The authors analyse data collected over a five year period at the nine sites of the German Ultrafine Aerosol Network (GUAN) in relation to new particle formation (NPF) events.  They compare the behaviour between different site types in terms of important variables such as nucleation frequency, nucleation rate, particle growth rates and implications for incremental ultrafine particles and CCN. They seek to place the results in the context of other European sites.   The research is carefully conducted with methods clearly described, and the results, coming from a high quality harmonised network, are of value to the research community.

Points which require further attention before publication include the following:

1. Line 213. Exclusion of months with data recovery <75% could bias the dataset more than leaving the data in or extrapolating to the full month, as if it excludes a partial month with high or low values, exclusion will cause bias.  There is also a question over the fact that site data collected over different periods (of years) are being compared without considering whether the shorter time periods are representative in relation to the longer datasets.  The issue of inter-annual variability is not explored in depth, and perhaps should be.
2. The explanations for lower nucleation frequencies at high altitude sites are not very convincing. Temperature is mentioned as a contributor to this effect, which is unlikely as lower temperatures serve to stabilise the clusters which are the precursors of new particles.  Given this fact and the likely lower condensation sink, the most likely explanation would appear to relate to gaseous precursor concentrations.
3. Line 286. The authors acknowledge that starting times are influenced by different lower cut points for the measurement sites.  This could have been harmonised by extrapolation of data back to 2nm diameter, which is arguably more realistic than using sizes closer to 10nm.
4. Line 177. The method of estimating the effect of NPF upon CCN concentrations appears to be flawed.  The method is cited as following two published papers, especially that of Ren et al. (2021).   The other reference, Kalkavouras et al. (2019) is incorrect in the reference list, but the correct paper is concerned primarily with kappa values, not this method of estimating CCN enhancements.  The clear implication of presenting the CCN enhancement is that the new particles are growing to CCN sizes, which at 4nm/h would take from 12 – 48 hours dependent upon the supersaturation. Taking a ratio of estimated CCN within and after the NPF event to CCN in the two hours before the event is reflecting the growth of particles which were already approaching CCN size as opposed to newly formed particles.  The authors may have been aware of this, but it is certainly not made clear to the reader.
5. The authors might also like to consider:

• The abstract states formation rates and growth rates but doesn’t state at what diameter.
• It would be useful to change J_nuc and GR_nuc in the text to J_10-25 and GR_10-25 as it is currently confusing.
• The statement in the introduction “Existing theories still cannot fully explain the fundamental chemical mechanisms of NPF events observed under diverse tropospheric environments and the result of field measurements are often controversial concerning the contribution of the chemical species to nucleation and growth of nanoparticles (Lee et al., 2019)” is not really true – I am not sure that the results of field campaigns are ever controversial, and our current understanding explains most of what we observe.
• There are differences between Leipzig-TROPOS and Leipzig-WEST despite them being very close to one another. NPF starts at the same time, but the frequencies, growth rates, and formation rates (as well as their seasonal trends) differ. Do the authors have any comments on this?
• It would be very useful to see the condensation sink plotted for each of these sites in the same way that GR and J are plotted.
• It would also be nice to see the size distributions plotted up. Perhaps a mean size distribution for each site on each event day/non-event day/undefined day.
• In the supplement, a mean contour plot on an NPF, undefined, and non-event day would be useful to see for each site too.

The MS reports, analyses and discusses data on the NPF occurrence frequency, particle formation rate, growth rate, starting time, contribution of NPF events to UF particle number concentration, to CCN concentration and to aerosol extinction coefficient obtained within the German Ultrafine Aerosol Network for 5 years. The topic of the MS is timely and of interest for the international scientific community. The experimental and evaluation methods were carefully deployed and realised. The obtained results are very valuable considering both the spatial and temporal scales. They are also put into an international context by comparing them with those from some other, mainly central European sites. Nevertheless, there are several limitations which should be corrected and some aspects in which the MS could be and should be extended.

Major concerns

1. The formation rate and growth rate were calculated for the diameter of 10 nm (Eqs. 1 and 2). Nevertheless, they are denoted as J_nuc and GR_nuc. This is misleading since the subscript nuc usually indicates the properties at the nucleation, i.e. at a diameter of ca. 1.5 nm. Since the dynamic properties can strongly change with particle diameter in this range, this practice is not acceptable. The authors should use J₁₀ and GR₁₀ instead of the present notation all over the MS. More importantly and a consequence of this, it should be considered and discussed in more detail that the comparisons of these data to the other European results were accomplished at different diameters, e.g. to 3 or 6 nm. It is also noted in this respect that 1) the mean diameter (L144) should be replaced by the modal diameter, and 2) it should be specified for which particle diameter Dp was the CoagS_Dp calculated.

2. The authors mention (L290) that the starting times (t₁) of the NPF event were determined at different particle sizes of 5, 10 or 20 nm (lower diameter limit of the measurement setups), which makes the discussion and comparison difficult. This particle diameter variability could be taken into consideration in the first approximation when the starting times are shifted to the critical nucleation diameter of ca. 2 nm using the GR values which are the closest to the actual particle diameter by a subtraction of e.g. t₁-(10-2)/GR₁₀ (for 10 nm). This modified starting time t_nuc could be compared more advantageously despite our knowledge on the GR in the diameter range from 2 to 3 or 5 nm is still inconclusive.

3. Local time was used in the work as the time base (L136, L288). This selection (as all the other options as well) has both advantages and limitations. It should be clarified weather the starting times in Fig. 8 were adjusted to clock change (expressed in UTC+1) or not. In relation to this, the reader may also wonder what the reasons were for selecting the 14-day smoothing in Fig. 8 and not something else.

4. The authors are requested to explain why they can show occurrence frequency data for the winter months (e.g. Fig. 3), while the column bars of _nuc and GR_nuc for winter are mostly missing in Figs. 6 and 7. Cf. lines 212-213.

5. The statement in L298-L299 is only partially acceptable. The occurrence and timing of the NPF events depend more sensitively on the ratio of the sources and sinks than on the sources (precursor emissions) alone. The authors may want to add new aspects on the effects of the higher CS in cities.

Minor comments

The Conclusion section is more a summary. It should be substantially shortened and the emphasis is to be put on the consequences of the actual results.

The authors may want to revisit their rounding off strategy at many places in the text and tables (e.g. 2.89 cm^–3 s^–1 in L25, Table 2), since the anticipated precision of these values seem exaggerated.

L31-L32. It is unusual to state that the particles are formed from precursors. Those chemical compounds (usually with longer atmospheric residence times) which yield the active players (usually with shorter residence times) in reactions are ordinarily called precursors. Thus, SO₂ is a precursor compound, whereas its gas-phase oxidation product of H₂SO₄ is the vapour that plays an active role in the nucleation process. Clarification is needed.

L32-L33. The sentence is misleading. Condensation does not increase the particle number concentrations, while coagulation decreases them. These processes do not lead to growing particle number concentrations. Reformulation is required.

L90 and L125. Check the citing format requirements of the journal.

L114. Replace “Aerosol PNSD was measured” by “Aerosol PNSDs were measured”, and similarly: “PNSD were generally measured” by “PNSDs were generally measured” (L118).

L158. The NSF was actually introduced and improve in Salma at al., 2017 and not in the reference cited.

The authors may want to write in the section title of 3.1 NPF “occurrence” frequency, and of 3.2 Growth and formation rate”s”.

L243-L245, twice. The order of the words and grammar should be checked: “the frequency of NPF event and undefined event are 14.5 % and 5.4 % for JFJ, respectively”. Consider: the frequencies of the NPF events and undefined events were 14.5 % and 5.4 %, respectively for JFJ.

L261. The correct formulation is (Kulmala et al., 2022a, 2023).

L342-L345. Remove the considerable repetition.

Caption of Fig. 8. Correct: “…the red dash and black dash line indicate…”.