Dear Colleagues,

Attached are comments and notes on your study for your consideration.

More importantly, however, the manuscript and study needs significant improvement before it can be published. Currently, beyond the fact that these reduced mechanisms produce different results, I do not find sufficient analysis presented that explains the differences between the response of ozone within these mechanisms. As you show that the box model replicates many of the features of the 3-D simulations, I am puzzled by the lack of use of this tool. In many places, you offer hypotheses for the different behavior (e.g. PAN chemistry is missing), but tests of these hypotheses (which seem easy to perform in the box model) are missing. As a result there is little here provide guidance for how these mechanisms should be improved.

Finally, I am left wondering if (as was shown by Fabien Paulot in his investigation of the response of tropical ozone to changes in the description of the alkylnitrate chemistry) most of the differences in ozone and tendency in the present study simply reflect differences in the NOx abundance. Again, this seems relatively easy to test -- the hypothesis is that the ozone product rate in the different simulations simply reflects different lifetime or concentration of NO. If this hypothesis is correct, it would provide a possible road map for how to proceed. One could determine the dependence of the NOx lifetime on the forcings of interest (e.g. how much isoprene is in the air or what the temperature is) using the best and most expensive mechanism and simply use these parameterizations in the climate simulations.

Dear Alex and colleagues,

Thank you very much for the care you have taken in the revised manuscript. I learned a lot reading the new version. The use of the Browne and Cohen formulation together with the analysis using isopleths has substantially helped me understand the different responses.

Paul