Role of wind, mesoscale dynamics and coastal circulation in the interannual variability of South Vietnam Upwelling, South China Sea. Answers from a high resolution ocean model

. The South Vietnam Upwelling (SVU) develops in the South China Sea off the Vietnamese coast under the influence of southwest monsoon winds. A very high resolution configuration (1 km at the coast) of the SYMPHONIE model was developed over the western coastal region of the South China Sea. A simulation was performed over the period 2009-2018 to study the functioning, variability and influence of oceanic circulation and hydrology in the coastal region, in 25 particular of the SVU. The realism of the simulation in terms of representation of ocean dynamics and water masses, from daily to interannual and coastal to regional scales, is assessed here in detail by comparison with available satellite data and 4 sets of in-situ observations. The interannual variability of the SVU is examined for its 4 main development areas: the southern (SCU) and northern (NCU) coasts, the offshore area (OFU), and the Sunda Shelf area off the Mekong Delta (MKU). For the SCU and OFU, our results confirm the driving role of the summer mean wind and the summer circulation 30 over the offshore area in the interannual variability of the upwelling intensity. They moreover reveal the impact of the spatial and temporal organization of mesoscale ocean structures and high frequency atmospheric forcing. For the NCU, the upwelling interannual variability does not seem to be related to regional scale forcing and dynamics, but is mainly determined by coastal mesoscale structures and circulation : similar summer wind conditions can be associated with very contrasting NCU intensities, and vice versa, depending on the circulation in the NCU area. Finally, our study reveals that 35 upwelling also develops off the Mekong Delta, with an interannual variability mostly determined by the summer wind and the wind-driven circulation over the SVU region. coastal and offshore upwellings and the intensity of the summer ocean circulation (eastward jet, AC/C dipole and 80 associated northward/southward boundary currents) induced by the summer monsoon wind. These studies moreover showed that upwelling sometimes develops along the northern part of the coast from 11°N to 15°N. Da et al. (2019) also revealed and quantified the important contribution of ocean intrinsic variability (OIV) in the SVU interannual variability in the offshore area, mainly related to the role of vorticity associated with mesoscale structures of strongly chaotic nature. Chen et al. (2012) moreover showed that tides and river plumes could be additional mechanisms involved in SVU 85 dynamics. for BoxOF (0.77) than for BoxSC (0.85). This shows that other factors modulate the interannual variability of OFU, even than SCU. also be due to the chaotic part of ocean dynamics, which were shown to modulate the wind-induced variability of upwelling over BoxSC, BoxOF and BoxNC. Our results above suggest that this OIV could be stronger for 535 BoxNC, whose interannual variability more influenced ocean by explaining the larger differences between simulated and observed NCU time series. intensity and position of cyclonic eddies north of the eastward jet also modulates OFU intensity. Our simulation also confirmed that upwelling can develop along the northern part of the Vietnamese coast, as shown for the first time by Da et al. (2019) from AVHRR SST satellite data and from a 1/12° simulation for summer 1998. The high 575 resolution of our simulation (~1 km over BoxNC) allowed to better understand which factors trigger the interannual variability of NCU. This variability is not driven by the same processes as SCU and OFU, i.e. intensity of summer wind and summer circulation, but can rather be explained by the interactions between wind and small mesoscale structures and circulation in the coastal area. In agreement with observations of Ngo and Hsin (2021), who concluded that conditions favorable to SCU and OFU were unfavorable to NCU, and vice versa, we first identified 2 situations. They correspond to 580 cases where strong (weak) summer wind and offshore circulation favorable (unfavorable) to OFU and SCU strengthen (weaken) alongshore currents and weaken (strengthen) offshore currents along the northern part of the coast, thus offshore advection of cold water that could result from Ekman transport, resulting in weak (strong) NCU. Second, we also identified situations where wind conditions favorable to SCU and OFU are also favorable to NCU, due to the development of multiple dipoles and eastward jets ; and the opposite (unfavorable for the 3 areas), due to weak wind and to an 585 alongshore northward current all along the coast. Similar summer wind conditions can therefore result in very contrasted NCU intensity, and conversely opposite wind conditions can result in similar NCU intensity. Last, our study revealed that upwelling also develops offshore the Mekong delta, along the western flank of the northeastward current. As for SCU and OFU, the interannual variability of MKU intensity is mostly driven by the 590 interannual variability of the wind-induced summer circulation in the offshore region. However, contrary to what is


Introduction
The scope of the study needs a more clear definition (ln.94-100). This research didn't start from zero. The role of the background current variability in the interannual variability of upwelling was already highlighted by Da et al. (2019). What was discovered before and what is focused in particular in the present study should be better introduced. This concerns the "processes", which were not clearly defined, and "scales" which are targeted.
Ln 99: The text should be reworded with respect to my previous comment. "The objective is … scientific" should be removed. Perhaps a clear definition of the numerical tool should be provided here. If it is different from the numerical model, it should be specified.

Section 2
The numerical model is briefly presented in this section. I think some terms require clarification. The first concerns "the biharmonic viscosity of momentum" and the second concerns "nudging". I assume the authors mean how the tidal motions were prescribed at the open boundaries. But the word tidal is missed in the text.
Ln 126: the authors use the term "zoom on the VN coast". I don't have impression that the technique of zoom was implemented in the model. This needs clarification.
Presentation of the data sources. Sometimes, the information provided is absolutely useless: for example the program code, the name of PI. On the contrary, some acronyms need clear definition such as IO.
The term "hydrological characteristics of water masses" is misleading. The temperature and salinity are used for water masse characterization. What role the hydrology plays (the freshwater input, as I understand the term) in modifying T,S characteristics is unclear.
My major concern is about the definition of the upwelling indicators and the choice of the reference box which area is much smaller than that of the corresponding upwelling box. The authors should justify their choice of the reference box size and the temperature values. When the size is small and the reference location is close to the upwelling region, the reference temperature is obviously dependent on the temperature observed during the upwelling event. To what degree this quantity is independent? This needs clarification as the results could be sensitive to the choice of the reference value. What could be the difference if an overall mean temperature (space and time mean) is used as the reference value?

Section 3
As I indicated above, the authors furnished an effort in analyses of the data from different sources and in validation of the modeling results. The results are convincing. However only spatial distribution of different quantities at the surface is used in comparison. But the model is three-dimensional and high resolution. A demonstration of the model capability in reconstructing the upwelling circulation (and related water properties) in the vertical plan can be an added value. This can support the choice of high-resolution in the horizontal and also in the vertical.
The authors often use the term "coastal scale". The word coastal is not appropriate for scale definition. The scale needs clarification.

Section 4
In this section the authors explore in detail the upwelling variability for each year. Nine years in total are considered with strong and weak upwelling events. The effect of the mean wind in interannual variability of upwelling was verified but this is not a novel result. Mechanisms which can explain large difference in upwelling intensity between years with a similar mean wind are identified.
I have two major points of criticism regarding this part of the study.
The first point concerns the interpretation of the variability of the wind stress at high time resolution which is presented as the "other factor modulating the wind induced interannual variability of SCU" (ln 354). Fig. 9,10 is not sufficient. The authors use CV, the coefficient of variation. But how it helps in quantification of the contribution of high frequency compared to low frequency variability? This needs clarification.

The idea behind seems clear, but what is less clear how to quantify and interpret the effect of highfrequency variability. A method or a metric should be used in this demonstration. A visual inspection of the wind stress curves given in
I see a small inconsistency in the interpretation of the "other factor modulating the wind induced variability". First, in this particular case, the high and low frequency variability of upwelling is windinduced. The physical process involved in upwelling generation is the same. Second, I cannot imagine how the high frequency can modulate the low frequency signal. Different averaging techniques can provide different values of the mean quantity. But this is not sought as modulation. The authors should find a better formulation.
The second point of criticism concerns the role of the wind stress and surface currents in upwelling variability in different years. From my point of you, the surface current velocity and the current velocity curl are tightly related to the wind stress curl (example of anti-symmetric eddies and the eastern current). I have impression that only the wind stress magnitude is used in analysis. The added value of the study will increase if analysis of the wind curl and perhaps the wind stress vector field can be introduced and comparison with current velocities be made. This will help in interpretation of the surface current variability. What part of the current variability is wind-induced? and what part is remotely induced? The choice of the method of quantification is an important issue. And this is related to the statement (used for the second time) "another factor of non-wind origin" controlling the variability of upwelling. A part of the background current variability, independent of the wind, should be clearly identified and characterized. This requires a method of identification. I didn't see a clear description of such a method in the manuscript.

Discussion and conclusions
Section 4.5 and section 5 (Conclusion) should be reorganized. Section 4.5 contains the discussion of the modeling results and should be entitled "Discussion". A part of section 5 also contains the discussion and should be relegated to "Discussion".
Conclusion section should contain the major and novel results in a condense form. The comparison with previous studies is already done (in principal) in Discussion section. The form is important. Please avoid sentences of five lines difficult to follow. Highlight what knew knowledge the study brought and in what it is different compared to the results of previous studies.

Technical corrections
Ln 30: move "driving" to different location … our results confirm the role of the … in driving the interannual … Ln 34: perhaps "structures of circulation" ? I would recommend replacing the word ability, when you talk about the model skill, by capability, in the whole text.
Ln. 103-105: When describing the paper structure, it is better to use the word Section, not Part.
Ln 45-46: Please reword the text concerning the CSS contribution.