This is a well written paper with reasonable conclusions. In particular, the SOM analysis seems to do a good job of selecting more strongly convectively influenced versus subsidence profiles. And the discussion around the importance of diagnosing the seasonal variation in the ozone trend is very useful to say. My main comment is that the paper seems longer than necessary due to using so many convective proxies, rather than simply just using rain itself (more in comment below). The technical standard of the paper was quite high, and I don't have any significant comments to make on these aspects of the paper.

The SHADOZ sondes also contain RH, and it would have been interesting to do a trend analysis of that quantity also, since it is also linked to convection. Could be issues with the data of course, especially in the upper troposphere.

The authors might have considered rain itself as a more direct proxy for convection than water column, cloud brightness temperature, OLR, velocity potential, etc. There may have been a bit of "over analysis" here in using all these proxies instead of going straight to rain. I think the paper would be more useful if a direct connection between ozone and rain could be established, since rainfall is more directly relevant. Presumably, some version of the TRMM dataset (3 hourly) would be appropriate. I realize rain is a more "noisy" variable, but it is also more interesting.

It might be borderline appropriate to call the ozone trend using 22 years of data a trend instead of some form of longer term variance. In this context, it also might be worth mentioning, or even showing, the absolute ozone trends (with MJO, ENSO variability included), since there could be changes in the fraction of the total rain/ozone variance that are part of these oscillations, so that there may not be an absolute trend in these quantities.

The positive ozone trends near the surface in the top panel of Fig. 6 from June onward look like they are confined to the boundary layer. Comparison with the stability could confirm. This would make more physical sense than saying "below 700 hPa". Could simply then say that, for whatever reason, BL increases in ozone are not being communicated to the free troposphere (maybe since BL increases are a local response to emissions increases).

This is a well-motivated and well-focused manuscript that investigates trends in tropospheric ozone from ozonesonde observations  at two stations in equatorial Southeast Asia and their association with concomitant changes in observed convection (as deduced from multiple remotely sensed cloud metrics and environmental proxies). This was one of the most polished papers I've received in review and I enjoyed reading it from start to finish. I applaud the authors for their crafting of an excellent narrative with appropriate concision, clear goals, and unexaggerated interpretations. The figure quality was also superb and each well justified and discussed. Congratulations on a wonderful effort that I believe is ready to be published as is, with only a single technical correction - "trends" on line 89 should be "trend".