp(−b·RHi). The exponents b are indicated in the figure.
Fig. 16
Probability density of vertical velocity in cloudy air (red), clear air (green) and all data (blue) for the time intervals 60≤t≤210 min (top panel, active convection) and 210≤t≤400 min (bottom panel, decaying convection).
Fig. 17
Probability density of ice crystal number concentrations for the time intervals 60≤t≤210 min (red) and 210≤t≤400 min (blue).
Fig. 18
Vertical profiles of potential temperature, entropy ice potential temperature, Brunt-Vaisala frequency and squared moist Brunt-Vaisala frequency, respectively, at times t=0 min and t=360 min, shifted to an equivalent height (i.e. start height).
Fig. 19
Potential temperature profile as obtained from the radiosonde (thick line) and eight realisations from the simulation at a corresponding simulation time (t=195 min). The profiles are shifted by Δθ=0.5 K for a better representation.