This paper presents results for surface ozone trends in the United States over the period from 1990-2023 and investigates the relationship between heatwaves and ozone over time.  Surface ozone trends and their relationship to heatwaves are an important topic in air quality.  This study incorporates novel approaches such as changepoint detection and a variety of statistical metrics to provide new insights on a long-standing topic.

General Comments

1. Either within the results or in the conclusions, it would be useful to expand on how the results of this study fit within the context of the large existing literature on US surface ozone trends. Please highlight what is new versus confirmation of previous findings.
2. On page 7, it is noted that all changepoint candidates are considered. However, we might have an a priori expectation that the changepoint will occur in a particular year (or range of years) given known changes in emissions.  What is the justification for considering all possible years as candidates?
3. Since the ozone changes are often related to emission changes in the text, it would be helpful to include a figure showing regional emissions timeseries. Also, could you apply the changepoint detection methodology to the emissions, and if so, would it yield similar results to ozone concentration changepoints?
4. Organization: Section 4.1 doesn’t seem specific to heatwaves
5. Some statements in the paper need clarification, as noted in the specific comments below.

Specific Comments

1. Abstract lines 9-10: please clarify: does this mean decrease in exceedances, or in exceedances during heat waves, or something else?
2. On page 4, 6 different temperature threshold definitions are listed. While there is value in considering more than one definition, the text might be clearer if this list were shortened or any sensitivity of the results to the choice of definition highlighted.  For example, do we expect TX95pct and TX90pct to yield different results?
3. Page 8, lines 4-5: define “average marginal effect”
4. Page 10 line 29: what does it mean that positive trends are decreasing? Do you mean they switch from positive to negative, or just become less positive?
5. Page 11 bottom: The current peak being in April does not, by itself , prove the cycle is moving toward Feb-March, and it isn't apparent from comparing the different epochs in Fig 4 that it's heading that way
6. Page 14 line 8-9: What about the negative arrows in Fig. 7?
7. Page 15 line 28: is there a physical mechanism, like a weaker response of O3 to T when NOx is lower?
8. Page 16 line 17-18: I suggest rewording this sentence to make it clear you are contrasting the 41% at the 10th percentile with the lower percentages at the higher percentiles
9. Page 16 line 21: Does “50^th/90^th percentiles” mean 50^th and 90^th percentiles?
10. Page 16 line 25-27: see general comment 3. This is a place where an emissions timeseries plot or changepoint analysis would be helpful to refer to.
11. Page 17 lines 4-5: It’s stated that heatwaves likely counteract the effectiveness of emission controls, but aren't the emission controls still effective at reducing the negative impact of heat waves?
12. Appendix A first sentence: Either split this sentence in two or add a conjunction/connection between “our focus has been on the seasonal extreme percentile” and “this section aims…”

This manuscript analyzes changes in ozone trends over the past almost 30 years in the contiguous United States. The authors focus on changepoint detection across various percentiles, and they investigate the impact of potential climate penalties that may offset ozone improvements. They find decreases of extreme ozone values since the 2000s during the warm months, while wintertime trends tend to be increasing. Heat waves increase ozone exceedance probabilities, and some sites in California may have been impacted by the ozone climate penalty during this timeframe.

The manuscript and analyses are strong, and this is a good addition to the ozone literature which introduces a statistical technique rarely used in atmospheric chemistry to the analysis of sparse ozone data. There are some minor clarifications and discussions that are needed, discussed below.

Page 3, Line 23: “summertime period in which the MDA8 value exceeds the thresholds of 70, 60, 50, and 35 ppbv…” Could you please explain your rationale behind using these ozone concentration cutoffs?

Page 4, Line 21: “An episode of heatwave event is detected if at least 3 consecutive days of temperature exceedances are found.” What is meant by a temperature exceedance here? Is this peak temperatures, average temperatures, etc.?

Page 5, Line 25 and Page 6, Line 2: “However, the possibility of incorporating two changepoints will be evaluated, if the junctures occur separately at around 2000 and 2010.” And “Based on the above discussions, for each monitoring site and season, we fit the trend model to all possible changepoint candidates (between 2000-2013)…” What is the rationale for choosing these years? Is it not possible that changepoints beyond these years could be present?

Page 10, Line 6: “This pattern can be attributed to strong ozone enhancements across the Northeast/Midwest in 2012.” Does the flattening of the trend in the most recent decade (2013-2023) remain flat if that anomalously high ozone values in 2012 are removed? I am concerned that this one spike early on in the recent time series is drowning out any other trend that may be occurring.

NOx Trends

A flattening of NOx trends after 2010 is mentioned in a few places (Page 10, Line 7-9; Page 11, Line 11-12). It is important to note that the Jiang papers mentioned in the manuscript are based on satellite data, which is sensitive to free tropospheric NOx and may not be fully representative of surface trends. A couple papers come to mind that suggest that NOx trends have not flattened out over urban areas (Christiansen et al., 2024; Silvern et al., 2019). The slowdown in NOx trends after 2010 is most noted over rural areas, where surface NOx is lower and thus background sources of NOx (soil NOx, free tropospheric NOx, lightning NOx) that remain relatively constant over time dominate the signal. Over urban areas, other evidence suggests that NOx emissions have continued to decrease. Since most AQS sites are in urban areas, I am not sure that the ozone trends after 2010 can be adequately explained by a deceleration in NOx decreases.

References

Christiansen, A., Mickley, L. J., and Hu, L.: Constraining long-term NO_x emissions over the United States and Europe using nitrate wet deposition monitoring networks, Atmos. Chem. Phys., 24, 4569–4589, https://doi.org/10.5194/acp-24-4569-2024, 2024.

Silvern, R. F., Jacob, D. J., Mickley, L. J., Sulprizio, M. P., Travis, K. R., Marais, E. A., Cohen, R. C., Laughner, J. L., Choi, S., Joiner, J., and Lamsal, L. N.: Using satellite observations of tropospheric NO₂ columns to infer long-term trends in US NO_x emissions: the importance of accounting for the free tropospheric NO₂ background, Atmos. Chem. Phys., 19, 8863–8878, https://doi.org/10.5194/acp-19-8863-2019, 2019.