General comments:

This study deals with the complex dependence of CCN activity on the hygroscopicity of organic aerosols. Since CCN activity determines the indirect effect of aerosols on radiative forcing, the subject is of great interest. Using chemical and size distribution measurements, the authors study the role of OA hygroscopicity on CCN activity, for different atmospheric conditions. The positive matrix factorisation method was used to recover the relative contribution of OA with different oxidation levels, showing that medium and low oxidised OA are predominant, and that their contribution varies as a function of the vertical transport of PBL to the site. The originality of this study lies in the use of a neural network model to predict the amount of CCN using aerosol size distribution data, the fraction of OA and a factor of PMF, and radiation. This innovative tool gave the best results compared with assumptions about global chemical composition. The authors conclude by stressing the importance of taking into account the complexity of the aerosol and in particular its internal/external mixing. The manuscript is well structured and well written. The conclusion is clear and the message to be retained is correctly underlined. I consider that the manuscript can be published, after minor revision and responses to the following points.

Specific comments:

Please add the equation which links κ, D_crit and SS at the CCN activation in section 3.2 (κ -Kohler theory). You could then cite it after (eg. L 435).

L 281. Could you explain why the eBC increase starts earlier than the inorganics and OA increase during the day ?

L 324. What are the wind direction and speed like during the two identified periods ?

L 365. What about the solar radiation diurnal profile role in the photochemical oxidation ?

L 413. What do you mean by “Time-dependent” ? If I’ve well understood, you described previously that in Scheme 3 specific κ values for LO-OOA and MO-OOA have been used to take into account their relative contribution at SNS, but these values are fixed, and not varying as a function of the time. Could you please clarify this ?

L 443. Could you add the D_crit values for SS= 0.2 and 0.6 % to compare it with D_crit = 72 nm ?

L 451 to 461. Did you observe a difference in the agreement between N_CCN observed and N_CCN calculated when considering the two time period separately (before and after June 26^th )?

Technical corrections:

Please check the units notation in the text and the figures (eg. µg m^-3 instead of µg/m³)

Figure 4 : please add a note on the shaded areas representing the PDF of each variable

L 489. “is associated with”

Figure 6: Please add a) and b) on the different panels.

L 513.: please give the values used for κ_IA and κ_BC

Figure 8a. : Could you sort data by f44 values such that the points with the lowest f44 are plotted in first and the points with the highest f44 and plotted over (ie. yellow points overlapping the blue points) ?

L 588. Slopes and correlation coefficients don’t correspond to the ones in Fig. 9a.

L 603. R² = 0.88 or 0.94 (cf. Fig. 9a) ?

Please find my comments in the attached file.