Sea-Level Fingerprints Due to Present-Day Water Mass Redistribution in Observed Sea-Level Data

Round 1

Reviewer 1 Report

Review of “Sea-level fingerprints due to present-day water mass redistribution in observed sea-level data” by L. Moreira and colleagues. 

 

The authors compare the sea-level fingerprint solutions with the altimetry- and gravimetry-based estimates of sea level data corrected for steric effects. They found that ~25% variability in the altimetry-based estimates of sea level in the vicinity of Greenland is explained by the sea-level fingerprints. They also show strong correlations between the altimetry-based solutions and sea-level fingerprints in the pacific subtropical basins. Similarly, they report a large correlation between the gravimetry-based solutions and sea-level fingerprints in several regions (e.g., in the south Atlantic and Greenland's vicinity). These results are useful as a search for unambiguous detection of sea-level fingerprints of modern-day on-land ice/water mass changes continues. The paper is well written and logically organized, but I find two main issues that need to be addressed before it can be accepted for publication.

 

First, the satellite gravimetry-based sea level data (i.e., GRACE data) represents the change in relative sea level (i.e., the change in ocean water column height). So, these data should be compared against the relative sea-level (RSL) solutions of sea-level fingerprints, not the absolute sea level (ASL) solution shown in Figure 1. The authors briefly mention on line 105 why they should compare GRACE data against the ASL fingerprints, but it is not convincing to me.

 

Second, the basin-scale comparison of the altimetry-based solutions and ASL fingerprints does not look promising, even though strong correlations are reported for the tropical pacific basins (~line 292). These basins are in the “far-field” from the on-land mass loss hotspots (correctly noted on line 380), where the fingerprint effects (well, the GRD effects) are nominal anyway. Therefore, the reported correlation does not necessarily imply the detection of fingerprints. It should be mentioned in the paper to avoid potential misinterpretations.

 

Some additional comments:

 

Line 15: static factors? I've never heard about this before. You mean “static factors” are also called “sea-level fingerprints.” It is confusing to me.

 

Lines 18-19: Be consistent in using “land water” or “terrestrial water” here and elsewhere.

 

Line 23 and elsewhere: It is not clear what the “atmospheric and oceanic loading” means. You should explain it in the main text.

 

Lines 46-54: It is too long a sentence to comprehend… Rewrite.

 

Line ~86: “…The two effects cause…” Here, you define the relative sea level (RSL), not the absolute sea level (ASL). You should differentiate RSL and ASL upfront in the manuscript and relate these to satellite altimetry and gravimetry data used in this analysis.

 

Line 90: greater that => greater than.

 

Line 92: I would say about 10 to 30%. Also, mention that relatively large variability in the “far-field solution” is mainly dictated by the rotational feedback.

 

Eq. 1 is not the sea-level equation. It is a mere definition of RSL.

 

Line 97: time => time period of interest.

 

Line 102: gravity => gravitational acceleration.

 

Lines 104-106: This sentence does not look grammatically correct. Rewrite.

 

Lines 119-122: Not sure about this reasoning. Why can you not simply subtract the global-mean value from the ASL fingerprints?

 

Lines 123-127: All these values are related to the global-mean value, right? Make it explicit.

 

Figure 1: I recommend showing the uncertainty map as panel b in the figure.

 

Line 136 and elsewhere: Due to different American/international conventions, write the date more explicitly, e.g., March 10, 2021.

 

Line 151: These abbreviations (MBT and XBT) are not defined.

 

Line 175 and elsewhere: “We tried to remove…” => “We remove…”. Avoid using “…tried to…” here and elsewhere.

 

Line 175: There has also been an investigation on sea-level acceleration at regional levels. See, for example, https://doi.org/10.1029/2019GL086528. I think it is worth citing this article.

 

Line 180: Can you explain why and how you provide GIA correction to the noted solutions? Did you subtract GIA-related geoid signals? You do not need to subtract GIA-related VLM signals from the altimetry-based sea-level solutions, right?

 

Line 196: “it makes us wonder…” => avoid flowery phrasings.

 

Line 264: And accordingly, we can state that => We conclude that. This is an important finding that is worth reporting in the abstract, in my opinion.

 

Lines 275-276: It may be useful to show the basin boundaries in one of the figures (e.g., Figure 1).

 

Figure 8 panel: Warn the reader that x-axes are in log scale, although you have mentioned it in the text.

 

Lines 346-349: You should expand the GIA correction procedure a bit more clearly.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Review on "Sea-level fingerprint due to present-day water mass redistribution in observed sea-level data" by Moreira et al.

Dear Editor,

This paper is the detection of relative sea-level change induced by changes in Earth Gravity, Earth Rotation, and visco-elastic solid Earth Deformation, in short sea-level fingerprints. The authors compare a modelled result by Adhikari et al. (2019) with two type of data that could be interpreted as relative sea-level data: C3S altimetry sea-level (steric corrected) and GRACE CSR RL06 mascon model. Limited correlation is seen in the altimetry data, where some of the ocean basins show better correlation than others. The GRACE data shows much better correlation with the sea-level fingerprint model. The steric component seems to be the major contributor of the altimetry data and therefore is still difficult to say if the sea-level fingerprint is seen in the altimeter data. It is better observed (with higher correlations) in the GRACE data. I find the idea very interesting, but I would like to see if the authors could redo the correlation study, but then with similar spectral signature in the observed and modelled signal.

I would recommend minor revision before this paper can be published.

R1: The modeled signal (figure 1) has much more smoothed appearance (only long-wavelength signal) than the observed altimetry and to lesser amount the GRACE data. I would propose that a spectral decomposition is applied to the different signals and that the correlation is only performed for the spectral bandwidth in which the modeled signal has significant power. This alone could explain the improved correlation in the GRACE data with respect to the altimetry data, as GRACE data has less high wavelength signals. This exercise might also reveal better correlation in the altimetry data.

R2: I would advise to make FIGURE 2 rotate 90 degrees, as the figures are now skewed and are more difficult to interpret with respect to the other figures of the paper.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Review of Moreira et al 2021 – Sea-Level fingerprints due to present-day water mass redistribution in observed sea-level data

Moreira et al (2021) present a statistical analysis with the aim of detecting the barystatic-GRD sea level fingerprint of contemporary mass change in both the sea level altimetry and GRACE ocean mass record, for the period 2005 – 2015. The study utilises well established previously published products to conduct the analyses. They find a statistically significant trend for certain ocean basins, whilst acknowledging the signal to noise ratio is still low. This paper presents an interesting analysis and would be of benefit to the wider community, but I believe major corrections are required to both some of the explanations provided in the text plus the figures in the manuscript. My broad comments and specific line by line comments are listed below.

Broad Comments

Some of the choices for corrections on the datasets don’t seem to be consistent (see the specific comments for more detail), for example the differing GIA correction on GRACE ocean mass data and the steric corrected altimetry. There may be perfectly valid reasons for this, but they don’t seem to be explained in the text. Additionally, I cannot see any discussion as to the uncertainties used on the GRACE product for ocean mass. Is it from the data centre? Is it derived from a spread of mascon solutions? Considering a Pearson’s correlation coefficient is being run on the data this needs to be explained to the reader.

Some of the figures are very unclear and require modification to make them appropriate for the data they are representing. For example, Figure 2 is the wrong orientation, making it very difficult to view. Additionally, diverging colour ramps are used to represent some quantities that are not diverging (see specific comments below).

Specific Comments

L29 – Authors notes the number of sea level budget (SLB) closure studies that have emerged in recent years but have provided no references to these. I realise they are stated later in the paragraph, but they should also be stated here as well.

L33 – 39 The author lists many applications of SLB closure but gives no examples in the literature for any of these applications – I would expect there to be some references in the sentence to support this point.

L46 – 54 This is a very long sentence and quite difficult to read, I would rephrase along the lines of:

 “While the global mean sea-level rise is now well explained by global mean ocean thermal expansion plus mass contribution from ice  sheet and glacier mass loss, at regional scale, sea-level trends depart from the global mean as a result of regional changes in steric sea-level (regional changes in ocean temperature and salinity), ocean circulation-related mass redistribution, and regional sea-level pat terns associated with the Glacial Isostatic Adjustment (GIA) as well as present-day land  ice melt and terrestrial water storage change. The latter are referred to as sea-level fingerprints (or barystatic-GRD fingerprints), referring to relative sea-level change induced by changes in Earth Gravity, Earth Rotation and visco-elastic solid Earth Deformation ”

L55-56 – “GRD fingerprints may become detectable” Do the studies referenced to this statement give a time estimate on when they may become detectable?

L78 – 79 The nomenclature for the Sea level Fingerprints should be the other way round. The Gregory et al [2019]paper attempts to standardise the terminology around this subject (of which in the past has caused confusion in the community with interchangeable terminology) and settles on ‘Barystatic-GRD fingerprints’ to refer to the sea level fingerprints this study is investigating. Therefore, this term should be used throughout the text..

L93 “Using as input ice melt” – this is poorly worded and implies that all ice melting across the globe, which is not the case. Sections of East Antarctica are gaining mass for example.

L101 -102 These equations do not read well in line with the text and should be separated and numbered as per equation (1)

L102 – Standard SI units should be used for gravitational acceleration (m s ^-2)

L116 – You mention the CSR and JPL fingerprints are similar – but how similar and what are the differences between them? The Figure in SM1 does not make for an easy comparison and would suggest a difference plot between the two products so it’s much easier to see discrepancies.

L124 – “proximity of the melting bodies” Sounds very colloquial, would be better to reword along the lines of “in the proximity of large ice mass losses”

L165 – What is the impact the choice of GIA model would have on your results?

L170 – ‘spatially interpolated’ What method was used to interpolate these diverse products to a common grid?

L181-182 – You are correcting the altimetry with a different GIA model to that of the GRACE ocean mass. Why is this the case? What are the differences between these models? If a comparison is to be made between products, then they should all be subject to the same common corrections?

L178 -182 – In my opinion this paragraph would be better suited in the discussion of the altimetry data being used in section 3.1. I understand it is separate as it relates to processing that the author has undertaken, but it reads quite disjointed.

L199 – ‘standard uncertainty’ not needed, 1.65 sigma would be suitable

L204 – 206 Is there any reason to why the range of trend estimates from different steric products is chosen to parameterise the uncertainty? Why not simply use the standard deviation spread between products?

L273-277 – It would be much clearer to have the basins as a plot in the supplementary material – it would also enable you to reduce the text in this region.

L297-298 - Unless this technique is particular to MATLAB, I don’t think it’s necessary to state the function used to conduct this analysis.

L342-343 I’m not sure that there is no circular reasoning in the approach. Even though only ocean mascons are being used in the comparison, the fact that the same mascon’s are being used for the land mass loading change means there would be an inherent relationship between two given conservation of mass. Plus, even with the CRI filter applied, there is likely to be a Gibbs phenomenon effect that will propagate into the ocean domain as a response to large changes in terrestrial mass. An example of this effect (whilst using GOCE instead of GRACE) can be seen in Fig 4 of Hughes and Bingham [2008]. Does the author know what the effect these phenomena may have on their comparisons?

L365 – “About 0.7” – I think a more precise number needs to be given here.

Figures

Fig 1 – Figure caption should include what type of trend this is  – linear, periodic etc…

Fig 2 – The figure is in the wrong rotation and therefore makes it hard to read, this needs rectifying.

Fig 3 – This figure is very unclear to the reader. The choice of a diverging colour bar is not appropriate for stating uncertainty values on the trends – as they are not naturally diverging around a 0 value. For example, this plot indicates divergence from the value 1.65 and accentuates the increase in errors in the coast. This figure needs to be changed with a more appropriate colour bar.

Fig 4 – Same issues as figure 3, there is a diverging colour bar used to represent a continuous uncertainty value that increases from 0. This makes it very difficult to discern what the plot is representing. The colour bar needs to be changed.

Fig SM2 – This figure caption is confusing in that it is stating two different significance levels, these need clarifying to the reader.

References Used in this Review

Gregory, J. M. et al. (2019), Concepts and Terminology for Sea Level: Mean, Variability and Change, Both Local and Global.

Hughes, C. W., and R. J. Bingham (2008), An oceanographer’s guide to GOCE and the geoid, Ocean Sci., 4(1), 15–29, doi:10.5194/osd-3-1543-2006.

 

 

 

Comments for author File: Comments.docx

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 4 Report

The manuscript ‘sea-level fingerprints due to present-day water mass redistribution in observed sea-level data’ presented a study on the spatial trend pattern of sea-level fingerprints. The altimetry-based sea-level data were used. As the sea-level rise is accelerating, the topic of this manuscript is very interesting. As regional sea-level rise rates usually different from the global mean sea-level rise rate, it is important to reveal the cause of the regional variations in the rates. Previous studies have reported several factors, for example, the earth quake, the land subsidence due to underground water extracting, large-scale ocean circulation shifting, etc. can cause the local sea level rises in a different rate from the other regions. In this manuscript, the sea-level fingerprints were examined. What’s the implication of the fingerprints to the local sea-level rise rate? For example, in the Pacific Ocean. Since the sea-level fingerprints show large spatial variations, it might be useful by adding more explanation on the meaning of such sea-level fingerprint pattern, and what’s the contribution of such study on understanding the accelerating sea-level rising rate.    

Figure 2 needs to be modified.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 3 Report

See attached document

Comments for author File: Comments.docx

Author Response

Please see the attachment.

Author Response File: Author Response.pdf