The work under review is a relatively straightforward regional sea level budget analysis based on the Scripps Argo product for the steric component, several gravity-derived and reanalysis-based manometric sea level  estimates, and the C3S altimeter dataset for sea level. I have a considerable number of issues that I wish the authors can address in a revision of the current draft. Major and minor comments are listed below, with reference to line numbers where relevant.

Major comments 

The paper deals with linear trends in many of the figures but there is no discussion of whether the trends are statistically significant (i.e., statistically different from zero at some confidence level). This seems to be a major shortcoming of the presentation and can be simply corrected by stippling or not plotting insignificant values in figures 1—7. The discussion should then be focused on statistically significant non-zero trends. 

In addition, all budget analyses need to carry at least a qualitative discussion of possible uncertainties in all the terms, in this case sea level, steric and manometric components derived from the different products. The authors use an ensemble of manometric estimates but they fail to provide a spread or some quantitative measure of uncertainty for their results. And some features in their results are treated lightly or ignored. For example, there is a clear issue with the major Tohuku and Sumatra earthquakes affecting the trends in the mascon solutions, but that is never mentioned. 

Last but not least, I find the discussion of many figures loosely presented and at places not really jiving with what I could reasonably infer from the figures themselves. Here are several examples of this issue:

325-326/ I would disagree. For example, halosteric term is larger than thermosteric term in parts of the South Indian Ocean, northeast North Pacific, and manometric term is larger than thermosteric in parts of the Southern Ocean. 

333-335/ I think the claimed similarity between manometric and thermosteric terms is hardly justified, and actually I can’t think of any reason why we should expect them to be similar.

336-339/ The text mentions strong residuals, but contrary to what it is claimed, in the eastern Atlantic there are relatively strong negative thermosteric trends, not fully compensated by positive halosteric trends,  which seem to give rise to the large residuals (sea level trends are fairly weak);

419-420/ Again, I am having a hard time seeing this claim in the figure. Actually, I see opposite signs in spatial trends of the ensemble mean reanalyses and GRACE in the North Atlantic as well.

If qualitative claims are going to be made, the authors should make sure they are clearly evident in the figures. In addition, quantitative analyses such as determining pattern correlation coefficients should be attempted when deriving inference about similarity between spatial trend patterns.

Minor comments 

51-52/  Note that “redistribution” can also lead to changes in steric component (you can redistribute density, not just mass).

66-60/ This sentence (or perhaps actually the full paragraph) almost equates global mean sea level budget with the issue of global mean sea level rise, but of course budget analyses can and should apply to all time scales, not just the linear trend, for both global mean and regional cases. A recent example of regional budgets for the seasonal cycle is found in https://doi.org/10.1029/2024EA003978.

101/ Delete “already in the ocean”.

141-143/ Sentence could be improved for clarity.

157-158/ Statement is not strictly true, as changes in global mean pressure can come from changes in the mean water vapor content of the atmosphere and those can affect the barystatic term. Likely a small effect, however!

168/ “GRDs” sounds weird! “GRD effects” or similar would read better, here and elsewhere in the text.

171-172/ I think this statement pertains to “absolute” sea level. Please clarify the text. Similar issues may apply to other parts of the text.

175/ Delete “sea level”; also “northeast coast of North America”.

205/ “which leads to”

179/ “based on observations”

213-214/ I think you are referring to the drift in the global mean, but statement needs to be clarified. 

217-218/ We need some estimate of these deep steric changes at the local scales of interest to this paper.

246/ “two filtering levels”? Please clarify this text.

254-255/ The GRACE and GRACE-FO records have many gaps. Please clarify in the text how those are handled in the analyses.

263-264/ Unclear sentence.

283-285/ Can the authors discuss, at least qualitatively, what sort of errors this may imply in the derived manometric trends?

287-293/ I am confused by the treatment of atmospheric loading corrections for GRACE. Of course, GRACE does not “see” atmospheric loading effects, if one has an inverse barometer behavior.  Effects of atmospheric loading would only be apparent in the global mean bottom pressure, but those are apparently removed in the present analyses. I think this needs to be clarified, to make sure corrections are appropriately applied.

329-333/ You start by calling out large residuals in the North Atlantic but those are in the eastern part of the basin. You go on to discuss the western part and south of Greenland, where residuals actually seem relatively small. This is somewhat confusing.

340-345/ It would be useful to include an extra panel with residuals calculated on the basis of steric trends only. Do they look better? Actually including such panel in figure 2 might justify that figure better. Otherwise, figure 2 is not needed, as those same two panels can be readily examined in figure 1.

374/ “everywhere” is an overstatement given results in figure 3c,d.

375/ Very unclear what inferences are being made in this text, given the previous discussion of results in figure 3. Please rewrite for clarity.

399/ Delete either “ ’s” or “the” before Camargo.

434-435/ “CIGAR also does not assimilate”

447/ “If we exclude FOAM”

459/ “assimilation in Figure 4 (e.g.,”

458-463/ You reference figure 6 but the text seems to be comparing CIGAR and C-GLORS results in figure 5?

461/ The largest differences are actually west of the Drake Passage.

463-465/ I don’t follow the corollary.

475/ Actually all reanalyses, not just CIGAR show positive residuals.

494/ “stripe”

495-497/ Very unclear what this sentence means. Please rewrite.

Figure 9/ This figure is hardly justified. It does not bring anything new to what is already discernible from Figure 8a. I don’t think the global EOF adds any relevant information to the discussion.

511-513/ I think the residual for “all but North Atlantic” case is also significant?

514/ The halosteric decrease is evident after 2016, not 2013-2014?

Review for: “Regional sea level budget over 2004-2022” by Marie Bouih, Anne Barnoud, Chunxue Yang, Andrea Storto, Alejandro Blazquez, William Llovel, Robin Fraudeau and Anny Cazenave  (https://doi.org/10.5194/egusphere-2024-3945).

The authors investigate the regional closure of 20-year sea level trend budget based on altimetry (total), ARGO (steric), GRACE/Reanalysis (manometric) sea level datasets and GRD fingerprints. The paper focuses on the differences between various manometric datasets. It examines the influence of the GRACE processing on the regional sea level budget and emphasises the need to apply the geocenter correction. The largest regional trend discrepancies occur in the North Atlantic, and the authors suggest that a spurious drift in the salinity measurements may be responsible.

The paper deals with an relevant topic and is well written. The introduction provides a good overview of the topic and cites the relevant literature. Equations could help to make it easier to find out which components were taken into account for which data sets to calculate the residual trends. The figures are clear and informative, however, some of them seem to be in the paper twice. If possible, the datasets and the corresponding versions should be cited unambiguously (e.g. doi for altimetry).

Unfortunatelly, none of the manometric components studies provides a sufficient closure of the regional trend budget.  Possible causes for the differences between the datasets should be discussed in more detail.  A discussion of the uncertainties of the trend estimates is missing for all data sets. With regard to the North Atlantic, the explanation for not closing the trend budget should be better justified. The figures suggest rather decadal variability between the gyres than consistent long-term drifts in the entire area.

Specific comments:

Lines 78-92:

The usage of the expressions regional, basin-scale, sub-basin scale and local scale is confusing and could even be inconsistent

Lines 141-143: Shouldn’t the barystatic component distribute according to the GRD fingerprints?

Line 180: Could you provide numbers for the small regional GIA & contemporary GRD trends?

Line 190-200: Please specify the data version (doi?). Some of the given information is abundant since it is not used (Topex side A drift, total sea level uncertainties)

Lines 217: Could you provide numbers for the regional deep steric contributions?

Lines 265-: Could you specify the differences between the  manometric data from ocean reanalysis? What is the uncertainty and why did you choose these models?

Line 271: Are all reanalyses based on ARGO data? Is it known how the salinity drift is handled in the individual reanalyses?

Lines 304-305: Figure 1 suggests that the spatial filtering of altimetry and mascon datasets is not consistent.

Lines 306-308: Are there systematic differences between the trends for these two periods?

Figure 2: Is there a difference to figure 1b and 1e?

Line 340-345, Couldn’t there be problems with the degree 2 terms of the GRACE-processing as well?

Lines 370-390: I would suspect that the ARGO-data, as well as the ocean reanalyis, are referenced to the centre of figure.  Even though the altimeter orbits were calculated relative to the centre of mass they may have been transferred  to the centre of figure somewhere on their way to the level 4  Copernicus sea level grids.

Figure 4: Trends differ quite a lot. What is the uncertainty of individual  trend estimates?

Line 458: You decide to focus on the CIGAR  model, but the results for CIGAR are not included in Table 2.

Lines 511-515: The halosteric component seems to be on the decline after 2015/2016 and to be anticorrelated with the thermosteric component. You might want to consider the budget for the subtropical and the subpolar gyre separately.

Why should the effects of a spurious drift in salinity measurements only be effective in the North Atlantic?

Figure 9:  If you suspect that the PC1/EOF1 of the halosteric signal is dominated by the North Atlantic signal, why do you perform a global EOF-analysis?