The interesting article presents a 23-year time series for recording Saharan dust at the Jungfraujoch GAW and ACTRIS station and compares different methods and instrumentation in terms of their characteristics for identifying Saharan dust. Possible conclusions on the climatology of Saharan dust events are presented from a cautious perspective.

The comparisons of methods and instruments are very valuable and underscore the importance of in situ dust detection and characterization.

Overall Comments:

* It is not always clear in the text whether this refers to the complete data set or to data with applied noise thresholds (eg. L 225 ff, L 585 f)

* The article focuses very much on Jungfraujoch data and the role of Mt. Helmos is not clear for the reader

* With regard to the noise threshold values, an evaluation should be carried out to rule out the possibility that this could introduce a certain influence or bias. Since the upper wavelength range (red or IR) will be the decisive criterion for the threshold value in both the nephelometer and the aethalometer (lowest scattering and absorption), a change in the AE also influences whether or not the noise threshold value is exceeded and hence, this might introduce a slight artifact.

Detailed suggestions:

L 109: as far as I know, AE31 measures at different wavelengths: 370, 450, 520, 590, 660, 880 und 950 nm

Diagram 1: The differences in detail are very difficult to identify. For example, AE could be restricted to the range from -2 to 5.

L 266f : The line implies that the hours of Saharan dust at Sonnblick are overestimated. There is no evidence to support this assumption.

L 282 ff: which thresholds were used for Ecotech Nephelometer

L 416: space after comma between February and May

L 445 ff: As mentioned in the article, the Flexpart method also has its issues, and it could also be events that were detected using coarse-mode particle concentration but do not contain Saharan dust at all.

L 571: 24 years or 23 years? I understand the difference between 24 calender years and 23 years time series but  you  Ok, not important

The submitted paper investigates Saharan Dust Events (SDEs) observed at the high-altitude monitoring station of Jungfraujoch in the Swiss Alps. The analysis combines various in-situ measurement techniques with CAMS reanalysis products and back-trajectory simulations to examine both the inter-annual and intra-annual variability of the frequency and intensity (in terms of mass concentration) of SDEs since 2001. Overall, the authors present a high-quality study that appropriately addresses all key aspects expected in a well-structured scientific paper. Therefore, I recommend the manuscript for publication after the authors have addressed the minor comments listed below.

1. Lines 47 – 59: I would suggest that the authors include a short discussion (2-3 sentences) acknowledging that the results may vary depending on the variable of interest, due to differences in their representativeness. This is attributed to the different representativeness of the selected variable. For example, aerosol optical depth (AOD) describes the total particle load throughout the atmospheric column, whereas ground-based in-situ measurements mainly reflect conditions within the planetary boundary layer. Consequently, elevated dust layers residing in the free troposphere may not be adequately captured by ground-based PM₁₀ measurements.
2. Line 65: Are you referring to the application of the Ångström formula to the spectral single scattering albedos (SSAs)?
3. Line 142 – 143: What do you mean by “The absorption coefficients were evaluated ….”? Are you referring to the number of SDEs mentioned in the following sentence?
4. Line 152: In order to classify a case as an SDE, is it required that all hourly SSA values within each of the defined time windows (e.g., 4 h, 6 h, etc.) be negative? Have you considered defining discrete time windows (e.g., 4–6 h, 6–12 h, 12–24 h, 24–48 h, and ≥48 h) to avoid overlapping?
5. Sections 2.2.2 and 2.2.3: It is not clear how the low-pass filter is applied. Is the first low-pass filter applied to the raw time series, the second to the already smoothed time series, and so on? How is defined the difference after the iterations?
6. Section 2.2.3: The CAMS ensemble is derived by nine or eleven models? (https://ads.atmosphere.copernicus.eu/datasets/cams-europe-air-quality-reanalyses?tab=overview). Which variable from the CAMS outputs is processed?
7. Section 2.2.4: I suggest clarifying more clearly why it is important for your study to utilize both FLEXPART and LAGRANTO-COSMO simulations. Why was FLEXPART run for such a long period (30 days)?
8. Line 195: Where is the KZ low-pass filter applied?
9. Lines 317-320: The agreement between the CAMS-based and coarse-mode-based SDEs is not very evident. Could the authors comment on the significantly higher number of SDE hours derived from the FLEXPART sensitivities compared to the other methods?
10. Figure 4: I strongly recommend condensing the discussion section and emphasizing the main findings of the analysis. The current discussion is rather lengthy, which makes it difficult for the reader to clearly identify the key outcomes of this analysis.
11. Lines 444-448: I believe this part of the text requires further clarification. It is not clear how air masses traveling over regions where fine aerosols are more likely to be present can result in higher coarse-mode concentrations at JFJ.
12. Section 3.1: I would suggest shortening this part of the manuscript to improve readability.
13. The measurements acquired at Mt. Helmos should be given greater emphasis in the analysis. I would expect the authors to include an intercomparison of SDE characteristics (e.g., frequency of occurrence, intensity) between the two mountainous stations. Such an analysis would provide valuable insight, considering that JFJ and Mt. Helmos are influenced by dust outbreaks originating from different source regions and driven by different atmospheric circulation patterns.