Dear Authors,

I am very sorry for the very long delays, but I am glad that your manuscript has now recieved excellent reports from two independent referees. Based on these reports, I am happy to decide that the manuscript is accepted with minor revisions.

In making your revisions, please take into account the recommendations of the two reports. There are three more points that need to be considered in your revision. First, δD is no isotopologue ratio, but a δ-value. You should introduce it as such and use the correct denomination. Then, it is not clear whether this is discussed in the cited literature, but what is surprisingly missing in the introduction of the current manuscript is a short (physical) explanation why δD is/should be an interesting tracer for convective vertical transport, ie beyond your phrase that " δD is strongly affected by vertical transport and convective processes" in line 43 ... The physics is only mentioned much later (section 4) and explicitly referred to when super-Rayleigh events are mentioned. Finally, the "Nature" variable seems to be ill defined. As used in Figs 4 and 6, it is a quantity that depends on ω. Given the units, however, it is a vertical velocity (ω) with "nature" being a qualifier. So you should label the axis ω_nature with the common units of a vertical velocity. Do I understand this correctly ? In this case, Nature(ω) should be replaced by ω_nature (or whatever) throughout the manuscript.

In what follows are some more suggestions that mostly aim at simplifying and improving the reading or at correcting typos.

l25 societal impact -> societal consequences
l25 Because observations of dD at the location and time of these unstable atmospheric conditions are unavailable, this indicates a unique remote impact of elsewhere available δD observations on the analyses of these convective events. you might considering to replace 'remote impact' by '(long-)distance effect' or 'effect at a distance'
l33 determines -> determine
l36 modify -> modifies
l40 realted-> related
l41 check whether it is not preferrable to delete 'the analysed'
l42 assimilaton -> assimilation
l44 this spectrally resolved radiation -> this radiation spectrally resolved
l46 repetitive use of provide, please change
l51 remove the reference to Craig. It does not give the value of V-SMOW.
l119 analyses' ?
Caption table 3: "what" -> "which" , "referring the" -> referring to the
Table 3 Skill calculation should be replaced by Definition of skill
Table 3, see comment to l159.
l159 The definition of skill is a little bit surprising as it is the negative of a relative deviation with respect to the reference. Maybe you should write it a such :
skill = - (exp-ref)/ref. It is evident that you want it to be positive when the relative deviation diminishes. Interestingly, you later use 'loss of skill' which is the negative of the skill and corresponds to a relative deviation.
l161 "informs on the relative change of the RMSD value obtained from an assimilation experiment with respect to the reference assimilation experiment". As mentioned just before, it is the negative of the relative change with respect to the reference.
l163 'of' is missing
l181,184 etc. Consistency with figure caption 1: 'cumulative distribution functions' should be abbreviated either "cdf" or "cdfs" throughout. It is preferrable to use cdf for the singular case and cdfs for the plural case.
l185 For q all -> For q, all; shifter -> shifted
l196 Because for this calculation, continuous time series are used -> Because continuous time series are used for this calculation,
l202 Skill -> Skills
l215 Sect.4. -> Sect. 4.
l235 observation of q, T, and δD -> observations of q, T, and δD
l286 these large amount -> this large amount
l289 until we consider 98% of all the ∆i,j values occurring for the considered ω-bin -> until we consider 98% of all the ∆i,j values occurring for a particular ω-bin
Fig 4. Please indicate quickly in the text (maybe around l285) why downward velocities (0.5 Pa/s) are higher than upward velocities (-1.5 Pa/s) ...
Fig 5. use 10^3, ... 10^9 on the coordinate scale instead of 1000, 100000 ... 1000000000.
l306 This reduction of the ω uncertainty is more pronounced for the ∆{q,T,δD} than for the ∆{q,T} distributions -> This reduction of the ω uncertainty is slightly more pronounced for the ∆{q,T,δD} than for the ∆{q,T} distributions
l307 and largest for high actual ω values. -> and largest for high absolute values of ω.
l324 together 441253 events, which is 441253/586800 = 75.20% -> together 441253 out of 586800 events, which is 75.2%.
l325 Nature -> nature
l325 only comprises 5363 events, i.e. only 5363/586800 = 0.91% -> only comprises 5363 events, i.e. 0.91%
l330 Each of the groups -> Each group
l346 are depicted as blue line in Fig. 6 -> are depicted as blue lines in Fig. 6
l351 later (l385), unstable conditions are characterized by (Nature(ω) < −0.2 Pa/s). Please be consistent in your definition of unstable events.
l354 However, for the infrequently occurring events corresponding to unstable atmospheric conditions, δD observations become at least as important as T observations. -> However, for the infrequently occurring events corresponding to unstable atmospheric conditions, δD observations become at least as important as T observations for all variables except for T itself.
l364 seems -> seem
l392 This real world data are only available -> This real world data is only available
l400 rare -> rarely
l401 are -> is
l417 we need atmospheric isotopologue enabled models that capture as much as possible details of a convective atmosphere. -> we need atmospheric isotopologue enabled models that capture as many details of a convective atmosphere as possible.
l421 cloud-free free troposphere -> cloud-free troposphere
l425 highly resolving models -> high resolution models
l437 observationsare -> observations are
l442 does does strongly further improve -> further strongly improves
l446 we use atmospheric vertical velocity -> we use the atmospheric vertical velocity
l454 indicates to -> indicates; elsewhere available
ll456 you might considering to replace 'remote impact' by '(long-)distance effect'
l457 & 461 'depleted signals' is misleading -> signals of (isotopic) depletion, or (isotopic) depletion signals
l458 the δD observations -> δD observations
l460 "achieving" sounds strange here. Maybe you should write "... as a promising indication that δD can have the same remote impact in the real that it has in our OSSE study for a simulated world."
l470 important contribution -> "important contributions" or "an important contribution"

Dear authors,

Thank you for taking into account all suggestions and corrections that have been proposed previously. The paper can now be published and this is the occasion to thank you again for submitting to Atmospheric Measurement Techniques. Please consider also the following technical corrections when preparing the document to be sent to the production office:

l 41 'ocean water' instead of 'Ocean water'
l 184 'Q_2' in italics
l 313 consider 'most negative' instead of 'strongest negative'
l 459 'This indicates to a' -> 'This hints to a'

With best regards,
Christof Janssen