essd-2021-464

The main issue I see, maybe not strictly related to the paper, is the fact that following the link https://ppdb.us.edu.pl it's not easy to find the data pertaining to this paper. It's not immediately clear how to reach Geonetwork or OPeNDAP HYRAX and the datasets descripted in Geonetwork don't contain the direct links to the data files. I would suggest to include the links in it. It was easier to follow the Zenodo links.

L41 to which specific "disparities" are you referring to? Please clarify.
L47 "is a well-studied…" L49 "glacier covered an area…" I suggest to use the past form.
L50 I would suggest to specify that Slyngfjellbreen is the south-east part of Werenskioldbreen.

Collective answer to the above comments: Corrected.
Table1 HMP45AC is from HOBO? Answer: No, it is Vaisala sensor. Corrected. L91 Specify the month(s) corresponding to the end of the ablation season, as done for the accumulation season.
L96 boreholes have been drilled "on" the stakes? You mean "at" or "near"? L111 SWE already defined at L93 L116 "and another from Pleiades…" L131 did you checked the meaningfulness of these trends?
Collective answer to the above comments: We have calculated the significance of the trend presented in this study using the non-parametric modified Mann -Kendall test (Hamed and Rao, 1998) with considering the effect of autocorrelation of time series. The slope of the trend was calculated using Sen's method (Sen, 1968). The test indicated the statistically significant increasing temperature trend in the period 2010-2016 (with the significance level alfa = 0.05 and p-value = 0.036) taking into account the 12-month seasonality the Sen's slope was 0.02. The text in the manuscript has been corrected.
L135 I would say that the height of the sun is the first responsible for daily variations.

Answer: Agree. Corrected.
L140 Could you provide estimates of albedo for snow and ice obtained from your measurements?
Answer: Based on the collected data on the upwelling and downwelling of shortwave radiation, calculating the albedo for ice and snow does not pose major problems. Nevertheless, the issue of albedo, its variability, and modelling is the subject of another paper which the authors work on, which is based on the data presented in this manuscript. So we are not going to present the results of albedo in this paper.
L155-163 I would like to see here some description on how the point mass balance is calculated. There is only a light sentence like "based on physical parameters of snow and ice".

Answer: The description has been extended.
Figure3 Please correct the y label "equivalent". Also, I would draw the zero line.

Answer: Corrected
Figure4 Whiskers are difficult to see. And please try to explain better what they represent. What does it mean "an error"?
Answer: The whiskers present the error estimated using the total differential function. The size of the measurement errors is smaller than the size of the points on the plot that's why it is hardly visible on the graph. L222 I'm not sure is correct to use the expression "increases by -014". I would use "decrease by 0.14" also for the summer balance.
Eq 2 and 3 What is n? What is deltat? I would like to see a clearer description of the equations.

Answer: The description has been extended.
L245 Melting is assumed to be zero when the temperature is positive?

Summary:
A data-paper serves usually a dual purpose and is somewhere between a technical report describing the measurements/ dataset and a research paper presenting detailed analyses. The manuscript addresses both, and describes datasets from Werenskioldbreen, a glacier in the Arctic archipelago of Svalbard, that cover both glacier mass balance measurements and operation of an automated weather station. These data have been made available and the manuscript describes the meteorological and glaciological measurements, presents some characteristics of each dataset and puts them in relation to each other. I applaud the authors for collecting and publishing these datasets from a generally datasparse environment such as the Arctic and I will recommend publication of the manuscript once some of the identified shortcomings have been rectified.
Description and presentation of data: although nominal sensor uncertainties are stated in Tab 1, I miss a more thorough discussion of data quality, clearly stated quality thresholds used for discarding data, procedures for gap filling etc. The problem of varying sensor hight above the surface is not at all addressed. Not much is learned from e.g. Fig 2 that presents 7 years of data in hourly resolution. Instead, values could be presented aggregated in monthly boxes to enhance figure readability (for instance Schuler et al. 2014). Adding supporting information about data completeness/ quality would give the reader a quicker overview than a listing of all gaps in the main text. Some figures present uncertainties, but it is never explained what these uncertainties represent (standard deviation, RMS error or other measures?). Datasets in the zenodo repository come in plain CSV format without explanatory readme-file or other metadata. Even with the manuscript at hand, it needs some intuition to identify the individual variables and which units the values are affiliated with. This must clearly improved to make the dataset useful. At the other repository https://ppdb.us.edu.pl, the meteorological data are available in netCDF format and contain all required metadata. This should be the case also for the data on zenodo. Especially given the difficulties finding the data at PPDB, that the other reviewer already commented on. Throughout the manuscript, the authors repeat statements about the significance of the dataset, but after all these statements appear more persuasive rather than convincing and would benefit from better support (or more modest wording). Related to the point above, the analyses to show the significance of the data should be improved; 10 years too short for climate trend analysis. Instead the data could be discussed in light of existing longterm series from Hornsund. Do trends agree? How do differences vary (seasonally, decade)? Such a discussion would provide good arguments for the necessity of these measurements, why should we need additional series if they are only linearly related to other operational series?
Answer: The methods of calculating errors were added. The chapter about the data preparation, quality thresholds and procedures for gap filling were refined. L16: since you refer to Cogley et al 2011, you should stick with its terminology: …consists of seasonal and annual surface mass balance measurements… ("surface annual winter, summer and net balance" is at least confusing if not inappropriate).

We've added also the information and analysis about the changes in the height of the sensors above the glacier surface. Thank you for your comment and for drawing attention to the technical shortcomings of the data repositories. Data records in zenodo have been improved. They are in analogues format both for the PPDB and in NetCDF. The direct links to datasets in Polish Polar
Collective answer to the above comments: Corrected.
L18: this is an example for the unsupported, persuasive wording mentioned above: "highquality and long-term datasets".the quality is not well assessed in the manuscripts and 10 years is not really long-term.

Answer: Agree. Corrected.
L18: …to serve as local forcing data… L24: "…and significantly expand our knowledge" not per se, but they may contribute to that. Such a bold statement would require a reference.
Collective answer to the above comments: Corrected. Answer: Damaged sensors were replaced as soon as a failure was noticedusually during spring or autumn service.
Tab 1: I understand that sensor level above surface will continuously vary, but at least you could indicate original height or the target heights at which you wish the sensors to be… Answer: Added.
L83: is this the measurement interval or the recoding interval?
Answer: measurement interval. It has been corrected in the text.
L87: The locations have been chosen to cover the elevation range… (combine with next sentence) L90: did you really use an ice core drill or the Kovacs stake auger? (same brand, different tool) Answer: Yes we use ice core drill. We didn't have the Kovacs stake auger in our inventory at the time. : what you refer to as "surface mass balance" seems to correspond to the "glacier wide mass balance". The first considers only mass changes at the surface of the glacier, the second is the integral over the surface area of the glacier, see Cogley 2011.

Collective answer to the above comments: Corrected
L155ff: which DEM did you use for which period? How did you deal with transitions between them?
Answer: We've used 2008 DEM for the calculations for 2008-2019 period and the new 2017 DEM for the latest 2020 WGMS report. We did not have access to this product before. There were no special transitions between these DEMs. Currently, we have collected more DEMs for the period 2008-2020 and earlier, on the basis of which we are going to prepare a data correction for WGMS and an article on the recalculation of the mass balance of Werenskioldbreen and Hansbreen. The text in the manuscript has been supplemented with relevant information.
L120: travel restrictions started in Mar 2020, so they cannot be blamed for data gaps in 2017-2020… Fig2: how much are the values influenced by changing sensor levels as the surface melts down/ sensors are relevelled? Add some information about completeness of the records (e.g. percentage of time). The shortwave radiation records could be analyzed in terms of albedo and its seasonal changes; adding potential clear sky radiation would give an impression about cloud effects.
For a melting snow/ glacier surface, the surface temperature is capped at the melting point and so is the outgoing longwave radiation (316 W m-2). Discuss why you observe upwelling radiation in excess of this value! Answer: Based on the collected data on the upwelling and downwelling of shortwave radiation, calculating the albedo for ice and snow does not pose major problems. Nevertheless, the albedo, its variability, and modelling of the albedo is the subject of another scientific paper which the authors are working on, that's way we do not want to present the results of albedo in this paper.

Increasing distance between the glacier surface and the sensors during the season is affecting on the air temperature measurements. Periodic measurements of the vertical temperature gradient between 0.5 and 4 m carried out at the AWS indicate that the air temperature in the atmospheric boundary layer changes by about 0.2oC per 1 m during the ablation season. Assuming the homogeneity of the surface around AWS, increasing the distance of the CNR4 sensor from the glacier surface should not affects its measurements,.
The values above 316 W m-2 of outgoing longwave radiation may be caused by the presence of the water under the station or the presence of sediment or cryoconite. Both of these situation were observed at AWS. In the he view of the CNR4 sensor (180 o ), there is also a mast with a logger, sensors and a solar panel, what cause distorts of the observations. These problems are difficult to eliminate. Table 2 and the text in the manuscript has been supplemented with relevant information.

Answer: Corrected.
L148: if the sensors are not robust enough for polar conditions why did you not use others that are better suited? Answer: We use high-quality sensors. However, in stations that are left on the glacier for long periods without servicing, they fail more quickly. The sentence has been corrected.
L150: …are of great value for solving specific scientific problems. Could you give some examples to support this claim?

Answer: Added
L161: no, you cannot measure these by only 1-2 visits, you do measure the balance! Answer: Agree, the text has been corrected.

L161: …on the physical properties…
Answer: The sentence has been changed.
L163: better description of error estimation is required.

Answer: Added
Fig 3: y-label should be " Water equivalent" Why are the stake values not sorted according to stake number? This is a confusing sequence… Answer: The stake values are sorted according to the height above sea level. The description has been added in the text.
L165ff: the observation that the variability of annual balances is dominated by the variability of summer balances is not new, add references. This is actually much better seen in Fig 6. Answer: Added L178: …was generally low.
L181: …for calculations of the glacier-wide surface mass balance.
L188: …was then averaged weighted by layer thickness.
L189: mention more about the method of Sturm (2010)

Answer: Corrected
Tab3: add the SWE values as well.

Answer: Added
How did the snow density vary in space/ with elevation?
Answer: There no correlations between snow density and the snow cover depth or elevation. This was also described in the manuscript. Answer: The coefficients of the regression and error equations have been added. The snow depth is very highly and positively correlated with the elevation (correlation coefficient was 0.85 for p < 0.05)), but there was no correlation between the bulk snow density with the elevation (-0.07) and snow depth (0.16). Base on the regression analysis of the SWE and the single parametersnow density (ρ), elevation and snow depth (hs), explains 13%, 62%, 96% of the SWE respectively. The variability in Fig5b. is dependent more on the snow depth than snow density. The list of average SWE values has been added to Table  2. L199: R2 = 0.62 is not "very high correlation" !! L206: as mentioned above, your "surface mass balance" corresponds to "glacier-wide mass balance" (Cogley 2011) In the following lines you fuse the model results and your measurements to infer a longterm evolution. This is questionable as long as you do not show the level of agreement between model and observations. However, this is not possible to derive since the two series do not overlap in time (model 1959-2002, measurements 2009- L247: better explain how the error estimate has been derived Answer: Added L248: better explain the gap filling: has the Hornsund data been adjusted using a regression or similar? Answer: Description has been added in the chapter on meteorological data. What value for DDF has been used? How has it been determined/ selected? How are eqs 2 and 3 used in the calculations? For the entire glacier? Or per stake location (and according adjustment of temperature records)?
Answer: For each season, a separate DDF was determined based on air temperature data and cumulative ablation calculated from glacier-wide mass balance. Equations 2 and 3 were used for the entire surface of the glacier. Answer: Additional description of the analysis and comparison of the data has been added to the manuscript (L320-348). The data in the database does not contain any completed gaps made with the use of data from Hornsund. Comparing such data with the data from Hornsund would be unjustified.
L284-306 do not belong to a quality-control section but rather to a separate section "dataset structure" Answer: New section has been added L307: the first sentence can be removed without losses.
L310-317: consider moving this to the "methods" section Answer: After adding a new section "Dataset structure" this paragraph fits here.
L318: what are the results of this comparison? Answer: References have been added. Further analyzes will be presented in the forthcoming article on the recalculation of the mass balance of Werenskioldbreen and Hansbreen.
L321-326: consider moving this to the "methods" section L348: "high-quality and long-term" …see comments above.

Answer to Editor comments:
You state that the files are CF and ACDD compliant. I ran one through 'cchecker.py' (the IOOS compliance checker) and there are several ACDD and CF issues. Please run a CF-compliance checker on your NetCDF files. Please use ISO 8061 date formats (YYYY-MM-DD).
ANSWER: Thanks for your comments. NetCDF files have been supplemented with additional information, the dates to ISO 8061 standard have been corrected and the problem with ACDD has been solved.
Line 32: I do not think "Svalbard airport" is the correct name.
ANSWER: The station is situated at the airport of Svalbard, which in Norwegian is called Svalbard lufthavn, and this is the name used in Norwegian station lists. In English-language publications, the name Svalbard Airport is used.

ANSWER: It has been corrected.
Line 245: Melting is assumed to be zero when air temperature is >= 0 °C ? Should this be <= 0?
ANSWER: This is a mistake. It has been corrected.
We hope that the changes we have made will be satisfying.