This manuscript presents a comprehensive, nationwide analysis of temperature inversions (TIs) over China from 2016 to 2021, leveraging high-resolution radiosonde data and collocated PM2.5 measurements. The study provides valuable insights into the spatiotemporal patterns of surface-based and elevated inversions and their distinct impacts on air pollution, with a particular focus on the differential mechanisms between northern and southern China. The topic is of high interest to the readership of Atmospheric Chemistry and Physics, and the observational analysis is generally robust. However, several aspects of the methodology and interpretation require clarification and strengthening before the manuscript can be considered for publication.

Major Comments

1. The authors classify Surface-Based Inversions (SBIs) as those with a base height (Hb) < 100 m and Elevated Inversions (EIs) as 100 m ≤ Hb ≤ 2000 m. While this follows some previous studies, the specific rationale for the 100 m threshold is not sufficiently justified. Given that this threshold directly influences the reported frequencies and subsequent correlation analyses with PM2.5， its appropriateness across all of China's diverse terrains (e.g., plateau stations, urban areas) should be demonstrated. The authors should provide a stronger justification, supported by literature or a sensitivity analysis, showing how the results might be affected by a different, physically-based threshold (e.g., related to the nocturnal boundary layer height).
2. The study effectively establishes a statistical association between TI parameters and PM concentrations. However, the assertion of a direct causal impact requires more support. The analysis does not fully disentangle the influence of emissions and synoptic-scale meteorology, which co-vary with inversion conditions. For instance, are the observed PM increases during SBIs primarily due to the inversion trapping locally emitted pollutants, or are the same large-scale stagnant conditions that cause the inversion also responsible for accumulating pollutants via regional transport? The authors should strengthen their causal interpretation by, for example, discussing the diurnal emission cycles or incorporating analysis of wind patterns and back-trajectories during specific inversion events to better attribute the pollution buildup.
3. The division of China into seven regions is a key aspect of the analysis, but the criteria for this specific partitioning are not clearly defined. The description in Section 2.1 mentions "integrated meteorological characteristics and major urban agglomerations," but this is vague. A more explicit justification is needed. For example, why is the Sichuan Basin grouped within the larger Southwest region rather than being treated separately, given its unique meteorology? The authors should provide a clear rationale or reference an established regional framework, perhaps including a map that overlays key topographic or climatic boundaries with the regional divisions.
4. The authors mention that temperature inversions primarily affect PM2.5 pollution in winter and are one of the causes of severe pollution events. However, this viewpoint is mainly based on nationwide averages and lacks a detailed analysis of specific case studies. To enhance the persuasiveness of the conclusions, it is recommended that the authors include case studies of several major city clusters, such as the Beijing-Tianjin-Hebei region, the Yangtze River Delta, and the Pearl River Delta. These city clusters face different climatic and geographic conditions, and analyzing specific cases would provide a more comprehensive understanding of the actual impact of temperature inversions on PM2.5 pollution, especially during severe pollution events.
5. One of the most intriguing findings is the negative correlation between EI strength and PM in southern China. The proposed explanation—a "transport-suppression mechanism" where EIs cap the boundary layer and isolate it from northerly pollutant inflow—is plausible but remains speculative. This hypothesis should be supported with more direct evidence. The authors could analyze wind direction and speed data during these EI events to show a reduction in northerly flow, or cite studies that have documented such a synoptic setup in southern China.

Specific Comments

1. Line 99-107 Does this section lack a description of Central China?

2. Line 211 I believe you are describing three figures rather than two. “(Fig S2, 4a)” changed to “(Fig S2, S3, and 4a)”.

3. Figure 2: The color scheme for the different regions in the bar plot is difficult to distinguish. Please use a more distinct and colorblind-friendly palette.

4. Figure 8: The presentation of the results in this figure is very dense. Consider breaking it down into two separate figures (e.g., one for 08:00 BJT and one for 20:00 BJT) for improved clarity.

5. Line 295-297: The location of 'Hetian' in southern Xinjiang, as discussed in the text, should be marked on Figure 8 to improve readability.

6. Line 334 Extra parentheses in citation “(Liu et al., 2022))” → Remove extra parenthesis

7. Line 354-356: The statement "a deep but weak inversion can be eroded by mechanical turbulence more readily than a shallow but intense one" is a key physical insight. This point could be emphasized earlier in the manuscript to better frame why strength is a more dominant factor than thickness.

8. Language and Abbreviations: The manuscript is generally well-written. A minor check for consistent use of abbreviations is recommended (e.g., "TI" is used throughout, but the full term is sometimes repeated unnecessarily).

This study provides a relatively comprehensive analysis of temperature inversion and its relationships with PM_2.5 concentrations nationwide based on radiosonde observations with very high vertical resolution. The results provide strong evidence that TIs could strongly affect the formation of air pollution. But the analysis is still shallow. More detailed information should be excavated from the very high-resolution observations. For the interpretation, although the authors showed some discussions, but some of them lack data support or evidences. The definition of TI strength and the discussions on the relationship between TI and PM_2.5 need further consideration. The writing of this manuscript also needs further refinement. Overall, in my opinion, with proper modification, this paper could be a good contribution to the journal Atmospheric Chemistry and Physics. Please see detailed comments below:

1. Is ~16 m significant?
2. Line 26, we don’t really say that PM₅ is a driver.
3. Line 39, specific weather systems could also trigger TIs.
4. Line 42, a reference is needed for the mentioned wind speed.
5. Line 43, warm advections do not necessarily be synoptic-scales.
6. Line 48, does this “ventilation” mean vertical dispersion?
7. Please provide the locations of all used air quality monitoring stations in Fig. 1. Please also provide terrain features to show if the 10 km threshold result in any large differences in terrain features.
8. Line 115, please provide the valid bottom height for radiosonde observations.
9. Maybe lapse rate is a better parameter for TI strength.
10. Line 162, it should be “a daily mean PM₅ concentration exceeding 75 ug m^-3can be identified as a pollution event”. However, the authors could use this threshold to identify pollution events, do not necessarily follow the official standards.
11. Line 171, if a TI event’s depth is more than 200 m, will it be counted twice?
12. Figure 2, it seems this 200 m threshold could not tell more details about the proportion of SBI. Please provide such information in line 172.
13. Line 180, “plausible”?
14. Figure 3, the consistent color bar conceals lots of useful information especially for SBI.
15. Line 191, grammar issue.
16. Line 192, please make the first factor clearer.
17. Line 201-202 & 208, shouldn’t these sentences be placed in the previous paragraph?
18. Line 202, when the term “peak” is used, please do not mention so many months. Please rephrase this sentence.
19. Line 211, “are more likely to occur”.
20. Line 214-219, the authors tried to explain the seasonal features of TI, but more detailed and in-depth analysis are needed (better with some evidences from meteorological data).
21. Line 232, this sentence is hard to understand.
22. Section 4.1.2, the TI strength defined in this paper is strongly related to the thickness of the TI. Thus, if a TI is very tick but with smaller temperature lapse rate, the strength could still be very strong. Please also see question 9.
23. It seems the SBI and EI have very close thickness. However, for EIs, especially for those occurred over higher altitudes, generally have thicker depth, even reach kilometers. For now, because we don’t have a full access to the data, we don’t know if such thick TI exists.
24. Please rephrase line 259.
25. The calculation of the thickness of TI when both SBI and EI exist was not clear.
26. There is an issue about the statistical analysis of the relationships between TIs and PM₅. For now, there are only two profiles of temperature per day for most of the stations, and the observing time are 08:00 and 20:00 BJT, which are usually accompanied by relatively low planetary boundary layer height. Such bias is more significant over western regions, since 08:00 BJT there means 06:00 or even 05:00 LST. Such low PBLH is usually the result of the lack of solar heating and is thought to be the major reason for trapping air pollutants during early morning and evening. Therefore, the identification of PBLH and the analysis of TIs and PM_2.5 under different PBLH is very important here.
27. Line 360, does the “intensity” here means strength? Please make these terms consistent.
28. Line 363-364, the path of incorporating real-time diagnostic information into nowadays operational models (mostly numerical models) is unclear. It is better to say, improve vertical resolutions and the forecasting skills of temperature inversions in numerical models could benefit the forecast of air pollution events.