General comment

This paper aims at addressing the radiative impact of contrails in the North Atlantic between 2016 and 2021. This is an important issue since contrails represent a non-CO₂ effect of potentially high magnitude. In order to select appropriate mitigation strategies, comprehensive knowledge of the most significant processes is needed. To this respect this paper provides an important contribution and I strongly recommend its publication in ACP.

The methodology is based on the use of several blocks:

• Air traffic information
• Meteorology data over a 6 years period
• Aircraft type and performance, mass, engine properties, emissions
• A contrail-cirrus prediction tool.

Each set of data required the use of various tools and methodologies, for instance to estimate aircraft engine emissions for cruise conditions from ICAO LTO dataset or to apply corrections to the ERA5 humidity fields. It looks to me that this represents a tremendous amount of work and the application of a strategy/methodology following already previously published work.

The paper is clear and very well written. It is an impressive work.

Specific Comment

Line 245-250: The mean ice crystal radius is smaller in wintertime relative to the summer. Less condensable matter is available (RHi is lower in figure S14e) probably explains this point. Mentioning that contrails are formed at lower temperature (“lower temperature with less condensable water”) is finally confusing since the effect of temperature is already included in the RHi which drives particle water uptake.

I probably missed a definition of the mean radius r_ice. Is it averaged along the contrail forming flight distance?

Supporting information line 260: there are quite a number of papers deriving saturation water vapour pressures. Has the choice of Sonntag (1994) been made for consistency with some other data, for instance in ERA5. Surely the choice of the parameterization used can modify RHi significantly and the predicted ISSR.  

Technical corrections

Line 51: Missing “.” after “…Heymsfield, 2017)”

Line 78: There are actually two ICAO 2021 references in this paper and they should be called differently. One of them is misplaced in the References section (line 504).

Line 122: delete “and” after “ensembles)”

Line 226: Correction “than it would have been” instead of “than it would been”

Line 425 and 427 correct Schumann & Graf

Supporting Information Line 441 Table S5 caption : spelling « annonymised »

Teoh et al. make use of detailed flight recordings over the North Atlantic ocean in combination with a number of parameterisations to estimate various aspects of the climate effects of contrails in this region. A specific aspect of their study is the estimate of the reductions in air traffic due to the COVID-19 pandemic. The study is of interest to the readers of Atmos. Chem. Phys. It is very diligently conducted and excellently written. I have only a few minor remarks that should be addressed before publication.

Minor remarks

l48 The authors should discuss the IPCC AR6 assessment.

l223 – 225: it is unclear what are local / regional effects and what are global ones. The net cirrus RF given in l224 in my understanding is regional, for the region of interest of the North Atlantic ocean. The contrail climate sensitivity of Kärcher (2018) is a global number. To estimate the cooling effect (0.05 to 0.07 K), was some effort made to extrapolate the RF globally? Or is there a reason to believe the cooling would be confined to the region where the RF occurs? Also I have trouble seeing where the 0.05 K lower bound comes from. Isn’t it rather 0.02 K?

l229 Where is this number seen in Table 1?

l262 A bit puzzling logic. The situation appears particularly frequently in summer, much more so than in winter, i.e. at times where a net cooling would be more likely.

l267 How is that possible? Shouldn’t there be more incoming sunlight in summer (cf Fig. S14i)?

l282 EF was introduced as an integral measure, why would one now normalize again by contrail length? Why not length and width and go for forcing?

l285 Fig. 3. The legend says “time of day” but really it is UTC, isn’t it? There is a large ambiguity on which time (in UTC) is sunrise and which time is sunset, given the breadth in longitude of the Atlantic Ocean. Else it would be useful to indicate the time (spans) of sunrise and sunset that are discussed in the text. As written above, I do not understand the usefulness of EF per length, why not stick to EF, or else omit this panel.

l321 same comment as above on SDR

l336 The half-sentence “below optically thick high-level cirrus” should better start a new sentence with the second argument/condition. But this is not so obvious. If the cirrus are optically thick, why would they not have the same effect in the solar as the optically thick low-level clouds?

l347 It is noteworthy perhaps that the cirrus that are neither strongly cooling nor strongly warming seemingly have a smaller absolute effect in either direction, they occur at smaller nvPM.

l371 Is there an explanation for this result? Is there a reason to believe the HRES resolution is better? Is it appropriate, or would still higher resolution lead to still smaller results?

l381 Is there some problem in the parameterisation that leads to this strong increase in contrail age simply because the input fields have a coarser temporal resolution?

l414 At the end of this uncertainty / sensitivity section, it would have been nice to do an overall uncertainty quantification by propagating all uncertainties to an overall uncertainty on the assessed RF.

l443 This could be a point where the overall uncertainty is reported.

Typos

l191 “ensemble”

l226 “would have been”

l328 “persist”