Transport processes of dust and water in the Martian atmosphere revealed by
the MMX Infrared Spectrometer (MIRS): Fast retrieval code for aerosol and
gaseous profiles for limb-sounding
H. Nakagawa, Graduate School of Science, Tohoku University, Japan (hiromu.nakagawa.c1@tohoku.ac.jp),
S. Aoki, Graduate School of Frontier Sciences, University of Tokyo, Japan,
T. Gautier, LATMOS, CNRS, Sorbonne Université, UVSQ-UPSaclay, Guyancourt, France ; LESIA, Observatoire de Paris, Université PSL, CNRS, Sorbonne Université, Université de Paris, Meudon, France,
A. Doressoundiram, LESIA, Observatoire de Paris, Université PSL, CNRS. Sorbonne Université, Université de
Paris, 5 place Jules Janssen, 92195 Meudon, France,
M. A. Barucci, LESIA, Observatoire de Paris, Université PSL, CNRS. Sorbonne Université, Université de Paris, 5 place Jules Janssen, 92195 Meudon, France,
S. Fornasier, LESIA, Observatoire de Paris, Université PSL, CNRS. Sorbonne Université, Université de Paris,
5 place Jules Janssen, 92195 Meudon, France,
P. Bernardi, LESIA, Observatoire de Paris, Université PSL, CNRS. Sorbonne Université, Université de Paris, 5
place Jules Janssen, 92195 Meudon, France,
J. M. Reess, LESIA, Observatoire de Paris, Université PSL, CNRS. Sorbonne Université, Université de Paris, 5
place Jules Janssen, 92195 Meudon, France,
T. Iwata, Institute of Space and Astronautical Science, Japan Aerospace Exploration Agency, Japan,
A. Spiga, Laboratoire de Météorologie Dynamique, Sorbonne Université, CNRS, France,
T. Bertrand, LESIA, Observatoire de Paris, France,
F. Montmessin, LATMOS, CNRS, Sorbonne Université, UVSQ-UPSaclay, Guyancourt, France,
A. C. Vandaele, Institut d’Aéronomie Spatiale de Belgique, Brussels, Belgium,
K. Ogohara, Faculty of Science, Kyoto Sangyo University, Japan,
T. Imamura, Graduate School of Frontier Sciences, University of Tokyo, Japan,
A. Mahieux, Institut d’Aéronomie Spatiale de Belgique, Brussels, Belgium ; The University of Texas at Austin,
Austin, Texas.
Y. Kasaba, Planetary Plasma and Atmospheric Research Center, Tohoku University, Japan,
H. Iwabuchi, Graduate School of Science, Tohoku University, Japan,
MIRS-MMX,
Mars Sub-Science Team members
provide near-infrared (IR) spectral maps in the
range from 0.9 to 3.6 micron with a spectral resoAbstract:
The MMX infrared spectrometer (MIRS) aims to
lution better than 20 nm for the Martian atmosclarify the temporal evolution of the atmospheric
phere, in addition to those of Phobos and Deimos
species and aerosols to reveal the rapid transport
(Barucci et al., 2022, EPS; Ogohara et al., 2022,
processes with a relatively high spatial resolution
EPS).
(typically 2-10 km) at the timescale of hours,
thanks to its equatorial orbit. To process such a
Mars observation by the MMX Infrared Spechuge spectral dataset for each map, a pretrometer (MIRS):
calculated look-up table of synthetic spectra is
The equatorial orbit of MMX offers some adcreated to retrieve physical parameters. We also
vantages for global mapping of the Martian atprepare a new fast retrieval code, JACOSPAR,
mosphere. Being placed on the same orbit as
for the limb-geometry dataset, to reduce the huge
Phobos around Mars, at altitudes ~6000 km, the
computation cost, and address the complexity of
spacecraft will have a ~7 hr orbit around Mars,
full light scattering processes in spherical geomeallowing it to complete a global mapping of
try. This study applies for the first time simultaMars. The well-controlled scanner system of
neous limb retrievals for both dust and water ice
MIRS enables the specific pointing of the inclouds. Under standard dust conditions, the dust
strument, thus allowing MIRS to obtain wide
and water-ice cloud effective radius and number
spatial coverage in hourly timescale. The spacedensities were successfully obtained at an altitude
craft’s slewing capabilities in combination with
range roughly below 40 km.
the instrument’s scanner will also be used during
observations of Mars to maximize the coverage
Introduction:
The MMX infrared spectrometer (MIRS) is an
of the MIRS from low to mid-latitudes within an
imaging spectrometer onboard the MMX JAXA
hour, or to obtain an almost global mapping of
mission. MMX (Martian Moon eXploration)
the sunlit hemisphere up to ~60° N/S latitude
(Kuramoto et al., 2022, EPS) is scheduled to be
with several orbits. These provide the first opporlaunched in 2024 with a sample return from
tunity to follow the temporal evolution of the atPhobos to Earth in 2029. MIRS will remotely
mospheric species (CO2, H2O, CO) and aerosols

(dust and clouds) to reveal the rapid transport
processes with a relatively high spatial resolution
(typically 2-10 km). Time-resolved pictures of
the atmospheric phenomena should be an important clue to understanding both the processes
of water exchange between the surface/underground reservoirs and the atmosphere
and the drivers of efficient material transport to
the upper atmosphere. Different observation
strategies will be possible to maximize either
temporal or spatial coverage of MIRS monitoring
of the Martian atmosphere from 30’ time resolution observation of a limited zone to complete
coverage in a few orbits.
Retrieval tools for MIRS observations:
One of the challenges to process such a high spatial resolution data will be its large data volume.
For a typical observation by MIRS, approximately 106 spectra are acquired for each map. Common retrieval techniques, such as retrievals using
line-by-line calculations with exact values for
geometric and atmospheric parameters for each
spectrum, will be computationally too expensive.
To solve this issue, we have been developing a
pre-calculated look-up table of synthetic spectra.
This method prepares a series of synthetic spectra
that are calculated in advance at tabulated grid
values of geometric and atmospheric parameters
(such as surface pressure, solar zenith angles,
emission angles, atmospheric temperature, surface albedo, aerosols abundance, etc). Forget et
al. (2007) and Spiga et al. (2007) demonstrated
that this method is robust and fast to retrieve surface pressure with MEx/OMEGA data. We have
applied this method to other molecules such as
water vapour and carbon monoxide.
MIRS will also perform limb observations to obtain information on the Martian atmosphere at
high vertical resolution, to better understand the
Mars atmosphere which is a considerably mutually coupled system between the surface, the lower
and upper atmospheres, and the space environment.
Fast retrieval code for aerosol and gaseous
profiles in the Martian atmosphere for limbsounding observation:
Limb-sounding observations are still largely unexploited, due to their huge computation cost,
and complexity of full light scattering processes
in spherical geometry. On the other hand, limbsounding is essential for understanding the mutual coupling system of the Martian atmosphere between the lower and upper atmosphere, as done
by MCS onboard MRO (e.g., Heavens et al.,
2018). In this study, we present a new retrieval
code used to invert the limb observations using a
Bayesian approach (Rodgers, 2000), to retrieve
the vertical profiles of dust and water ice density

and their particle size.
In this scheme, the forward model, JACOSPAR,
is a full radiative transfer code that accounts for
multiple scattering of the sunlight photons by the
atmospheric aerosols, to model the radiances
with high precision (Iwabuchi, 2006; Mahieux et
al., 2019). We aim to retrieve vertical profiles for
dust and water ice cloud effective radius and
number density, which all show clear absorption
and/or scattering structure in the considered
wavelength region. JACSOPAR uses the backward the Monte Carlo and dependent-sampling
method to efficiently calculate multiple wavelength regime (Marchuk et al., 1980). It calculates the scattering for a given number of wavelength values and interpolates the radiance at the
other wavelengths. JACOSPAR accounts for the
instrumental field of view in its calculations.
JACOSPAR also computes precise analytical
Jacobians relative to the radiances with respect to
the absorption and scattering extinction profiles.
They are used to derive the Jacobians to the aerosols number density and effective radius, which
are used in the Bayesian algorithm. We compute
the aerosol’s single scattering albedo, phase function and extinction coefficients using the Mie
theory (Wiscombe, 1980), for altitude constant
modified-gamma size distribution taken from
Kleinbohl et al. (2009), using the refractive index
of dust and water ice from Wolf and Clancy
(2003) and Warren (1984), respectively. We implemented the Bayesian algorithm approach developed by Rodgers (2000) using the LevenbergMarquardt method. Based on the a priori atmosphere obtained from the GEM-MARS (Neary et
al., 2018), we fit the logarithm of the different
aerosol’s number densities and effective radius,
assuming temperature and pressure conditions
obtained from Mars Climate Database (MCD)
(Millour et al., 2018) for the latitude, longitude,
time and local solar time observation mean value.
This is the first time such an algorithm is applied
to the limb retrievals for both dust and water ice
simultaneously. Although the gaseous profiles
can also be retrieved by this method, this will be
done in near-future development. Here we focus
on the retrievals of aerosols number density and
effective radius.
The forward model was validated by comparing
it with other existing codes. The comparison implied that the derived radiances agree within 3 %
in the nadir-geometry case, and within 1 % in the
limb-geometry case. The retrieved accuracy and
the sensitive altitude range were evaluated by the
synthetic spectra given by known profiles of dust
and water ice. Under standard dust conditions,
the dust and water ice effective radius and number densities were successfully obtained for altitudes below ~40 km (Figure 1). In this case, the
retrieved number densities of dust and water ice

were distributed within roughly ~10-2 /cm3 accuracy (~100%) of the true values. The retrieved effective radius of dust and water ice also show
good agreements within ~20% of the true values.
The code needs further evaluation under various
atmospheric conditions to study the sensitive altitude range and accuracy of the retrieved parameters. In addition, the effect of surface albedo
needs to be evaluated. We will discuss the comparison with the recent study by D’Aversa et al.
(2022). This code is intended to be applied to the
inversion of the NOMAD (Vandaele et al., 2015)
onboard ExoMars TGO limb observations, in addition to MIRS onboard MMX.

doi:10.1029/2009JE003358.
Wolf and Clancy (2003), J. Geophys. Res., 108,
5097, doi:10.1029/2003JE002057.
Warren (1984), Applied optics, 23, 1206.
Neary et al. (2018), Icarus, 300, 458-476.
https://doi.org/10.1016/j.icarus.2017.09.028
D’Aversa et al (2022), Icarus, 371, 114702.
https://doi.org/10.1016/j.icarus.2021.114702
Kuramoto et al (2022), Earth Planets Space, 74.
https://doi.org/10.1186/s40623-021-01545-7
Ogohara et al (2022), Earth Planets Space, 74:1.
https://doi.org/10.1186/s40623-021-01417-0

References:
Barucci MA et al (2021), Earth Planets Space
https://doi.org/10.1186/s40623-021-01423-2
Forget et al. (2007), J. Geophys. Res., 112,
doi :10.1029/2006JE002871.
Heavens et al. (2018), Nature astronomy, 2, 126-132.
https://doi.org/10.1038/s41550-017-0353-4
Iwabuchi (2006), J. Atmos. Sci., 63, 2324.
Spiga et al. (2007), J. Geophys. Res., 112,
doi :10.1029/2006JE002870.
Rogers (2000), Inverse methods for atmospheric
sounding, World Scientific.
Mahieux et al. (2019), Symposium on Planetary
Sciences 2019, Sendai, Japan.
Marchunk et al. (1980), Springer Series in Optical
Sciences.
Millour et al. (2018), Science Workshop “From Mars
Express to ExoMars”, ESAC, Madrid, Spain.
https://www.cosmos.esa.int/documents/1499429/1
583871/Millour_E.pdf
Winscombe (1980), Applied Optics, 19, 9.
Kleinbohl et al (2009), J. Geophys. Res., 114,

Figure 1. Comparison of dust and water ice clouds number density between (i) true value, (ii) a priori,
and (iii) retrieved profiles (top). Jacobian profiles of dust and water ice clouds number densities (bottom).