We thank the reviewer for their helpful and constructive comments. Our replies are included below in italics with the reviewer comments, marked by Reply.

This manuscript presents a huge dataset obtained during an observationnal atmospheric campaign in March 2021 in Norway including a lot of complementary measurements (aerosols, water isotopes) at different locations using facilities at different sites. The description of the acquisition and calibration methods is done carefully (but could be sometime a bit reduced to avoid repetitions). While no interpretation of the data is presented (as expected for a ESSD paper), the complementarity of the measurements is presented which opens the way to future studies.

This manuscript is in the scope of ESSD, the data are usefull for further studies of the cloud microphysics and the data quality is well assessed so that future users can do the best of this dataset. The manuscript can thus be accepted with only minor revisions as detailed below.

• In the abstract and several times during the manuscript, the spatial representativeness of vapour isotopes measurements is mentionned. This is an important added value of this study with several instruments running in parallel (despite the leak on one of the inlet lines). It would be nice to have also a statement in the « discussion – conclusion » on the spatial representativeness to have an idea of the scale of the processes which are studied here.

Reply: We will add a statement to the discussion/conclusion section to specify the scale of spatial representativeness addressed within our study.

• In the introduction, the potential complementarity between INP and water isotopes is strongly underlined for process studies. However, later in the text, the INP data are not much described and it also lacks from the last section. Also the fact that the INP data are not made publicly available (but upon request) rise question on the usefullness of this dataset. I encourage the authors to detail more in a revised version the potential use of the INP data and its link with water isotopes data, not only in the introduction.

Reply: There are other studies underway making joint use of the INP and isotope data from the ISLAS2021 campaign. We will more clearly highlight the combined use of INP and isotope data in the revised manuscript.

- l. 52,you already write about the d-excess while it is only defined 5 lines below.

Reply: We will remove the mentioning of d-excess from this sentence.

- l. 101-105  and after. You describe very well the sampling strategy but the main scientific question addressed by this campaign is missing. It would be nice to explicit it.

Reply: L. 103 states the main aim, but we agree that it will be an advantage to also clearly state the main scientific question. We will revise the text accordingly.

- Table 3. I was wondering why the data acquisition could not be run continuously.

Reply: The aerosol lidar system at ALOMAR requires one to open the main hatch of the building, and thus a sufficiently long precipitation-free period to operate the laser system without risking that water enters the building. This will be mentioned in the revised manuscript.

- Paragraph starting from l. 259 : it would help to have some ideas here on the sampling frequency and invoke the possibility of post-deposition effects (which are adressed later but it would be nice to introduce them here)

Reply: Good point. In the revised manuscript, we will add a sentence that specifies the typical duration until sampling, mentioning the potential for post-deposition effects to modify the precipitation signature.

• l. 342-343 : I was wondering why the CRDS analysers could not operate continuously

Reply: As is shown in Fig. 5, all 4 analyzers were operating for most of the campaign period, with the exception of the Tromsø installation, which was sampling room air until the installation was fixed. In the revised manuscript, we will update the text to clarify that the analyzers were running ~80% of the time, and specify more cleary that the installation problem limited the simultaneous available data.

- Caption of Figure 5. Can you explicit how you evaluate the data quality for the different instruments ?

Reply: We will amend the caption of Fig. 5 accordingly.

- l. 375 : Why did you change the standard on 21 March 2021 ?

Reply: The reservoir bag for that water standard was running low and was therefore exchanged.

- l. 385-387 : Can you explicit why the analytical uncertainty for d18O is 1.25 permil on l. 386 and on l. 387, you mention that the total propagated uncertainty is 0.14 permil for d18O. In general, for this entire section, I had difficulty to understand how the number for uncertainties are evaluated. Since there are a lot of repetitions in this section (uncertainty calculation for each analysers), I would suggest to rewrite the section giving more details on the uncertainty calculation (or present some graphs) and then avoid repeating for all analysers except when there is a specificity as for the anlyser with a leak in the inlet line. A table to summarize the uncertainties for each analyser may also help reducing repetitions.

Reply: Unfortunately, the wrong uncertainties were given for δD and δ18O in L. 386, this will be corrected in the revision. We will consider to write a general section about the uncertainty calculation and propagation, and collect those in a table to facilitate reading of this section.

- l. 453 – 454 : I do not understand the sentence – especially, I do not understand the use of the first sample. Probably rewriting a bit this sentence would help clarifying.

Reply: We will remote this sentence from the revised manuscript, and shorten the text that follows, as the reference in the previous sentence is sufficient to document the sample handling.

- Figure 9 : It would help to have a label for the x-axis.

Reply: A date label will be added to the x-axis.

- Caption of Figure 11 : not sure that it is clear for the reader that d_ALOMAR is the d-excess since it was not defined as such in the manuscript before.

Reply: In the revised manuscript, we will introduce the symbols in the caption in the text explaining Figure 11.

- l. 616-617 : can you better describe the « another step at about 12 :30 UTC »

Reply: This refers to the step-like temperature change at about 12:30 UTC for the black line in panel b. This sentence will be rephrased for clarity.

- l.645 – 651 : it would be interesting to further detail the processes at play when comparing dD in precipitation and dD « expected from equilibrium fractionaton ». I find this section lacks from explanations on what has been calculated and how we can interpret processes such as exchanges between snow and water vapor or vertical advection. No reference is given also to support this interpretation. Please complete this section.

Reply: We will revise this section using the concept of equilibrium vapour as applied for example by Graf et al., 2019.

• Figure 13 : I think that there are some problems of coherency between the figure and the caption (caption describes panels from a to d while there are panels from a to e on the figure).

Reply: We will correct the panel labels for this figure in the revised manuscript.

References

Graf, P., Wernli, H., Pfahl, S., and Sodemann, H.: A new interpretative framework for below-cloud effects on stable water isotopes in vapour and rain, Atmospheric Chemistry and Physics, 19, 747–765, https://doi.org/10.5194/acp-19-747-2019, 2019.

We thank the reviewer for their helpful and constructive comments. Our replies are included below in italics with the reviewer comments, marked by Reply.

This manuscript documents an extensive observational effort during the ISLAS2021 campaign in coastal Northern Norway, combining water vapour and precipitation stable isotope measurements with aerosol, INP, and meteorological observations. The paper is generally well-structured, and the dataset is of clear value to the community. However, I have several substantive concerns and recommendations that should be addressed.

1) The Introduction are presented largely in parallel rather than converging towards clearly articulated goals. It remains unclear what the central scientific questions motivating ISLAS2021 are. The authors should explicitly state primary scientific questions, and clarify how the observational strategy was designed to address them. This would greatly strengthen the coherence of the manuscript.

Reply: A similar remark has been made by reviewer #1. The scientific questions are are currently entangled with the main objectives. We will rewrite the introduction following the suggestion of the reviewer.

2) A key strength of ISLAS2021 is the simultaneous measurement of water isotopes, aerosols/INPs, and meteorology. However, beyond the Introduction, the manuscript largely treats these components separately, with limited synthesis explaining how they jointly constrain atmospheric processes. The authors should include a short synthesis section or paragraph explicitly describing: how the different measurement components complement each other; which combinations of variables are expected to be most informative; and which limitations currently prevent a full joint analysis.

Reply: The Discussion is currently combined with the Conclusions section (Sec. 7). In the revised manuscript, we will separate the Discussion and Conclusions sections, allowing to bring forward the synergy between different types of observations made during the ISLAS2021 field campaign as suggested by the reviewer.

3) Spatial representativeness of vapour isotope measurements is repeatedly mentioned as a key motivation for the distributed network. However, the manuscript does not clearly summarize what spatial and temporal scales can realistically be probed with the available data. The authors should provide a qualitative (or semi-quantitative) assessment of the representativeness and scale of the observations, explicitly linking this to the network geometry and data availability.

Reply: We think that a detailed assessment of the spatial representativeness will be beyond the scope of this data description paper. Such an assessment would involve, for example, Lagrangian transport model calculations to assess when airmasses from one station reach the next one. We do currently briefly discuss the vertical and horizontal representativeness of the water vapour isotope measurements in Sec. 5.5 and Sec. 7. In the revised manuscript, we will follow the reviewer's suggestion and add a separate, semi-qualitative discussion on the aspect of representativeness.

4) The manuscript applies four different calibration strategies for the four CRDS analyzers, with large differences in uncertainty, particularly for the Coast site. While the authors describe these strategies in technical detail, the implications for data consistency and inter-site comparability are not adequately addressed. The authors must provide a concise, synthetic comparison of all analyzers, ideally in a single table, including calibration method and uncertainty ranges. The manuscript must explicitly state whether inter-site vapour isotope gradients can be meaningfully interpreted given the differing uncertainty levels. If no cross-site harmonization or bias assessment was attempted, this must be clearly acknowledged and justified.

Reply: All analyzers are comparable across sites, and the bias correction at site Coast is given in the manuscript. There is a typo in the δ¹⁸O and δD uncertainty for site ALOMAR which may have prompted the reviewer's comment (see reply to reviewer #1).  We acknowledge however, that the writing of Section 4.1 could be more clear and concise, as also pointed out by reviewer #1. We will follow the suggestion of both reviewer and use a table to present the uncertainty of all analyzers in the revised manuscript.

5) Although uncertainty estimates are provided for individual instruments, the manuscript lacks a high-level synthesis that translates these uncertainties into practical implications for data use. The authors should include a synthesis section that explicitly addresses: Whether observed isotope variability typically exceeds measurement uncertainty; What temporal and spatial scales can be robustly investigated using this dataset; Which analyses are likely to be uncertainty-limited.

Reply: This comment addresses a similar point as in comment #3. We will include discussion of this aspect in the Discussion section, which will be added in the revised manuscript.

6) A significant strength of the manuscript is the detailed calibration discussion (Section 4.1), but it also reveals substantial differences in uncertainty across sites. Please summarize the spatial consistency and reliability of the data more explicitly in the main text (e.g., Section 5), and guide users on how to interpret or filter the lower-quality measurements. Are there any recommendations for users on which periods or sites are most reliable for process studies?

Reply: Information about which periods are suitable for different analyses is partly contained in Sec. 3.1 and the corresponding Fig. 3. While it is difficult to give general recommendations, we will consider adding example recommendations for use cases to the Conclusions section.

7) The case study is a key element to demonstrate dataset utility but currently lacks depth. Please expand the analysis to more fully demonstrate how the SWI and INP data co-evolve across multiple sites (e.g., Bergen–Tromsø–Andenes). Consider including a figure showing temporal evolution of δ18O, δD, and d-excess during a representative IOP, ideally aligned with meteorological context (e.g., pressure, IWV, radar reflectivity). Is there any evidence of isotope distillation or below-cloud processes that you can highlight?

Reply: We chose to limit the analysis of the case study period for this data paper to not overload this data description paper with information. Some of the information requested in the comment is already given in Fig. 12 and 13. To give more concrete indication of below-cloud effects, we will redraw Fig. 13 using the equilibrium vapour concept (Graf et al., 2019), rather than a constant offset value.

8) The dataset includes a valuable INP record using the DRINCO method, but the presentation is minimal. Please expand on the sampling frequency, sensitivity, detection limits, and types of INPs targeted (biological vs mineral dust?). Indicate whether INP data have been used previously or how they compare to other Arctic datasets (e.g., Ny-Ålesund, ARM sites).

Reply: Most of the additional information requested by the reviewer is in the reference Gjelsvik et al. (2025). In the revised manuscript, we will provide some additional details on the method and sample collection. We will also add a brief additional analysis of the interrelation between INP and water isotope data data in the revised manuscript.

1. Line 535: Fig. 8a shows notable discrepancies between precipitation rates from the Parsivel2 disdrometer and the Andenes AWS, particularly during high-intensity periods (IOP1 and IOP5). Please clarify the likely causes (e.g., wind-induced undercatch for the AWS gauge, phase-dependent biases during snow/graupel, disdrometer sampling/quality-control limitations, siting/height differences, and any post-processing corrections). If possible, quantify expected uncertainty ranges for each sensor and indicate whether the differences affect subsequent isotope interpretations.

Reply: We agree that this is an interesting aspect that can be mentioned in the context of Fig. 8. We will include a sentence that summarizes possible reasons for the different response, such as undercatch of solid precipitation for the Geonor station, but also artifacts that are in particular during cases of snow for the Parsivel2. The uncertainty of the precipitation amount does not induce uncertainty of the isotope composition of the precipitation, unless one is interested in the amount-weighted isotope composition, as for example in hydrological applications.

 

2. Section 5.4: The transect analysis is valuable, and the manuscript notes potential artefacts (e.g., exposure time/evaporation in boxes and spillover for T3–T4). However, the conclusion that Coast measurements are “representative across the Lofoten archipelago” would benefit from clearer quantification. Please (i) define representativeness (which isotope metrics, which conditions), (ii) provide summary statistics (mean bias/spread vs distance; results with/without T3–T4), and (iii) discuss how limited spatial sampling and event-to-event variability constrain generalization beyond the sampled transects, particularly when inter-comparison with other products (e.g., satellite- or reanalysis-based gridded datasets) is needed to assess spatial consistency.

Reply: We will rephrase the text in this section by specifying that we consider representativeness in terms of the inter-event variations in δD being larger than the within-event variation. Figure 10 is meant to represent such summary statistics visually to indicate the potential for a more quantitative analysis. We are not aware of satellite or reanalysis-based datasets that would allow us to make statements about the spatial representative of water isotopes in surface snow or precipitation. We feel that while interesting, an extensive discussion of these aspects is beyond the scope of a dataset description paper, as it would venture into a quantitative evaluation and dataset interpretation, rather than indicating dataset potential.

 

3. Line ~605: You state that the strongest negative vapor isotope gradients occur during mCAO periods when surface fluxes and non-equilibrium fractionation are strongest. This is plausible, but please add physical support: e.g., show/quote the coincident surface flux or stability indicators during mCAO and clarify the mechanism linking enhanced fluxes to the sign/magnitude of ∆δD and ∆d-excess. Since you use 10-min averages due to a few-minute site time offset, please also provide a brief sensitivity check (e.g., 10 vs 30 min) to demonstrate the classification result is not an artefact of averaging choice.

Reply: We have repeated the analysis in Fig. 10 using 30-min time averaging and find almost the exact same results. We will briefly state in the revised manuscript that the detected gradients are quantitatively very similar when averaging at a 30 min interval. The argument for the stronger gradient is that the near-surface vapour is actively influenced by surface evaporation flux, which takes place under stronger non-equilibrium conditions during mCAOs (e.g. Duscha et al., 2021; Sodemann et al., 2024). The positive d-excess signature is an integrated signature from water vapour advection over the open water upstream, and the CAO index means that the ocean is warmer than the atmosphere. In the absence of local surface flux and atmospheric profile measurements over the open ocean, we use citations from relevant literature to support our claim here.

 

4. Section 6 (IOP2 case study): The IOP2 multi-instrument synthesis is very informative, including the period where the MRR stops recording precipitation while Parsivel/AWS indicate higher rates. Please add a brief cross-IOP context: are the key isotope–meteorology linkages seen in IOP2 also observed in other IOPs (especially IOP1 and IOP5)? Even a compact table/paragraph summarizing which features repeat versus which are event-specific would help users interpret representativeness.

Reply: With regards to other IOPs, we already note in the discussion of Figs. 6 and 8 that some similarities, but also differences between IOPs are present. It is difficult to fully address the reviewer's request without diving deeper into the scientific analysis of the dataset that we consider consistent with a data description paper. However, we will note in Sec. 6 commonalities between several IOPs, namely that the precipitation δD is more negative than vapour in IOP1 and IOP5, and that we see co-variations in the aerosol size distribution with δD. We will state that these co-variations are not uniform across all events, but clearly motivate further investigation of the interrelation of aerosols and water isotopes.

 

5. Discussion: The manuscript suggests the dataset can support improvements in NWP/model prediction for high-latitude mixed-phase precipitation. Please specify what aspects are most directly constrained by these observations (e.g., cloud liquid/ice partitioning, precipitation formation efficiency, low-level evaporation/sublimation, boundary-layer mixing) and outline a concrete example of model evaluation (which variables to compare, at what time/vertical resolution, and how isotope information adds value beyond standard thermodynamic fields).

Reply: Thank you for this valuable suggestion. Some of the requested information is already provided in the introduction, and in the rephrased scientific objectives. The Conclusion section also provides an extensive list with specific examples suggesting where aerosols/INPs and water vapour isotope measurements can be used in combination to improve numerical models. We believe this already corresponds to what the reviewer requests, and therefore decided not to repeat some of same information in a Discussion section.