General comments:

Jeske and Tost describe improvements of the CVTRANS submodel. CVTRANS is an existing convective tracer transport algorithm based on mass fluxes from a convection scheme using a bulk approach. The novelties in this manuscript are additions to the equations in CVTRANS to prevent turbulent entrainment/detrainment from being eliminated and the implementation of a convective exchange matrix. Results from applications of the new CVTRANS are presented showing the changes in terms of convective tracer transport for 2 different time periods and for selected regions. This is an interesting study and fits into ACP. The paper is well structured.

My main concern regarding this manuscript is the missing of an evaluation of the main development, the modifications in CVTRANS related to entrainment/detrainment. Are the corresponding changes in the transport matrix well-founded? The English language is sometimes difficult to understand. The manuscript is rather long and should be shortened.

Specific comments:

Titel: The historical climate trend resulted in changed convective transport patterns in model simulations.

I would suggest to be more specific in the title: ...changed vertical transport patterns in climate model simulations.

Page 1, line 6: The statement that convection transports material less efficiently to higher layers, but transports more to the tropopause region seems contradictory to me.

Page 1, line 15: rapid vertical transport time scales -> rapid vertical transport or short vertical time scales.

Page 3, line 61: MESSy is an interface --> MESSy is a software framework ... (see also www.messy-interface.org)

Page 3, line 68ff: I suggest to provide a complete overview of the submodels used (similar as in  Tost et al., 2010, their Table 1)

Page 3, line 72: Are SSTs and sea ice really nudged or are they prescribed?

Page 5, line 18: detainmen or entertainment -> detrainment or entrainment

Page 6, line 149: How large is a small number of tracers?  < 10?

Page 7, line 169: Where does the 6.96% come from? I looked for it in the main diagonal, as described in Fig. 1, but probably the color bar does not discriminate enough between the values.  What is meant by "large-scale subsidence is a strong process"?

Page 7, line 173ff: I would suggest that pressure values should be rounded to full hPa.

Page 7, line 182: I am surprised that the Tiedtke scheme can provide mass fluxes above the tropopause.  Is the convective height in accordance with the height provided by the Tiedtke/Nordeng convection parametrisation? Please comment.

Page 8, Fig.2c: How can it be that from level 14 (origin level) the upward transports (in CVTRANSturb)  is smaller than from the origin layer 13? In your Fig.2c there is a jump visible, which is also present in Fig. 2b (same levels), but smaller in level 13 (it is the dark red area above the diagonal with values in the thousandths of a % range). The results from CVTRANSturb seem unrealistic. Could you please comment.

Page 8, line 208: I don't understand the reference to Tiedtke regarding "maximised turbulent mixing".  What is meant by that (I could not find it the Tiedtke publication)?

Page 9, line 212: “CVTRANSturb is not efficient enough transporting material to the upper troposphere”. This statement is a validation of the transport in CVTRANSturb, which however is not justified. See my main concern in the general comments.

Page 10, line 228: Couldn't one have shown the convective transport matrix only for the cases where convection has occurred? And differentiated according to the 3 types of convection (deep, mid, shallow)?

Page 10, line 238: the effects by downdrafts are striking

Page 11, line 258: I doubt that convective overshooting is present in the Tiedtke-Nordeng convection scheme. Did you find overshooting events in individual deep convective events? This should be clarified and the text adapted accordingly. May be this is an effect of averaging the tropopause height and should not be interpreted as “overshooting”. Please check.

Page 12, line 269: Is it the increased penetration of convection that leads to a rise of the tropopause? According to Meng et al., 2021 it is the warming/cooling of the troposphere/stratosphere that modifies tropopause height (https://doi.org/10.1126/sciadv.abi8065 )

Page 13, symetric -> symmetric

Page 14, Fig. 7: You state “crosses” in the figure, but I only see filled circles!

Page 15, line 302: I find it somewhat unfortunate to speak of "zonal trends", since in principle this also includes spatial trends (along a zonal band). Suggestion for a section title: Tropical and extra-tropical trends

Page 15: line 311: I doubt, that stratospheric levels are reached, can you please check the height from CVTRANS results with the actual tropopause height from the convection parametrisation?

Figs. 9 and 10 Caption: omit “in the tropics”

Page 17, line 329 – 337: This text passage is difficult to follow. Could you please summarize the important differences with respect to the global and the NH case?

Page 18, line 340ff: Did you mean: the definition of the upper troposphere is defined as the vertical region between the level of the tropopause (in hPa)  + 150 hPa?

Page 18, line 350: The authors evaluated the results of CVTRANSturb by a bold comparison between 2 different variables (convective rain against medium transport into the upper troposphere by CVTRANSturb). This is not convincing. Others, I would generally have expected that an evaluation of CVTRANSturb already took place earlier in the results section.A comparison of convective rain rates would only make sense, if convective rain rates of EMAC were presented, but an evaluation of Tiedtke/Nordeng rain rates is probably not the topic of the paper.

Page 21, line 414: …to identify climate change in convective activity.

Page 22, line 439: Are sea surface temperatures and sea ice nudged or prescribed?