Groundwater Storage Variations across Climate Zones from Southern Poland to Arctic Sweden: Comparing GRACE-GLDAS Models with Well Data
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThe article compared observation well data and GRACE-GLDAS model in Poland and Sweden, and made some analysis. However, the article has many shortcomings. It lacks innovation in methodology and lacks highlights in conclusions. The article reads more like a survey report than an academic paper. Therefore, I recommend rejecting the submission.
Specific comments:
1. Line 112-116: The authors state that the purpose of the study is to increase the understanding of differences (if any) in the accuracy of GWS change predictions obtained from the GLDAS groundwater model or from the GRACE/GLDAS connection under different (hydro climatic) conditions, potentially identifying limits of their related application and potential improvement needs. However, GRACE-GLDAS can obtain groundwater data, but can the GLDAS model obtain groundwater changes separately? Additionally, the article lacks coupling analysis between different hydrological climate conditions and groundwater changes. Simply selecting data from different regions does not enhance the understanding of groundwater changes under different hydrological climate conditions.
2. Line 126-127 and Figure 1: The location of the river and two basins needs to be marked in Figure 1. In addition, Figure 1 has a low resolution and the marked numbers for basin 3 and 4 are not displayed completely.
3. Line 217-218: To the best of my knowledge, GFZ does not have mascon product. The publishing agencies for mascon products are JPL, CSR, and GSFC. Therefore, this is a major mistake in the article.
4. Line 271-272 in Figure2: I did not find the green line in Figure 2, please check.
5. Results: The entire results section seems like a survey report, lacking effective explanations for the differences between the GRACE-GLDAS and observation well data, and lacking coupling analysis with hydrological climate distribution.
6. Conclusions: The conclusion of an article should highlight the methods employed, the comparisons made, and the conclusions drawn. It should emphasize the key aspects of the methodology and the findings, rather than simply listing survey results. If one were to only read the conclusion, they might not fully understand the scope of the authors’ work in the article.
Comments on the Quality of English LanguageThe English expression is poor, and authors need to improve their English writing skills.
Author Response
Please see the file attached
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsReview report
for the manuscript entitled “Groundwater storage variations across climate zones from southern Poland to Arctic Sweden: Comparing GRACE-GLDAS models with well data” submitted to the Remote Sensing Journal
The topic of the manuscript entitled “Groundwater storage variations across climate zones from southern Poland to Arctic Sweden: Comparing GRACE-GLDAS models with well data” is suitable for the publication in the Remote Sensing Journal.
The manuscript shows the fit of groundwater storage variations determined using GRACE and GLDAS models to groundwater level variations obtained from measurements in wells over the area of Poland and Sweden in 2003-2022. In particular, the results obtained show that JPL_NOAH_GWSA model (the combination of GRACE JPL and GLDAS NOAH) is not suitable for reliable reproduction of groundwater storage variations for the area investigated. It is due to overestimation of the component obtained from GLDAS NOAH, being the sum of the soil moisture, the snow water equivalent and the biomass of plant canopy. In contrast, the GFZ_CLM_GWSA model (the combination of GRACE JPL and GLDAS CLM) reflects groundwater dynamics for Poland and southern Sweden very well. It is confirmed by cross correlation coefficients of 0.8 level. However, for the sub-Arctic and Arctic northern Sweden the GFZ_CLM_GWSA doesn’t fit to the observations in wells. Cross correlation coefficients are equal 0.2-0.3. According to the authors conventional methods for deriving groundwater storage anomalies and their seasonality ceases to be reliable in presence of considerably infiltration-blocking ground frost and permafrost.
The agreement of trends of groundwater storage and groundwater level anomalies depends on the area. For Poland and sub-Arctic and Arctic Sweden they generally agree. For south Sweden there is the disagreement in trends that according to the authors can be caused by anthropogenic pressures. The observation wells are often located in eskers used for water supply.
The validation studies described in the article are needed to increase the understanding of spatio-temporal limits of remote sensing model applicability, not least in data-scarce sub-Arctic and Arctic environments where processes are complex due impacts of snow and (perma)frost.
The results are acceptable for publishing, however, the manuscript needs revision and editing. The following general comments should be taken into consideration:
1. Language: please be more clear and precise in your statements (e.g. „For the seasonal dynamics in sub-Arctic and Arctic climate zones…” Seasonal dynamics” of what quantity?), check all abbreviations, add the explanation of the missing ones, rephrase some sentences/phrases (please see the attached manuscript remotesensing-2956744-peer-review.pdf)
2. Figures: Please review the drawings Fig. 1, 7, 8, 9 (details in the attached manuscript remotesensing-2956744-peer-review.pdf)
3. References: Double reference in [12]. If possible, give up non-English language items. Unify the citation style of doi numbers.
4. I propose replace the phrases "CCF function/functions" occurring in the text with "CCF", because the abbreviation CCF means cross correlation function.
5. Please explain why to study correlations with shifts up to 20-something months. Isn't it enough up to a maximum of one year?
6. The sentences are often very long, making it difficult to read the text understandably. For example “The aim of this paper is to assess the correlation of groundwater level changes obtained from direct measurements in wells with groundwater storage changes calculated using GRACE observations and GLDAS models in Poland and Sweden, recognizing that such validation studies are needed to increase the understanding of spatio-temporal limits of remote sensing model applicability, not least in data-scarce sub-Arctic and Arctic environments where processes are complex due impacts of snow and (perma)frost.”
Other specific comments are added in the text of the manuscript (remotesensing-2956744-peer-review.pdf).
Comments for author File: Comments.pdf
Quality of English language is acceptable. The text is understandable.
Author Response
Please see the file attached
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsThis paper mainly assesses the correlation of groundwater level changes obtained from direct measurements in wells with groundwater storage changes calculated using GRACE observations and GLDAS models in Poland and Sweden. The quality of this paper is recommended for publication. Notably. Several comments follow.
#1. The introduction should be improved with previous research work, and explain why you choose these two models.
#2. Figure 2 should redraw, as it doesn’t show the full details.
Author Response
Please see the file attached
Author Response File: Author Response.pdf
Reviewer 4 Report
Comments and Suggestions for AuthorsPlease find below my comments related to the manuscript titled:” Groundwater storage variations across climate zones from 2 southern Poland to Arctic Sweden: Comparing GRACE-3 GLDAS models with well data “.
The paper presents a comparison of groundwater level changes with GLDAS outputs from the GFZ and JPL analysis centers.
The overall approach is not novel, however it is always useful to have comparison works of various open datasets and know their restrictions and advantages. From this point of view the present work is interesting.
There are two main issues in this work that should be carefully addressed:
a) For long term trend assessment and comparison authors should first remove the seasonal terms from all datasets and then compare the linear trends
b) The spatial aspect of similarities of the examined data sets is not touched at all. Authors just use the spatial means and compare them. It would be much more useful to see the spatial pattern of agreement of the examined datasets. E.g. JPL’s is found to perform poorly, but maybe this is only valid for specific areas.
I would also suggest that authors should publish their whole dataset along with their work. This would support open science and provide useful data to scientific community to replicate authors’ work and expand the research findings.
Detailed comments can be found in the annotated manuscript attached.
Comments for author File: Comments.pdf
Minor English editing is required
Author Response
Please seethe file attached
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Comments and Suggestions for AuthorsIn fact, the author has made some modifications based on previous comments and made some improvements to the article, but the overall quality of the article has not been significantly improved, the revised manuscript using annotation mode looks confusing. In addition, I believe the most important thing is the method used in the article, Grace-GLDAS vs. observation wells data, which has been widely used for many years and cannot be considered as the 'state-of-the-art methodology' expressed in the article. Therefore, I do not have more specific comments and tend to reject the article or make revisions based on the comments of other reviewers.
Comments on the Quality of English LanguageModerate editing of English language required
Author Response
Thank you for the feedback. Indeed, from a methodological viewpoint, Grace-GLDAS has been used for many years and we surely see the reviewer’s point that the technique as a whole is not brand new – in the manuscript, we rather consider a spectrum of modules that were released at different times during the past years. What we tried to convey (and could have expressed better) is the broader scientific perspective, where the method is being used in a large number of recent and current research studies, including those that study key environmental changes in cold regions (not least the Arctic), which are subject to extended periods of ground frost. We specifically relate to an existing knowledge gap in such cold regions (that also are generally data-scarce), namely that remote sensing methods have not been thoroughly validated against observations. We hence see our inclusion of quality-checked observation data from Swedish cold regions in our total dataset as one of the main novel aspects of the work, which hopefully can contribute to, for instance, an increased understanding of the region’s associated on-going environmental changes.
To avoid misleading statements, we have not used the “state-of-the-art” expression in the manuscript itself, nor in the present, revised versions either, considering the above concerns.
Indeed, we carefully considered and responded to the other review comments.
Reviewer 4 Report
Comments and Suggestions for AuthorsI went through the revised work titled: “Groundwater storage variations across climate zones from southern Poland to Arctic Sweden: Comparing GRACE-3 GLDAS models with well data”
The manuscript is improved but I would recommend that authors put a few details on long term trend estimation, like what they have put in their response letter (see below highlighted text), including reference:
“The calculations performed used the approach of simultaneous determination of the trend and seasonal components, using the model:
𝑮𝑾𝑺𝒎𝒐𝒅𝒆𝒍=𝒂𝟎+𝒂𝟏𝒕+𝒂𝟐𝐬𝐢𝐧(𝛚𝐭)+𝒂𝟑𝐜𝐨𝐬(𝝎𝒕),
where the coefficients 𝒂𝟎,𝒂𝟏,𝒂𝟐 𝒂𝒏𝒅 𝒂𝟑 were determined on the basis of linear regression using the least squares method.
According to the literature, this approach is correct, see e.g.
“Modeling trends and periodic component in geodetic time series: a unified approach” by M. Calvano et al., 2015, DOI 10.1007/s13209-015-0134-1”
In my initial review, I had raised the issue of spatial context, but authors did not touch it. The spatial aspect should be somehow covered, at least authors should provide some maps of the spatial trends in TWS, just to show which areas are more severely impacted by changes.
In my initial review I had suggested provision of the whole dataset. Authors replied that they have provided the links, but actually they did so only for the open remotely sensed and modeled datasets, which are already publicly available. They did not provide the groundwater level dataset they analyzed. I strongly suggest that authors put a Data availability section at the end of their work and put links for all datasets used, including the groundwater level data. They can publish them in some repository (I think mdpi provides such an option but also, they can put them in Mendeley Data, or any other repository). This is important to support the integrity and reproducibility of the work presented but also would promote open and fair science and give the opportunity to many other researchers to use them. Also, citations of manuscripts providing open datasets are expected to raise.
Comments on the Quality of English LanguageMinor spell check is required
Author Response
Dear Reviewer,
Thank you for your support and suggestions. Here are our answers:
I went through the revised work titled: “Groundwater storage variations across climate zones from southern Poland to Arctic Sweden: Comparing GRACE-3 GLDAS models with well data”
The manuscript is improved but I would recommend that authors put a few details on long term trend estimation, like what they have put in their response letter (see below highlighted text), including reference:
“The calculations performed used the approach of simultaneous determination of the trend and seasonal components, using the model:
????????=??+???+?????(??)+?????(??),
where the coefficients ??,??,?? ??? ?? were determined on the basis of linear regression using the least squares method.
According to the literature, this approach is correct, see e.g.
“Modeling trends and periodic component in geodetic time series: a unified approach” by M. Calvano et al., 2015, DOI 10.1007/s13209-015-0134-1”
Answer:
Thank you very much for your suggestion, we have put the above formula in the “Theory” chapter of the manuscript.
In my initial review, I had raised the issue of spatial context, but authors did not touch it. The spatial aspect should be somehow covered, at least authors should provide some maps of the spatial trends in TWS, just to show which areas are more severely impacted by changes.
Answer:
Thanks for sharing thoughts and suggestions on the spatial issue. The above referred initial review suggested that: If we would look closer into the spatial variations, perhaps we can understand why JPL_NOAH_GWS performed so poorly (as specifically said in the review: “E.g. JPL’s is found to perform poorly, but maybe this is only valid for specific areas”). We have now analyzed this question about local JPL performance and spatial trends further. In principle, many groundwater parameters indeed can show considerable spatial heterogeneity, as implied by this review comment. However, in the additional analysis, we saw that - in fact - the critical phase shift relative to observations was present in all local groundwater data. The wells hence behaved similar in this aspect; all of them had similar seasonal peaks and low-points that were distinctly different (shifted) from the pattern shown by the JPL_NOAH_GWS model.
We have to average over area since the GRACE original resolution is 3 degrees x 3 degrees and considering smaller area shouldn’t be performed. That is why everywhere values are averaged (we mean wells, GLDAS and GRACE data), taking into account smaller or point analyses is not correct.
In my initial review I had suggested provision of the whole dataset. Authors replied that they have provided the links, but actually they did so only for the open remotely sensed and modeled datasets, which are already publicly available. They did not provide the groundwater level dataset they analyzed. I strongly suggest that authors put a Data availability section at the end of their work and put links for all datasets used, including the groundwater level data. They can publish them in some repository (I think mdpi provides such an option but also, they can put them in Mendeley Data, or any other repository). This is important to support the integrity and reproducibility of the work presented but also would promote open and fair science and give the opportunity to many other researchers to use them. Also, citations of manuscripts providing open datasets are expected to raise.
Thank you for the feedback. Links are provided in appendix
