The paper addresses the interesting topic of the Asian tropopause aerosol layer (ATAL). The presented results provide interesting information regarding the source regions and the transport pathways of the ATAL based on balloon-borne measurements and two Lagrangian transport models driven by different reanalysis data (ERA5, ERA-interim) and model parameterisations/setups. The investigation of the ATAL is a topic of many studies during the last years. The added value of the present paper is the thorough intercomparison between the results obtained with different simulation scenarios (with different Langrangian model employed, the use of different reanalysis data -ERA5 vs ERA-interim-, vertical coordinate -kinematic vs diabatic approach- and convective parametrisation).

In my opinion, since the publication focuses mainly on the comparison between the different simulation scenarios, this should be reflected in the title. Overall the manuscript is well written and I recommend its publication in ACP after providing some clarifications and addressing some minor issues.

Specific comments

I think that it is better to use the term "model boundary layer" instead of the term "model boundary". (at least when it is mentioned for the first time).

The Asian highlands is not a continuous region. It would be interesting to mention the percentage of air masses with origin only from the Thibetan Plateau.

l45: "up to 360K". It would be useful to clarify here (for readers less familiar with the subject) that you are referring to potential temperature.

ll120-122 "The hybrid coordinate ζ ... above around 300 hPa and 380 K, respectively. "Please rephrase to "The hybrid coordinate ζ is near the surface an orography-following sigma coordinate and transforms continuously into potential temperature at higher altitudes above around 300 hPa (or 380 K).

l229 "transport pathways from two regions with two very different land-cover properties converge, the Tibetan Plateau or the Indo-Gangetic plain." please correct to "Tibetan Plateau and the Indo-Gangetic plain"

l240 and Fig. 1d: please correct the typo "Arabian See" -> "Arabian Sea"

l245 "from the Bay of Bengal more air parcels can enter the ASMA directly from the maritime boundary than from the Arabian Sea". Can you elaborate on this and provide an explanation? Also why the mean transport time from the MB into the UTLS is higher compared to the AS?

ll253-254: "Wide agreement can be found with ERA5 even when models, integration step-sizes, and vertical velocities are varied." and l268: "Model scenarios driven with ERA5 show very similar results." The scenario with extreme convection parameterisation (E5-kin-M-ECP) is also driven by ERA5 reanalysis data. It's better to mention this here (e.g. Model scenarios driven with ERA5 show very similar results except when employing the extreme convection parameterisation)

According to Figs 3b and 4c under the scenario with ECP the contributions from SE Asia increase. Please elaborate more on this in the text.

ll300-301: " This leads to... Indo-Gangetic plain." Please elaborate more on this.

ll313-314: Please rephrase (begin the sentence with "For ERA-Interim...")

l322: " However, our approach of the ECP has to be considered as an upper limit for convective transport." Can you elaborate more on this? In the Appendix F You mention that "the parametrization can be improved to avoid spurious parameterized convection events over the Persian gulf and the Red Sea". It was not possible to further imporove the parametrization? I mention this because I am concerned about the large differences between the scenario with ECP and the other scenarios.

Appendix F: Please provide a table with a short description of the scenarios shown in Fig F1. (similar to Table 2)

parametrization vs parameterisation: Please choose one spelling for consistency.

This paper’s goal is to trace the source regions and pathways of air that forms the ATAL over Nainital, India in August 2016 using two different Lagrangian transport models (CLaMS & MPTRAC) utilizing two different meteorology sources (ERA-Interim & ERA5), and two different vertical coordinates (kinematic and diabatic) to create an ensemble of results to measure the robustness source locations.  This paper adds significant value in that it compares all of the results of these models, meteorology sources, and vertical coordinates to identify the most likely source regions and pathways of air that forms the ATAL.

I fully agree with the other reviewers comment that since this study focuses so heavily on the comparison of different simulation scenarios, the modeling and transport aspects of this work should be reflected in the paper title in some way. 

I think this paper is well written and thoughtfully laid out and recommend publication in ACP after a few very minor clarifications and revisions.

Line 61-62, it may be better to say that ‘These calculations rely on reanalysis data and their ability to adequately resolve convection and diabatic vertical ascent in the ASMA.’  since no model ‘correctly’ resolves convection.

Line 84, It doesn’t appear that the authors are dealing with any explititly modeled convection in this paper.  The convection in all of the models described in this paper use parameterized convection.

Line 133-139, Could reference Appendix F in this paragraph.

Line 210, can also get a third - albeit less frequent - ASMA mode over the western Pacific Ocean near Japan (from Honomichl and Pan, 2020). 

Line 317, maybe a better way to say is that the parameterization of convection performs better in ERA5 than it does in ERA-i

Line 319-320, Is the ERA5 convective parameterization missing the whole onset of convection or is the convection present but just as deep compared to the ECP?  The additional information could be a good Segway to the comment that ECP is considered an upper limit for convective transport.

Line 368, several 10% is a little confusing.  I’m not entirely sure what you mean by it.

Line 451, parameterizes convection better than ERA-i

Line 451, “it might still underestimate”