Review of Estimating oceanic vertical velocities in a wind-influenced coastal environment

Summary:

The manuscript presents new observations of vertical velocities using a novel ADCP instrument (JULIO; Judicious Location for Intrusion Observation), deployed on the 100 m isobath offshore Marseille in the Gulf of Lion. The dataset covers two periods, 2012–2015 and 2020–2023, with a temporal resolution of 30 minutes. A limited intercomparison with other instruments measuring vertical velocities over a few hours demonstrates that all three instruments capture the same order of magnitude but different variability.

The paper focus on the assessment of upwelling and downwelling events, relating them to the horizontal flow and regional wind forcing, as well as the temperature gradient and SLA, which may be modified by vertical velocity events.

Before this, the authors examine the distribution of vertical velocities across the study period, highlighting the dominance of downwelling across the years. They identify most of these negative velocities at night, suggesting a biological origin, and propose a filtering method to remove this contribution and retain only those velocities of physical origin.

After applying this filter, the authors analyse two upwelling events (2012 and 2022) and two downwelling events (2014 and 2021). The upwelling events, both lasting two days, were associated with wind speeds larger than 15 m/s, with a net vertical flow of 138 and 86 m/day, respectively. The authors demonstrate that the two upwelling events are related with rapid decreases in temperature, with gradients exceeding the temperature changes experienced in the region over several months. The downwelling events, lasting 12 hours, are also related to intense winds exceeding 15 m/s, and appear to occur at a peak in SLA. The authors obtain a negative vertical flow of -465 m/day (2014) and -346 m/day (2021).

General comments:

• Presentation of a time series of vertical velocities measured by a novel ADCP instrument (JULIO) in the Gulf of Lion during the periods 2012-2015 and 2020-2023.
• Development and application of a methodology to remove biologically induced vertical flow.
• Clear methodological explanation of the filtering procedure, with figures that illustrate the step-by-step methodology.
• Clear and useful figures showing the results presented in the draft and the conclusions drawn from them
• Insightful effort made to achieve a good representation of the distribution of vertical velocities measured in the water column analysed as a function of time and its relationship with other variables such as horizontal circulation in the region and the consequences on other physical variables such as temperature and SLA.

Specific comments:

1. The manuscript repeatedly states that JULIO has measured vertical velocities “since 2012”. This phrasing is misleading because the dataset is discontinuous, with a gap between 2015 and 2020 (see Table 1 and Figure 10). This limitation should be explicitly acknowledged in the abstract and the rest of the manuscript, and do not consider the analysis of more than a decade of data.
2. The comparison between JULIO ADCP, FF-ADCP and VVP is presented as a central result, but the conclusions are too general and not the core objective of this paper. The conditions for the comparison should be highlighted and clarified, including the choice of periods used for JULIO ADCP and FF-ADCP of around 2 hours while VVP data is limited to17 minutes. Could more detailed insights be extracted from this intercomparison? For example, how each method captures upward vs. downward velocities, or how the sign of w evolves in time?
3. The manuscript does not discuss the intrinsic measurement errors of the instrument. What are the uncertainties associated with the velocity estimates (e.g. Figure 3)? Given the small magnitude of the measured vertical velocities in the intercomparison, this information is essential.
4. The manuscript is somewhat disorganised. Very specific event analyses are introduced early, without giving a proper explanation of this choice, while a broader overview of the JULIO time series only appears near the end (Figure 10). A more logical structure would be to first present the general time series (Figure 10 or even Figure S18) after Figure 3 and then focus on the specific events.
5. The annual probability distributions in Figure 3 should be interpreted with caution. According with Table 1 and Figure 10, only 2021 and 2022 provide complete annual coverage, so interannual comparison are likely dominated by sampling differences and seasonality of the vertical flow. The statement on lines 234 and 235: “boxplots highlight that 50% of the values are between -5 and 5x10^(-3) m s-1” should be framed as an order of magnitude, not a strict threshold. The manuscript should also clarify that the variability in the number of observations is due to short time series than a different sampling frequency (this is mentioned only later, on Line 316 but should be brought to the reader much earlier).
6. The criteria for selecting the upwelling and downwelling events, as well as the period used for the identification (Line 302) are nor well justified. The chosen events do not appear to be the most intense in Figure 10. For example, in U2022, why do the authors chose two days for the consideration of the upwelling? Do the authors use the wind velocity to identify these events? If so, how sensitive are the results to these choices? For U2012, the extreme wind event above 15 m s-1 is shorter than for 2022 and lasts less than 2 days. I don’t understand the choosing of the window because it starts after a 15 m s-1 measurement, but it is not contained in the considered period. Without a statistical analysis of the entire time series, the representativeness of these four events is questionable, and the conclusions should not be general.
7. The average upwelling and downwelling values reported in the abstract and the conclusions are extremely sensitive to the chosen filter parameters and averaging window. Moreover, they are based on a very small subset of events (one single event of 2 days and 12 hours respectively) relative to 5-year dataset.

Technical corrections:

The manuscript requires careful proofreading, especially regarding comma usage. Below are some specific corrections and suggestions:

• Line 11: “associated with” (not “associated to”)
• Line 12: Use consistent notation for vertical velocity, here you use lower-case w, but during the rest of the draft you refer as W.
• Line 12: “Hence,” (add comma).
• Line13: “high-frequency”
• Lines 34 and 42: “of a few” (not “of few”)
• Line 35: “allow estimating” instead of "allow to estimate”
• Line 38: “Frajka-Williams et al. (2011) were”
• Line 39: “Other 3-month” (not “Other 3-months”)
• Line 44: change “since decades” to “for decades”
• Line 50: change "electro-magnetic” for “electromagnetic”
• Line 50: Clarify whether APEX is an acronym?
• Line 61: change “ship mounted” for “ship-mounted”
• Line 64: change “inter annual” for “interannual”
• Line 84: “downwelling and upwelling events” (not “downwellings and upwellings events” )
• Lines 94 and 196: Could you maintain the same format for coordinates?
• Line 97: change from "vertical velocities measurements” to “vertical velocity measurements”
• Lines 100, 119, 270: “W was…, W is …, W was” W is referred as plural in the rest of the manuscript
• Line 102: Add comma “Generally, W have very low intensities”
• Line 112: change from “also impact” to “also impacts”
• Line 116: Introduce the JULIO acronym at the first mention, not in line 135
• Line 117: change “multiyears” to “multi-year”
• Line 117, title Table 1, 191 and 409: “time series” (not “time serie”)
• Lines 122 – 24: No space before a question mark in English.
• Line 127: Introduce the FF-ADCP acronym at the first mention
• Line 144: Which is the sampling frequency of JULIO? I think it is written in line 230, can you put it here instead?
• Line 150-157: Missing reference
• Line 160: Justify parameter choices with references.
• Line 164: Define “MIO”.
• Lines 188 to 190: Missing commas before numbers?
• Figure 2: Could the authors thicken the dashed lines in JULIO and FF-ADCP panels? In the FF-ADCP case, I think that thickening the lines would improve the readability of the different colours. Clarify meaning of lines in caption (depth units, time periods, shading). Could you add also the time period of each methodology in the caption, will be easier for the reader than to search for it in the body text.
• Line 245: Remove double parenthesis in reference: “((Heywood, 1996))”.
• Line 247: Missing space in “(Fig.6 top panel)”.
• Figure 4 caption: There is not solid lines in the figure, do you mean dotted lines?
• Line 252: remove space “value :”
• Figure 6 caption: Could you add the period of study as in Figure 4?
• Line 260: Replace “penultimate panel” with “panel c)”.
• Line 264: Could you add a reference? This affirmation is for the climatology. Is always constant during the year or it has a seasonal cycle?
• Line 266: Reference needed for the 15 m s-1 threshold
• Line 275: Use American spelling “analyzed” consistently
• Line 277: Add space in “(Fig.8)”.
• Line 284: I do not think that it is necessary to add <W_2012 >, if you do, please add it in the other cases <W_2022 >.
• Figure 7. Improve colour contrast of the lines U2012 and U2022. Could you add the years 2012 and 2022 above each column as title? It would help with reading. Could you add letters to the panels in the figure to make it easier to refer to them in the text? Adding a horizontal line at 15 m s-1 wind panels could help to define better the average region.
• Figures 7 and 9. The authors are using a sequential colormap for a variable with positive and negative values. I recommend using a diverging colormap to improve readability and the understanding of the discussion and conclusions.
• Figure 10 caption: Clarify whether horizontal lines correspond to +- 5 mm s-1?
• Figure 10: Ensure consistency between Figure 10 and Table 1 (dates for 2013–2014). Time series during 2013 and 2014 do not correspond with the dates of Table 1 (Time series 2). In the table is written 03-26-2013 as initial date, but the time series in Figure 10 show that the values start at the end of September 2013.
• Line 321: Add missing space and remove dot: 15m.s−1
• Line 324: Add missing units: 0 m s−1
• Line 325: Add missing space: 5 x 10^-3m s−1
• Line 327: remove s in “exhibits”
• Line 343: Consider citing Jacox et al. (2018) on coastal upwelling.
• Line 346: Missing r in “occurences”
• Line 360: Justify choice of a two-month time series? Consider seasonal climatology
• Line 371: Meridional component during D2014 event seems much more variable than for D2121 event. Maybe it is because of the colormap chosen does not allow for a clear view of positive and negative values as I mentioned above.
• Line 391: Remove s in “works”
• Lines 390-396: The authors should detail that they are talking about the four events analysed and only generalised to the entire period of study after demonstrating that they are representative events. Could you compute the correlation coefficient between wind velocity and vertical velocity?
• Line 401: remove s in “medians”
• Line 405: missing s in “type”

General comments

This manuscript presents a study of estimating physics-driven vertical velocities in the Northwestern Mediterranean Sea, a wind-influenced coastal environment. The authors took measurements using JULIO ADCP, FF-ADCP, and VVP, and developed a new methodology to filter out biology-induced signals. Upwelling and downwelling events associated with strong winds were then identified, and the results validated with satellite data which include SST and SLA.

Overall, this manuscript contributes to advancing our understanding of the different mechanisms of vertical velocities. The literature review, datasets, and the filtering methodology are outlined and described clearly. Nevertheless, the results can be enhanced with more quantitative analysis. More importantly, further improvements are required to enhance the overall presentation, clarity and precision of the paper. As such, I will recommend this manuscript for publication after the following comments are addressed.

Specific comments

1. Line 224: avoid using the root word “significant” in scientific writing if no statistical tests are conducted. Appropriate statistical tests should be conducted before concluding that the results of all measurement methods are not significantly different from one another. Additionally, the use of scientific notation is not needed if the value is zero (refer to 0.0 x 10^-3).
2. A sensitivity analysis of the wind speed threshold helps in validating the 15 m s^-1 choice. This is relevant since as mentioned in line 346, Berta et al. (2018) have shown that wind speeds above 10 m s^-1 are strong enough to drive upwelling in the authors’ area of study.
3. Lines 11–13: the authors in the abstract highlight how the order of magnitude of W depends on the spatiotemporal scale of its analysis and also the need for high-frequency measurements. The statements are not only unclear, but also feel random and out of place because there is hardly any discussion on how different sampling frequencies affect the final results of the study.
4. Section 3.6 and Figure 10 should be placed before Section 3.5 and Figures 7–9 for better structure and flow. Describe the general trend and features (e.g. the total number of upwelling and downwelling events identified in the entire time series) before moving on to describe the specific events in detail.
5. I also suggest adding the adjective “physics-driven” in the title to better reflect the scope of your study.
6. The organisation of the manuscript can be further polished and requires a thorough proofreading. There are many short paragraphs that are 1–2 sentences long (e.g. lines 54, 93, 339, 350, 387, 389, 404) that could instead be integrated with other relevant text to produce longer coherent paragraphs.
7. Some descriptions and ideas are repeated often, thereby reducing the brevity of the manuscript. One notable example is the Discussion section, where some paragraphs are largely a rehash of the results presented in section 3.5. To improve the structure of the paper and make it more concise, here are some key suggestions:  
   • Move lines 96–99 to section 2.2. 
   • Move section 3.6 in front of section 3.5 (as mentioned earlier). 
   • Move the bulk of the description in lines 324–333, lines 354–363, and lines 365–374 to section 3.5, and reduce the repetition.
   • Lines 337–338: sentence is phrased like a summary statement that is placed awkwardly in the middle of the Discussion section. I suggest removing it.

Technical corrections

Line 1: change to “oceanic vertical velocities (W)”.

Lines 11–12: Write “w” in caps.

Line 13: remove “spatiotemporal”.

Line 18: change to “allowed for”.

Line 22: awkward use of colon.

Line 24: the word “very” is redundant since “complete” is already an absolute adjective.

Line 27: change from “during the last decades” to “over the decades”.

Lines 35, 40, 364, 383: the phrase “allow to” is used awkwardly, please rephrase.

Lines 41, 76, 88, 234, 326, 334, 367, 373: I suggest against starting a sentence with “indeed”. It is a filler word that hardly adds or change any meaning to the overall flow.

Line 66: awkward use of colon. The part after the colon is also repetitive and redundant.

Line 72: “…, seasonal variability affecting its width, …” – awkward phrasing.

Line 121: no need for a new paragraph.

Line 144: change to “information”.

Line 149: remove the extra blank line.

Line 156: add a space between number and unit (i.e. 4 m).

Line 168: add a space between number and unit (i.e. 2 Hz).

Line 183: spell out 8 and write one as a numeral. For items other than units of time or measure, use words for cardinal numbers less than 10. Add a space between number and unit (i.e. 200 m).

Line 188: attitude or altitude?

Line 195: change to “10 m”.

Line 210: change to “resulting in”.

Line 238: change to “example”. 

Line 240: present your results in prose instead of a bullet point list.

Line 243: remove “a”.

Line 267: remove “respectively” in parentheses.

Line 268: awkward use of the word “completed”. Sentence is thus unclear.

Line 275: change from “of” to “in”. Use “of” to describe the magnitude.

Line 285: change to “ small decrease of”.

Line 295: change to “24-day long”.

Line 302: change to “variability of”.

Line 331: change from “upwelling and downwelling W” to “upwelling and downwelling events”.

Line 348: remove “a”. 

Line 365: change to “With a shorter”.

Line 383: sentence is unclear with some awkward phrasing.

Line 402: “arise the question of 2012” – unclear.

Line 403: write “w” in caps.

Line 404: change to “for the first time”.

Lines 406–407: add a space between the number and kHz.

Figure 1: intensity or velocity? Add a scale bar. Are the locations of the VVP and FF-ADCP the same as the JULIO mooring? If not, label them on the map.

Figure 2: the font size of the title is too small.

Figure 6: in the caption, the panel labels (a) to (d) should precede the description of each figure panel, not after.