INTERNAL DYNAMICAL VARIABILITY IN THE NASA AMES MARS
GENERAL CIRCULATION MODELS
A. F. C Bridger, Department of Meteorology & Climate Science, San Jose State University, San Jose, CA, USA
(alison.bridger@sjsu.edu), M. A. Kahre, and R. J. Wilson, NASA Ames Research Center, Moffett Field, CA,
USA.

Introduction: General Circulation Models
(GCMs) have been used for decades in order to understand and predict how the larger-scale motions
and associated structures and phenomena develop
and behave. One use of GCMs involves varying external forcings in order to understand the atmosphere’s sensitivity to forcings. This can include
(Earth) sea-surface temperatures or CO2 levels, and
(Mars) dust storms or the surface albedo field. Interpretation of results depends on an understanding of
the internal variability of the GCM.
We seek here to extend an earlier study [1] to
characterize the internal variability of the NASAAMES MGCM. There are two incarnations of this
model: (a) the Legacy model; and (b) the new FV3based version. The Legacy model has been under
development for decades and is now “frozen”. It was
recently released for general use [3], [2]. The newer
FV3-based version is built around the finite-volume
dynamical core developed at NOAA/GFDL. The two
models share common physics packages but have
different dynamical cores. Our earlier study utilized
an earlier version of the Legacy model. Our first step
in the current study is to repeat those earlier experiments using the most recent Legacy code. After that,
we will conduct parallel experiments with the new
model, and thus determine whether variability we
find is characteristic of both models.
Both the terrestrial and Martian atmospheres have
structures which accomplish poleward heat transfer.
From a zonal- and time-average perspective, these
include the Hadley cell, and both stationary and transient eddies. The net poleward heat transfer is required (i.e., forced by) the pole-to-equator net
radiative heating gradient on each planet. The breakdown amongst the various components is less well
understood. For example, is there any a priori expectation that the Hadley cell should be responsible for
the bulk of the heat transport (on either Earth or
Mars)? We established earlier [1] that this breakdown can vary from year-to-year (Y2Y) in multi-year
simulations: this can be viewed as one characterization of internal variability in an MGCM.
Procedure: Our first task is to repeat the earlier
analysis but with the final/frozen form of the Legacy
code. The chief difference between this and the earlier code used is in the updated physics packages,
many of which were informed by new observations.

Assuming we find year-to-year variability in the core
dynamic behavior (which we do), the physics codes
give us the opportunity to see whether variability is
suppressed/enhanced when we add e.g., clouds or a
water cycle. The model is run for several years under
these conditions, assuming MY31 dust opacity which
is allowed to repeat every year. For now, we choose
to examine the Ls 270 season, since this is characterized by strong transient and stationary wave activity
(e.g., as observed since Viking and via modeling).
For each year of the simulation, we extract 30 sols of
data centered around Ls 270. We extract meridional
winds and temperatures, and use them to construct
the various poleward heat flux fields outlined in [4]
(their Eq 4.10). Specifically, the total heat flux can
be broken down to show that the contributions of: (i)
the Hadley cell (time- and zonally-averaged circulations); (ii) stationary eddies; (iii) transient eddies;
and (iv) transience in the Hadley circulation. Not
included in this formulation are contributions (if any)
from the condensation flow associated with cap formation/sublimation.
Results: Each quantity discussed above is computed for the 30 sols of data. Stationary eddy heat
fluxes are plotted for years two and three in Fig. 1.
Although the same general patten exists, there are
significant differences in amplitude [O(10-20%) below 10 Pa].

such as the presence/absence of a water cycle, the
dust cycle, and surface conditions such as albedo.
These studies will inform the overall importance of
the internal variability (which is important in terrestrial models but might not be for Mars).

Figure 1. Latitude-height distribution of poleward heat
fluxes due to stationary waves in years 2 (upper) and 3 of a
Legacy MGCM simulation. Warmer shades indicate higher
values.

Transient eddy heat fluxes are plotted for years
two and three in Fig. 2. Differences here are as pronounced as above. Note that this includes activity at
very low levels around 60N.

Summary: Simulations of the Martian atmosphere over multiple years show interesting Y2Y variability in one field examined – poleward heat flux
(we have earlier shown Y2Y variations in surface
stress/dust deflation [1]). Whether these variations
have any practical importance – e.g., in understanding the generation of planetary dust storms or in mission planning – remains to be established. However,
an understanding of the dynamical variability in a
GCM is necessary in order to fully understand a
model’s response to an imposed forcing (on Earth for
example, the response to enhanced greenhouse gases;
on Mars, for better understanding Y2Y variations in
the dust cycle).

References:
[1] Bridger & Haberle, 2004: Intrinsic variability in
dust deflation potential in multi-annual MGCM
simulations. Division of Planetary Sciences Annual
Meeting, Louisville, KY.
[2] Haberle et al, 2019: Documentation of the
NASA/Ames Legacy Mars Global Climate Model:
Simulations of the present seasonal water cycle. Icarus, 333, pp 130-164.

Figure 2. As Figure 1 but for transient eddies.
These early results confirm the presence of Y2Y
variability in the Legacy MGCM. Meanwhile, the
zonally- and time-averaged circulations are essentially the same Y2Y (not shown). The differences in
eddy heat fluxes (stationary and transient) exist despite this. Our results could mean that there are
more/fewer storms Y2Y, or that they have weaker/stronger amplitudes, or that their vertical and/or
horizontal structures vary Y2Y, allowing more/less
efficient heat flux (all of which will be examined).
We are continuing to examine Y2Y variability in
the Legacy code via longer simulations. We plan
next to conduct similar experiments with the new
FV3-based model. Provided both models show similar degrees of internal variability, we plan to then
conduct studies of: (i) the total heat flux Y2Y; (ii)
heat fluxes before/after the solstitial pause; (iii) fluxes in the southern winter; and (iv) whether Y2Y variations respond to (strengthen? diminish?) conditions

[3] NASA Ames Legacy Mars GCM Virtual
Workshop Tutorial, November 2-4, 2021.
https://www.nasa.gov/mars-climate-modelingcenter-ames/MarsGlobalClimateModelTutorial: Mt.
View, CA.
[4] Peixoto & Oort, 1992: Physics of Climate. American Institute of Physics.