Summary

This study focuses on the aerosol, cloud, and boundary layer properties during the ARM EPCAPE field campaign by contextualizing the analysis in the different synoptic patterns,  classified using self-organizing maps.  Overall, this paper presents an interesting perspective by linking the in situ and remote sensing data of aerosol and clouds, and thermodynamics data collected during the field campaign with the background synoptic pattern, which is key in determining the cloud types and the aerosol transport. However, I believe further clarification is needed regarding the analysis. I recommend accepting this paper with the revisions suggested below.

Major comments

1. The writing and structure.

• The results section is quite long. Please consider adding a brief introduction at the start of Section 3 to show how each subsection is connected.
• The SOM classified 9 clusters, which are hard to keep track of their numbers throughout the manuscript. I suggest the authors may consider naming each cluster based on its pattern, so it would help the physical interpretation and make the paper easier to follow.

2. Equations for deriving Nd and LWP based on satellite data

• Please add references for the equations used for calculation, and evaluation results for how these retrievals work.
• These two equations use different underlying assumptions, with the formula for Nd assuming an adiabatic cloud with LWC increasing linearly with height, while the equation for LWP assumes a vertically homogeneous cloud with constant LWC. Given the focus on low clouds in this study, the LWP assumption might not be proper, as the observed LWC profiles typically increase with height. Please explain the assumptions here and how this would affect the results.
• Please explain how the constants for Nd are derived here, as it will depend on the ambient temperature.

Minor Comments

Line 20. It seems confusing to have “9 regimes”, but 4 points listed.

Line 30. Please be more specific about what “smaller-sized particles” mean.

Line 99-100. I suggest that the authors note the dominant cloud type in the abstract as well.

Line 137. To avoid confusion with the minus sign before the numbers for frequency, I suggest using parentheses here instead of hyphens.

Line 143. Please clarify how these rain flags are derived. Does this mean that the LWP results presented later are for non-precipitating clouds only? If so, this should be noted in the figure captions and main text.

Line 149-151. If LWP data have been filtered to exclude raining time periods, does the Nd derived here only apply to non-precipitating clouds?

Line 207. Are the times here in UTC or local time?

Line 287-290. Do the SOM maps based on data for all the months, or only for data in April-Oct of these 11 year period? If so, please clarify in the text and also in the caption of Figure 1.

Line 306. Please explain briefly what the silhouette score is and how one should interpret the values.

Line 411. Please double-check the temperatures given in the text, which are inconsistent with Figure 4a.

Line 502. Are the low clouds below 2km?

Line 586, Figure 9. MWR3 data has been filtered based on rain flags. Are there similar filters for the satellite data?

Line 596. If the aerosol data is a merged product based on both SMPS and APS, should the larger sizes also include coarse mode?  

Line 606. Please clarify why only these two levels are chosen for analysis, given that EPCAPE has CCN data measured at 6 supersaturation levels.

Line 624. Perhaps just revise the order of Figure 11 and Figure 12, given that Figure 12 is presented first.

Line 624-630. The results in Figure 12 are quite interesting, but some more explanation is needed. Line 626 states, “enhanced turbulence promotes aerosol growth into accumulation mode aerosols”. Is there a decrease in Aitken mode aerosol concentration with increasing TKE to suggest that the enhancement in the aerosols larger than 100nm is due to enhanced aerosol growth? Or is that just because stronger TKE is associated with enhanced aerosol production, such as sea spray, or aerosol transport from the coast?

This study investigates aerosol, cloud, and boundary-layer properties during the warm season (April–October 2023) of the ARM EPCAPE campaign by linking in situ observations, satellite retrievals, and ERA5 reanalysis with synoptic regimes identified using Self-Organizing Maps. Overall, the manuscript presents an interesting and valuable framework for examining how large-scale circulation and boundary-layer dynamics modulate cloud structure and aerosol properties along the coastal Northeast Pacific. However, several aspects of the methodology and interpretation require further clarification. I therefore recommend publication after these have been adequately addressed.

Major comments:

I am somewhat confused about the temporal setup of the SOM analysis. The SOM training was performed using daily ERA5 data from 2012–2022, whereas the subsequent analysis focuses specifically on the April–October 2023 EPCAPE period. Could the authors clarify whether the SOM training included all months or only April–October periods? If all months were included, please discuss the rationale for this choice and whether the inclusion of winter circulation regimes may influence the identified clusters relevant to the stratocumulus season.

In case the full year data has been used, I also suggest examining the monthly occurrence frequency of the identified regimes. Although the SOM input fields were normalized to reduce the influence of seasonality, presenting when the different regimes occur most frequently could provide additional climatological context and help better interpret the physical mechanisms associated with the observed cloud, aerosol, and boundary-layer characteristics.

The manuscript identifies a coherent clockwise evolution of the synoptic regimes resembling the progression of weather disturbances (4, 1, 2, 3, 6, 9, 8, and 7). Given this physically meaningful sequence, could the authors clarify why the clusters were not presented in this order throughout the figures and discussion? Presenting the regimes following their identified evolution may improve the readability and physical interpretation of the results. In addition, assigning physically descriptive names to the regimes, rather than repeatedly referring to cluster numbers, could make the manuscript substantially easier to follow.

The study domain includes significant orographic influences, while the SOM is trained using 975 hPa wind fields. Could this affect the resulting SOM classification?

Could the authors further justify the use of daily data for the SOM clustering analysis? Some of the processes discussed throughout the manuscript, such as the land–sea breeze circulation and boundary-layer coupling/decoupling, are inherently subdaily processes. It would therefore be helpful to discuss whether using higher temporal resolution data (e.g., 6-hourly fields) could potentially better capture physically distinct circulation regimes.

The manuscript applies several filtering criteria, including the removal of precipitating cases and profiles with LWP < 5 g m⁻². Could the authors clarify how much of the dataset was excluded by these filters and whether the amount of removed data differs across the identified regimes? In particular, it would be useful to know whether certain regimes are characterized by a substantially larger fraction of precipitating clouds, as this could potentially influence the reported inter-regime differences in cloud and aerosol properties.

In addition, the rationale for excluding precipitating cases could be further clarified. Since coalescence scavenging and wet deposition are important components of aerosol–cloud–precipitation interactions, removing precipitating periods may also remove physically meaningful aerosol and cloud processes relevant to the interpretation of the observed regime differences. A brief discussion of the implications and potential biases introduced by these filtering choices would strengthen the manuscript.

Minor comments:

Abstract: The manuscript refers to the identification of nine synoptic regimes; however, only four regime numbers are explicitly listed!

L354: I suggest considering the Estimated Inversion Strength (EIS) in addition to LTS, as EIS is often a more physically representative stability metric for marine stratocumulus environments (Wood & Bretherton, 2006) and may help better explain the observed differences across the regimes.

L287-290: The manuscript states that the SOM input fields were deseasonalized using a 30-day moving mean and standard deviation in order to reduce the influence of seasonal variability and emphasize synoptic anomalies. While this approach is reasonable, it would be helpful if the authors could further justify this preprocessing choice and discuss how sensitive the identified regimes are to the deseasonalization procedure.

Figure 4. It is unclear whether the differences between the profiles (e.g., Fig. 4) are statistically significant or simply visual. Please clarify whether any statistical tests were performed to support these differences.

Figure 11: You could strengthen the back-trajectory analysis by also considering the vertical structure of the trajectories. Vertical back trajectories could provide additional insight into the role of free-tropospheric transport, entrainment processes, and the vertical separation between marine and continental air masses across the different regimes. This may help better interpret the observed differences in cloud properties and aerosol characteristics.

In addition, I suggest combining the trajectory analysis with the aerosol composition measurements (mass concentrations of organics, sulfate, nitrate, etc.). Linking transport pathways with aerosol chemical composition could help better constrain the likely source regions and provide a more physically based discussion of the dominant aerosol sources influencing each regime.

L766: The statement that “Nd drivers can be more explicitly observed in Figure 12” should be revised, as the relationships shown in the figure represent correlations/associations between variables and do not necessarily demonstrate physical causation or “drivers.” A more cautious wording distinguishing association from causality would strengthen the interpretation.

Finally, the manuscript discusses relationships between several variables (e.g., BLH, TKE, CCN, Nd, etc.).  A correlation analysis across all data, independent of the weather regimes, would provide a clearer and more robust assessment of these relationships.