MODELING THE HDO CYCLE WITH A GLOBAL CLIMATE MODEL DURING THE MY34 “DUSTY” SEASON.
M. Vals, (margaux.vals@latmos.ipsl.fr), L. Rossi, F. Montmessin, F. Lefèvre, Laboratoire ATmosphères, Milieux,
Observations Spatiales, Paris, Guyancourt, France, F. Gonzalez-Galindo, Instituto de Astrofı́sica de Andalucı́a-CSIC,
Granada, Spain, A. Fedorova, M. Luginin, Space Research Institute (IKI), Moscow, Russia, F. Forget, E. Millour,
Laboratoire de Météorologie Dynamique, Paris, France, O. Korablev, A. Trokhimovskiy, A. Shakun, Space
Research Institute (IKI), Moscow, Russia, A. Bierjon, Laboratoire de Météorologie Dynamique, Paris, France, L.
Montabone, Laboratoire de Météorologie Dynamique, Paris, France, Space Science Institute, Boulder, CO, USA.

Introduction
The D/H ratio observed in a planetary atmosphere is a
proxy for the ratio of the current water reservoir over
the initial water reservoir of the planet. The current D/H
ratio measured in the Martian atmosphere is at least five
times that of the Vienna Standard Mean Ocean Water
(VSMOW) [1],[2],[3],[4]. This high value of the martian D/H ratio, derived from the HDO/H2 O abundance
ratio, is a precious indicator of the large escape of water
from the martian atmosphere along time. Apart from the
mass difference between both isotopes, the differential
escape of H and D comes from the preferential photolysis of H2 O over HDO [5] and the Vapor Pressure Isotope
Effect (VPIE) that produces an isotopic fractionation at
condensation [6],[7],[8]. Modeling the HDO cycle and
the different processes, which lead to the D/H ratio at
escape in the upper atmosphere, is key to understand
the history of water in the Martian atmosphere. Rossi
et al. 2021 [9] have implemented the HDO cycle in
the updated version of the LMD Mars Global Climate
Model (GCM) [10], based on previous work conducted
by Montmessin et al. 2005 [11] accounting for the fractionation by condensation. The NOMAD and ACS instruments onboard ExoMars Trace Gas Orbiter (TGO)
spacecraft have recently provided unprecedented observations of the HDO amount and D/H ratio profiles in
the martian atmosphere [12],[13],[14],[15], motivating
models development and validation [9],[16]. In particular, TGO data cover the second half, commonly named
the “dusty season”, of the martian year (MY) 34, which
includes a Global Dust Storm (GDS) (Ls 180-230◦ ) and
a regional dust storm (Ls 315-330◦ ). These exceptional
events have been proven to be of particular importance
in the hygrogen escape inventory, as the large load of
dust in the atmosphere leads to an increase of temperature, which fosters the transport of water vapor to the
upper atmosphere [12],[17],[18],[19],[20].

HDO modeling
Using previous work by Montmessin et al. 2005[11],
Rossi et al. 2021[9] have introduced the representation
of the HDO cycle in the LMD Mars Global Climate

Model (GCM)[10]. In this last version of the model,
dust is represented by a semi-interactive scheme[21],
which simultaneously allows the free evolution of the
vertical distribution of dust, going through the different
physical parametrizations, and ensures the match of the
integrated column dust opacity to the observed values,
as compiled by Montabone et al. 2015,2020[22],[23],
for the radiative transfer calculations. In this context, the
same simplified cloud model version as in Montmessin
et al. (2005)[11] was used, meaning water vapor is
turned into ice as soon as saturation is reached. Vals
et al. 2022[24] have recently adapted these implementations to the complete representation of the water ice
clouds, including microphysics and radiative effect of
clouds [25],[26]. Microphysics refer to the parametrization of the different processes of formation of the clouds
such as nucleation of the ice particles on dust particles,
water ice growth and dust scavenging implemented in
the model by Navarro et al. 2014[26] and necessary to
allow the occurrence of supersaturation in the model.
The effect of kinetics in the fractionation by condensation process, relying on the difference of diffusion of
HDO and H2 O molecules in a saturated atmosphere,
has also been included, so as the photodissociation of
HDO and the photochemical reactions of the deuterated
species, implemented in the model by F. Lefèvre and F.
Gonzalez-Galindo, based on [27],[28],[29], accounting
for the effect of fractionation by photolysis.

Model sensitivity to physical parametrizations
In Vals et al. 2022[24], GCM simulations are confronted
to temperature, H2 O and saturation occultation profiles
derived from the TGO/ACS instrument provided by A.
Fedorova with the NIR channel [18]. The impact of
each different implementation in the model mentioned
above on the D/H cycle is summarized by Figure 1[24].
It shows the H2 O and HDO mixing ratios, and the D/H
ratio at 80 km altitude, as computed by the model, and
interpolated at the ACS measurements coordinates, for a
simulation of reference [REF], where all physical packages (radiatively active clouds, microphysics, kinetics
and photochemistry) have been activated, and for all
other simulations, in which each physical process stud-

ied has been separately turned off. This study confirms
the importance of taking into acccount these different
physical processes to significantly improve the comparison to observations, and it particularly emphasizes
the major role of representing the state of supersaturation (see Figure 1). Comparisons of the corresponding
D/H ratio profiles between GCM and observations are
shown and analysed in the companion paper Rossi et al.
2022[30].

Figure 1: H2 O, HDO and D/H ratio GCM outputs interpolated
to the ACS spatial and temporal coordinates at 80 km altitude,
as a function of solar longitude (Ls). Left column: northern
hemisphere, right column: southern hemisphere. First panel
corresponds to the data coverage in latitude and local time.
Second row displays the H2 O volume mixing ratio in logarithmic scale. Third row displays the HDO volume mixing
ratio in logarithmic scale. Fourth row contains the D/H ratio.
GCM outputs are shown for the reference simulation [REF],
a simulation in which the radiative effect of clouds has been
turned off [RAC OFF], a simulation in which the supersaturation has been turned off [supersat OFF], a simulation in which
the effect of kinetics has been turned off [kinetics OFF] and
a simulation in which the photochemistry has been turned off
[photochem OFF]. ACS water vapor volume mixing ratio are
shown in the second row (black dots).

from around 60 km up to the higher altitude ranges,
where the model presents much higher amount of water vapor than observations between Ls 210-240◦ in both
Hemispheres, whereas it fails at representing the peak of
water vapor observed around Ls 270-300◦ , in particular
in the Southern Hemisphere.

Figure 2: H2 O mixing ratio at 40, 60 and 80 km altitude
ranges, as a function of solar longitude (Ls). (red) GCM outputs of the reference simulation interpolated to the ACS spatial and temporal coordinates. (blue) ACS/MIR observations,
Belyaev et al. 2021[31]. (grey) ACS/NIR observations, Fedorova et al. 2020[18]

Modeling water cycle during the MY34 GDS
Although the new implementations have considerably
improved the comparison with observations in regards
to water vapor amount, which has a direct impact on the
D/H ratio, some discrepancies remain. In particular, Figures 2 and 3 display the amount of water vapor at three
different altitude ranges, respectively at 40, 60, 80 km,
and at 80, 100, 110 km, obtained by the model reference
simulation compared to the ACS/NIR and ACS/MIR
observations[18],[31], still during the second half of
MY34. Model results and observations are in good
agreement at 40 km, but this is not the case any more

Figure 3: Same layout as Figure 2 at 80, 100 and 110 km
altitude ranges.

Discussion
Figure 4 displays GCM H2 O mixing ratio and temperature profiles compared to ACS observations[18] between
Ls 235-245◦ , in the tropics. Although temperature profiles are quite in good agreement, the GCM is clearly
showing a wetter atmosphere than observed from around

REFERENCES

60 km altitude. Figure 5 is showing some ice mass load
profiles measured by Luginin et al. 2021[32] in the same
period of time and latitude ranges and compared to the
corresponding GCM ouputs. This Figure reveals that
water ice clouds observed above 60 km are not reproduced by the model. These comparisons suggest that
the model is missing the representation of condensation
occuring during the GDS and later on at the upper altitudes. The suspected reason for that is the lack of dust
nuclei in the model, as it is well known that GCMs have
difficulties to correctly represent the vertical distribution
of dust, especially during dusty events, where sub-grid
scale convection processes are not well represented. Indeed, Neary et al. 2020[19] and Daerden et al. 2022[16]
chose to prescribe the dust profile during the GDS and
achieved good comparisons of water vapour amount and
D/H ratio profiles.

Conclusions and perspectives
We have improved the representation of the HDO cycle
in the LMD Mars GCM by adapting it to the full water
cycle of the model including microphysics and radiative effect of clouds, by completing the modeling of the
fractionation by condensation with the effect of kinetics,
and by adding the complete photochemistry of HDO and
deuterated species. The main impact of these different
implementations is the representation of supersaturation
state, which considerably improves the comparison of
water vapor amount between model results and observations during the “dusty” season of MY34. However,
some discrepancies remain in the upper altitude ranges,
where the model predicts a much wetter atmosphere than
observed in both hemispheres. The poor representation
of the dust vertical distribution is suspected to play a
huge role in this aspect and will be further investigated.
Further results and interpretations on the comparison between model ouputs and TGO/ACS observations will be
presented at the conference.

References
[1] Owen et al.: Science, 240 (4860), 1767-1770, 1988
[2] Encrenaz et al.: A&A, 612 , A112, 2018.
[3] Krasnopolsky: Icarus, 257 , 377 - 386, 2015
[4] Villanueva et al.: Science, 348 (6231), 218â^221,
2015.
Figure 4: Profiles of water vapor and temperature between Ls
= 235-245◦ and latitude 25◦ S and 25◦ N as measured by ACS
(black line) and computed by the GCM (red line).

[5] Cheng et al.: Geophysical Research Letters, 26 (24),
3657-3660, 1999
[6] Krasnopolsky: Icarus, (2), 597-602, 2000.
[7] Fouchet et al.: Icarus, 144 , 114-123, 2000.
[8] Bertaux et al.: Journal of Geophysical Research:
Planets, 106 (E12), 32879-32884, 2001.
[9] Rossi et al.: grl , 48 (7), e90962, 2021.
[10] Forget et al.: JGR, 104:24,155–24,176, 1999.
[11] Montmessin et al.: Journal of Geophysical Research, 2005.
[12] Vandaele et al.: Space Science Reviews, 214 (5),
80, 2018.
[13] Vandaele et al.: Nature, 568 (7753), 521-525,
2019.

Figure 5: Profiles of water ice mass load between Ls = 235245◦ and latitude 25◦ S and 25◦ N as measured by ACS (black
line) and computed by the GCM (blue line).

[14] Korablev et al.: Space Science Reviews, 214 , 7,
2018.
[15] Korablev et al.: Nature, 568 (7753),517-520, 2019.

REFERENCES

[16] Daerden et al.: Journal of Geophysical Research
(Planets), 127 (2), e07079, 2022.

[25] Madeleine et al. : Geophys. Res. Lett., 39 , 23202,
2012.

[17] Aoki et al.: Journal of Geophysical Research
(Planets), 124 (12), 3482-3497, 2019.

[26] Navarro et al.: Journal of Geophysical Research
(Planets), 2014.

[18] Fedorova et al.: Science, 367 (6475), 297-300,
2020.

[27] Lefèvre et al.: Journal of Geophysical Research
(Planets), 126 (4), 2021.

[19] Neary et al.: Geophysical Research Letters, 47 (7),
e84354, 2020.

[28] Cheng et al.: Journal of Chemical Physics, 120 (1),
224-229, 2004

[20] Chaffin et al.: Nature Astronomy, 5 , 1036-1042,
2021.

[29] Chung et al.: Nuclear Instruments and Methods in
Physics Research A, 467 (2002), 1572-1576, 2001.

[21] Madeleine et al.: Journal of Geophysical Research
(Planets), 116 , 11010.669, 2011

[30] Rossi et al.: Journal of Geophysical Research
(Planets), under revision, 2022.

[22] Montabone et al.: Icarus.2014.12.034700, 2015.
[23] Montabone et al.: Journal of Geophysical Research: Planets, n/a(n/a), e2019JE006111, 2020.
[24] Vals et al.: Journal of Geophysical Research (Planets), under revision, 2022.

[31] Belyaev et al.: Geophysical Research Letters, 48
(10), 2021.
[32] Luginin et al.: Journal of Geophysical Research
(Planets), 125 (11), e06419, 2020.