Using Landsat Image Series to Identify and Characterize Persistent Oceanographic Structures in a Dynamic Marine Protected Area (North of San Jorge Gulf, Argentinian Patagonia)

Round 1

Reviewer 1 Report

Please find attached.

Comments for author File: Comments.pdf

Author Response

Response to reviewer #1

This manuscript utilizes Landsat-8 imagery, with a resolution of 30 meters, to identify persistent submesoscale structures in the northern San Jorge Gulf from 2015 to 2021. After identification, authors characterized and classified the structures and then studied tide current, wind and seasonal stratification in the area.

 

Q1. However, while the manuscript contains an abundance of technical information, the authors have not provided sufficient detail regarding their methods. For instance, while they used tidal current data from the TPXO model, it remains unclear how this data was used or implemented. Furthermore, the authors stated that they masked land and cloud, but they did not describe the methodology employed in this process.

R1: The “Data and Methodology” section has been completed with details on the characteristics of the tidal current and wind variables, including some references where they are described in detail. In addition, details of the procedure followed to perform the sunglint correction and the land and cloud masks have been added. Finally, at the end of the “Data and Methodology” section, a paragraph has been added explaining the type of treatment given to the forcing variables together with the images. We hope that these changes have clarified the section.

 

Q2. Additionally, Figure 2 appears to lack significance, as there is no comparison with other measurements.

R2: We have kept this figure because we believe it plays an important role in the work: it is a small catalog that shows, in a single image, the diversity of structures in the study area (showing its high dynamism) and, on the other hand, it is the reference image for the entire subsection 3.1, avoiding the reader to go through images 3 to 6 in order to follow this subsection.

 

Q3. In Figures 3 through 6, each image has its own velocity of tidal current. However, it is unclear where the location of these tidal current velocities is. Moreover, if the authors are examining tidal current images, they should consider comparing their images with data from the TPXO model.

R3: The reviewer is right, the location of the tidal current nodes was only marked in Figure 1, but it is much clearer to add it in each of the figures. In the current version of the MS this information is included in all figures where zoom makes it possible. Regarding images showing the tidal current (vector) field, or the spatial mean in a given area (as has been done in other works), they are not shown or compared with because they were not used. In our case, the time data series used corresponded to 4 atlas nodes inside the study area, chosen to be close enough to the structures analyzed in detail and far away enough from the emerged areas so as not to be affected by them.

 

Q4. Lastly, it is not clear how the authors measured wind speed in the area. Further explanation is needed to clarify this aspect of their methodology.

R4: Thank you for your comment. The wind was not measured but obtained from the ERA-5 Reanalysis, as mentioned in the “Data sources” subsection. This point has been clarified as explained in R1.

 

Reviewer 2 Report

The author analyzed 80 images spanning the years 2017-2021 and manually identified structures in the study area. Related variables including tides, bathymetry, wind, and seasonal temperature stratification were considered as potential forcing variables. The relationship between the identified structures and the investigated variables was discussed. However, the stated relationship is supported by weak methods and insufficient evidence.

In the method section 3.2.3, it is not clear how the forcing parameters are related to the identified structures, i.e. how these relationships are established and identified, and if such relationships can be quantified and compared. In the results section, there is no strong evidence how this relationship quantitatively varies among the investigated variables, and why the tide is identified as the most significant driving force. Is the relationship a strong correlation in occurrence time? If so, how do you determine the causality from the correlation relationship? How do you identify the causal relationship from a single image but not a series of images in a time span (such as a few hours)?

The effect of bathymetry is not investigated and discussed in detail.

Author Response

Response to reviewer #2

Question 0. The author analyzed 80 images spanning the years 2017-2021 and manually identified structures in the study area. Related variables including tides, bathymetry, wind, and seasonal temperature stratification were considered as potential forcing variables. The relationship between the identified structures and the investigated variables was discussed. However, the stated relationship is supported by weak methods and insufficient evidence.

Response 0: The reviewer is right that there is no statistical analysis behind the results shown in the MS. In our case the work was approached with a descriptive approach, reviewing for each of the images the characteristics of the observed structures and the value of the variables related to them. This approach, although it is more tedious and has the bias of the researchers observing the structures, allows us to extract a large amount of information and, in our case, also allows us to observe consistent patterns over time and the study area. On the other hand, although no particular statistical test was performed, or any model was adjusted (usual approaches when working, for example, with ocean color sensor images, where the approach is pixel by pixel), the results of the work show tables (1-4) that present the number of images (or % of them) corresponding to each of the situations described, which in itself represents a descriptive analysis.

 

Q1. In the method section 3.2.3, it is not clear how the forcing parameters are related to the identified structures, i.e. how these relationships are established and identified, and if such relationships can be quantified and compared.

R1: The “Data and Methodology” section has been completed with details on the characteristics of the tidal current and wind variables, and has references where they are described in detail. In addition, details of the procedure followed to perform the sunglint correction and the ground and cloud masks have been added. Finally, at the end of the Data and Methodology section, a paragraph has been added explaining the type of treatment given to the forcing variables together with the images. We hope that these changes have clarified the section.

 

Q2. In the results section, there is no strong evidence how this relationship quantitatively varies among the investigated variables, and why the tide is identified as the most significant driving force. Is the relationship a strong correlation in occurrence time? If so, how do you determine the causality from the correlation relationship? How do you identify the causal relationship from a single image but not a series of images in a time span (such as a few hours)?

R2: As we mentioned above, a paragraph has been added in the Material and methods section of the document explaining in more detail the methodology followed by the authors. In this case, no statistical tests have been performed that would allow, on the basis of these tests, to define the relationships as significantly strong. However, as shown in Tables 1-4, the persistence percentages of the described patterns show very few exceptions to them, so this correlation between the tidal current and the described meso- and sub-mesoscale structures has been considered strong (as expressed in the results). There is no in situ data included in this work that would allow, without a doubt, to define the causality of this relationship. However, the high persistence of the observed patterns, as well as the review of other usually important variables (such as bathymetry or winds) allow us to venture this causality. For this not to be true, there should be a confounding variable (not reported so far in the area or in similar works in other areas) highly correlated with the tidal current that we are not considering. On the other hand, other work conducted in the area and cited in this MS reached very similar conclusions using in situ data (Pisoni et al., 2023, under revision). In addition, an experiment still in progress (unpublished) using surface drifters shows in the vicinity of Pan de Azucar Island, the influence of the tidal current in the formation of small vortices southwest of the island (see figures below).

On the other hand, there are multiple papers in the literature that deal with this subject in regions with large tidal amplitudes, such as our region (see for example: Wolanski et al. 1984; Pattiaratchi et al. 1987) and associate the detected structures to the interaction of currents with headlands and islands, but there are also studies that use numerical simulations of oscillatory currents and observe structures such as those described in this work (see for example: Signell & Geyer, 1991; Guo et al, 2020).

Signell, R. P., & Geyer, W. R. (1991). Transient eddy formation around headlands. Journal of Geophysical Research: Oceans, 96(C2), 2561-2575.

Guo, B., Ahmadian, R., Evans, P., & Falconer, R. A. (2020). Studying the wake of an island in a macro-tidal estuary. Water, 12(5), 1225.

The detailed inspection of the 80 images supports the hypothesis of tidal currents being the main forcing effect behind the development of these structures although, probably, wind could affect their size or shape but not originate structures on top of the dynamics ruled by tides. Otherwise, structure locations would not be found as consistently as described in this work with the regularity of tidal currents.

 

Q3. The effect of bathymetry is not investigated and discussed in detail.

R3: As in the case of wind, bathymetry was evaluated in those cases where the tidal current (main forcing of all observed structures) did not fully explain the observed patterns, as in the case of Tova Island (see Fig. 6 and Table 4). Of course, the reviewer is right, that a more quantitative analysis could be addressed in this case as well, but this would require fitting a model by performing bathymetry sensitivity simulations, of great interest but far from the objectives of this MS.

Round 2

Reviewer 2 Report

The comments have been well addressed.