The paper describes formulation and an application of a new extension to the Meso-NH mesoscale atmospheric cloud-resolving model. The extension enables the model to capture the dynamics of water isotopic composition across gaseous, liquid and solid phases of water.  A single-moment mixed-phase bulk scheme is used. Three stable isotopologues are considered: light water, semiheavy water HDO, and heavy-oxygen water (Oxygen-18). The example application uses a two-dimensional setup from a prior Meso-NH study (1994) simulating a tropical squall line observed during a field project back in 1980-ties.

I consider the presented material matching well the journal scope, and useful for the community. In the following comments, there are numerous suggestion for how to improve the description, but overall I find the paper well written and balanced. My review focuses on the microphysics description and the isotope "basics", for I have less experience in interpretation of the isotopic features of large-scale weather systems or the hydrological implications.

Abstract:

• suggest rephrasing "present-day phenomena such as squall lines..." as it suggests that a squall-line is something unique to present day
• the abstract would benefit from more specifics regarding what type of model is presented in the paper: single-moment mixed-phase bulk, kinetic fractionation considered, deuterium and oxygen-18 isotopologues tracked, ...

Section 1:

• despite defining isotopic delta only in eq. (9) - 5 pages later, it is already used on page 1/line 11, suggest rephrasing to (deuterium and oxygen-18)
• line 25: "So far, modeling studies have relied on models with parameterized convection" seems to contradict later statement on existing cloud-resolving models or large-eddy simulation models
• line 31: worth mentioning as well the DESCAM isotope module: Hiron & Flossman 2020 (DOI:10.1029/2019JD031753) and commenting on the type of microphysics description employed in each model (single-moment, multi-moment, bin; liquid- vs. mixed-phase, which isotopologues tracked, ...)
• line 41: typo: "third", not "thirst"

Section 2:

• first paragraph: among the 17 literature references provided for Meso-NH, several have no relevance for the presented development, suggest referring to references given within Lac et al. 2018, and focusing in this paragraph only on the model components which have further usage in the text;
• line 60: does the per-mass-of-dry-air definition also apply to (specific) number concentrations?
• line 69: use \textsuperscript{th} instead of math mode to avoid italics
• line 89: would "ablate" instead of "decay" be a better suited word here? (see, e.g., DOI:10.1029/2019MS001689)
• line 93: this line needs elaboration as it is not understandable without getting acquainted with the referenced paper and it is unclear what "suppressed" implies here?

Section 3:

• despite the "implementation" in the title, there is no discussion of implementation in this section, it rather deals with model formulation, what about "Formulation of the Meso-NH-ISO water stable isotopologues model"?
• line 98: suggest highlighting that the approach implies that enabling the isotopic fractionation dynamics does not alter any of the model results that would be obtainable with the fractionation disabled (IIUC)
• line 101: "duplicated twice" sounds like quadrupled, but here you get three entities, right?
• line 102: "tracked" or "considered" rather then "implemented"
• eq. (7): nabla is missing
• line 105: the paper deals with 2D simulation, hence suggest omitting "3D" here
• line 108: "source species" sounds like referring to the source term, better "light species"?
• line 118: VSMOW ratios are defined for atoms, not molecules, worth commenting on it (https://web.archive.org/web/20200729203147/https://nucleus.iaea.org/rpst/documents/VSMOW_SLAP.pdf)
• eq. (9): inclusion of the 1000 factor in the definition is misleading - this is just a unit for reporting the values; analogously, the definition of the mixing ratio is "m_v/m_d" and not "1000*m_v/m_d" even though you report it in g/kg; and rain rate definition does not feature 60 or 1000 multipliers for second-to-minute and meter-to-mm conversions; I am aware that this is common in hydrological literature, but still suggest just commenting that the common unit is per mille (see also eq. (1) in the above-linked IAEA Reference sheet); same for unit of d-excess in Table 2 (should be dimensionless) and delta units in Tables 3 & 4 (dimensionless).
• eq. (10): explain the "meteoric water line" concept, elaborate on the 8 factor (cf. page 14/line 316!) and refer to Craig 1961 (DOI:10.1126/science.133.3465.1702) where it was first given
• eq. (10): suggest using a short dash instead of "minus" for typesetting "d-excess" 
• subsection 3.2: suggest highlighting the relation with vapour (super)saturation
• subsection 3.2.1: the handling of the Bergeron process should also be elaborated on here, and the implication for isotopic fractionation highlighted
• eq. (16): according to Table 2, D_v should be in SI unit m²/s, but the value 0.211 yields a result in cm²/s (it is correct in the code, see also a note code quality below)
• eq. (23): the text above suggests that Pruppacher & Klett provided the isotopic formula, which is not the case, suggest simply referring to eq. (14) above or citing a work which covered isotopologue mass transfer (e.g., Gedzelman & Arnold 1994, DOI:10.1029%2F93JD03518)
• line 206: "interact" -> "interacts"

Section 4:

• line 223: suggest changing "fully observed" to "subject to comprehensive radar observations" (also in line 427)
• line 224: "simulated with ... numerical experiments" -> "simulated in ... numerical experiments" (?)
• line 230: "short-live" -> "short-lived"
• line 233: it is unclear if 40% concerns mass or number
• line 246: "open" contradicts "periodic", and this is important to define if the isotopes are simulated in an open or closed reservoir (also in the context of later discussion on lack of steady state in vapour isotopic composition) 
• line 246: clarify if "no lateral advection" refers to Y dimension?
• line 251: elaborate on what are the implications of using a 3D turbulence scheme in a 2D simulation
• line 254: is "Steady-state" a good title for a section concluding that "isotopic composition of water vapor does not reach equilibrium at the end of the simulation"
• line 263: "moist and" -> "moist air"
• line 276: the text suggests that there is information about ice in the figure, but given that "total condensed mixing ratio" likely sums all five non-vapour water species, it is hard to interpret the figure. Suggest plotting liquid and solid water separately, and adding plot with relative humidity wrt ice 
• Figure 3: y-axis label should be "r_v" not "q_v" if it is consistent with the caption, otherwise the caption is inconsistent with the text, and the symbol tables need addition of specific humidity
• Figure 4 caption: "kJ" not "kj"
• line 293: "Isotopic results" sounds awkward to me, and perhaps "Simulated isotopic composition of ambient vapour and rain water"?
• line 312: "We also probably detect rain evaporation within downdrafts" suggests as if it was not possible to confirm, while running the simulation with a given process disabled should provide a clear way of confirming it. There are several features commented as occurring "probably" or "likely"; given the low computational cost of a 2D simulation, I expect that some of these are verifiable by exploring simulation results with the processes in question toggled on or off.
• line 344: "minima" instead of "minimums" (as with datum->data)
• line 409: use "\Delta t" for the time step, not "dt"

Section 5:

• line 431: "non-isotopic steady state" is not clear (also line 434), please rephrase
• line 433: rephrase "region on development"
• line 436: "well captures" -> "captures well" 

Table 1:

• values for "c" and "C" are missing $\times$ symbols (3 occurences)

Supplement:

• Figures in the Supplement are raster graphics (screenshots) not suitable for publication
• according to GMD guidelines (https://www.geoscientific-model-development.net/submission.html): "Supplementary material is reserved for items that cannot reasonably be included in the main text or as appendices." (videos, very large images, maps, CIF files, as well as short computer codes...). Please incorporate the supplementary figures into the main body of the manuscript, especially that these are repeatedly mentioned in the text and required to follow the narrative.

Maths, symbols, nomenclature:

• the term "isotope" applies to atoms, while "isotopologue" to molecules, hence suggest replacing "Water stable isotopes" with "water stable isotopologues" (incl. the title and keywords)
• number concentrations have a per-mass definitions in the paper, which contradict the common per-volume understanding of the term, "specific concentration" would be a better name
• use upright font for label-like non-value subscripts, e.g. $\rho_\text{a}$, not $\rho_a$, etc (D_v, k_a, L_s, L_v, m_r, N_c, N_r, r_c, r_r, r_v, R_c, R_i, R_l, ...)
• both "isotope ratio" and "isotopic ratio" terms are used in the paper, please unify, see e.g. https://list.uvm.edu/cgi-bin/wa?A2=ind2110&L=ISOGEOCHEM&P=R40040
• it is unclear from the text if "total water" includes all isotopic species or only the light isotopologue?
• semiheavy water is sometimes labelled as HDO, and in other instances as HD¹⁶O, suggest unifying

Code:

• the model name and version given in the paper title (Meso-NH-ISO v1.0) are not consistent with the code archive (MNH-V5-5-0), and are not mentioned at the linked website (http://mesonh.aero.obs-mip.fr/mesonh55); the code archive must include the same labelling as the paper title
• the Zenodo archive Barthe 2026 lists Creative Commons as license, but the CC licenses are not applicable to software (https://creativecommons.org/faq/#can-i-apply-a-creative-commons-license-to-software), and the contents of the archive and paper text indicate CeCILL-C license.
• the text indicates that the Zenodo archive contains "complete code of Meso-NH v5-5-0 including the water isotope scheme", but Zenodo metadata indicate: "Fortran routines to be inserted in Meso-NH V5-5-0 to use the water isotope scheme."
• the Zenodo metadata indicate three authors: Barthe, Francois and Vimeux; the reference entry in the paper lists just one
• the large archive size of 234 MB is due to inclusion of "convenience copies" of external libraries and tools, which is neither expected nor described anywhere, and I'd consider it a malpractice as the libraries are straightforward to obtain. Examples:
  • 84.7 MB: MNH-V5-5-0/pub/ncl-6.4.0.tar.gz
  • 18.8 MB: MNH-V5-5-0/src/LIB/netcdf-c-4.7.4.tar.gz
  • 16.2 MB: MNH-V5-5-0/src/LIB/oasis3-mct_v3.tar.gz
  • 12.0 MB: MNH-V5-5-0/src/LIB/hdf5-1.12.0.tar.gz
  • 11.0 MB: MNH-V5-5-0/src/LIB/eccodes-2.18.0-Source.tar.gz
  • 8.6 MB: MNH-V5-5-0/src/LIB/RAD/ecrad-1.4.0.tar.gz
  • 6.5 MB: MNH-V5-5-0/pub/ga-5-0-3.tar.gz
• Moreover, the MNH-V5-5-0/bin_tools folder contains 43 MB of binary files which are not portable (e.g. would work only on Linux systems with outdated libpng12.so.0 present), and also not expected to be included in a source archive.
• aiming to double check the correctness of the diffusivity formula (16) constant value given with non-SI unit in the paper, and expecting a single code location for the constant, I have spotted a troubling amount of code duplication:
  • src/src_isotopes/isotopes_tendencies.f90:528: ZDV(:) = 0.211E-4 * (ZZT(:)/XTT)**1.94 * (XP00 / ZPRES(:))
  •  src/src_isotopes/ice_adjust.f90:625: ZDV(:,:,:) = 0.211E-4 * (ZT(:,:,:)/XTT)**1.94 * (XP00 / ZPRES(:,:,:))
  • src/SURFEX/canopy_blowsnw_subl.F90:78: ZDV(:,JK)  = 0.211E-4 * (PT(:,JK)/XTT)**1.94 * (XP00 / PP(:,JK))
  • src/MNH/lima_tendencies.f90:478: ZDV(:) = 0.211E-4 * (ZT(:)/XTT)**1.94 * (XP00 / PPABST(:))
  • src/MNH/ice4_tendencies.f90:430: ZDV(JL) = 0.211E-4 * (ZT(JL)/XTT)**1.94 * (XP00 / PPRES(JL))
  • src/MNH/lima_adjust.f90:530: ZDV(:) = 0.211E-4 * (ZZT(:)/XTT)**1.94 * (XP00 / ZPRES(:)) 
  • src/MNH/lima_adjust.f90:823: ZDV(:) = 0.211E-4 * (ZZT(:)/XTT)**1.94 * (XP00 / ZPRES(:))
  • src/MNH/lima_adjust.f90:935: ZDV(:) = 0.211E-4 * (ZZT(:)/XTT)**1.94 * (XP00 / ZPRES(:)) 
  • src/MNH/lima_notadjust.f90:353: ZDV(:) = 0.211E-4 * (ZZT(:)/XTT)**1.94 * (XP00 / ZPRES(:))
  • src/MNH/subl_blowsnow.f90:376: ZDV(:) = 0.211E-4 * (ZZT(:)/XTT)**1.94 * (XP00 / ZPRES(:))
  • src/MNH/rain_ice_elec.f90:2455: ZDV(:) = 0.211E-4 * (ZZT(:)/XTT)**1.94 * (XP00 / ZPRES(:)) 
  • src/MNH/ch_aqueous_tmicice.f90:1059: ZDV(:) = 0.211E-4 * (ZZT(:)/XTT)**1.94 * (XP00 / ZPRES(:)) 
  • src/MNH/lima_cold_slow_processes.f90:324: ZDV(:) = 0.211E-4 * (ZZT(:)/XTT)**1.94 * (XP00 / ZPRES(:))
  • src/MNH/lima_mixed.f90:440: ZDV(:) = 0.211E-4 * (ZZT(:)/XTT)**1.94 * (XP00 / ZPRES(:)) 
  • src/MNH/rain_ice_slow.f90:120: PDV(:) = 0.211E-4 * (PZT(:)/XTT)**1.94 * (XP00 / PPRES(:))

From software engineering standpoint, the repeated definition of constants across the codebase (15 times) and the lack of code reuse (i.e., copy-pasted lines instead of shared function definition) are both anti-patterns. This approach leads to bug-prone code, technical debt, high software maintenance costs, and hindered opportunities for future developments.

References:

• fix DOI in Caniaux et al. 1994 (should be 10.1175/1520-0469(1994)051<2046:ANSOTS>2.0.CO;2)
• fix DOI in Chang et al. 1993 (should be 10.1175/1520-0469(1993)050<0161:TDMSOS>2.0.CO;2)
• fix DOI in Chong et al. 1987 (should be 10.1175/1520-0493(1987)115<0670:ATSLOD>2.0.CO;2)
• fix DOI in Gamache & Houze 1981 (should be 10.1175/1520-0493(1982)110<0118:MAMAWA>2.0.CO;2)
• fix DOI in Hall & Pruppacher 1976 (should be 10.1175/1520-0469(1976)033<1995:TSOIPF>2.0.CO;2)
• fix DOI in Meyers et al. 1992 (should be 10.1175/1520-0450(1992)031<0708:NPINPI>2.0.CO;2)
• fix DOI in Redelsperger & Lafore 1988 (should be 10.1175/1520-0469(1988)045<1334:ATDSOA>2.0.CO;2)

HTH

Review of "Meso-NH-ISO v1.0: a water stable isotopes scheme in the non-hydrostatic mesoscale atmospheric model Meso-NH. Application to a 2D West African squall line" by Barthe, Vimeux, Risi and Francois, Manuscript egusphere-2026-548

Summary: This paper introduces Meso-NH-ISO, a water-isotope-enabled version of Meso-NH, a numerical model for the study of mesoscale meteorology.  The implementation of water isotopes is described, and a 2D simulation of an observed squall line case is presented.  The original squall line case didn't include observations of isotopes, so water isotope observations from other West African squall lines are used to inform (qualitatively) the performance of the water isotope implementation and simulation.

Assessment: The paper does its job well, introducing the model, its implementation and performance in a demonstration simulation, while highlighting the water isotopic signals in rain evaporation and how they relate to different regions/phenomena within squall lines.  I have some questions about details of the isotopic implementation below, specifically the Bergeron process and the kinetic fractionation of deposition onto ice, and ask for clarification from the authors.

Recommendation: Minor Revisions.  The comments below are minor, but I would be curious to read the responses to my queries about the details of the isotopic implementation.

Minor Comments (5/112 means page 5, line 112, REFERS TO TRACK CHANGES VERSION):

Reference: Looking at this analysis of rain evaporation made me think of the bin model implementation of H218O in Hiron and Flossman (2020, https://doi.org/10.1029/2019JD031753).  They have a plot of isotopic composition as a function of droplet size in their figure 8.  Not sure that paper needs to be cited here, but I find it interesting to think about how the broad range of radii in the rain category of a bulk scheme might have a variety of isotopic compositions in reality.

4/85: The arrow representing the Bergeron process in Figure 1, which connects cloud liquid water directly with cloud ice, causes some concern.  My understanding of the Bergeron process is the transfer from cloud liquid water to cloud ice occurs through the vapor phase due to the differing saturation vapor pressures of liquid and ice.  Cloud liquid evaporates (presumably at saturation with liquid and in isotopic equilibrium, since the scheme uses saturation adjustment), and deposition onto ice occurs in supersaturated conditions with fractionation as described in section 3.2.3.  If this is the way that the Bergeron process is implemented, please adjust the arrow in Figure 1.  If not, please explain in the reply how the Bergeron process is implemented and whether that representation of the process differs from what I've described here.

7/168: The exponents in the formula for dynamic viscosity should be "-5", not "5" as is shown in the pdf.

8/191: Add "equilibrium" when describing alpha_l/v

8/195: Couldn't c^{'iso} be described more simply as "c^{'iso} = c' (D/D_iso)^(1/3)"?  This simpler definition might help the reader focus on the differences between the light and heavy water equations.

sec 2.3.2-2.3.3: I'm curious about how the ventilation factor affects the kinetic fractionation of water isotopes as a function of drop or particle size.  If we defined alpha_k = D_iso*f_iso/(D*f) = (Diso/D)^n, the exponent n would tell us how strongly the different diffusivities impact the isotopic transfer between vapor and condensate across those particle sizes.  Presumably, n~1 for small particles and decreases from there.  Stewart (1975, https://doi.org/10.1029/JC080i009p01133) suggests such a formulation in the paragraph after their equation 1.

9/eqn 31: I compared this formulation of alpha_k with equation B26 in Blossey et al (2010), whose b and S_s seems equivalent to the present manuscript's F and (S_i+1), respectively.  The two formulas don't quite match, making me wonder whether the (S_i - 1) terms in the numerator and denominator of equation 31 should actually be (1 + S_i), since S_i is defined as the supersaturation over ice just after equation 26.  Please double check.  If I've misunderstood something, please do clarify.  I'm trusting that this is coded correctly in the code regardless.

10/246: Move the sentence about open boundary conditions to the the end of the paragraph on lines 239-240, since (in my mind) the boundary conditions should be included in the description of the model domain.  Could the impact of inflow at the boundaries on water and its isotopes be described briefly?  Something like: "Where inflow occurs at the domain boundary, the temperature, water vapor mixing ratio and its isotopic composition are taken from the values in the initial soundings at that height."  I think an additional sentence would make more clear the impact of the open boundaries.  Also, I didn't quite understand what was meant by "There is no lateral advection."  Perhaps, clarifying that would also help the reader.

10/sec 4.3: While the water vapor mixing ratio and the isotopic composition of precipitation do reach a quasi-steady state from hours 6-8, the rain rate itself is changing with time (by about a factor of two) during this period.  While the quasi-steady-state of some variables is interesting, the simulation isn't really in steady state, so perhaps a different name for section 4.3 would work better, maybe "Quasi-steady-state regime" or "Mature squall line behavior" or something like that?

12/290: "Fig. 4c and _4d_".  In the figure itself, the fourth panel is labeled "e", which is consistent with the text but surprising to this reader.

13/305: Is this the first time (\delta^{18}O_p)_{eq} appears?  I feel that this variable and (d_p)_{eq} appear so often in this paper that they deserve numbered equations defining them, perhaps back in section 3.  I realize that equation 11 does this implicitly, but for the uninitiated, perhaps the definition could be made explicit.

14/316: The diffusivities of HD^16O and H_2^18O differ but only slightly, by about 3 per mille of each other (e.g., Merlivat and Jouzel, 1979, bottom of p. 5030).  The larger fractionation factor of HDO relative to H218O makes it likely that the jump in isotopic composition (expressed as delta) between rain and vapor is larger (by roughly that factor of eight) for HDO, leading to fluxes that are roughly eight times larger, with a slight modification because D_HDO > D_H218O.  Perhaps, this could be rephrased in terms of fluxes rather than as "eight times faster.

14/317-319: Re "rain evaporation leads rain to be more enriched ..."  Graf et al (2019, https://doi.org/10.5194/acp-19-747-2019) nicely talks about this and could be a good reference here or wherever the d18O_p - (d18O_p)_eq quantity is introduced, because Graf spends a lot of time thinking about how that and the d_excess equivalent behave.

18/373-376: Lee and Fung (2008, https://doi.org/10.1002/hyp.6637) analyze how much the near-surface layer will affect the isotopic composition of precipitation to understand its dependence on drop size and rain rate.  Perhaps, such arguments could apply here.

19/Fig 8: Please add thin, possibly dashed line showing the zero values in panels a-c, including a vertical one in panel c.  This would make it easier for the reader to see the sign changes of disequilibrium and enriching/depleting rain-vapor exchange.  If the authors don't like lines, at least add tick marks on the right-hand and possibly top axes.

20/Fig 9: Do the high evaporation ratios happen only in regions where the mixing ratio is very small?

21/431: "... despite the non-isotopic steady state of the near-surface water vapor ..."  As noted above in my comment about section 4.3, I don't think the transient behavior of the isotopic composition of near-surface vapor is a problem.  I would suggest rephrasing in a more neutral tone: "While the isotopic composition of near-surface vapor continues to evolve late in the simulation while the water vapor mixing ratio and the isotopic composition are quasi-steady, the isotopic results are actually ..."

21/435: "minor problem". As above, I don't think this is a problem.

SI/Fig S1: This is repeated in the pdf I have.  Please check.

SI/Fig S2: Could additional panels be included showing the d18O and d excess of all non-precipitation condensed water?  This could strengthen the argument about why the precipitation just below the melting layer is more depleted in the stratiform region.

SI/Fig S6: Same comments about zero lines as for Fig 8 above.

Typographical suggestions:

9/206: "interacts"

12/284: "snapshots"

21/432: "This is also _true_ earlier in the simulation ..."

22/457: "The _plans for_ this implementation ..." Is a simpler phrasing.