Characteristics of the Greenland Ice Sheet Mass Variations Revealed by GRACE/GRACE Follow-On Gravimetry
Round 1
Reviewer 1 Report
The topic is very interesting and helpful to understand the ice mass variation in the Greenland. However some critical problems need to be addressed further before I recommend it to be published by this journal.
Comments for author File: 
Comments.pdf
Author Response
Point 1: My main comment concerns the reasonability of the method of non-cumulative mass variations. For most of GRACE applications, we are used to looking at the “mass anomaly” by subtracting the mean value during the time spans, which reflect mass variations respect to a fixed (static) time point. The status of the target can be directly recognized by its positive/negative face value, i.e., positive value represents mass accumulation and negative value represents mass loss.
Response 1: Thanks for the comment. We totally agree that most of GRACE applications generated mass anomalies by subtracting a mean field value. These mass anomalies reflect mass change relative to the mean field value. If the mass anomaly is positive, we can say that mass is accumulated relative to the mean field value. If it is negative, we then say that mass is lost relative to the mean field value. As the mean field value varies during different timespans, such mass anomaly actually reflect a relative mass change. As most of GRACE applications were focused on the long-term trend of mass change time series, mass anomalies by subtracting a mean field value were commonly used.
Point 2: But the non-cumulative mass change in the paper has completely different physical meaning, which represents monthly relative mass variation relative to last(previous) month. It is a dynamic index, for every month, the reference point is different. The face value of the non-cumulative mass thus can’t indicate the status of the ice sheet, for example, positive value may also represent mass loss as long as the melting in the latter month is less than the former one, but glaciers were actually melting. Thus the said “mass gain” and “mass loss” in the paper are unreliable.
Response 2: Thanks for the comment. There must be some misunderstanding. As we know, GRACE/GRACE-FO observations detected cumulative mass change (van den Broeke et al., 2009). It is not an absolute mass increase/decrease at certain month. The non-cumulative mass change approximately represents absolute mass increase or decrease during the period from last month to current month. It has nothing with a fixed reference point. It does reflect absolute mass increase or decrease at each month, which will directly influence the status of the ice sheet. One can consider the case like the following: assuming that the abolute mass of the GrIS is at the month and is the selected mean field value during certain timespan, indicates cumulative mass change observed by GRACE/GRACE-FO at the month and reflect the actual mass increase or decrease occurred from the month to the month. The physical meaning of is straitforward, which is the base for the Mass Buget Method(MBM). The MBM can also generate cumulative mass change using simulated monthly precipitation, meltwater runoff, sublimation, erosion/deposition due to drifting snow and ice discharge, which agreed well with those from GRACE (van den Broeke et al., 2009). Therefore, our method works for better revealing the characteristics of the GrIS.
Here it is the reference: van den Broeke, M.R., J. Bamber, J. Ettema, E. Rignot, E. Schrama, W.J. van den Berg, E. van Meijgaard, I. Velicogna, and B. Wouters, 2009. Partitioning recent Greenland mass loss, Science, 326, 984-986.
Point 3: On the other hand, the findings and conclusions are quite bland and expected. It does confirm
the validity of the datasets and the computational models used, but that is not surprising,
especially now that such validity has already been confirmed for GRACE time and time again in the
literature.
Response 3: Thanks for your comment. As you mentioned, most of GRACE applications used mass anomaly by subtracting a mean field. Here we compute non-cumulative mass change time series and we confirm the feasibility of our method using the MBM. To my knowledge, we report for the first time that spring mass loss became more frequent since year 2015, and autumn mass gain occurred more frequently after year 2014. We find that in recent two decades, mass accumulation time series show a positive trend (7.43 Gt/yr), while a negative trend of 7.03 Gt/yr can be found from mass loss time series. We also quantify the impact of factors on monthly mass change of the GrIS. Our results illustrate the advantage of calculating non-cumulative mass change on revealing the characteristics of the GrIS.
Point 4: As readers may want more geophysical mechanism, but section 3.4 is obviously flabby
and unpersuasive, the authors just draw the time series of the components of SMB while their
contribution to GrIS’ mass variations are absent. Section 3.4 undoubtedly needs adequacy
changes, for e.g., the correlation between runoff and temperature is meaningless, any arbitrary
time series with annual oscillation will all show high correlation with them, there is nothing
mysterious about these.
Response 4: Thanks for your comment. Our study is focused on analyzing the characteristics of the GrIS revealed by GRACE/GRACE-FO during the past two decades. In section 3.4, we show monthly time series of each component of the MBM in order to better understand the factors dominating the monthly GrIS mass variations. We compare the magnitudes of total precipitation, meltwater runoff, sublimation and erosion/deposition as well as ice discharge. We do compute the correlation coefficient between monthly meltwater runoff time series and variations of 2 m temperature. And we quote a previous study to illustrate the complex relationship between these two factors.
To quanify the contribution of each component, we add one paragraph and Figure 8 in Section 3.4. Modifications can be found in lines 417-439.
In section 3.5, considering that summer mass change dominates mass loss of the GrIS at the corresponding year, we analyze the connection between summer mass loss and summer NAO index. A fully explanation of the geophysical mechanism is another topic and it would be further studied in future. Thanks again for all your comments.
Author Response File: 
Author Response.docx
Reviewer 2 Report
The authors monitored the mass changes in the Greenland Ice Sheet (GrIS) with GRACE/GRACE-FO data in recent 20 years, and then analyzed its variations in different months and seasons with reasonably explanation. This manuscript is full of content and clear logic, I think it will make useful contribution on regional mass change researches after some improvements:
1) Chapter 3.1, is there any paper to support your conclusions?
2) Line 152~158, you can make linear regression fitting on monthly mass variations of each month and mark the trends like Figure 5b, which may express your views more clearly.
3) Line 182~188, please cite some published papers to verify your results.
4) Chapter 3.4, can you quantify the contribution of different factors on GrIS mass changes combine with other data. This may be one of the highlights of your manuscript.
Author Response
Point 1: The authors monitored the mass changes in the Greenland Ice Sheet (GrIS) with GRACE/GRACE-FO data in recent 20 years, and then analyzed its variations in different months and seasons with reasonably explanation. This manuscript is full of content and clear logic, I think it will make useful contribution on regional mass change researches after some improvements:
Response 1: Thanks for your positive response. Yes, we cannot agree more that this study will be useful for researchers to better understand the characteristics of the GrIS mass variations at multiple spatio-temporal scales, which could contribute to researches related to regional mass change over the GrIS.
Point 2: 1) Chapter 3.1, is there any paper to support your conclusions?
Response 2: Thanks for the comment. Do you mean the consistency between mass change time series from GRACE/GRACE-FO and the MBM? Sure, there was a paper which confirmed their consistency using cumulative mass change time series during the GRACE period (van den Broeke et al., 2009). Here we compute and compare non-cumulative mass change time series from both methods, aiming at directly showing the characteristics of mass change, especially at monthly and seasonal timescales. As you know, cumulative mass change time series are commonly used to analyze the long-term mass trend. We should keep in mind that the essence of calculating cumulative mass change time series is the same as that of computing non-cumulative mass change time series. Modifications can be found in lines 167-168.
We also cite a paper and add two sentences to demonstrate the numerical differences on estimating the most negative and the second negative mass change of the GrIS. Modifications can be found in lines 217-221.
Point 3: 2) Line 152~158, you can make linear regression fitting on monthly mass variations of each month and mark the trends like Figure 5b, which may express your views more clearly.
Response 3: Thanks for the suggestion. As the amplitudes of mass change in each month are visible during the study period, and there are too many colors indicating each year in Figure 2, we think it is not necessary to add the linear trends here. We do update Figure 2 with some modifications in line 241.
Point 4: 3) Line 182~188, please cite some published papers to verify your results.
Response 4: Thanks for the comment. In Lines 179-180, we provide the website of polar portal season report provided by the Danish Arctic research institution. Sentences in Lines 182-188 are rewritten according to these season reports.
Point 5: 4) Chapter 3.4, can you quantify the contribution of different factors on GrIS mass changes combine with other data. This may be one of the highlights of your manuscript.
Response 5: Thanks for the comment. This is a really good one. We add one paragraph and Figure 8 to quantify the impact of each factor on monthly mass variations of the GrIS. Modifications can be found in lines 417-439. Many thanks for all your comments.
Reviewer 3 Report
1, I think the expressions in Figure 3 and Table 1 are duplicated.
2, The final analysis should be attributed to the change in sea level height for each season.
3, There should be a horizontal line at the x-axis in Figure 2, which is not clear in the current figure.
4, Figure 2 should be represented (a), (b), (c), (d) for each subplot.
5、The horizontal axis of Figure 2 should have a title item.
6、For the analysis of Greenland ice cap mass, the most important thing should be the analysis of the overall ice cap mass trend. This study has calculated the ice cap mass for each season, please elaborate the important background significance.
7、The captions in Figure 5 have obvious formatting errors, e.g. " a)";
8、For the reliability of the overall analysis, it is recommended to include altimetric data, etc.
9、In line 76, the mission should be "missions" in this place.
10、In line 89, please pay attention to the grammar of the sentence.
11、Please check the grammar of the whole article.
Author Response
Point 1: I think the expressions in Figure 3 and Table 1 are duplicated.
Response 1: Thanks for the comment. Do you mean the contents in Figure 3 and Table 1 are duplicated? If my understanding is correct, there should be a misunderstanding. In Figure 3, we compute total mass variations in each season, with the patterns of seasonal mass variations directly shown. In line 268, as the total mass variation in each season includes mass gain and mass loss, we quantify the contribution of mass gain and mass loss at monthly timescale in each season in Table 1. Thus, one could get more details about seasonal mass varitions, e.g., there were slight mass gain occurred in summer during the past two decades, such as the summer in 2006, 2009 and 2013 as well as 2019. We modify the Caption for Figure 3 and Table 1. Please find it in lines 286-289.
Point 2: The final analysis should be attributed to the change in sea level height for each season.
Response 2: Thanks for the comment. In this study, we are focused on analyzing the characteristics of the GrIS’s mass variations at multiple spatio-temporal scales, it is not necessary to express seasonal mass variations in sea level height. When extracting the long-term trend from cumulative mass change time series of the GrIS, it would be necessary to evaluate its impact in sea level height.
Point 3: There should be a horizontal line at the x-axis in Figure 2, which is not clear in the current figure.
Response 3: Thanks for the carefully checking. This is a good one. Yeah, we update Figure 2. Modifications can be found in line 241.
Point 4: Figure 2 should be represented (a), (b), (c), (d) for each subplot.
Response 4: Thanks for pointing it out. Modifications can be found in line 241.
Point 5: The horizontal axis of Figure 2 should have a title item.
Response 5: Many thanks. Modifications can be found in line 241.
Point 6: For the analysis of Greenland ice cap mass, the most important thing should be the analysis of the overall ice cap mass trend. This study has calculated the ice cap mass for each season, please elaborate the important background significance.
Response 6: Thanks for the comment. We agree that the overall ice mass long-term trend would be quite important for reflecting the status of the ice sheet, predicting its future evolution and searching for potential connection with climate change. As we know, the long-term trend is extracted from monthly cumulative mass change. Since we do not have long enough monthly mass change time series, the long-term trend is actually contaminated by inter-annual variations. The monthly cumulative mass change time series include short-term changes which may significantly influence the mass trend during a relatively short period. For instance, the short-term mass changes in year 2013 cause a slowdown on the mass trend of the GrIS. Whether there would be more cases like the short-term changes in 2013? We need observations as long as possible and it is necessary to analyze the characteristics of the GrIS’s mass variations.
Point 7: The captions in Figure 5 have obvious formatting errors, e.g. " a)";
Response 7: Thanks for pointing it out. Modification can be found in line 364.
Point 8: For the reliability of the overall analysis, it is recommended to include altimetric data, etc.
Response 8: Thanks for the comment. In this study, we compare non-cumulative mass change time series from GRACE/GRACE-FO and the MBM (which adopted independent data sources). The consistency can confirm that our results are reliable. If altimetric data is used, there could be some issues, e.g., the snow/ice density problem and the coverage of altimetric data.
Point 9: In line 76, the mission should be "missions" in this place.
Response 9: Thanks for pointing it out. Modifications can be seen in line 85.
Point 10: In line 89, please pay attention to the grammar of the sentence.
Response 10: Thanks for carefully checking. Modifications can be found in lines 98-99.
Point 11: Please check the grammar of the whole article.
Response 11: Thanks for the comment. We carefully check the whole manuscript and make modifications in lines 164, 193, 227. Thanks again for all your comments.
Author Response File: Author Response.docx
Reviewer 4 Report
In this manuscript, the authors compute monthly non-cumulative mass variations of the Greenland Ice Sheet (GrIS) in order to avoid potential bias between GRACE and GRACE-FO derived mass variations, aiming to analyze characteristics of the GrIS variations at monthly, seasonal and annual timescales.
The manuscript presents new and relevant findings of interest for the international scientific community. The topic of the manuscript fits well into the scope of the journal Remote Sensing.
The work is very well presented. The manuscript is clearly and concisely written and all figures are well prepared. The methods are clearly described and the results are well presented. Clear and relevant conclusions are drawn from the presented material.
In my view this is a very good, relevant and well presented manuscript which can be accepted in its present form.
Author Response
Point 1: In this manuscript, the authors compute monthly non-cumulative mass variations of the Greenland Ice Sheet (GrIS) in order to avoid potential bias between GRACE and GRACE-FO derived mass variations, aiming to analyze characteristics of the GrIS variations at monthly, seasonal and annual timescales.
Response 1: Thanks for the comment. Yes, this is exactly what we studied in the manuscript.
Point 2: The manuscript presents new and relevant findings of interest for the international scientific community. The topic of the manuscript fits well into the scope of the journal Remote Sensing.
Response 2: Thanks for the comment. We totally agree that the topic is appropriate for the journal Remote Sensing, and the new findings of the manuscript will contribute to researches related to the GrIS’s mass variations in the scientific community.
Point 3: The work is very well presented. The manuscript is clearly and concisely written and all figures are well prepared. The methods are clearly described and the results are well presented. Clear and relevant conclusions are drawn from the presented material.
Response 3: Thanks for your positive response. We carefully check the sentences and figures in the manuscript. One paragraph and Figure 8 are added to quantify the impact of each factor on monthly mass variations of the GrIS in Section 3.4.
Point 4: In my view this is a very good, relevant and well presented manuscript which can be accepted in its present form.
Response 4: Many thanks for the support.
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
The authors has made efforts on improving the manuscript. It sounds there is too much acronyms in the text, it will be good to attach a acronym list. I recommend it to be published by the journal.
Author Response
Point 1: The authors has made efforts on improving the manuscript. It sounds there is too much acronyms in the text, it will be good to attach a acronym list.
Response 1: Thanks for the comment. It is a good one. Yeah, we add a table in Appendix A.
Point 2: I recommend it to be published by the journal.
Response 2: Many thanks for the recommendation!
Reviewer 3 Report
Thank you for your answers and modifications. I think this research is very meaningful. Please carefully check the grammar of the article before online.
Author Response
Point 1: Thank you for your answers and modifications. I think this research is very meaningful. Please carefully check the grammar of the article before online.
Response 1: Thanks for the support. Yes, we carefully check the whole manuscript. Thanks again.
