SEASONAL AND DIURNAL VARIATION IN THE VERTICAL
STRUCTURE OF THE MARTIAN NITRIC OXIDE NIGHTGLOW LAYER
Zachariah Milby, Nicholas M. Schneider, Sonal K. Jain, Kyle Connour, LASP, U. Colorado, USA
(nick.schneider@lasp.colorado.edu) Francisco González-Galindo, Instituto de Astrofísica de Andalucía (IAACSIC), Granada, Spain, Franck Lefèvre, Laboratoire Atmosphères, Milieux, Observations Spatiales
(LATMOS), CNRS, Sorbonne Université, UVSQ, Paris, France, François Forget, Laboratoire de Météorologie
Dynamique/IPSL, Institut Universitaire de France, Sorbonne Université, École Normale, Jean-Claude Gérard,
Laboratoire de Physique Atmosphérique et Planétaire (LPAP), Université de Liège, Liège, Belgium
Introduction:
Nitric oxide (NO) nightglow is an emission feature that can provide insight into the dynamics of the
Martian middle atmosphere (Figure 1). It is an atmospheric airglow phenomenon produced by the
chemi-luminescent relaxation of excited NO molecules and is the brightest middle-ultraviolet (MUV)
spectral feature in the Martian nightside atmosphere,
with the occasional exception of aurora (Schneider
et al., 2018). NO forms when carbon dioxide and
molecular nitrogen dissociate on the dayside, the
atoms get transported to the nightside and the resultant nitrogen and oxygen recombine. The reaction
leaves the resulting molecule in an excited state
which decays radiatively by giving off one photon.
Because the brightness is proportional to the recombination rate of nitrogen and oxygen atoms, it responds to the dayside photo-dissociation rates of
carbon dioxide (CO) and molecular nitrogen (N2), to
Hadley circulation from the dayside to the nightside
and to local downward flux into the lower atmosphere (Schneider et al., 2020). It is a phenomenon
distinct from NO dayglow, a solar-driven fluorescence proportional to the number density of NO
molecules rather than an instantaneous reaction rate
(Stevens et al., 2019).
The first observations of NO nightglow on Mars
came from data obtained by the Spectroscopy for the
Investigation of the Characteristics of the Atmosphere of Mars (SPICAM) instrument on the Mars
Express (MEx) spacecraft. Bertaux et al. (2005) reported the first detection of NO nightglow in Mars
year (MY) 32 during southern hemisphere winter.
They interpreted their results using simulations from
the LMD-MGCM and proposed that the density of
nitrogen is the limiting factor for NO nightglow
emission because it is predicted to be orders-ofmagnitude lower in abundance than that of oxygen.
SPICAM-based studies by Cox et al. (2008) analyzed vertical profiles with a 1-D model, showing
that they could be adequately explained by downwelling of atomic species or by eddy diffusion. They
also confirmed model predictions that the southern
winter pole would be brighter with a lower altitude
emission peak. Further studies by Gagné et al.
(2013) and Stiepen et al. (2015) analyzed nightglow
observations from additional orbits, but were hin-

dered in their statistical analyses by the limited dataset.
Methods:
We performed a comprehensive study of the variability of Martian nitric oxide (NO) nightglow’s
brightness and altitude. We used MAVEN Imaging
Ultraviolet Spectrograph (IUVS) limb data gathered
between Mars years 32 and 36. The IUVS instrument
is an imaging slit spectrograph with a scanning mirror that moves perpendicular to the slit, gathering
spectral information with two spatial axes. The NO
nightglow band system is captured by the MUV detector which has a wavelength range of 175 to 340
nm, a spectral resolution of 0.65 nm and good sensitivity over the wavelengths with the strongest NO
nightglow bands. MAVEN’s eccentric orbit slowly
evolves in latitude and local time, giving IUVS the
ability to collect data over a wide range of physical
conditions unavailable to previous studies. During
limb observations the motion of the scanning mirror
generates vertical profiles, capturing spectra at different altitudes (Jain et al., 2015). In addition to the
limb scans obtained during periapse passage (used in
(Stiepen et al., 2017)), we use limb scans obtained
from the inbound and outboard orbit segments just
before and after periapse.
General circulation models (GCMs) provide insights regarding the middle atmosphere and NO
nightglow. GCMs are the primary way by which we
understand the Martian middle atmosphere, as previous observations of this region are limited (Barnes et
al., 2017); consequently, any observational constraints we can provide will help further refine the
insights they provide. Previous studies of the Martian NO nightglow (Bertaux et al., 2005; Gagné et
al., 2013; Stiepen et al., 2015, 2017; Schneider et al.,
2020) interpreted their results using Laboratoire de
Météorologie Dynamique Mars GCM (LMDMGCM) simulations. While many aspects of
nightglow are matched by the model, persistent discrepancies remain. Most of these previous studies
found that the simulations did not accurately predict
some of the observed characteristics of the nightglow
layer—the model almost always predicted the peak
altitude of the layer to be tens of kilometers higher
and around half as bright as observations indicated.

The model is apparently underestimating some key
physical components, or otherwise incorrectly accounting for NO nightglow, that may have implications for the broader planetary circulation.
Results:
We investigated how the nightglow varied with
latitude and season. Our multi-year analysis showed
no systematic year-to-year variability of the NO peak
brightness or altitude. We found that in low-to-mid
latitudes, the nightglow was significantly brighter
and found at lower altitudes in the aphelion season
than the perihelion season. The nightglow was significantly brighter over the winter poles, with the
north pole being the brightest and peaking at lower
altitudes than than the southern pole. This confirms
the latitude-dependence in peak altitude and peak
brightness reported by Stiepen et al. (2017) and
Schneider et al. (2020), though with additional coverage. Our model simulations occasionally agreed
with these observations but generally underpredicted
the brightnesses. They consistently produced peak
brightnesses 15–20 km higher than our observations.
We also explored the diurnal cycle of the nightglow.
Our observations suggest the mesosphere is relatively steady during the aphelion season. The perihelion
season was more dynamic, producing a nearmidnight enhancement in the brightness. Our simulations did not predict the relatively calm aphelion
season that we observed. They produced some of the
features we observed during the perihelion season
but most notably did not predict the local time
ehancement that we noticed.

in the fit rather than geophysical variation in the longitude of the observations. This wave speed is slightly faster than the one reported by Schneider et al.
(2020), but within one standard deviation, and both
are consistent with the DE2 wave. The peaks were
also at approximately the same planetary longitudes
and the wave near 0° longitude exhibits the largest
fractional deviation in the early evening hours, suggesting we were able to detect the same equatorial
wave structures in the limb observations. The simulations in figure 2 (b) show a wave-3 structure as
well. However, the simulated wave propagates eastward at a slower 3.3 ± 0.9 ° h−1 and doesn’t exhibit
the same peak brightness near 0° longitude in the
observations.
We also observed a remarkable wave-3 structure
in the altitude of the peak limb brightness (Figure
2b), also smoothed in longitude by the same 30°wide kernel as the brightness. The maxima of the
wave structures in (a) and (b) do not coincide, but
appear appear to be offset from each other by about
45°. This offset of the waves is especially noticeable
in figure 2 (c) which shows the wave structures corrected for the 5 ° h−1 DE2 wave motion by removing the longitudinal shift relative to midnight. If
brightness and altitude were perfectly correlated or
anticorrelated, first-order explanations might present
themselves. But the 45° shift suggests a complex
interplay of multiple dynamical influences. Unfortunately, the model does not replicate much of the
wave-3 structure let alone its amplitude, so this
leaves a challenge for future modelers.
References:

Atmospheric tides are an insightful metric for
identifying atmospheric forcings and revealing how
the atmosphere has responded. Stiepen et al. (2017)
first reported the detection of a wave-3 structure in
the peak brightness of the nightglow. Schneider et al.
(2020) were able to detect similar waves and analyzed them using the broad spatial and temporal coverage of the imaging data. They interpreted their
results using LMD-MGCM simulations and they
determined the waves to be primarily diurnal, nonmigrating eastward-propagating wave-2 structures
(called DE2) with a small semi-diurnal westwardpropagating wave-1 component (SW1). The central
peak of their observed equatorial wave structure
moved eastward at a rate of 4.7°h−1.
Figure 2 (a) shows a pronounced wave-3 structure in the peak limb brightness between −45° and
45° latitude, smoothed in longitude by a 30°-wide
kernel. This confirms the existence of the waves
reported by (Stiepen et al., 2017) and Schneider et al.
(2020). We used least-squares fits to the peaks of the
waves to calculate an average eastward wave motion
of 4.2 ± 0.8 ° h−1. Note that unlike previous statistics in this paper this uncertainty represents the error

Bertaux, J.-L., et al., (2005). Science, 307(5709),
566–569. doi: 10.1126/science.1106957
Cox, C., et al. (2008). Journal of Geophysical
Research, 113(E8). doi: 10.1029/2007je003037
Gagné, M.-È., et al.. (2013). Journal of Geophysical Research: Planets, 118(10), 2172–2179. doi:
10.1002/jgre.20165
Jain, S. K., et al. (2015). Geophysical Research
Letters,
42(21),
9023–
9030.
doi:
10.1002/2015gl065419
Schneider, N. M., et al. (2020). Journal of Geophysical
Research:
Space
Physics,
125
(e2019JA027318). doi: 10.1029/2019ja027318
Stiepen, A., et al. (2015). Geophysical Research
Letters, 42(3), 720–725. doi: 10.1002/2014gl062300
Stiepen, A., et al. (2017). Journal of Geophysical
Research: Space Physics, 122(5), 5782–5797. doi:
10.1002/2016ja023523

Figure 1. The NO nightglow recombination process.

Figure 2. Wave propagation with local time between −45° and 45° latitude. (a) shows observed peak limb
brightness for a given local time bin and (b) shows the same for the simulations. In both, dashed lines show fits
to the peak brightnesses (c) shows the absolute difference from the average peak altitude and (d) shows the same
for the simulations. The grey dashed lines are replicated from the panels above, showing the peak brightness and
altitude are out of phase. (e) shows the brightness and altitude data from (a) and (c) averaged across all local
times, normalized and corrected for DE2 wave motion by removing the longitudinal shift relative to midnight.
(f) shows the same for the simulations in (b) and (d).