Modelling the water isotopes distribution in the Mediterranean Sea using a high-resolution oceanic model (NEMO-MED12-watiso-v1.0): Evaluation of model results against in-situ observations

Ayache, Mohamed; Dutay, Jean-Claude; Mouchet, Anne; Tachikawa, Kazuyo; Risi, Camille; Ramstein, Gilles

doi:https://doi.org/10.5194/gmd-2023-237

Preprints

https://doi.org/10.5194/gmd-2023-237

Preprints

Submitted as: model description paper

10 Jan 2024

Submitted as: model description paper |

| 10 Jan 2024

Status: a revised version of this preprint is currently under review for the journal GMD.

Modelling the water isotopes distribution in the Mediterranean Sea using a high-resolution oceanic model (NEMO-MED12-watiso-v1.0): Evaluation of model results against in-situ observations

Mohamed Ayache, Jean-Claude Dutay, Anne Mouchet, Kazuyo Tachikawa, Camille Risi, and Gilles Ramstein

Abstract. Stable water isotopes (δ¹⁸O_sw and δD) have been successfully implemented for the first time in a high-resolution model of the Mediterranean Sea (NEMO-MED12). In this numerical study, model results are compared with available in-situ observations to evaluate the model performance about the present-day distribution of stable water isotopes and their relationship with salinity on a sub-basin-scale. There is good agreement between the modelled and observed distributions of δ¹⁸O_sw in the surface water. The model successfully simulates the observed east-west gradient of ¹⁸O_sw characterising surface, intermediate and deep waters. The results also show good agreement between the simulated δD and the in-situ data. The δD shows a strong linear relationship with δ¹⁸O_sw (r² = 0.98) and salinity (r² = 0.94) for the whole Mediterranean Sea. Moreover, the modelled relationships between δ¹⁸O_sw and salinity agree well with observations, with a weaker slope in the eastern basin than in the western basin. We investigate the relationship of the isotopic signature of the CaCO₃ shell (δ¹⁸O_c) with temperature and the influence of seasonality. Our results suggest a more quantitative use of δ¹⁸O records, combining reconstruction with modelling approaches.

Received: 08 Dec 2023 – Discussion started: 10 Jan 2024

Download & links

Preprint (PDF, 8354 KB)

Supplement (282 KB)

Download & links

Mohamed Ayache, Jean-Claude Dutay, Anne Mouchet, Kazuyo Tachikawa, Camille Risi, and Gilles Ramstein

Status: final response (author comments only)

RC1:
'Comment on gmd-2023-237', Antje Voelker, 07 Feb 2024
Ayache and co-authors present the first high-resolution modeling study for water isotopes in the Mediterranean Sea. As a first attempt to relate their results to future paleoceanographic applications, they apply their water isotope model outcomes to calculate ∂¹⁸O in marine carbonate. Overall, this is an interesting and novel study and, in my opinion, fits well into GMD. As someone working with water isotopes in sea water, I am very happy to see such studies advancing our knowledge. The manuscript is well written and the figures all informative and needed. The results are relevant and future attempts to go towards a fully coupled ocean-atmosphere model should be of great interest for the scientific communities interpreting speleothem and lacustrine paleo-records in the Mediterranean region.
The science presented is sound, although I am not an expert in climate models and therefore cannot fully judge if the model description is sufficient and can be reproduced based on the information given. From my reading I would say both criteria are sufficiently fulfilled.
I do not have major comments for the manuscript and believe minor revision will address the points I am making below. Some relevant changes might arise from the additional in-situ data I am pointing out in the specific comments, but those will not change the overall outcome of the study. One caveat I see in the manuscript is that Nile river run-off is never mentioned and discussed. For the sapropel research (mentioned in the manuscript) and tracing influences of NW African monsoon rainfall in paleoclimate studies, but also in the modern hydrological cycle that is an important process

Specific comments:
Line 10: as a paleoceanographer I understand where you want to go with the phrase “CaCO₃ shell” but not every reader will be aware that you referring to planktonic foraminifera shells here. So, the text needs to be amended here to be understandable for every reader.
Line 83: if you just want to focus on paleoceanography, you need to add Sea after Mediterranean. However, I believe you can go further and say Mediterranean (region) paleoclimate as the modeling results should also be relevant for studies of speleothems and lacustrine sediments, besides paleoceanographic studies (that would also go beyond foraminiferal calcite shells). You actually hint to the broader potential impact in line 319!
Line 121: verify bouquin AIEA; this reads like a placeholder text for a missing reference. It might also be IAEA.
Line 141: please provide reference for the standard isotopic values.
Line 199: there exist additional/newer in-situ observations in the buffer zone west of the Strait of Gibraltar and one additional station in the Alboran Sea:
Voelker, A.H.L., Colman, A., Olack, G., Waniek, J.J., Hodell, D., 2015. Oxygen and hydrogen isotope signatures of Northeast Atlantic water masses. Deep Sea Research Part II: Topical Studies in Oceanography 116, 89-106, doi: 1016/j.dsr2.2014.11.006. With the raw data available in Pangaea, e.g. Voelker, Antje H L; Colman, Albert Smith; Olack, Gerard; Waniek, Joanna J; Hodell, David A (2015): Oxygen and hydrogen isotopes measured on water bottle samples during EUROFLEETS cruise Iberia-Forams. PANGAEA, https://doi.org/10.1594/PANGAEA.831462

Benetti, M., Reverdin, G., Aloisi, G., Sveinbjörnsdóttir, Á., 2017. Stable isotopes in surface waters of the Atlantic Ocean: Indicators of ocean-atmosphere water fluxes and oceanic mixing processes. Journal of Geophysical Research: Oceans 122, 4723-4742, doi: 1002/2017JC012712. With the data included in Reverdin, G., Waelbroeck, C., Pierre, C., et al., 2022. The CISE-LOCEAN seawater isotopic database (1998–2021). Earth Syst. Sci. Data 14, 2721-2735, doi: 10.5194/essd-14-2721-2022. https://www.seanoe.org/data/00600/71186/

Voelker, A.H., 2023. Seawater oxygen and hydrogen stable isotope data from the upper water column in the North Atlantic Ocean (unpublished data). Interdisciplinary Earth Data Alliance (IEDA), doi: https://doi.org/10.26022/IEDA/112743

and in the Alboran Sea itself: Voelker, Antje H L (2017): Seawater oxygen isotopes for Station POS334-73, Alboran Sea. Instituto Portugues do Mar e da Atmosfera: Lisboa, Portugal, PANGAEA, https://doi.org/10.1594/PANGAEA.878063

Line 251: Voelker et al. (2015, DSR II) obtained a lower slope of 0.32 for surface waters in the NE Atlantic with a strong bias towards subtropical waters (see their figure 11a). Craig and Gordon (1965) also observed a slope of 0.22 for the Atlantic’s subtropical to tropical waters. So, your MedSea slopes fit well to those observations.
Line 275: you could check the model’s performance in the buffer zone west of the Strait of Gibraltar as much of the new data listed above include dD measurements.
Line 284: in a general sense, you could compare d-excess trends with Benetti, M., Reverdin, G., Pierre, C., Merlivat, L., Risi, C., Steen-Larsen, H.C., Vimeux, F., 2014. Deuterium excess in marine water vapor: Dependency on relative humidity and surface wind speed during evaporation. Journal of Geophysical Research: Atmospheres 119, 2013JD020535,doi: 10.1002/2013JD020535. That reference might also fit in the discussion in line 377.

Technical corrections:
Line 5: define what sw in ∂¹⁸O_sw stands for: sea water or surface water? If sea water, the more common practice is to just use “w” for water.
Line 14: O is missing
Line 19: correct spelling to “include”
Line 48: replace input with inflow
Line 49: replace into with in. Later in the sentence, correct the word order to Levantine Intermediate Water.
Line 142: salinity is nowadays only given as a number (as correctly, done, for example inline 234); so, PSU should be deleted here.
Line 201: I assume you mean eastern and not western basin as all the Gat et al. (1996) data are from the eastern basin.
Line 207: EMed and WMed as acronyms should be defined.
Line 209: if you write western Mediterranean instead of WMed, Sea should be added behind Mediterranean.
Lines 228-229: check the longitudes given for the eastern and western Med, respectively. If referring to the WMed, there should also not be a negative sign before the 6°E.
Lines 365-366: add the article the before EMed/WMed, respectively.
Figures: chosen color scheme: many of the figures include a red to green color range with symbols overlain in such colors. So, for color blind people it will be impossible to correctly read some of the figures. The author might want to check, if plotting in a different color range would be possible.
Review by Antje Voelker, 7^th February 2024
Citation: https://doi.org/10.5194/gmd-2023-237-RC1
- AC1: 'Reply on RC1', Mohamed Ayache, 12 Apr 2024
  
  The comment was uploaded in the form of a supplement: https://gmd.copernicus.org/preprints/gmd-2023-237/gmd-2023-237-AC1-supplement.pdf
  
  Citation: https://doi.org/10.5194/gmd-2023-237-AC1
RC2:
'Comment on gmd-2023-237', Anonymous Referee #2, 29 Feb 2024
General comment
Ayache et al. present the implementation of stable water isotopes (d18O and dD) in the high-resolution regional ocean model NEMO-MED12. The simulation of such isotope proxies in climate models is very useful for past climate reconstruction and to better understand climate processes recorded in the water cycle. Ayache et al. performed a simulation for present-day conditions and evaluate their results with available isotopic observations in seawater and marine calcite. They also investigate the relationship of isotopes with salinity.
There are not so many studies on isotope modeling in the ocean, even more in a regional model. Moreover, the Mediterranean Sea is interesting in several points of view: many data, a strong east-west contrast in oceanic evaporation, a relatively short residence time… The article is easy to follow, and the analyses are sound. The figures could be improved, especially the used color scales, and some details on the description of the simulation are missing. Moreover, the discussion section is not really a discussion, yet, but more a summary of the results. After addressing these minor points, detailed below, the article of Ayache et al. could be published in GMD.

Specific comments (rather minor revisions)
Some details on the simulation are missing. Especially, what is spinup time? How was it performed? On line 154, it is said that LMDZ-iso simulation outputs for the period 1990-2020 were used as isotope boundary conditions? What does it mean exactly? That the authors performed a simulation 30 years between 1990 and 2020 with the forcings of the corresponding year? Or that the authors used a climatological average of the LMDz-iso 1990-2020 simulation as boundary conditions, in order to perform a simulation of several decades (so with the same conditions all along the simulation)? What does it involve in terms of bias in isotopic modeled results compared to the observations?

Still about the experimental design, one important aspect for the calculation of d18Ocalcite from modeled d18Osw is the forcing for surface temperature conditions. As said by the authors, the surface temperature conditions do not come from LMDZ-iso but from an ERA-40 relaxation term applied to the ARPERA heat flux. It means that inconsistencies between d18O of freshwater fluxes and temperature are possible. Could the authors elaborate on this aspect? Could they evaluate the potential biases on the ocean temperature and so on the modeled d18Ocalcite?

The discussion section is not really a discussion but more a summary of the results at the current state of the paper. Here are some topics the authors can discuss: How are the results NEMO-MED-wiso compared to global ocean models or coupled models? Are they improved thanks to the high resolution of NEMO-MED-wiso? The authors talk about coupling as a perspective, but what is possible to do with this model given that it is a regional model, not a global one? How can it bring new useful insights for paleoclimate applications except by putting as boundary forcings the atmospheric fields from paleoclimate global simulations (i.e., offline)? Can this model be used to improve global climate models? The seasonality aspect on d18Ocalcite is interesting, could you elaborate more on this aspect?

The green-to-red colormap used in several figures is not appropriate for colorblind people and should be changed.

The d18Ocalcite dataset is not described in the method section (section 2.5).

The difference between R96 and R144 is described very briefly. To show the difference map between R144 and R96 for both d18Osw and applied isotope freshwater fluxes could help to understand better what does (not) happen. Could the remapping from LMDZ-iso grid to NEMO-MED-wiso one partly explain this non-difference? See technical comments below.

Minor technical comments:
Line 14: O is missing in d18O.

Line 16: (d18O-S relationship) can be removed.

Line 40: “high resolution regional ocean model, yet.”.

Line 57: Replace that by which.

Line 67: remove “as an oceanographic tracer”.

Line 76: We use isotope fluxes from…

Line 110: The term isotopologue should be used at the beginning of the paper (line 14). Then you can say you use the term isotope instead.

Line 110: high-resolution

Line 121: replace bouquin AIEA by the appropriate IAEA reference.

Lines 142-143: Table S2 are…

Section 2.3: sea major comment about spinup time and simulation length.

Line 157: remove Risi et al., 2010b.

Section 2.5: see major comment about the description of d18Ocalcite dataset.

Line 230: pseudo-salinity results or standard modeled salinity?

Section 2.3: please change salinity by pseudo-salinity where needed to avoid misunderstanding between the modeled standard salinity of NEMO-MED and the pseudo-salinity described in this paper. Change salinity by pseudo-salinity in the title too.

Line 238: spatial slope?

Line 241: between observed salinity and d18Osw…

Lines 259-265: this part has nothing to do with the d18O-pseudo salinity relationship. It should be removed, except if you can show a change in the relationship when using R96 or R144 LMDz-iso fields.

Section 3.3: I think the part on dD can be removed. It’s similar to d18o and there are not so many data. Figure 7 should be removed too.

Lines 266-277: can be removed.

Line 278: remove (d-excess= δD - 8* δ18Osw, Dansgaard, 1964) as you already said it at the beginning of the paper.

For the Figure 9, it could be interesting to see the depth profile of d-excess too. Are there some data to compare with in EMed (according to Figure 7)? Then you could maybe elaborate a little bit more for the section 3.3.

Figure 1: typo in “Precipitation” in plots d, e, and f. Also in the legend of the figure.

Figure 6: could you show the difference R144-R96 in the d18Osw, but also in the applied isotope freshwater fluxes. It could help to understand the little difference between the two simulations and to elaborate a little bit more (for now, the results are described in 6 lines at the wrong place (lines 259-265)).

Figure A1: Apply different scales for EMed and WMed d18Osw.

Legend of figure A2: average vertical profiles.
Citation: https://doi.org/10.5194/gmd-2023-237-RC2
- AC3: 'Reply on RC2', Mohamed Ayache, 12 Apr 2024
  
  The comment was uploaded in the form of a supplement: https://gmd.copernicus.org/preprints/gmd-2023-237/gmd-2023-237-AC3-supplement.pdf
  
  Citation: https://doi.org/10.5194/gmd-2023-237-AC3
RC3:
'Comment on gmd-2023-237', Allegra N. LeGrande, 29 Feb 2024
Ayache et al NEMO iso review

Water isotope tracers are indeed a useful way to track the water cycle, and this study seeks to provide for high resolution insight into the Mediterranean ocean.

The authors of this study include expert isotope modelers, so the work is on the whole very solid.

I have Mostly few questions about the specifics of implementation and the write up.

for someone who is not a water isotope modeler, the casual inclusion of shorthand / jargon without explanation needs to be expressly defined. I.E., 𝛿¹⁸O_sw or 𝛿D (also—shouldn’t you write 𝛿D_sw to be consistent?) or CaCO₃ or 𝛿¹⁸O_c all need to be defined – what does the delta mean. What do the subscripts mean. Some of the equation rendering has broken down maybe on the author’s side, maybe on the Copernicus side.

In the write up of previous work in the med for isotopes, the authors may (not?) be aware that there is almost certainly a mistake in the 𝛿D values of Gat as they vary much much less than 𝛿¹⁸O_sw – probably the original source should be sought out for that validation.

When it is said that ‘we use fluxes’ from LMDZiso – that is surface water isotope fluxes? How are fluxes from rivers handled? Do you use observed isotope values or simulated ones? (Do the simulated river values closely approximate the measured ones?) If no measurements are available, what was done instead?

On page 5, they say “it is common to transport the isotopic ratio rather than the individual isotope…” then later “and pseudo-salinity fluxes”. I don’t know NEMO that well, but I am going to guess they are saying in a round about way that this ocean model has a rigid lid instead of a free surface. They should say either way. Because most isotope models do not in fact transport around concentrations of isotopes, they transport around mass. Sure – some models do not actually conserve mass – they are forever having to reimplement water isotopes in their code because they have virtual moisture or salt fluxes. Anyhow, those who can do indeed transport around mass not concentration. The per mil isotopic composition is determined on post-processing. Why? This is done so that the isotope / tracer code can have an exact replica of ‘water’ from the non-tracer code and this tracer can be 1:1 compared throughout the entire model to made sure mass isn’t being gained/lost anywhere spuriously. Isotopic composition comes into play because SMOW is defined and fractionation at phase changes is defined. This is, in general, simpler for an ocean model where the mass of water is simply (MO – S), but if you have a rigid lid, then you have virtual mass fluxes of isotopes. Clarity for this point is required.

The ‘interpolated to 20 min time step’—does this mean that actual rainfall and weather systems otherwise are regressed and then passed to the model at this finer time step, or is the daily value simply applied/scaled at the 20 minute interval. I would guess that if you are using some sort of nudged version of LMDZiso that there is useful information at a finer timescale (i.e., if its been nudged at 3 hour timesteps, why not interpolate from 3hr->20min) – otherwise you’ll miss the finer temporal resolution features. You wouldn’t need to store all of LMDZiso values at that timestep—just those in your domain.

I’m still confused about the pseudo-salinity tracer. Please explain

Page 6:: the present day values seem awfully low. CO2 of 348ppm – I rarely encounter PhD students anymore born in a world with CO2 this low.

NEMO-MED12 grid is jargon that I don’t understand

Still confused on L165-170 how the isotopic composition for the rivers was determined. It sounds like you are saying that the isotopic composition of river discharge = local grid box precipitation isotopic composition (which would be wrong of course). Can’t you use observations or use d18Oriver from LMDZiso (or another isotope enabled model). Since you have already established that the Med is an evaporative basin, you might expect that d18Oriver to be a bit enriched compared to d18Oprec… (Places downriver or downhill in a P>E location you would expect d18Oriver to be a bit depleted compared to d18Oprec… ) But the Med, and particular places like the Nile, you definitely should expect some evaporation to strip out the light isotopes of the river.

Can you write up the E-W surface d18Osw context from obs ? Maybe putting observed d18Oriver would make for a better gradient. (The baseline composition is set by your SMOW definition—I’d worry less about that.)

For deriving d18O-S relationships – can you put yours in context of the LMDZiso? Would you expect NEMOiso to differ that much given that you are prescribing your end member from the coupled model? Is this a useful section?

For section 3.3 – can you please check the Gat96 comparison. Does it make sense?

For the d18Ocalcite discussion, what is the correlation between d18Oc and temperature temporally and spatially. For interannual variability, does the inclusion of d18Osw confound the correlation. Also—you are presuming surface dwelling foraminifera. Maybe its interesting to look at species specific d18Oc.

There are some existing SWING comparisons of different isotopic compositions for different groups. Maybe for your next paper you could pull those in, but for this one, you should at least mention and speculate if it would be useful.
Citation: https://doi.org/10.5194/gmd-2023-237-RC3
- AC2: 'Reply on RC3', Mohamed Ayache, 12 Apr 2024
  
  The comment was uploaded in the form of a supplement: https://gmd.copernicus.org/preprints/gmd-2023-237/gmd-2023-237-AC2-supplement.pdf
  
  Citation: https://doi.org/10.5194/gmd-2023-237-AC2

Mohamed Ayache, Jean-Claude Dutay, Anne Mouchet, Kazuyo Tachikawa, Camille Risi, and Gilles Ramstein

Supplement

https://doi.org/10.5194/gmd-2023-237-supplement

Mohamed Ayache, Jean-Claude Dutay, Anne Mouchet, Kazuyo Tachikawa, Camille Risi, and Gilles Ramstein

Viewed

Total article views: 329 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
239	60	30	329	22	17	17

HTML: 239
PDF: 60
XML: 30
Total: 329
Supplement: 22
BibTeX: 17
EndNote: 17

Views and downloads (calculated since 10 Jan 2024)

Month	HTML	PDF	XML	Total
Jan 2024	75	23	6	104
Feb 2024	68	8	6	82
Mar 2024	40	15	10	65
Apr 2024	56	14	8	78

Cumulative views and downloads (calculated since 10 Jan 2024)

Month	HTML	PDF	XML	Total
Jan 2024	75	23	6	104
Feb 2024	68	8	6	82
Mar 2024	40	15	10	65
Apr 2024	56	14	8	78

Viewed (geographical distribution)

Total article views: 319 (including HTML, PDF, and XML) Thereof 319 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 26 Apr 2024

Short summary

Water isotopes (δ¹⁸O, δD) are one of the most widely used proxies in ocean climate research. Previous studies using water isotope observations and modelling have highlighted the importance of understanding spatial and temporal isotopic variability for a quantitative interpretation of these tracers. Here we present the first results of a high-resolution regional dynamical model (at 1/12° horizontal resolution) developed for the Mediterranean Sea, one of the hotspots of ongoing climate change.


Total:	0
HTML:	0
PDF:	0
XML:	0