General comments:

The authors used altimetric data and an ocean reanalysis product to evaluate changes in the path and destabilization point of the Gulf Stream between 1993 and 2021. They demonstrated that the destabilization point has moved by 1400 km in longitude and 300 km in latitude over this period, in a process that is likely linked to the North Atlantic Oscillation variability.

Most of the results shown in the manuscript have already been shown by previously published papers, like the westward shift in the GS destabilization point (Andres, 2016), the relationship between NAO and GS path (Joyce et al, 2000), the changes in EKE (Chi et al, 2021) and seasonal meridional position (Tracey & Watts, 1986; Sato & Rossby, 1995). What seems novel is the recent eastward migration of the destabilization point following the westward migration described by Andres (2016), as well as the meridional shifts in the destabilization point and their seasonality. These points, however, are not highlighted as the main contributions of this manuscript (neither in the abstract nor during the results and discussion).

In addition, much of the discussion relies on points raised by Andres (2016) and the authors leave out of the discussion references that are extremely relevant to this topic, like Lilibridge & Mariano (2013) and Rossby et al (2014), which are only referenced for the sake of the definition of the GS path, or Dong et al (2019), which is only referenced to corroborate results. Other manuscripts like Renault et al (2016) and Seidov et al (2019) that could also contribute to the discussion are not referenced.

Overall, I think that the manuscript has a few interesting points, but most of its results and discussion rely heavily on previously published work. It could also benefit if underexplored points were highlighted in the results and discussion. For example, the conclusions shown in lines 247-251 seem relevant, but it was not clear to me how they were made, since the previous paragraph was confusing and lacked statistics and graphic visualizations. Although I do not think that this manuscript is a relevant contribution to the scientific understanding of the Gulf Stream understanding at this stage, I believe that it can make an impact if (1) the results that actually represent novel knowledge are highlighted (including visual and statistical representation) and (2) the discussion significantly improves by connecting the results to the existing knowledge rather than simply corroborating the results with previously published manuscripts.

Specific comments:

The most important method for GS path identification (25cm SSH contour) used in the manuscript is not referred to in the introduction. Why is that? Furthermore, the concept of GSNW, which is highly explored in the introduction, is not related to the GS path identification via altimetry. From the introduction, I would expect that the manuscript focused on the isothermal analysis of the GS path (i.e., analysis of GSNW), which is not the case. I would suggest modifying the introduction so that it more closely reflects the concepts and methods explored in the body of the manuscript.   

Lines 156-158: The description of “matching time series” must be backed up by graphical/statistical representation, similar to what was presented for the relationship between the wintertime NAO index and destabilization point lat/lon.

Lines 162 and 171: The use of “before“ and “after” is ambiguous. In line 162, it refers to time, so do you mean before and after the change in trends rather than the change in the destabilization point itself? In line 171, does it refer to time or space? If space, then it should be used “upstream and downstream” instead.

Lines 171-175: I did not understand what this paragraph describes, even compared with Fig 3b. What is the distribution of monthly mean pathways? Is it shown? What is the spatial extent of the standard deviation? How do you relate the signal in SSH to the signal in temperature? I suggest that this paragraph be re-written for clarity.

197-199: Are the meridional locations of the destabilization point really different between seasons? This is not clear from the map nor from statistics, which are not given.

Lines 210-219: There is plenty of literature reporting the influence of GS rings on the MAB, including those showing impacts on ecosystem changes and heatwaves, which could be explored in the discussion.

Lines 235-246: I do not see how this paragraph converses with the rest of the manuscript. What do you mean by “homogenous variations”? Why didn’t the authors add any figures showing these results? From the following paragraph (lines 247-251) it seems like these results may be important to the overall understanding of the manuscript, but they are not shown.

Technical corrections:

Line 11: “Shift” instead of “displace”.

Line 13: In the abstract, the authors mention the use of “outputs from a numerical model”, but what they really use is a reanalysis product.

Line 28: “Gulf Stream play” instead of “Gulf Stream to play”.

Line 57: “Northward shift” instead of “north shift”.

Revise names in citations such as Rosby (should be Rossby) in line 77 and Kurchitser (should be Curshitser) in line 114.

Line 78: What do you mean by “ADT (SSH)”? Following the pattern in the manuscript, acronyms in parentheses refer to the term immediately before, but this is not the case here.

Line 86: “Following Andres (2016)” instead of “According to Andres (2016)”.

Lines 100-101: Joyce did not use the 12°C isotherm at 450m to define the GS path in either of their cited papers.

Line 112: “75°W” instead of “75°N”?

Line 113: “Gulf Stream Extension” instead of “GSNW”?

Line 153: ‘Reverses” instead of “inverses”.

Line 153: After applying a 5-yr running mean to a time series that runs until 2021, the final year is 2019, not 2021.

This manuscript presents an analysis of Gulf Stream variability using gridded altimetry and an ocean reanalysis product. The primary focus is the location of the destabilization point first discussed by Andres (2016), and this manuscripts main result is an extension of Andres’s calculation to later years. Andres (2016) had shown that the destabilization point was moving upstream, but Sáchez-Román et al. find that the destabilization point started moving back downstream around 2015. Further, the motion of the destabilization point is highly correlated with the NAO with a lag of 1-year. These are both fairly remarkable results, even if the manuscript’s methodology is not particularly novel. A paper focused on investigating and explaining these results in greater detail would have been very interesting. Unfortunately, the present manuscript lacks focus and is not clearly written or organized. It also includes a great deal of extraneous and often unsupported material that makes it unsuitable for publication in its current form.

1. Major Comments
   
   There are long passages where results are quoted, but there are no figures or anything else to back them up. These results are practically meaningless without evidence. If the results are worth discussing, show a figure. If not, do not discuss them. The following passages should either be removed or be revised to include evidence:
   
   Lines 111–131
   
   Lines 235–251
2. The usage of the term “Gulf Stream” is inconsistent, both internally and with oceanographic nomenclature. The current following the path shown in figure 1 is properly considered the Gulf Stream, but many things are called the “Gulf Stream” in the text that are not the Gulf Stream.
   1. Lines 33–34: It is stated that the Gulf Stream “becomes the North Atlantic Current”. The North Atlantic Current and the Gulf Stream are geographically distinct. A fraction of the water that flows through the Gulf Stream eventually flows into the North Atlantic Current, but a large amount flows elsewhere as the North Atlantic Drift or as part of the Gulf Stream’s recirculation gyres. Importantly, most of the water from the Gulf Stream ends up recirculating in the North Atlantic Subtropical Gyre and does not flow into the North Atlantic Current.
   2. Line 10: “The Gulf Stream transports warm waters into the subpolar eastern North Atlantic … ” The current that transports warm water into the subpolar North Atlantic is the aforementioned North Atlantic Current. The Gulf Stream does not have a direct subpolar connection.
   3. Lines 112–113: The Gulf Stream does not become the Gulf Stream North Wall (GSNW). The GSNW is part of the Gulf Stream, so the Gulf Stream cannot become the GSNW.
3. The usage of Gulf Stream North Wall (GSNW) also inconsistent and occasionally incorrect. The GSNW is the strong temperature front on the northern flank of the Gulf Stream. It is not “the path described by the Gulf Stream Extension” (lines 34–35) as it is north of the main core of the Gulf Stream. As noted in Chi et al. (2019), the GSNW and the main core of the Gulf Stream do not necessarily even follow the same path.
   
   Lines 70–71: The Gulf Stream, not GSNW, converts from a stable to an unstable jet. The GSNW is a front, not a jet. Fronts are often associated with jets through thermal wind, but they are not the same thing.
4. While gridded altimetry comes on a daily 1/4º grid, it does not have a temporal resolution of one day or a horizontal resolution of 1/4º. At the latitude of the Gulf Stream, the ground tracks used to construct the gridded produce are located approximately 2º apart and sampled roughly every 10 days. The values between the tracks and sample times are “filled in” using optimal interpolation. This produces smooth-looking fields, but can also invent spurious features and give the false impression of high precision. Ballarotta et al. (2019) estimates that gridded altimetry has an effective resolution of 150–200 km (e.g., 1.5º–2º) in the Gulf Stream region. As such, reporting locations obtained from gridded altimetry with sub-degree precision (as on line 107) is not meaningful. The authors reference Ballarotta et al. (2019) in noting that gridded altimetry misses some mesoscale features, but don’t appear to acknowledge that the coarse resolution of altimetry may affect their estimates of the location of, for example, the destabilization point.
5. Lines 94–95: Without information about how the confidence interval computed there’s no way for the reader to determine if they trust it.
6. Figure 2b: The value given for 1994 in figure 2b is approximately 5º to the west of that in Andres (2016) figure 3. What is the source of this disagreement?
7. Section 4: This section is labeled “Discussion and conclusions”, but it is almost entirely more results rather than a discussion or a conclusion.
8. Section 4.1: Figure 2b shows that the destabilization points migrates by more than 10º on interannual timescales. Against this background, the ~1º seasonal shifts are not meaningful. Indeed, the confidence intervals in figure 4e are barely non-overlapping.
9. Lines 259–263: The fact that Andres (2016) found a 5 year lag between the destabilization point does not justify a 5-year running mean filter. A centered running mean has no effect on the phase of a time series, so a 5-year lag in the original time series would remain a 5-year lag in the smoothed time series.

Minor Comments

1. Lines 26–27: The authors should clarify that the AMOC accounts for 90% of the heat transport at 26.5ºN (the latitude of the RAPID array). There is no reason to expect that this holds true at other latitudes.
2. Lines 27–28: It is not clear what “subpolar planetary heat exchange” means. Is it heat exchange on a planetary scale?
3. Line 49: Not clear what “instability” means in this context. Why are lateral shifts more associated with instability than meandering?
4. Line 62–63: Joyce used the 15ºC isotherm at 200 m depth. As written, the text suggests that the 15ºC isotherm is equivalent to the 200-m temperature.
5. Line 63: Joyce’s index locates the GSNW which, as noted previously, is not necessarily colocated with the region of strong flow.
6. Figure 3: It would help to indicate the regions discussed in the text on the figure.
7. Line 173: The range over which the standard deviation decreases between the two periods is tiny—only 1º, which is below the effective resolution of gridded altimetry. What is the reader supposed to take away from this result?
8. Lines 210–211: Should “low-frequency remarkable shift” be “remarkable low-frequency shift”?
9. Lines 211–203: This is a very confusing sentence. What is widespread along a larger fraction of the North Atlantic? Larger fraction than what?
10. Lines 224–225: How do we see from the results presented here that the shift of the destabilization point was accompanied by a weakening of the jet?
11. Lines 230–231: How do changes in EKE show that surface mesoscale diffusivities are largely influenced by climate variability? 
12. The final paragraph of the manuscript (Lines 270–279) does not seem to follow from the results presented in the paper. Indeed, it mostly summarizes background material and would fit better in the introduction. It would be worth, however, ending the paper with a proper conclusion.

Technical Corrections

1. The word “isobath” is used to mean depth, but it does not. An isobath is a contour of constant distance between the surface and the bottom. It is equivalent to a topographic contour. Replace with “depth” on lines 62 and 100.
2. Replace “inverses” with “reverses” on line 153.
3. Replace “inverses” with “reversal” on line 216.

References

Ballarotta, M., and Coauthors, 2019: On the resolutions of ocean altimetry maps. Ocean Sci., 15 (4), 1091–1109.

Chi, L., C. L. P. Wolfe, and S. Hameed, 2019: The distinction between the Gulf Stream and its North Wall. Geophys. Res. Lett., 46 (15), 8943–8951.