General Comments

This paper uses the closed-loop integration method, together with OMI and TROPOMI satellite observations and ground-based monitoring data, to systematically analyze the spatiotemporal variations of tropospheric NO₂ column concentrations, surface concentrations, and boundary transport fluxes over the Tibetan Plateau from 2005 to 2024. Given the important role of the Tibetan Plateau in regional and global climate systems, this study focuses on the key issue of transboundary NO₂ transport. A random forest approach is also introduced to assess the relative contributions of the westerlies and the South Asian monsoon at different altitude layers to the transport flux. Overall, the results contribute to a deeper understanding of regional pollutant transport processes and their controlling factors and provide references for studies on cross-regional coordinated emission reductions. This study employs a series of advanced datasets and innovative methods, and the relevant conclusions hold scientific value and provide useful reference. This paper fits the scope of ACP, and I recommend publishing it on ACP after some issues have been addressed.

Some aspects could be further refined to enhance the readability and rigor of the manuscript.

1. I noticed that the authors did not include a section on the validation of the machine learning results in the main text, although I saw it in the appendix. I believe this is a crucial part and the validation parameters (such as R2 and RMSE) should be presented in the main text. I recommend including a table in the main text to demonstrate this.

1. There are minor grammatical and expression issues in the manuscript, which are suggested to be revised to improve overall readability.

Specific comments:

1. Page 2, Line 25: “…currently operated in TP, Taking Tibet as an example…” → “…currently operated in TP. Taking Tibet as an example…”

1. Page 5, Line 29: “To ensure strict temporal and spatial consistency, …” should be “To ensure strict spatiotemporal consistency, …”.

1. Page 6: The formula formatting here is incorrect; the author should adjust it. The same issue appeared in the supplement.

1. Page 7, Line 1: “…some grids in this region that previously exhibited mixed fluxes shift from internal efflux to external influx…” The definition of “mixed fluxes” is not sufficiently clear.

1. Clarify the meaning of “mixed fluxes” in the text, for example, whether it refers to net fluxes close to zero or to spatiotemporal alternation.

1. Page 8, Line 2: A large number of references are cited (Guo et al., 2022; Kong et al., 2023; Zheng et al., 2024; Zhang et al., 2025), but the specific phenomena or mechanisms corresponding to each reference are not clearly indicated in the text.

1. Clearly indicate the specific phenomena or conclusions corresponding to each reference when citing, to avoid citation stacking.

1. Page 8, Lines 16-21: The text repeatedly mentions “high-value” or “high-value source regions” → it is recommended to define it upon first occurrence. Is it referring to PSCF values above a certain threshold? Please clarify.

1. Page 9, Line 2: “…pivotal role over TP…” → “pivotal role of the TP” is more accurate.

1. Section 3.3 Local emissions and regional transport of NO₂ over the TP

1. Page 9, Line 7: “…have exhibited exhibited distinctly different spatiotemporal evolution patterns…” → “…have exhibited distinctly different spatiotemporal evolution patterns…”.

1. Page 9, Lines 16-17: “…resulting in further spatial expansion of NO₂ hotspots(Choudhary et al., 2021) …” → “…resulting in further spatial expansion of NO₂ hotspots (Choudhary et al., 2021) …”.

1. Page 9, Line 22: “one of cleanest regions” → “one of the cleanest regions”

1. Page 14, Line 38: “…pristine background region, the TP is increasingly…” → “…pristine background region, is increasingly…”

Figure comments

1. Figs. 3–5: The figures and text are a bit small. Making both bigger would improve readability.

The authors analysed the transport of NO2 over the Tibetan Plateau using OMI and TROPOMI observations in a ML-model that they have published before. They found a seasonality and a spatial gradient in the NO2 columns over the TP with increasing NO2 levels. 

The topic can be interesting but I currently miss several aspects in this research: The chemical lifetime of NO2 was not studied, validation of the results is missing and the effect on air quality is not very convincing. Below I will give a more detailed account of my concerns.

When presenting transport of NO2 a discussion of its chemical lifetime is expected. The relative short lifetime of hours for NO2 distinguishes it from a passive tracer and also from CO. The transport of NO2 was modelled based only on concentrations and wind fields, while the chemical lifetime was not taken into account. If the temperature was take into account in the ML model, at least a simple link could have been made with the chemical lifetime, which is temperature dependent.

The authors show that the incoming flux and outgoing flux over the TP is more or less equal, which is puzzling. Does this mean that there is no relevant chemical loss, or is the loss coincidently equal to the local emissions over the TP, or do other chemical processes (conversion with PAN) play a role? Maybe the chemical lifetime is especially long for this region? These questions are not answered or even discussed.

It is difficult to judge the reliability of the results when no validation is presented. There are apparently 30 sites with in-situ observations that could be used for comparison at ground level of the model, but they are largely ignored. I am also surprised that the model has layers from 1000 hPa (in steps of 25 hPa) as shown in Figure 9, while the TP has typically ground pressure levels of less than 600 hPa, which put the Figure in a completely different perspective.

The conclusion that this study is important for air quality is difficult to understand. Figure 1 shows concentrations that are too low to be visible with the used colour scale. The authors claim that the ground observations are showing a decreasing trend but are unreliable. I would like to know why they are considered unreliable. The authors also conclude that the transport from polluted areas takes place on higher levels in the atmosphere. Would this mean that they do not affect the surface level? On top of this, the receptor area of the NO2 transport in China has already high concentration levels due to local emissions, so the effect would also be very limited for this region.

While focus of this study is the TP, this region is completely white in Figure 1 with no visible NO2 concentrations. There are also no other figures of observed concentrations either from satellite or the ground. This provides no reference for any conclusions about the relevance for air quality.