This study investigates the high offshore ozone events in the Houston area by combining modelling study with data from the TRACER-AQ) field campaign. This is a very interesting study and the paper is in general well written. I only have a few relatively minor comments for the authors to address -

• Indeed, wind conditions can be very important meteorological factors affecting the high ozone events, but I wonder if it is possible to examine other meteorological conditions? For example, some studies have shown that certain meteorological conditions such as fumigation can lead to the “touch-down” of air pollutants from above and thus significantly enhance their surface concentrations. This can be quickly examined by looking at the atmospheric stability or vertical profile of temperature. Of course, I understand these temperature data may not always be available, but worthy to check.

• In abstract - “The region-wide increase of long-lived VOCs through advection not only leads to more O3 production under a NOx-limited regime but also …” – this sentence is confusing (if it’ indeed in NOx-limited regime, then the impacts of VOC change should be quite limited?), so I suggest rewriting this part.

The manuscript has presented a modelling analysis of the high ozone events over the offshore areas of Galveston Bay and the Gulf of Mexico observed by the TRACER-AQ field campaign. Six episodes with MDA8 ozone levels above 70 ppbv were analyzed using the surface site, ship, and balloon measurements interpreted by the WRF-CAMx modelling system and its process analysis tools. The results demonstrated the important role of chemical ozone production over the offshore areas from precursors emitted from adjacent land, and then high ozone could be transported back to land causing ozone exceedances.

The findings of this study are important for understanding the ozone variations in Houston and surrounding regions. The manuscript is overall well-organized and presented. The topic well fits the scope of ACP. I thus recommend acceptance for publication after the following minor comments have been addressed.

Specific comments:

1) Section 2.2, Emissions in the model

A more detailed description of the emissions may be required in this section. Besides anthropogenic emissions, does the CAMx model include any natural emissions, such as biomass burning, soil, and lightning? If not, how would these emissions affect the model results? Any ship emissions were included in this region?

2) Section 2.2, Line 113

We usually thought that simulations with nudging or with reinitializing would be better. Can you comment more here? Were there significant differences among the simulations, and with no nudging or reinitializing the results were statistically significant better?

3) Section 2.2, Line 141

It is not clear from Figure S1 that the redistributed emissions performed better. Can you please provide any comparison metrics to show this feature?

4) Section 3.1, Model evaluation with ozonesonde

The large model underestimates relative to ozonesonde measurements for the episode days may need some more discussion. The model underestimates appear to be persistent throughout the low troposphere up to 5 km as shown in Figure 4. It is not clear whether the model missed any other processes in addition to the wind factors. Please provide some further clarification.

5) Section 3.2, Page 10, Line 238

It is unclear why you examined the reactions of HO2 self-reaction and OH reaction with NO2. How were their rates linked with the production of ozone? A more detailed explanation in terms of chemistry is needed here.

6) Caption of Figure 8, what is CDT?