MEASURING THE 13C/12C IN CO2 IN THE LOWER ATMOSPHERE OF
MARS WITH NOMAD/TGO: CHALLENGES AND INTERPRETATION
G. Liuzzi, G. L. Villanueva, S. W. Stone, S. Faggi, V. Kofman, NASA Goddard Space Flight Center, Greenbelt, MD, USA (gliuzzi@american.edu), S. Aoki, Graduate School of Frontier Sciences, The University of Tokyo, Kashiwa, Japan, J. Alday, School of Physical Sciences, The Open University, Milton Keynes, UK, L.
Trompet, A. C. Vandaele, Royal Belgian Institute for Space Aeronomy, Belgium.

Introduction:
The two infrared spectrometers onboard the
ExoMars Trace Gas Orbiter (TGO) spacecraft, the
Atmospheric Chemistry Suite (ACS) and the Nadir
and Occultation for MArs Discovery (NOMAD)
have been providing observations since April 2018
that are a unique source of information to unveil the
vertical structure of the atmosphere of Mars in a level of detail never previously seen. In this work, we
focus on understanding how the NOMAD data can
be useful to constrain the isotopic composition of C
in CO2. This has been measured on several occasions
on the surface [1], and in near-surface atmospheric
samples [2,3], and is of great interest as CO2 is the
main C reservoir in the atmosphere of Mars, allowing
to trace the history of atmospheric loss and differences with other inner solar system bodies. Additionally, proper characterization of the uncertainties of
derived isotopic ratios is a very important proxy towards a better understanding of the instrument limitations and performances.

Data and analysis:
NOMAD is a 3-channel spectrometer working in
the spectral ranges 0.2-0.65 μm (UVIS channel) and
2.2-4.3 μm (SO and LNO channels). The data taken
by the Solar Occultation (SO) channel measurements
in the infrared channel is used in this work. SO is an
echelle grating spectrometer with relatively high
spectral resolution (λ/dλ~17,000), and an Acousto
Optical Tunable Filter (AOTF) used to instantaneously switch the observed diffraction orders, with a
typical cycle of 1 second. This yields a vertical sampling of less than 1 km from the surface to 250 km of
altitude. While typically NOMAD cycles through 5
or 6 diffraction orders at every occultation, this setup
is often not adequate to measure 13C/12C. Expected
differences between this isotopic ratio on Earth and
the one on Mars are indeed very small (< 0.1, or
equivalently 100‰), therefore it is important, when
attempting to measure it, to pinpoint the other parameters that can create a bias in the measured isotopic ratio. Because each diffraction order measured
by NOMAD covers only a limited number of spectral
lines with similar degrees of saturation, and mostly
similar temperature dependencies, the most effective
way to quantify simultaneously the temperature,
12
CO2 and 13CO2 is to choose numerous diffraction

orders measured simultaneously, with spectral lines
with similar degrees of saturation. For these reasons,
the data we use in this work are the Full Scan data,
where a large set (or all) of diffraction orders are
acquired at a few altitudes. Despite the very coarse
vertical sampling (3 to 25 km), these scans allow to
collect data in several diffraction orders that are not
measured frequently, maximizing the information
gathered for 12C16O2, 13C16O2 and temperature. For
this work, we have used full scans acquired between
Apr/2018 (LS ~160 MY34) and Dec/2021 (LS ~137
MY36), in particular the spectra from orders 141
(3170-3200 cm-1), 145 (3255-3285 cm-1), 146 (32783308 cm-1), and 148 (3332-3363 cm-1), for a total of
~2700 spectra per order. From order 148 the average
rotational temperature along the line of sight can be
derived with an accuracy of about 5 K [4], which
then can be used to derive the column density at a
specific tangent altitude of 12CO2 (order 141) and
13
CO2 (orders 145 and 146). In all these orders, various lines with different J are present, providing proper constraints on the abundances of the two isotopes.
The accuracy of the retrievals is also possible thanks
to recent calibration efforts for the AOTF transfer
function and the ILS of the SO channel, which are
fully parameterized with the diffraction order [5].
Retrievals are performed using the Planetary Spectrum Generator (PSG, [6,7]), a full general purpose
radiative transfer code, which contains a dedicated
retrieval module based on Optimal Estimation and is
configured with all the instrumental parameters of
NOMAD SO natively accounted for. The use of diffraction orders that are close in frequency minimizes
altitude biases, as all the orders will be simultaneously acquired within ~200 meters, effectively obtaining
well collocated measurements.
The intensity of the lines in the observed orders
are particularly suitable to obtain meaningful values
for the C isotopic ratio in the interval between 10 and
50 km in altitude. Below 10 km, the combination of
the atmospheric opacity and line saturation will degrade the accuracy, while above 50 km the Signal-toNoise Ratio (SNR) of the lines will be too low to
obtain significant values.
Results and discussion:
Once all the retrievals are completed, we filter
the values by excluding values with low SNR and
those spectra where the intensity of the spectral lines
is smaller than 5 times the radiometric noise. Once a

stringent selection is done, a total of ~2000 isotopic
ratio single values are selected. Figure 1 shows the

mittances across an order as they vary with the CO2
column with the radiometric noise, as seen in Figure
3.

Figure 1. Temporal series of the retrieved values
of 12CO2 (red, from order 141), and the 13CO2 (order
145 and 146, blue and green respectively). Enhancements of both isotopes due to the GDS can be
seen at the beginning of the series.
temporal series of the retrieved values of the two
isotopes of CO2 in terms of scaler of the density as
prescribed by co-located GEM model profiles [8].
The abundance of the two isotopes with respect to
the GEM model are the same within the dispersion of
the points, and we see that the CO2 density is for
both isotopes enhanced during the dusty season (in
particular the GDS during MY34). There does not
appear to any significant variability of the C isotopic
ratio with location, season, and atmospheric state.
However, because of the limited temporal coverage,
it is difficult to derive specific seasonal trends. The
most significant product of this dataset is an average
vertical profile of the C isotopic ratio. This is shown
in Figure 2, together with the full set of valid points.
The color of each point represents the difference
between the temperature as given by GEM and the
one derived using the data of order 148. A value of
on the x-axis equal to one means that the measured
isotopic ratio is equivalent to the Earth Vienna Pee
Dee Belemnite (VPDB) for carbon, 13C/12C = 1.123
× 10−2). The single points show no significant trend
in terms of temperature vs. isotopic ratio, meaning
that the information about temperature and the abundances of the two isotopes are independent of each
other.
The observed dispersion in the single values increases with altitude because the SNR of the spectral
lines decreases. One of the most important questions
to assign the correct uncertainty to the average
13 12
C/ C at each altitude is the determination of
whether the observed dispersion in the single spectrum results is of stochastic nature, meaning that it is
not dominated by systematics, rather it is related to
the limitations of the instrument and the accuracy of
spectroscopy. To answer this question, we have conducted a specific analysis that quantifies the expected
uncertainty of a single measurement depending on
the observed column of CO2 and the prescribed radiometric noise. By using the simulated curve of
growth of the lines of the two isotopes in the observed orders, we compare the median of the trans-

Figure 2. Full set of retrieved values of 13C/12C
(x-axis, 1 meaning VPDB standard) as a function of
the altitude (y-axis) and related error bars. The color
of each point represents the retrieved difference in
temperature along the line of sight with respect to the
GEM model. The expected dispersion of the points
as derived from the analysis of the curve of growth
(see text and Figure 3) is shown by the dashed lines.
This allows, for every column amount of CO2 (namely, a certain altitude), quantification of the uncertainty in the derived column as related to the saturation
of the lines (spectroscopic limitations) and the instrument limitations (resolving power, continuum
effects). It can be clearly seen that the uncertainty
diverges for columns below the one corresponding to
an average altitude of 50 km, consistent with the observed spread of the points in Figure 2. This analysis
enables quantification of the expected dispersion of
13 12
C/ C, which is also reported in Figure 2.
The fact that more than 90% of the points, at every altitude, fall within the derived 3-sigma error from
the analysis of the curve of growth justifies the fact
that the observed spread of the points can be treated
in a stochastic fashion, meaning that averaging several measurements to achieve higher precision is
possible. By doing that, we quantify the isotopic ratio
in the lower atmosphere (10-25 km) is 1.027+/-0.030
VPDB, largely consistent with the value observed by
MSL in surface minerals and the atmosphere. The
value is fairly constant throughout the atmosphere
with oscillations within the error bars. Upper atmosphere values (>50 km) do not show any inconsistency with the values reported using ACS observation
above 60 km [9].

Figure 3. Simulated curve of growth for the
CO2 lines in order 146. The x-axis shows the column amount of CO2 along the line of sight, while the
y-axis shows the average transmittance across the
order (net of the median of the transmittance, to account for the continuum effects). The vertical error
bars correspond to the radiometric noise scaled by
the number of pixels (320 for SO), and the red lines
correspond to the average column at 5 and 65 km of
altitude.
13

Conclusions:
Observations by NOMAD have been used to
measure an average vertical profile of the CO2 C
isotopic ratio in the lower atmosphere of Mars. The
isotopic ratio shows no significant seasonal or altitudinal trends. Besides the values reported in this
work, this analysis has explored the main instrumental factors limiting the precision of the measurement.
We conclude that the precision of the measurement is
essentially limited by the radiometric SNR, and by
the precision with which the spectral continuum is
quantified, which pays an essential role when spectral lines are close to the saturation. If several lines
(namely, several diffraction orders) are used in the
analysis, the accuracy of the atmospheric temperature
is not the main limiting factor.

References:
[1] House et al. (2022), “Depleted carbon isotope
compositions observed at Gale crater, Mars”, PNAS,
doi: 10.1073/pnas.2115651119.
[2] Webster et al. (2013), “Isotope Ratios of H,
C, and O in CO2 and H2O of the Martian Atmosphere”, Science, doi: 10.1126/science.1237961.
[3] Mahaffy et al. (2013), “Abundance and Isotopic Composition of Gases in the Martian Atmosphere from the Curiosity Rover”, Science, doi:
10.1126/science.1237966.
[4] Liuzzi et al. (2021), “First Detection and
Thermal Characterization of Terminator CO2 Ice
Clouds With ExoMars/NOMAD”, Geophys. Res.
Lett., doi: 10.1029/2021GL095895.
[5] Villanueva et al. (2022), Under review,
Geophys. Res. Lett.
[6] Villanueva et al. (2018), “Planetary Spectrum
Generator: An accurate online radiative transfer suite
for atmospheres, comets, small bodies and

exoplanets”, J. Quant. Spectrosc. Rad. Trans., doi:
10.1016/j.jqsrt.2018.05.023.
[7] Villanueva et al. (2022), Handbook of the
Planetary Spectrum Generator, ISBN: 978-0-57836143-7.
[8] Neary & Daerden (2018), The GEM-Mars
general circulation model for Mars: Description and
evaluation, Icarus, doi:10.1016/j.icarus.2017.09.028.
[9] Alday et al. (2021), Isotopic Composition of
CO2 in the Atmosphere of Mars: Fractionation by
Diffusive Separation Observed by the ExoMars
Trace Gas Orbiter, J. Geophys. Res., doi:
10.1029/2021JE006992.