A NETWORK OF SURFACE PRESSURE OBSERVATIONS
CONSTRAIN AEROSOL FORCING IN MARS CLIMATE MODELS

TO

R. J. Wilson, NASA Ames Research Center, Moffett Field CA, USA, (Robert.j.Wilson@nasa.gov), M.A. Kahre,
NASA Ames Research Center, Moffett Field CA, USA.

Introduction: The evolving distribution of
radiatively active dust and water ice clouds plays a
major role in modulating the seasonal and
interannual variation in the thermal forcing of the
Martian atmosphere. Thermal tides are the globalscale atmospheric response to the diurnally varying
thermal forcing due to aerosol heating within the
atmosphere and radiative and convective heat transfer from the surface. The tide includes westward
propagating (sun-synchronous) waves driven in
response to solar heating, as well as nonmigrating
waves that result from zonal variations in the
thermotidal forcing caused by variations in topography and surface thermal properties and in the distribution of aerosols (dust and water ice clouds).
The migrating tides are of particular interest,
since they are generally directly responsive to the
aerosol distribution. However, distinguishing these
tides from the mix of additional nonmigrating tides
is difficult with only a limited number of surface
observations. For example, Figure 1 shows the seasonal evolution of the amplitude of the diurnal (S1)
and semidiurnal (S2) harmonics of surface pressure
observed by the Rover Environmental Monitoring
Station (REMS) aboard the MSL rover Curiosity in
Gale crater (4.5°S, 137°E) for Mars Years 33 and 34.
The figure highlights the very close correlation between the amplitude of S2 and the evolving global
column dust opacity during the MY34 global dust
storm. By contrast, the observed S1 response is less
easily interpreted. It is surprisingly strong at the start
of the regional dust storm at Ls = 323° while the S2
response remains relatively weak, consistent with the
global opacity.
It has become evident that radiative forcing by
water ice clouds contributes significantly to the
thermal balance particularly in the aphelion season,
although details are still not well-constrained. Another poorly constrained aspect of aerosol forcing is
the presence and impact of vertical variation of dust
(detached dust layers) on thermal forcing. We suggest that a modest improvement in the spatial coverage of diurnally-resolved surface pressure can be an
effective way of isolating tide modes that can be
diagnostic of aerosol forcing in the Mars atmosphere.
The recent acquisition of surface pressure data by
the Mars2020 mission in Jezero crater (18.5°N,
77.2°E) provides longitude coverage in the tropics,
that complements the 5-year MSL record, the fouryear record at Viking Lander 1 (VL1 at 22.5°N,
312°E), and the 1+year record at InSight (4.5°N,

135°E). The notable lack of interannual variability in
Martian climate during the aphelion season (Ls=0135°) allows these sets of lander data to be considered as a 4-station tropical network (Figure 2).
This presentation will describe an approach by
which a global perspective of the evolving diurnal
surface pressure response can be gained through the
use of high-resolution Mars global climate model
(MGCM) simulations. MGCMs include parameterizations that yield a thermal forcing field from distributions of dust and water ice clouds. Simulated atmospheric temperature and surface pressure are then
obtained consistently as the model responds to the
thermal forcing. The extent to which the simulated
tide fields correspond to observations provides insight into how well the thermal forcing field is represented. A better understanding of the global and
local scale influences on the diurnal variability of

Figure 1. (a) The seasonal variation of the diurnal tide
amplitude S1 at the MSL site for MY34 (blue) and
MY33 (cyan). Amplitudes are normalized by the local
diurnal-mean surface pressure. (b) As above but for the
semidiurnal tide, S2 at the MSL site. The tropical (30°S30°N) zonal-mean column dust visible opacity variations (from DGDM climatology) are shown as red and
magenta lines for MY34 and MY33, respectively. The
opacity has been scaled ( scale= 0.9 + 0.75) to emphasize the close correlation with S2.

surface pressure is critical for detailed comparisons
between atmospheric models and observations.
Thermal Tides and Aerosol Forcing: An
observed tide harmonic, Sn, at a lander site

Figure 1. (a) The seasonal variation of the diurnal tide

represents the sum over all zonal wavenumbers,
including the corresponding migrating component
and additional eastward and westward propagating
nonmigrating components. Migrating tides include
DW1, SW2, TW3, and QW4, respectively for the
westward
propagating
diurnal,
semidiurnal,
terdiurnal, and quad-diurnal migrating tides. There is
a roughly linear relationship between the SW2 amplitude and the dust column optical depth, which
makes this mode an effective proxy for globally
integrated thermal forcing. Moreover, due to the
meridionally broad and vertically deep structure of
the dominant Hough mode associated with SW2, the
pressure response is relatively insensitive to the
details of the vertical and latitude distribution of
thermal forcing. By contrast, the surface pressure
response of the migrating diurnal tide DW1 is weaker for a vertically extended dust distribution than for
a more shallowly confined distribution with equivalent column optical depth. Thus, elevated dust layers
or water ice clouds could have an influence on the
observed S1 response.
The most prominent nonmigrating tides are the
resonantly enhanced, eastward propagating diurnal
and semidiurnal Kelvin waves, DE1 and SE2, with
zonal wavenumbers 1 and 2, respectively. These
waves are forced by zonal wave 2 and 4 components
of thermal forcing, including the influence of topography and aerosol [Wilson and Hamilton, 1996].
Discussion of the influence of longitude variations of
aerosol on DE1 is deferred to a later section.

Figure 2. (a) The simulated spatial variation of normalized diurnal pressure variability, S1, in % of diurnal-mean
surface pressure for Ls= 105°. The locations of MSL,
InSight, Mars 2020, VL1 and VL2 are indicated. The
prominent zonal wave 2 modulation is due to the interference between the westward (DW1) and eastward (DE1)
diurnal tide components.

MGCM Simulations: We have been using a
high-resolution version of the NASA Ames MGCM
to simulate diurnal variability in surface pressure at
scales ranging from that of Gale crater to the planetary scale, thus distinguishing the pressure signature
of local topographically driven circulations from that
of the global tide. The results from the predecessor
GFDL version of the model are presented in Wilson
et al. (2017). More recent simulations of the MSL

tides during the MY34 global dust storm are described in a manuscript in preparation.
The NASA Ames MGCM uses a finite volume
dynamical core in a cubed-sphere geometry, which enables very high resolution simulations on a relatively
uniform grid. The physics included in the MGCM are
described in Kahre et al. (this meeting). Briefly, we
include many options for the handling of dust and
water ice clouds, including highly controlled prescriptions for their distributions and radiative effects.
We have performed such simulations (C48 and
C384) to examine the local and global-scale surface
pressure responses. The C384 simulation has a resolution of 0.25° x 0.25° (~15 km). We have annual
simulations at both resolutions. As we show below,
the MGCM does reasonably well at capturing many
of the tidal components discussed here.
The spatial distribution of normalized S1 amplitude at Ls=105° in a C384 simulation is shown in
Figure 2. The large-scale response is dominated by a
combination of DW1 and DE1, yielding the prominent zonal wave 2 modulation in diurnal tide amplitude. In the same season, there is a strong zonal
wave 4 pattern in the S2 response (not shown), seen
in previous simulation studies [Wilson and Hamilton,
1996; Guzewich et al., 2016].

Figure 3. Box plots of equatorial tide mode amplitudes
from the C384 simulation of the aphelion season derived from a space-time spectral analysis identifying
eastward and westward propagating diurnal (a) and
semidiurnal (b) components. The horizontal axis is
zonal wavenumber. The vertical extent of the blue
boxes spans the range of the middle two quartiles of
amplitude in the Ls=0-150° season, while red stars represent the outlier values. The DW1/DE1 pair dominates
the diurnal tide (S1) variability while the SW2/SE2 pair
dominates the semidiurnal tide (S2) variability.

Figure 3 shows the results of a space-time analyis
of the simulated equatorial surface pressure field,
indicating the range of amplitudes of the eastward
and westward propagating components of the
equatorial diurnal and semidiurnal pressure fields in
the Ls = 0-150° season. It is evident that the
dominant contributions to S1 structure and seasonal
variability are associated with the DW1/DE1 pair of

tide modes, with smaller contributions from other
waves, including DW2 and DW3. Similarly, the
SW2/SE2 pair dominates the structure and
variability of S2 in the aphelion season, with a lesser
contribution from SW1. The meridional structures
for both pairs of tide modes are quite broad so that
changes in component amplitudes and phases would
be reflected in similar tide changes throughout the
tropics. MGCM simulations support the expectation
that the migrating tide (SW2) accounts for the
majority of the S2 response during dustier seasons.

Figure 4. Simulations of DE1 and SE2 equatorial
tide amplitudes with realistic dust scenarios for
MY24 and MY26 (green and red, respectively). The
black curves show results using zonally averaged dust
scenarios

Figure 4 illustrates the sensitivity of the DE1 response to zonal variations in dust. Two simulations
were carried out using the seasonally-varying but
zonally averaged dust column scenarios for MY24
and MY26 compiled by Montabone et al. (2015).
These yield nearly identical and smoothly varying
seasonal variations in DE1 and SE1 in spite of the
differing histories of regional dust storms in the two
Mars years. The seasonal variation of the resonantlyenhanced Kelvin waves (DE1 and SE2) show a
strong preference for the two solstice seasons, with a
clear emphasis on the Ls = 90° season. Re-running
the simulations with the fully variable dust scenarios
yields
substantial
variability
theSE2
DE1
that
Figure
4. Simulations
of DE1inand
tideresponse
equatorial
realistic dust
for MY24
is amplitudes
associated with
with changes
to thescenarios
zonal wave
2 distriand MY26
(greendust.
and red,
respectively).
black
bution
of column
These
variations The
are particucurves
show
results
zonally
averaged
dust
larly
strong
at the
onsetusing
of the
pre- and
post-solstice
scenarios.
regional storms. The 3 storm events associated with
the Chyrse channel lead to significant amplification
of DE1, while the Isidis flushing event at Ls=208° in
MY26 results in the phasing of the zonal wave-2
component of the dust pattern that suppresses the
DE1 response. It is likely that rapid DE1 amplification is present at the start of the two MY34 dust
storm events seen in Figure 1. There are also systematic, though weaker influences on the DE1 response in the aphelion season. While these simulations did not include water ice clouds, it is reasonable to anticipate that zonal variations in the tropical
water ice cloud belt would induce changes as well.

Aphelion Season Observations: The amplitudes and phases of diurnal and semidiurnal pressure
harmonics at the 4 network sites are shown in Figure
5. The synchronized decline and increase in S2 amplitude at VL1, INS, and MSL centered about Ls
=90° is due to the destructive interference between
SW2 and SE1 at the lander longitudes, separated by
180 degrees. By contrast, S2 at the longitude of
Jezero crater has a relative maximum around solstice
due to constructive interference between SW2 and
SE2. Figure 6 shows a comparison of simulated and
observed S1 amplitude. Similar comparisons can be
made with the amplitudes and phases of S2 and higher order harmonics. The simulation also permits
consideration of the constituent tide modes. In this
simulation the model tends to overpredict the amplitude of S1. This is attributed to an overly strong
DW1 response, which likely is due to the absence of
upper-level heating associated with radiatively active
water ice clouds. The diurnal Kelvin wave (DE1) is
also likely somewhat too strong, as suggested by the
mismatch in amplitude (and phase, not shown) at
Jezero crater. The simulation indicates that the difference in S1 amplitude between MSL and INS is
largely due to the large-scale influence of the Mars
dichotomy on DW1 and the local-scale influence of
Gale crater; both enhance the response at MSL
(4.5°S) vs InSight (4.5°N).
Ongoing Research and Conclusions: We are
developing a fitting procedure where the network
observations of the amplitudes and phases of S1 and
S2 are used to estimate the evolving amplitudes and
phases of DW1, DE1, SW2, and SE2 in the actual
Mars atmosphere. We take advantage of the robustness of the predicted meridional structure of these
key modes to be fitted, which enables a least-squares
fit for their amplitude and phases as they evolve
through the aphelion season. We have also been
examining the sensitivity of the tide response to
aspects of the aerosol forcing like the strength of the
water ice cloud heating, and the vertical distribution
of dust and water ice clouds. Thus, we anticipate that
our estimates of the amplitudes of these diagnostic
tide modes will provide useful guidance on assessing
the radiative forcing by dust and water ice clouds.
References
Guzewich, S.D., et al. (2016), Atmospheric tides in Gale
crater, Icarus, 268, 37-49,
Kahre et al., (2022). 7th MAMO Conference, Paris, France.
June 14-17, 2022.
Montabone et al., (2015), Eight-year climatology of dust
on Mars. Icarus.
Wilson, R.J., and K.P. Hamilton (1996), Comprehensive
model simulation of thermal tides in the martian atmosphere, J. Atmos. Sci. 53, 1290-1326.
Wilson, R.J., J.M. Murphy, and D. Tyler (2017), Assessing
atmospheric thermal forcing from surface pressure data; Separating thermal tides and local topographic influence Sixth International Workshop on the Mars At-

mosphere: Modeling and Observations, Granada,
Spain,
Jan
17-20,
2017.
http://www-

mars.lmd.jussieu.fr/granada2017/abstracts/wilson
_granada2017.pdf

Figure 5. (a) Observations of the normalized diurnal harmonic amplitude of surface pressure, S 1 at 4 sites on the
Martian surface. These include Mars2020 (JZO, MY36, in cyan), MSL (MY33, in red), InSight (INS, green) and
Viking Lander 1 (VL1, 4 years, in black) (b) Semidiurnal tide amplitude, S2. (c) Diurnal tide phase. (d) Semidiurnal
tide phase.

Figure 6. (a) Comparison of simulated S1 tide amplitude (black line) with observations by MSL in Gale crater (in
brown). Also shown are the simulated seasonally-varying DW1(red) and DE1 (green) amplitudes at the latitude of
the lander. The magenta curve (R1) is the amplitude contribution attributed to the localized topographic response (as
defined in Wilson et al. 2017). Note that wave phase need to be considered in combining wave component contributions.

to the tide signal at a given longitude. (b) Simulation results for VL1. (c) Simulation results for InSight. (d) Simulation results for Mars2020 in Jezero crater.