Summary

This paper looks at cirrus clouds that formed in the extratropical transition layer using observations and model simulations. The authors look at the case study as Muller (2015), but include observations from a second set of instruments towed by the aircraft, in addition to the aircraft observations, which allows them to show the different gradients in and out of the cirrus clouds. The authors also run high-resolution simulations to determine the origin of air and the physical processes responsible for forming the cirrus and modifying the gradients in the extratropical transition layer. They show that radiation and turbulence are important for increasing PV above the cirrus, and microphysics for decreasing PV below.

This is an interesting study with interesting results but needs some clarification of figures and discussion before publication.

Corrections

• All figures
  • There is an inconsistent font size and it is sometimes quite small and difficult to read. I assume this is because you have images of various widths that are being resized. Please make the figures to a consistent width, so that the font size is consistent and readable.
  • The labelling of panels below each panel is confusing at first. A particular example is Figure A2, where the labels look like they correspond to the panel below.
• Section 2.1. For someone not intimately familiar with observational equipment, this was difficult to follow. There are a number of undefined acronyms, although I don’t think it is important to spell out the acronyms as long as the names of the instruments are attached to a citation or an explanation of what they do. FSSP, ICH, and MCH are not explained. This might be better presented as a table with columns for: quantity measured, sensor name, sensor reference, platform (Learjet or TOSS), sensitivity.
• Similarly in section 2 there needs to be explanations or references for what the IFS is and what R3BX means.
• Can you mark the start and end point of the leg you are analysing on figure 2. It’s not clear from the text exactly which part you are talking about. I think it is the roughly straight section from (6.7E, 55.05N) to (7.3E, 55N), given the discussion of figure 4, but I’m not sure. I then struggled to line this up with 3a. In figures 2 and 3a, there is a section with both Learjet and TOSS in cirrus, but in figure 2 this is surrounded by TOSS being continuously in cirrus, and in 3a TOSS is discontinuously in cirrus or there are no measurements?
• You state that the TOSS has a constant theta of 320K, but it shows about as much variation as Learjet.
• Figure 4. There are a lot of overlapping points in this figure, so it’s not clear whether the colour show is representative of the average or just the last point plotted. It might help to split each panel into two figures (one for each leg/distinct air mass), which would also back up the discussion which says that one mixing line corresponds to a different flight leg.
• L314 – It would be good to define the “measurement area” here. I assume it is the grid point range described in figure 5, but it would be better to also say it in the text. Also add a visual reference, such as the box in Figure 7, to figure 2 to give context to the size of the area in comparison to the flight leg.
• L227 - “This is in slight contrast to ERA5 derived PV in Fig. 3”. It’s hard to say if the model and ERA5 are actually inconsistent based on comparing figures 3 and 5. Figure 3 is exactly co-located with the flight track, whereas figure 5 is for all points in a box and the horizontal lines show substantial variability in PV at that level.
• L235 - “Notably, the dynamical tropopause in the model is about 500 m lower than the cloud top suggesting the presence of ExTL cirrus consistent with observations”. This is based on box averages so isn’t clear evidence. Particularly when you show in figure 6c that air parcels with cirrus on average have lower PV around the tropopause level (a higher tropopause). I would suggest to delete this sentence, because the following discussion of figure 6 showing regions with ice and PV>2 is more convincing
• Figure 6a – Similarly to figure 4, there are many points overlying each other, so it is not clear whether the colour is in any way representative and what sort of density of points occupy each region.
• Figure 6 b/c – There are two different blue lines in figure and legend not mentioned in the caption. Is the cyan line a subset of the blue line, for points with PV>2. The lines appear identical. Does clouds below 10km mean cloud tops below 10km or are there grid points included in both the blue and green lines where the clouds go across 10km?
• Fig. 7 – As with figures 4 and 6, the overlapping trajectories means the colour in the figure is probably not representative. The idea to focus on model trajectories that are stratospheric and have cirrus, rather than exactly co-locating model data with observed cirrus seems entirely justified anyway, so I would suggest removing Figure 7 and the discussion in section 4.1
• L285 - “The extensive cirrus cloud seen in Fig. 7 emerges in the 12 h before the arrival of the trajectories in the measurement area and arises by a combination of moist ExTL air over southern Germany and strong lifting of up to 1000 m (median: ≈ 500 m) during its northward propagation.” – What is this based on?
• Figure 8 – What does “substantial accumulated ice nucleation rates” mean specifically? (Same for L292). Depending on what you mean by this, this could account for your “not shown” conclusion on L289
• L300 - “If considering the parcel’s initial PV for constructing the vertical profile, the same structure emerges with small variations (not shown)”. Why not show this as an extra panel in Figure 8?

• Figure 9
  • Explain the green line and red hatching showing in (c)-(f), and the white lines, red hatching, and dots in (b) in the caption.
  • What is the difference between filled and unfilled hatches in (b)
  • Why is the red hatching slightly different in (b) compared to (c)-(f)?
  • Figure 9(a) is used as evidence against troposphere-stratosphere exchange, but the grey shading only shows the 25-75^th percentiles of PV, so there could still be a large amount of air parcels that fit the criteria. Also, as I understand it 9a shows the full set of trajectories that arrive in the measurement area, not just the cirrus subset. What fraction of these air parcels actually belongs to the cirrus subset?