ArticlePDF Available

Abstract and Figures

Numerical modeling tools can be used for a number of reasons yielding many benefits in their application to planetary upper atmosphere and ionosphere environments. These tools are commonly used to predict upper atmosphere and ionosphere characteristics and to interpret measurements once they are obtained. Additional applications of these tools include conducting diagnostic balance studies, converting raw measurements into useful physical parameters, and comparing features and processes of different planetary atmospheres. This chapter focuses upon various classes of upper atmosphere and ionosphere numerical modeling tools, the equations solved and key assumptions made, specified inputs and tunable parameters, their common applications, and finally their notable strengths and weaknesses. Examples of these model classes and their specific applications to individual planetary environments will be described.
Content may be subject to copyright.
Space Sci Rev (2008) 139: 107–141
DOI 10.1007/s11214-008-9401-9
Neutral Upper Atmosphere and Ionosphere Modeling
Stephen W. Bougher ·Pierre-Louis Blelly ·
Michael Combi ·Jane L. Fox ·Ingo Mueller-Wodarg ·
Aaron Ridley ·Raymond G. Roble
Received: 29 February 2008 / Accepted: 9 June 2008 / Published online: 15 July 2008
© Springer Science+Business Media B.V. 2008
Abstract Numerical modeling tools can be used for a number of reasons yielding many
benefits in their application to planetary upper atmosphere and ionosphere environments.
These tools are commonly used to predict upper atmosphere and ionosphere characteris-
tics and to interpret measurements once they are obtained. Additional applications of these
tools include conducting diagnostic balance studies, converting raw measurements into use-
ful physical parameters, and comparing features and processes of different planetary at-
mospheres. This chapter focuses upon various classes of upper atmosphere and ionosphere
numerical modeling tools, the equations solved and key assumptions made, specified inputs
and tunable parameters, their common applications, and finally their notable strengths and
weaknesses. Examples of these model classes and their specific applications to individual
planetary environments will be described.
Keywords Planets · Thermospheres · Ionospheres · Numerical modeling
S.W. Bougher (
) · M. Combi · A. Ridley
Atmospheric, Oceanic and Space Sciences Department, University of Michigan, Ann Arbor,
MI 48109-2143, USA
e-mail: bougher@umich.edu
P.-L. Blelly
CESR, Toulouse, France
J.L. Fox
Wright State University, Dayton, OH 45435, USA
I. Mueller-Wodarg
Imperial College London, London, UK
R.G. Roble
National Center for Atmospheric Research, Boulder, CO 80309, USA
108 S.W. Bougher et al.
1 Introduction and Scope
The arrival of new measurements of the neutral upper atmospheres and ionospheres of
various solar system planets and moons over the past 4-decades from various spacecraft
missions has been astounding (e.g., see Mueller-Wodarg et al. 2008; Witasse et al. 2008;
Johnson et al. 2008, and other chapters from this book). These measurements have been
used to characterize the structure and dynamics of these atmospheric environments and to
compare them to one another. A corresponding evolution of modeling tools, from simple
to complex frameworks, has occurred over the same timeframe. These tools are commonly
used to predict upper atmosphere and ionosphere characteristics and to interpret measure-
ments once they are obtained. This chapter focuses upon various classes of upper atmosphere
and ionosphere numerical modeling tools, the equations solved and assumptions, specified
inputs and tunable parameters, their applications, and finally their notable strengths and
weaknesses. Examples of these model classes and their specific applications to individual
planetary environments will be described.
1.1 General Uses/Benefits of Modeling Tools
Numerical modeling tools can be used for a number of reasons yielding many benefits in
their application to planetary upper atmosphere and ionosphere environments. First and fore-
most, modeling tools are commonly utilized to understand the processes that maintain ob-
served atmospheric structures and drive their variations over various timescales (e.g., solar
cycle, seasonal, diurnal, etc.). Time varying inputs (e.g., solar) are often specified to drive
model simulations and monitor the resulting variations among the simulated fields. These
same model simulations can also be examined to provide a diagnostic analysis of the indi-
vidual terms of the solved equations; e.g. thermal and momentum balances. Such diagnostic
studies provide valuable insight into the underlying processes that are responsible for the
time variable features of the atmosphere. Spatially or temporally limited measurements are
commonly used to constrain model simulations in an effort to construct reasonable pre-
dictions outside available dataset domains and/or time periods. This effort places available
observations in a more general/global context. Such model predictions can also be used to
motivate new measurements and/or conduct more thorough data analysis studies combining
existing datasets.
Model simulations can also be utilized in data processing to facilitate the conversion of
raw measurements into useful physical parameters. A good example involves the analysis of
aerobraking accelerometer measurements (e.g. Withers 2006;Tolsonetal.2007). Raw ac-
celerations are typically calibrated to yield mass densities and corresponding scale heights.
The estimation of temperatures from these density scale heights requires independent infor-
mation about the composition of the thermosphere (e.g. relative abundance of atomic and
molecular species). Global thermospheric general circulation model (TGCM) simulations
for Mars are available to provide a first estimate of these global abundances, enabling neu-
tral temperatures to be estimated from scale heights (see Sect. 4.2.5 below).
The comparative approach to investigating planetary upper atmospheres is becoming in-
creasingly fruitful as new information from various planet atmospheres is assimilated using
state-of-the-art modeling tools (e.g., Bougher et al. 2002). A comparison of the basic fea-
tures of the structure and dynamics of planetary upper atmospheres and ionospheres can
often be understood by examining the implications of their fundamental planetary parame-
ters (e.g., Bougher et al. 1999a, 2000, 2002; Rishbeth et al. 2000b). Such analysis can be
used to guide new model simulations (e.g., to estimate the relative importance of individual
Upper Atmosphere–Ionosphere Modeling 109
processes), and to subsequently interpret completed model simulations. Recent studies have
also shown that substantial advances in our understanding can be realized by investigating
common aeronomic processes across various planetary environments. A common model-
ing framework, modified to incorporate planet specific fundamental parameters, provides a
useful platform for examining the relative importance of these common physical processes.
Finally, model predictions for planetary upper atmospheres and ionospheres with limited
or no measurements are often made based upon our experience with previously successful
model frameworks encompassing similar physical processes (see Sect. 1.2).
1.2 Usefulness and Shortcomings of the Earth Paradigm
Terrestrial modeling frameworks and their assumptions have typically been used to launch
new simulations of other planetary upper atmospheres and ionospheres. This terrestrial par-
adigm is both useful and hazardous at the same time. The primary benefit of the Earth
paradigm can be realized for planetary upper atmospheres having similarities in their funda-
mental planetary parameters, basic processes, and vertical domains (atmospheric regions).
Simulations across these similar planetary environments can be effectively used to exam-
ine the relative importance of common aeronomic processes. A good example is the deter-
mination of the relative importance of CO
2
15-micron emission as a cooling agent in the
thermospheres of Venus, Earth, and Mars (e.g., Bougher et al. 1999a, 2000, 2002). How-
ever, planet specific assumptions are often applied when casting the model equations to
be solved and the physical formulations employed. Furthermore, fundamental planetary
parameters may be so different that application of a terrestrial model framework may no
longer be appropriate. A good example of the latter is the application of traditional Earth
thermospheric models to the upper atmosphere of Saturn’s moon Titan (see Sect. 4.2.3).
Here, the assumption of constant gravity over the Titan thermospheric domain (600–
1500 km) is not sufficient to characterize the extended atmosphere associated with this small
body.
In short, care must be taken when applying an existing modeling framework to a new
planetary environment. A review of the key equations to be solved and all supporting as-
sumptions must be made in light of the important processes to be incorporated and the
vertical domain to be addressed (see Sect. 2).
1.3 Roadmap for Chapter
This chapter describes various numerical model classes (and representative model tools) that
are typically used in simulations of the upper atmospheres and ionospheres of planets. As-
sumptions about the model equations to be solved by the different model frameworks can be
classified by the number of moments carried in the solution of the Boltzmann equation (see
Table 1 and Sect. 2). Both 1D and multi-dimensional model frameworks are employed, both
for the upper atmosphere (UATM) and whole atmosphere (WATM) environment (see Ta-
ble 1). One-dimensional models (see Sect. 3) are commonly used to thoroughly test detailed
aeronomic processes (e.g., thermal, diffusion, and chemical) before the addition of global
winds in a multi-dimensional model framework (see Sect. 4). This progression from 1D to
multi-dimensional models follows a development strategy involving increasing complexity,
internal self-consistency, and expanded temporal plus spatial coverage. Finally, modeling
frontiers and key problems for further research are described in Sect. 5.
110 S.W. Bougher et al.
Table 1 Classes and Examples of Upper Atmosphere Modeling Tools
Dimensions 5-Moment 8-Moment 13-Moment
1D (UATM) Earth (Roble) Earth GITM Mars
Venus (Fox) Mars GITM
Mars (Fox) Titan GITM
3D (UATM) Earth TIE-GCM Earth GITM
Earth TIME-GCM Mars GITM
Earth CTIM,CTIP, CMAT Titan GITM
Venus TGCM
Mars TGCM
Jupiter JIM
Jupiter TGCM
Saturn STIM
Titan TTIM
3D (WATM) Earth WACCM Mars GITM
(MWACM)
Mars MGCM-MTGCM
Mars LMD-MGCM
TGCM (Thermosphere General Circulation Model); GITM (Global Ionosphere Thermosphere Model);
WACCM (Whole Atmosphere Community Climate Model); TIE-GCM (Thermosphere Ionosphere Elec-
trodynamics General Circulation Model); TIME-GCM (Thermosphere Ionosphere Mesosphere Electro-
dynamics General Circulation Model); CTIM (Coupled Thermosphere Ionosphere Model); CTIP (Cou-
pled Thermosphere Ionosphere Plasmasphere Model); CMAT (Coupled Middle Atmosphere Thermosphere
Model); JIM (Jupiter Ionosphere Model); STIM (Saturn Thermosphere Ionosphere Model); TTIM (Titan
Thermosphere Ionosphere Model); MWACM (Mars Whole Atmosphere Climate Model); MGCM-MTGCM
(Mars General Circulation Model Mars Thermosphere General Circulation Model); LMD-MGCM (Labo-
ratoire de Meteorologie Dynamique Mars General Circulation Model)
2 Moment Solutions of the Boltzman Equation: Applicability Concerns for Upper
Atmosphere Models
There are numerous different ways to obtain the transport equations needed to model neutral
atmosphere and ionospheric behavior. The one which starts from basic physical principles
uses the so called moments of the Boltzmann equation. In the Boltzmann approach one is
not interested in the behavior of individual gas particles, but the gas is described by the
velocity distribution function f
s
(r, v,t), where the subscript s, denotes a given species, s,
r
s
is the spatial location, v
s
is the velocity and t is time. It can be easily shown (e.g., Schunk
and Nagy 2000) that the integro-differential equation for f
s
is:
∂f
s
∂t
+v
s
.f
s
+a
s
.
v
f
s
=
δf
s
δt
(1)
where
s
is the velocity gradient and δf
s
t is a short hand way to write the change in
f
s
due to collisions. In order to deduce the transport equations of macroscopic parameters
one first defines velocity moments of this distribution function. The most commonly used
macroscopic parameters of interest are:
n
s
(r, t) =
−∞
d
3
v
s
f
s
, (2)
Upper Atmosphere–Ionosphere Modeling 111
u
s
(r,t)=
−∞
d
3
v
s
f
s
v
s
/n
s
, (3)
T
s
=
m
s
3kn
s
d
3
v
s
f
s
(v
s
u
s
)
2
, (4)
q
s
=
m
s
2
−∞
d
3
v
s
f
s
(v
s
u
s
)
2
(v
s
u
s
), (5)
p
s
=
m
s
3
−∞
d
3
v
s
f
s
(v
s
u
s
)
2
, (6)
P
s
=m
s
−∞
d
3
v
s
f
s
(v
s
u
s
)(v
s
u
s
), (7)
τ
s
=P
s
p
s
I, (8)
where n
s
is the number density of species s, u
s
is the drift (mean) velocity, T
s
is the temper-
ature, q
s
is the heat flow vector, p
s
is the scalar pressure, P
s
is the pressure tensor and τ
s
is
the stress tensor.
In order to obtain the appropriate transport equations for these macroscopic parameters
of interest one multiplies the Boltzmann equation with the appropriate function of velocity
and then integrates over all velocities. Now a very important point should be noted that
the resulting, so called “transport equations”, do not lead to a closed system of equations.
Namely an equation governing the moment of order k contains the moment of order (k +1).
As an example the equation for density contains the drift velocity. A number of approaches
have been suggested in order to achieve closure. The most commonly used method finds an
approximate expression for the distribution function which allows closure and the evaluation
of the collision term.
The so-called 13 moment approximation is the most complex set of equations that have
been used so far in neutral atmosphere and ionosphere modeling. In this approach, f
s
is
approximated as a truncated expansion about the Maxwell Boltzmann distribution function
in terms of density, drift velocity, temperature, stress tensor and heat flow vector. The name
13-moment approximation comes from the fact that the gas is described in terms of 13 pa-
rameters (n
s
= 1, u
s
= 3,T
s
= 1, q
s
= 3
s
= 5). In most cases the 13 moment equations
(see Schunk and Nagy 2000) are too complicated to be able to be solved in a comprehensive
global model. The 8-moment equation neglects the stress tensor and the 5 moment equa-
tions neglect both stress and heat flow. The Navier-Stokes equation are obtained from the
13-moment equations by assuming that the collision frequency is very high and dropping
all q
s
and τ
s
terms, except those that are multiplied by the collision frequency. In this ap-
proximation q
s
and τ
s
can be expressed in term of n
s
, u
s
and T
s
. The most common set of
equations used in global models are these Navier-Stokes ones. However, it should be remem-
bered that in this approximation the distribution function must be close to a Maxwellian.
3 Representative 1-D Neutral and/or Ion Models
3.1 Earth
A global mean model of a planetary atmosphere is useful for the development of a self-
consistent aeronomical scheme, determining the vertical resolution necessary to determine
the basic atmospheric and ionospheric structure, time constants of physical and chemical
112 S.W. Bougher et al.
processes and a host of other properties of the atmosphere and ionosphere. Such a model
is numerically fast for long-time integrations, and easy to test and analyze the sensitivity
of the atmosphere to physical and chemical processes, such as eddy diffusion, chemical
reactions, branching ratios, radiation to space and many other parameters. One can easily
conduct a large number of numerical simulations to develop an understanding of the impor-
tant processes responsible for atmospheric structure and thus make it easier to transfer the
important processes to other higher dimensional models.
A 1D model that was important for the development of the series of National Center
for Atmospheric Research (NCAR) TGCMs was developed by Roble et al. (1987)forthe
thermosphere and ionosphere. It was further extended to include the mesosphere and upper
stratosphere by Roble (1995). A thorough description of the model has been given in the
1995 paper. The model was designed to be fully consistent internally with only specification
of temperature and composition at the lower boundary at 10 mb (30 km) and the upper
boundary in thermal and diffusive equilibrium (500–700 km). It is a time dependent model
generally run to steady state for global average forcing from solar EUV radiation, the aurora,
and specified eddy diffusion. It was primarily used to develop the aeronomical scheme for
the series of TGCMs and to test ideas and parameters. It has been used for numerous studies
by colleagues and students to test various ideas. One example of a simulation for solar
medium conditions is shown in Fig. 1.
This terrestrial 1D code has also been used to study global change in the upper at-
mosphere by Roble and Dickinson (1989) and more recently been used by Qian et al. (2006,
2008) to study decadal changes in satellite drag and ionospheric structure.
3.2 Venus and Mars
One-dimensional models have been constructed of the thermospheres and ionospheres of
Mars and Venus since the first radio occultation measurements of the electron density pro-
files were reported from early flybys and orbiters. Lodders and Fegley (1998)havereviewed
missions to these planets up to 1998.
There are two basic types of one-dimensional models: photochemical equilibrium (PCE)
models, in which transport is ignored, and those that include (usually) vertical transport of
species. While in the former type of model, the densities at each altitude can be computed
independently, in the latter type of model numerical coupling of each altitude to those above
and below it must be taken into account. Altitude profiles of neutral species in thermospheres
cannot be modeled with the PCE approximation, although ion density profiles in the high
density lower peak regions may be.
All of the early (and most of the subsequent) radio occultation electron density pro-
files measured by radio science experiments of both Mars and Venus have exhibited two
peaks: an upper F1 peak, which is produced by absorption of the main portion of the EUV
(150–1000 Å), and a lower E peak which is produced by the absorption of soft X-rays
(e.g., Kliore et al. 1967;Stewart1971). Curiously, however, the ion density profiles derived
from the Viking RPA measurements showed no lower peak (e.g., Hanson et al. 1977).
Most of the early models of the ionospheres of Mars and Venus that were designed to fit
radio occultation electron density profiles, assumed PCE (e.g., McElroy 1967, 1968a, 1969;
Stewart 1968, 1971). The Venus ionospheric model of Kumar and Hunten (1974), how-
ever, was a hybrid model, in which the heavy ions were assumed to be in PCE, while the
lighter ions were subject to transport. Shimazaki and Shimizu (1970) constructed a number
of models of the Martian ionosphere that included transport by diffusion and eddy diffusion,
which were compared to the electron density profile measured by Mariner 4. Interestingly,
Upper Atmosphere–Ionosphere Modeling 113
Fig. 1 NCAR 1D global mean model simulations for solar medium conditions. (a) Neutral, ion, and electron
temperatures (30–400 km); (b) Neutral temperatures (40–120 km); (c) Heating terms of thermal equation
(30–400 km); and (d) Cooling terms of the thermal equation (30–400 km). Individual curves for (a): Tn,
Te, Ti: neutral, electron and ion temperatures; T
ns
: neutral temperature from MSIS-90. Individual curves
for (c): SRC: Schumann-Runge Continuum; SRB: Schumann-Runge Bands; Q
T
: total heating; Q
J
: Joule
heating; Q
A
: auroral heating; O
3
:O
3
heating; O(
1
D): heating from O(
1
D) quenching; Q
NC
: neutral chem-
istry heating; Q
IC
: ion chemistry heating; Q
EI
: ion–electron heating. Individual curves for (d): Q
T
:total
heating; K
e
: cooling rates from eddy thermal conduction; CO
2
: 15-micron cooling; NO: 5.3-micron cool-
ing; K
m
: cooling rate from downward molecular thermal conduction; O(
3
P): oxygen fine structure cooling.
ZP (left vertical axis, log pressure scale used by the NCAR codes)
McElroy (1968b) noted the difficulty of constructing model ionospheres when the only in-
formation was in the form of radio occultation electron density profiles, without the benefit
of in situ measurements of ion and neutral densities. The lack of in situ measurements has
also plagued interpretation of the electron density profiles of the outer planets for many
years.
The PCE approximation becomes inaccurate above the boundary where the lifetime of a
species due to chemistry (τ
c
=1/L),whereL =L/n is the specific loss rate, L is the total
chemical loss rate and n is the number density of the species, becomes longer than that due to
vertical transport. If the main transport process is by diffusion, the lifetime of a species due to
transport is given approximately as τ
D
H
2
/D,whereD is the diffusion coefficient of that
species and H = kT /mg is the scale height. In this expression, k is Boltzmann’s constant,
while T,g,andm are the altitude dependent temperature, acceleration of gravity, the mass
114 S.W. Bougher et al.
of the atmospheric species, respectively. For ions, the scale height is H
i
=k(T
i
+T
e
)/m
i
g,
where T
i
is the ion temperature, and T
e
is the electron temperature.
If the major transport process (for neutrals) is mixing, the lifetime of a species against
transport is given by τ
K
H
2
avg
/K,whereK is the eddy diffusion coefficient, H
avg
=
kT/m
avg
g,andm
avg
is the average mass of the constituents. It is remarkable that the eddy
diffusion coefficient is on the order of 1.0 ×10
13
/n
0.5
cm
2
s
1
,wheren is the total num-
ber density in the lower thermospheres for both Mars and Venus (cf., Krasnopolsky 1982;
vonZahnetal.1980).
The boundary where diffusion of a neutral species becomes more important than mix-
ing is known as the homopause. In fact, however, that altitude is different for each species.
In early models of the thermospheres of Venus and Mars, a single homopause altitude was
assumed (e.g., McElroy 1967, 1969; Kumar and Hunten 1974; Chen and Nagy 1978). Be-
low the homopause the thermosphere is considered to be completely mixed for chemical
tracers/inert species; above the homopause, these species density profiles are assumed to be
distributed according to their own scale heights. Species that are formed photochemically do
not exhibit this behavior. Because of the availability of in situ measurements of the neutral
densities, and the computing power that is available today, even in one-dimensional models,
the homopause approximation is neither necessary nor used widely.
While PCE approximations are almost never used to compute the density profiles of
minor neutral thermospheric species, such models of ion density profiles continue to be
used for specific purposes, including studies focused on the electron density peak regions
(e.g., Cravens et al. 1981; Fox and Dalgarno 1981; Kim et al. 1989; Martinis et al. 2003), or
for airglow calculations of processes that originate near the electron density peaks (see, for
example Fox 1992, and references therein).
The Viking I and II probes carried neutral mass spectrometers through the Martian
atmosphere, and so the major neutral densities in the thermosphere at low solar ac-
tivity were measured in situ for the first time (e.g., Nier and McElroy 1976). Early
ionospheric models based on these measurements included, for example, those of McEl-
roy et al. (1976), Fox and Dalgarno (1979), and Chen et al. (1978). While the latter
model included vertical transport. The former two were photochemical equilibrium mod-
els. The in situ measurements of the Pioneer Venus orbiter and probes enabled more ac-
curate models of the Venusian ionosphere (e.g., Chen and Nagy 1978;Nagyetal.1980;
Fox 1982).
In the terrestrial ionosphere, the absolute maximum in the electron density profile is an
F
2
peak, which appears near 300 km. At this altitude, the chemical lifetime of the major ion
(O
+
) is approximately equal to that of transport. At high altitudes in the Venus ionosphere,
O
+
becomes the most important species, yet models show that it forms a peak that is not
(or is barely) visible in the electron density profile. On Mars, the O
+
density forms a peak
at high altitudes, but thus far measurements indicate that densities of O
+
are everywhere
smaller than those of O
+
2
(e.g., Hanson et al. 1977). Where the major loss is by transport,
photochemical equilibrium models do not reproduce the profiles of ions, such as O
+
and
other (mostly) atomic ions.
More sophisticated one-dimensional thermosphere/ionosphere models have included
both chemistry and transport by molecular and eddy diffusion (for neutrals), and ambipo-
lar diffusion (for ions), and do not assume a fixed homopause (e.g., Nagy et al. 1980;Fox
1982, 2004; Krasnopolsky 2002). The one-dimensional thermosphere-ionosphere models of
Venus and Mars of Shinagawa and Cravens (1988, 1989) have also included magnetic fields.
Although one-dimensional models have limitations, mainly that horizontal transport by
convection is ignored, they are simple enough that many species and reactions among those
Upper Atmosphere–Ionosphere Modeling 115
Fig. 2 Venus 1D simulations for solar maximum conditions: (a) Neutral densities; (b) ion densities and
electron density. From Fox (unpublished)
species can be included in them. For example, the current Mars and Venus models of Fox and
coworkers (e.g., Fox and Sung 2001;Fox2004; Fox and Yeager 2006) contain 23 species
and more than 200 reactions. These models include photoionization and excitation, photo-
electron impact ionization and excitation, and electron-impact excitation of twelve neutral
background species, and photodissociation, photodissociative excitation, and photodissocia-
tive ionization of all of the molecular neutral background species. Cross sections for all these
processes are required input for the models.
An example of a one-dimensional Venus model for high solar activity is shown in
Fig. 2. Figure 2a illustrates the background density profiles of 12 species; Fig. 2bshows
the computed ion density profiles. In this model an upward velocity boundary condition
of 2 × 10
5
cm s
1
was imposed on most ions, since the ionosphere of Venus is eroded by
day-to-night ion transport. O
++
ions are assumed to be in photochemical equilibrium. Us-
ing these models, density profiles of metastable species, such as O
+
(
2
D),O
+
(
2
P),N(
2
D),
N(
2
P),O(
1
S),andO(
1
D) can be computed. These species can participate in reactions that
are not available to ground state species, and are important to the energy balance in the
thermosphere/ionosphere. Quenching of these species produces local heating, and radiation
produces cooling. The radiation rates of shorter-lived excited species may be computed to
produce profiles of airglow intensities.
Using one-dimensional models, the altitude-dependent heating efficiencies can be readily
computed. The heating efficiency is defined as the ratio of the local heating rate to the
solar energy absorbed at a given altitude. Computations show that for Mars and Venus (Fox
1988; Fox et al. 1995), the most important heat source near the ion peak is dissociative
recombination of O
+
2
,
O
+
2
+e O +O +E (9)
where the product O atoms may be in various combinations of electronically excited states
and the energy released /E varies with the branching ratios (e.g., Kella et al. 1997). At
low altitudes, photodissociation and quenching of metastable species are more important.
Exothermic chemical reactions are of secondary importance, and electron impact is of minor
116 S.W. Bougher et al.
importance over the entire thermosphere. For Venus, the altitude dependent heating efficien-
cies range from about 16% at low altitudes for the “lower limit” model to 22% for the “best
guess” model. At high altitudes the heating efficiencies increase slightly with altitude up to
185 km. On Mars, the “best guess model” yields heating efficiencies of about 21 ±2% from
100 to 200 km, although at low altitudes the lower limit model shows heating efficiencies of
about 16%.
4 Representative Multidimensional Models
4.1 Model Classes to be Addressed
Table 1 summarizes the classes of multi-dimensional models that we will consider in
Sects. 4 and 5. Notice that multi-dimensional model development is following two im-
portant trends. First, existing upper atmosphere models are being extended to encom-
pass the entire atmosphere domain (ground to exosphere) for Earth (NCAR WACCM)
and Mars (Michigan-MWACM). The LMD-MGCM code for the Mars lower atmosphere
has also be extended upward into the thermosphere. These efforts reflect the availabil-
ity of both lower and upper atmosphere datasets for these planets. Coupling processes
(thermal, chemical, dynamical) linking these atmospheric regions are important to ad-
dress with these “whole atmosphere” modeling tools. Second, model frameworks are be-
ing developed and exercised using 8-moment and 13-moment solutions of the Boltzman
equation. The motivations here are at least twofold: (a) to incorporate a non-hydrostatic
treatment for improvement of the simulation of vertical velocities (Ridley et al. 2006;
Deng et al. 2008), and (b) to capture the physical processes that bridge the collisional and
non-collisional regions near the exobase (Boqueho and Blelly 2005).
Under hydrostatic equilibrium, a typical assumption used in most global planetary mod-
els, the pressure gradient in the vertical direction is exactly balanced by the gravity force
(Deng et al. 2008). However, for large vertical velocities, acceleration terms in the vertical
momentum equation cannot be ignored, and the basic balance is no longer hydrostatic. Two
likely examples of these conditions are realized for the sudden intense enhancement of high
latitude Joule heating for the Earth’s thermosphere-ionosphere (Deng et al. 2008), and that
of Jupiter as well. If the hydrostatic assumption is relaxed, the vertical momentum equa-
tion can be expanded to include additional acceleration terms: (1) forces due to ion-neutral
and neutral-neutral friction (when each constituent is solved independently), (2) centrifugal
and Coriolis forces, and (3) non-linear advection terms. The altitude variation of gravity
is easily accommodated in this framework. Vertical propagation of a non-hydrostatic “dis-
turbance” results in acoustic waves and non-hydrostatic gravity waves (Deng et al. 2008).
Care must be taken to either damp or accommodate these waves in non-hydrostatic mod-
els. In short, global non-hydrostatic models are needed to address phenomenon associ-
ated with large vertical winds in planetary upper atmospheres. New planet specific global
thermosphere–ionosphere models are being developed and validated for this purpose, based
upon the Global Thermosphere–Ionosphere Model (GITM) for Earth (Ridley et al. 2006;
Deng et al. 2008).
4.2 Representative GCM Model Descriptions/Results
4.2.1 Earth (NCAR TGCMs)
A series of TGCMs have been developed at NCAR over the past 30 years with each model
incorporating new processes of the coupled thermosphere-ionosphere system. The histori-
Upper Atmosphere–Ionosphere Modeling 117
cal development of the TGCM suite of models is discussed in Bougher et al. (2002). Self-
consistent temperatures, neutral-ion densities, neutral dynamics, and self-consistent electro-
dynamics are contained in the TIE-GCM (Richmond et al. 1992). This code was then ex-
tended to include the mesosphere and upper stratosphere to become the TIME-GCM (Roble
and Ridley 1994). These two codes are now the basic upper atmosphere models at NCAR.
The TIE-GCM is used to explore thermosphere-ionosphere-electrodynamic interactions and
the TIME-GCM has the same processes but extended to include aeronomic processes asso-
ciated with the mesosphere and upper stratosphere and to examine physical and chemical
interactions between upper atmospheric regions.
The TIME-GCM is a self-consistent model of the upper atmosphere extending between
30 km and 500 km altitude. It has been used for comparison and interpretation of satellite,
rocket and ground-based data for many years by a wide variety of scientific colleagues,
post doctoral fellows and graduate students. The TIME-GCM was initially designed for a
5
latitude and longitude grid in the horizontal and 2 grid points per scale height in the
vertical with a 5 minute time step. This “coarse” horizontal and vertical resolution was
later refined for specific model applications (see below). The TIME-GCM solves for the
neutral gas temperature, winds and constituents of the thermosphere, mesosphere and upper
stratosphere both major and minor species self-consistently. It also solves for the ionospheric
plasma properties of electron and ion temperature, electron density and ion composition,
and the electric field, plasma drift and resulting magnetic perturbations. The most recent
description of the model is given by Roble (2000).
In the late 1990s, the TIME-GCM was extended to use lower boundary data at 10 mb
(30 km) to force the variability propagating upward from the lower atmosphere and to study
couplings between the lower and upper atmospheres (Roble 2000). This included specifica-
tion of tides, gravity waves, planetary waves and other disturbances on a daily basis so that
continuous simulations of the upper atmosphere could be made for realistic daily simula-
tions that would be used to compare with observational time averaged campaign or satellite
orbital tracking data.
In order to examine the feasibility of developing a model that extended from the ground-
to-exosphere the TIME-GCM was flux coupled to the NCAR community climate model
CCM3(Khieletal.1998) at the boundary between the models near 10 mb. This allowed in-
formation between the upper and lower atmospheres to be exchanged simulating the entire
atmosphere, troposphere, stratosphere, mesosphere and thermosphere/ionosphere. This cou-
pled model simulated a strong stratospheric warming, mesospheric cooling, thermospheric
warming at high latitudes that was generated spontaneously from planetary waves propagat-
ing upward from the troposphere (Liu and Roble 2002, 2005). These studies indicated that
a continuous model from the ground to exosphere could be developed to examine coupling
aspects between regions of the whole atmosphere. This was a precursor for the development
of the WACCM that is discussed in Sect. 5.1.2.
The TIME-GCM resolution was limited by computer power in the 1990s and early 2000s.
While this was sufficient for a number of upper atmosphere studies, it was inadequate to rep-
resent the shorter wave migrating and non-migrating tides and planetary wave propagation
and dissipation. One temporary solution was to double the amplitudes of the Global Scale
Wave Model (GSWM) forcing at the lower boundary to get agreement with observational
data. Subsequent simulations showed that it was not necessary to double the amplitudes but
rather to improve both the vertical (0.25 scale height) and horizontal (2.5
×2.5
) resolu-
tion. These comparisons are shown in Fig. 3. With a double resolution, large tides developed
without the need to double GSWM amplitudes. This exercise clearly illustrates the need to
118 S.W. Bougher et al.
Fig. 3 NCAR TIME-GCM model simulations for solar moderate conditions. Meridional tidal winds (VN)
in the 80–150 km region at local noon for: (a) standard GSWM forcing, (b) doubled GSWM forcing, and
(c) standard GSWM forcing but with 2.5
horizontal resolution
match the model resolution (both vertical and horizontal) with the phenomena being exam-
ined. This new high-resolution model resolved a number of problems with simulation/data
comparisons.
Work is continuing on evaluating the impact of the new resolution on a number of previ-
ous studies, such as the migrating and non-migrating tides, auroral dynamics, thermospheric
densities and satellite drag and ionospheric F-region dynamics as well as a number of air-
glow studies. A web site that describes the model and some important simulations is: http://
www.hao.ucar.edu/modelling/tgcm/tgcm.html.
Validation of other planetary GCM codes (see Sects. 4.2.24.2.9) may also require sen-
sitivity tests to confirm the model resolution necessary to obtain converged solutions; i.e.
finer vertical (e.g. 0.25 scale height) and horizontal (e.g. 2.5
×2.5
) resolution may also
be required to obtain converged solutions for important desired applications.
4.2.2 Earth (CTIM, CTIP, CMAT codes)
Another major atmosphere general circulation model for Earth is the CTIP model developed
jointly by groups in the UK and US. The code originates from two initially separate mod-
els for the thermosphere (Fuller-Rowell and Rees 1980, 1983) and high latitude ionosphere
(Quegan et al. 1982) which were later coupled (Fuller-Rowell et al. 1987, 1996) for two-
way coupling between the neutral and ionized regimes of the upper atmosphere, a version of
the model often referred to as CTIM. A self-consistent plasmasphere model for the regions
Upper Atmosphere–Ionosphere Modeling 119
equatorward of around 30
geomagnetic latitude was added by Millward et al. (1996a)to
form CTIP. More recently, the original bottom boundary (80 km, 0.01 mb) of the CTIM
model was lowered by Harris (2001) into the stratosphere (30 km, 10 mb) in order to in-
clude the stratospheric and mesospheric chemistry and full vertical dynamical and chemical
coupling. This extended version of the model is referred to as the CMAT model.
What distinguishes these from other thermosphere/ionosphere models is primarily the
fact that ionospheric calculations are carried out along magnetic field lines rather than be-
ing treated on the same spherical grid as the neutral gases. This has proven to be a most
useful approach since one key issue is how to treat upper boundary conditions for ions and
electrons, in particular the plasma fluxes along field lines in the topside ionosphere where
field-aligned transport forms a dominant process affecting the distribution of O
+
and H
+
ions and electrons. The plasma flux boundary condition becomes important for calculations
of ionospheric densities when the boundary is located within a few scale heights above the
density peak. The particular choice of flux tube coordinates for plasma in CTIP/CTIM elim-
inates this difficulty. At high latitudes field lines are open and extend to around 10000 km
altitude in the model, while at low latitudes flux tubes in CTIP are closed. With plasma
densities near 10000 km being negligible compared with ionospheric densities, a zero flux
boundary condition can safely be assumed, while in the regime of closed flux tubes both
ends of a field line are within the photochemical domain of the ionosphere (in opposite
hemispheres), allowing the simple boundary condition of photochemical equilibrium.
The CTIP, CTIM and CMAT models have been used extensively over the past decades
to understand the global morphology of the thermosphere and ionosphere both through
purely theoretical studies and in comparisons with observations. Studies have investigated
(amongst others) the morphology of responses to geomagnetic storms (Field et al. 1998;
Fuller-Rowell et al. 2002), thermospheric composition and dynamics (Fuller-Rowell 1998;
Rishbeth and Mueller-Wodarg 1999), semiannual variations in the ionosphere (Millward
et al. 1996b; Rishbeth et al. 2000a), effects of tidal forcing (Millward et al. 2001;
Mueller-Wodarg et al. 2003) and NO chemistry (Dobbin et al. 2006).
4.2.3 Michigan GITM Codes
Earth GITM. The GITM code (Ridley et al. 2006) was designed from the bottom up with
flexibility in mind for every aspect of modeling upper atmospheres of planetary systems.
The grid system within GITM is fully parallel and is quite versatile. Users can run 1D cases
at a specified latitude and longitude (which is set at run-time in the input file) or 3D cases
with almost any latitudinal and longitudinal resolution that the user wants (once again, set at
run-time). GITM can run on a single processor machine or multi-processor machines. It has
been run on 256 processors resolving the upper atmosphere with a 1.25
latitudinal by 2.5
longitudinal resolution. For testing, GITM has been run with 10
by 20
resolution. This
flexibility in the resolution allows rapid development of the model and facilitates testing of
new physics within the code.
The main feature that differentiates GITM from all other coupled ionosphere ther-
mosphere models is the easing of the hydrostatic assumption within the vertical momentum
equation—the pressure does not have to balance with gravity (although it almost always
does). In regions in which there are non-hydrostatic forces (e.g., the auroral zone), large
vertical winds can develop (Deng et al. 2008). In addition, the GITM vertical momentum
equation allows gravity to be dependent on altitude, instead of constant, which is crucial
for small bodies with extended atmospheres, such as Titan. GITM utilizes an altitude grid,
which is also different than other upper atmosphere models. The resolution in the vertical
direction is stretched such that it is approximately 1/3 of a scale height at code initiation.
120 S.W. Bougher et al.
Fig. 4 Earth GITM simulation at 400 km. Horizontal wind vectors are superimposed upon mass density
color contours (1.15 to 5.15 ×10
12
g/m
3
)
An additional feature of GITM is the ability to turn on and off physics through the input
file. Different source terms (such as Coriolis, Joule heating, solar EUV heating, and thermal
conduction), can be turned off through the input file. This allows users to conduct numerical
experiments in which they self-consistently determine the effects of different source terms
on the coupled nonlinear system.
In order to allow GITM to be utilized for more than a single body, very little was hard-
coded into the core of GITM. Each atmospheric and ionospheric constituent is specified in
a planet specific module, such that the advective core need only loop over the number of
species to determine the hydrodynamic behavior of the atmosphere. Other common source
terms, such as solar EUV inputs, and inter-species frictional drag in the vertical direction, are
handled in a similar manner, allowing the code to be adapted to another planet very easily.
More specific source terms, such as radiative cooling, need to be coded for the particular
problem and are easily linked to GITM through hooks.
At Earth, GITM has an extremely flexible high-latitude energy input module. This al-
lows users to try different electric fields and particle inputs to drive GITM. At other planets,
this can be easily adapted to different types of forcing. Models of electron precipitation and
electric potential can be added with little difficulty. These will then be utilized to calcu-
late electric fields and ion and electron velocities. Joule heating and ion-neutral momentum
coupling are then self-consistently calculated.
Figure 4 shows results from the Earth-based GITM at 400 km altitude. The vectors show
the thermospheric neutral winds, while the coloring is the thermospheric mass density. The
neutral winds roughly follow a two-cell convection pattern at high latitudes, due to the strong
forcing by the ion convection and aurora. On the dusk-side (left), the neutrals are able to
form a completely closed cell, while on the dawn-side (right), the cell is less well defined.
Upper Atmosphere–Ionosphere Modeling 121
This is because on the dusk-side, the Coriolis force is in the same direction as the circulation
pattern, so the closed cell is accentuated, while on the dawn-side, the flow is inhibited, since
the Coriolis force opposes it. At very high latitudes and at the tail end of the convection cells
on the night side (i.e., just over Canada), there is a large enhancement in the thermospheric
mass density. This is caused by Joule heating in the auroral zone, where there are extremely
large electric fields and strong conductivities. During this strong driving period, the density
peaks at the poles, while during quieter times, the mass density peaks at lower latitudes.
Titan GITM. A new application of the GITM framework was recently developed (Bell et
al. 2006;Bell2008) in an effort to interpret and place in a global context new Cassini INMS
neutral and ion density and inferred temperature datasets of the Titan upper atmosphere
(e.g., Waite et al. 2005). The GITM framework was chosen to capture the unique physics of
the Titan upper atmosphere that requires: (a) variable gravity, (b) calculated fluxes of ma-
jor species out the top of the atmosphere, and (c) variable Saturn magnetospheric forcing.
Density gradients (yielding fluxes) were specified at the upper boundary in order to sim-
ulate measured Cassini INMS CH
4
density profiles; the self-consistent feedback of these
fluxes upon temperatures was also included. The Titan GITM code was designed to span
500 to 1500 km in the Titan upper atmosphere, covering the region below the homopause
(800 km) to just above the exobase (1300 km). Solar EUV forcing (heating, photo-
dissociation, photo-ionization) is dominated by N
2
and CH
4
solar absorption between 1.6–
170.0 nm. Hydrogen cyanide (HCN) rotational band infrared radiative cooling is incorpo-
rated and represents the dominant IR cooling agent in the thermosphere. A self-consistent
treatment of chemical production and loss processes for 4-major, 7-minor, several isotopes,
and thermally active (e.g. HCN) constituents is incorporated, based upon the scheme out-
lined by DeLa Haye (2005). Major photochemical ions are limited to key species: N
+
2
,N
+
,
CH
+
3
,H
2
CN
+
,andC
2
H
+
5
. Finally, a differentially, super-rotating lower boundary is specified
at 500 km, in accord with ground-based observations (Hubbard et al. 1993). Horizontal
and vertical distributions of simulated temperatures and densities have been successfully
compared to specific Cassini orbit measurements, especially profiles of key isotopes (e.g.,
Bell 2008).
Mars GITM. The need for a ground-to-exobase GCM at Mars has motivated another new
application of the GITM framework. Existing Mars lower and upper atmosphere datasets
need to be interpreted and connected using such a “whole atmosphere” model framework
(e.g., Bougher et al. 2006b). This is the first extension of the GITM framework over a wide
range of altitudes encompassing both upper and lower atmosphere processes.
A prototype MWACM code using the GITM framework has been developed, enabling
initial 1D and 3D simulations to be conducted over 0–300 km for specific solar cycle, sea-
sonal and dust conditions at Mars. Specifically, the terrestrial GITM code (Ridley et al. 2006)
was adapted to include Mars fundamental parameters, constants, and key radiative processes
in order to capture the basic observed features of the thermal and dynamical structure of the
Mars atmosphere from the ground to 300 km. For the Mars lower atmosphere (0–80 km),
an efficient (fast) radiation code was adapted from the NASA Ames MGCM code to the
framework of the MWACM. This now provides MWACM solar heating (long and short
wavelength), aerosol heating, and CO
2
15-micron cooling in the LTE region of the Mars
atmosphere below 80 km. For the Mars upper atmosphere (80 to 300 km), a fast formu-
lation for NLTE CO
2
15-micron cooling was implemented into the MWACM code, along
with a correction for non-LTE (NLTE) near-IR heating rates (80–120 km). In addition,
a thermospheric EUV-UV heating routine (based upon a CO
2
dominated atmosphere) was
adapted to the MWACM framework. Finally, detailed neutral-ion chemistry was recently
incorporated above 80 km, based upon Mars TGCM reactions and rates (see Sect. 4.2.5).
122 S.W. Bougher et al.
For the entire atmosphere, the MWACM dynamical core solver was modified to work
with the new terrain following coordinate system. The Martian terrain is now being incor-
porated into the MWACM code making use of Mars Global Surveyor MOLA topographic
data files. NASA Ames MGCM CO
2
condensation, and boundary layer routines will be
added below 80 km. At the surface, global empirical maps of albedo and thermal inertia will
be supplied to the radiation calculations. Initial simulations indicate this extended model is
stable and captures the basic observed temperatures and expected wind structures through-
out the Mars atmosphere.
4.2.4 Venus (NCAR VTGCM)
The large-scale circulation of the Venus upper atmosphere from 90 to 200 km (upper
mesosphere and thermosphere) is a combination of two distinct flow patterns: (1) a rel-
atively stable subsolar-to-antisolar (SS-AS) circulation cell driven by solar (EUV-UV-IR)
heating, and (2) a highly variable retrograde superrotating zonal (RSZ) flow (see reviews
by Bougher et al. 1997, 2006a, 2006b; Schubert et al. 2007). GCMs have proven useful
for synthesizing the available Pioneer Venus, Magellan, Venus Express, and ground-based
density, temperature, and/or airglow datasets and thereby extracting these upper atmosphere
wind components (see reviews by Bougher et al. 1997, 2006a; Schubert et al. 2007).
The Venus TGCM (VTGCM) is a 3D finite difference hydrodynamic model of the Venus
upper atmosphere that is based on the NCAR terrestrial TGCM. The VTGCM has been
documented in detail as revisions and improvements have been made over nearly 2-decades
(see Bougher et al. 1988, 1990, 1997, 1999a, 2002; Bougher and Borucki 1994; Zhang et al.
1996).
The modern VTGCM code (e.g., Brecht et al. 2007;Rafkinetal.2007; Bougher et
al. 2008) calculates global distributions of major species (CO
2
, CO, O, and N
2
), minor
species (e.g. O
2
, NO, N(
4
S), N(
2
D)), and dayside photochemical ions (CO
+
2
,O
+
2
,O
+
,and
NO
+
). These constituent fields are all consistent with the simulated 3-D temperature struc-
ture and the corresponding 3-component neutral winds. The VTGCM model covers a 5
by 5
latitude-longitude grid, with 46 evenly spaced log-pressure levels in the vertical, ex-
tending from approximately 80 to 200 km at local noon. Dayside O and CO sources arise
primarily from CO
2
net dissociation and ion-neutral chemistry; the latter utilizes the ion-
neutral chemical reactions and rates of Fox and Sung (2001). Simplified catalytic ClOx and
HOx reactions can also employed to specifically improve the chemical sources and sinks for
O and CO below 120 km (e.g., Bougher and Borucki 1994).
Formulations for CO
2
15-micron cooling, wave drag, and eddy diffusion are incorporated
into the VTGCM (see Bougher et al. 1999a;Brechtetal.2007). In particular, CO
2
15-
micron emission is known to be enhanced by collisions with O atoms, providing increased
cooling in NLTE regions of the upper atmosphere (see Bougher et al. 1994; Kasprzak et
al. 1997). VTGCM CO
2
15-micron cooling is parameterized as described by Bougher et al.
(1986), making use Roldan et al. (2000) exact cooling profiles at reference temperatures and
atomic oxygen abundances. The collisional O–CO
2
relaxation rate adopted for simulated
15-micron cooling is 3 ×10
12
cm
3
/s. In addition, near-IR heating rates are incorporated
using modern offline look-up tables from Roldan et al. (2000). These parameterizations
provide strong CO
2
15-micron cooling that is consistent with the use of EUV-UV heating
efficiencies of 20–22%, which are in agreement with detailed offline heating efficiency
calculations of Fox (1988).
The VTGCM is typically run to examine Venus thermospheric structure and winds for
solar maximum, moderate, and minimum EUV-UV flux conditions, corresponding to terres-
trial F10.7-cm indices of 200, 110-130, and 68-80, respectively. In addition, the VTGCM
Upper Atmosphere–Ionosphere Modeling 123
Fig. 5 VTGCM predicted
exospheric temperature variations
for dayside, equatorial
conditions. Texo 95 K for
F 10.7170 units.
Tmax =325 K; Tmin =230 K.
Temperatures at 140 km are also
plotted: 200 to 240 K
is designed to calculate O
2
visible (400–800 nm), O
2
IR (1.27 microns) and NO ultravio-
let (198.0 nm) nightglow distributions for comparison with various Venera, Pioneer Venus,
Venus Express, and/or ground-based measurements (e.g. Bougher et al. 1997, 2006a, 2008;
Brechtetal.2007). These night airglow layers, and the controlling global circulation pat-
terns being traced, span 90 to 150 km. Figure 5 illustrates the VTGCM predicted dayside,
equatorial exospheric temperature variation, in substantial agreement with available Pio-
neer Venus and Magellan observations (e.g., Keating et al. 1980; Keating and Hsu 1993;
Kasprzak et al. 1997). Figure 6 illustrates the VTGCM exobase circulation pattern and the
underlying asymmetric thermospheric temperature structure for solar moderate conditions.
Night airglow distributions respond to this mean circulation pattern, with maxima that oc-
cur on average near (or just beyond) midnight and close to the equator (e.g., Bougher et al.
2006a, 2008).
However, existing GCMs are presently unable to reproduce the significant variations in
observed diurnal density, temperature, and airglow fields utilizing a unique set of wind fields,
eddy diffusion coefficients, and wave drag parameters (Bougher et al. 1997, 2006a). This
problem may reflect missing physical processes or inputs, e.g., exospheric transport above
180–200 km, upward propagating planetary waves, and limited gravity wave constraints for
formulating wave breaking (Bougher et al. 2006a).
4.2.5 Mars (NCAR MTGCM)
The Mars TGCM (MTGCM) is a finite difference primitive equation model that self-
consistently solves for time-dependent neutral temperatures, neutral-ion densities, and three
component neutral winds over the Mars globe (e.g., Bougher et al. 1999a, 1999b, 2000,
2002, 2004, 2006b;Belletal.2007). The MTGCM code is adapted from the NCAR TGCM
framework (see Sect. 4.2.1).
The modern MTGCM code contains prognostic equations for the major neutral species
(CO
2
,CO,N
2
, and O), selected minor neutral species (Ar, NO, N(
4
S), O
2
),andseveral
photochemically produced ions (e.g. O
+
2
,CO
+
2
,O
+
,andNO
+
below 180 km). All fields are
124 S.W. Bougher et al.
Fig. 6 VTGCM simulated exospheric temperatures and superimposed horizontal neutral winds for Equinox,
F 10.7 =200 conditions (similar to early Pioneer Venus observations). Dayside equatorial temperatures reach
305 K. Nightside minimum temperatures drop to 100 K. Maximum horizontal winds reach 220 m/sec
(70–90% of the sound speed in the upper thermosphere) across the evening terminator. The average altitude
for this slice is 180 km
calculated on 33 pressure levels above 1.32 µbar, corresponding to altitudes from roughly
70 to 300 km (at solar maximum conditions), with a 5
resolution in latitude and longitude.
The vertical coordinate is log pressure, with a vertical spacing of 0.5 scale heights. Key
adjustable parameters which can be varied for MTGCM cases include the F 10.7-cm index
(solar EUV/UV flux variation), the heliocentric distance and solar declination corresponding
to Mars seasons. A fast NLTE 15-micron cooling scheme is implemented in the MTGCM,
along with corresponding near-IR heating rates (Bougher et al. 2006b). These inputs are
based upon detailed 1D NLTE model calculations for the Mars atmosphere (e.g., López-
Valverde et al. 1998).
A simple dayside photochemical ionosphere is formulated for the MTGCM, including
the major ions. Key ion-neutral reactions and rates are taken from Fox and Sung (2001);
empirical electron and ion temperatures are adapted from the Viking mission for various
solar conditions. The ionization rates required for the production rates are calculated self-
consistently, making use of specified solar EUV fluxes. Nightside ions are not yet simulated,
but will require either day-to-night ion drifts from modern magnetohydrodynamic models
(e.g., Ma et al. 2004), or energetic electron precipitation sources of nightside ionization (e.g.,
Fillingim et al. 2007).
The MTGCM is driven from below by the NASA Ames Mars MGCM code (Haberle et
al. 1999) at the 1.32-µbar level (near 60–80 km). This coupling allows both the migrating and
non-migrating tides to cross the MTGCM lower boundary and the effects of the expansion
and contraction of the Mars lower atmosphere to extend to the thermosphere. The entire
atmospheric response to simulated dust storms can also be calculated using these coupled
Upper Atmosphere–Ionosphere Modeling 125
Fig. 7 MTGCM exospheric
temperatures as a function of Ls
(season) and solar cycle
(F 10.7-cm index). Dayside
(LT =1500) equatorial
conditions are displayed. Curves
indicated: F 10.7 =175–200
(top), 110–130 (middle), and
70–80 (bottom)
models. Key prognostic variables are passed upward from the MGCM to the MTGCM at
the 1.32-µbar level at every MTGCM grid point: temperatures, zonal and meridional winds,
and geopotential heights. These two climate models are each run with a 2-minute time step,
with the MGCM exchanging fields with the MTGCM at this frequency.
This coupled configuration has been validated using an assortment of spacecraft observa-
tions, including Mars Global Surveyor, 2001 Mars Odyssey, and Mars Reconnaissance Or-
biter thermosphere and/or ionosphere data sets (Bougher et al. 1999b, 2000, 2004, 2006b).
Figure 7 shows dayside, equatorial exospheric temperature variations predicted by the MT-
GCM over the solar cycle and Mars seasons. These variations are in reasonably good agree-
ment with available aerobraking and orbital drag measurements (e.g., Keating et al. 1998,
2003; Forbes et al. 2008; Mazarico et al. 2007). Exobase neutral horizontal winds and tem-
peratures for solar maximum, southern summer solstice conditions are also illustrated in
Fig. 8. In general, the global wind patterns simulated by the MTGCM reveal very strong
summer-to-winter inter-hemispheric Hadley circulations that are consistent with observed
NO nightglow (Bertaux et al. 2005) and winter polar warming (Bougher et al. 2006b) fea-
tures. Finally, no downward coupling is presently activated between the MGCM and the
MTGCM. However, the impacts of lower atmosphere dynamics upon the upper atmosphere
are dominant (see Sect. 5.1.3).
4.2.6 Jupiter (JIM)
The JIM code is a general circulation model calculating the global thermosphere and
ionosphere on Jupiter above the 2 µbar level (set to 357 km above the 1 bar level). This
model, the first of its kind for Jupiter, was developed in the UK by Achilleos et al. (1998)
and based largely on the thermosphere model of CTIM (see Sect. 4.2.2). Solar heating and
photoionization are calculated self-consistently by solving the Beer–Lambert law. The ther-
mospheric dynamics are calculated assuming all non-linear terms in the momentum equa-
tion, including ion drag, and ion dynamic calculations include diffusion and neutral drag.
Calculations of the distribution of major neutral species H, H
2
and He include transport by
winds and diffusion as well as simple ion-neutral photochemistry above the homopause.
Unlike CTIM, ions in JIM are calculated on the same spherical pressure grid as the neutrals,
which assumes a 10
horizontal resolution. The principal ions considered are H
+
,H
+
2
and
H
+
3
, with more complex organic molecules which occur mainly below the homopause being
126 S.W. Bougher et al.
Fig. 8 MTGCM simulated exospheric temperatures and superimposed horizontal neutral winds for
Ls =270, F 10.7 =175 conditions (similar to late Mars Odyssey aerobraking observations). Dayside subso-
lar latitude (25 S) temperatures reach 320 K, nightside minimum temperatures drop to 145 K. Maximum
horizontal winds reach 550 m sec
1
(slightly in excess of the sound speed). The average altitude for this
slice is 215 km
treated through a generic single molecule in the ion recombination reactions. A magnetic
field is included for the calculations of ion dynamics and transport. In the auroral zones
JIM allows for the inclusion of particle precipitation from the magnetosphere to calculate
ionization and energy deposition. At high latitudes an electric field is included to simulate
the auroral electrojet (Achilleos et al. 2001). Studies with JIM have helped investigate the
morphology of thermosphere ionosphere coupling on Jupiter, in particular the dynamics at
high latitudes (Millward et al. 2005), electrodynamic coupling between the thermosphere
and the auroral ionosphere (Achilleos et al. 2001) and ionospheric conductivities (Millward
et al. 2002a, 2002b).
4.2.7 Jupiter (NCAR JTGCM)
The proper characterization of Jupiter’s upper atmosphere, embedded ionosphere, and au-
roral features requires the examination of underlying processes, including the feedbacks of
energetics, neutral-ion dynamics, composition, and magnetospheric coupling. Coupled ther-
mosphere and ionosphere GCM models (with magnetospheric inputs) can be used to address
these feedbacks.
A Jupiter TGCM (JTGCM) code has now been developed and exercised to address
global temperatures, three-component neutral winds, and neutral-ion species distributions
Upper Atmosphere–Ionosphere Modeling 127
Fig. 9 JTGCM. This schematic summarizes the various thermosphere, ionosphere, magnetosphere processes
and their couplings that are presently incorporated into the JTGCM code. Case 1 (auroral forcing alone),
Case 2 (auroral forcing plus moderate ion drag/Joule heating), and Case 3 (auroral forcing plus strong ion
drag/Joule heating) processes are identified corresponding to JTGCM simulations presented in Bougher et al.
(2005). Future planned upgrades and processes are also indicated (orange shading)
(Bougher et al. 2005; Majeed et al. 2005, 2008). This code is based upon the NCAR TIE-
GCM framework. Neutral temperatures, 3-component neutral winds, major neutral species
(H, H
2
, He), and major ion species (H
+
,H
+
2
,H
+
3
) are simulated. The domain of this JTGCM
framework extends from 20-mbar (capturing hydrocarbon cooling) to 1.0 ×10
4
nbar (in-
cluding auroral and Joule heating processes). A 5×5
horizontal grid is utilized, along with
a 0.5 scale height vertical resolution. Auroral and Joule heating processes are incorporated,
both of which contribute to maintaining the neutral temperatures and driving the global
winds. Joule heating is produced by the action of a prescribed plasma drift pattern that is
consistent with the convection electric field derived from the Voyager constrained magne-
tospheric model of Eviatar and Barbosa (1984). Simulated auroral electrojet ion wind mag-
nitudes approach 1.5–3.0 km sec
1
. The benchmark JTGCM was fully spun-up and inte-
grated for >50 Jupiter rotations, thereby achieving steady state solutions above the 1.0-
µbar level.
Initial results from three JTGCM cases incorporating moderate realistic auroral heat-
ing, ion drag, and moderate to strong Joule heating processes were presented by Bougher
et al. (2005). Figure 9 summarizes the various thermosphere-ionosphere-magnetosphere
processes and their couplings that are presently incorporated into the JTGCM code. The
neutral horizontal winds at ionospheric heights vary from 0.5kms
1
to 1.2 km s
1
(ap-
proaching the sound speed), atomic hydrogen is transported equatorward, and auroral ex-
ospheric temperatures range from 1200–1300 K to above 3000 K, depending on the mag-
nitude of Joule heating. Figure 10 shows JTGCM simulated equatorial temperature profiles,
and comparisons with multispectral and Galileo ASI observations (Majeed et al. 2005). The
128 S.W. Bougher et al.
Fig. 10 JTGCM equatorial
temperature profiles are shown in
comparison with corresponding
temperature profiles from JIM
(curve E) and in-situ
measurements from the Galileo
ASI probe (curve A). Curve B
assumes JTGCM auroral heating
alone (no Joule heating). Curve C
is from the “best” JTGCM
simulation that incorporates
auroral plus 15% joule heating
conditions. Curve D is from the
JTGCM simulation that assumes
auroral plus 30% joule heating
conditions. Remotely sensed
(multi-spectral) temperature
observations are also displayed
(see key). From Majeed et al.
(2005)
best fit to the Galileo data implies that the major energy source for maintaining the equa-
torial temperatures is due to dynamical heating induced by the low-latitude convergence
of the high-latitude-driven thermospheric circulation. Simulated Joule heating can be quite
large, requiring a scaling factor (15%) to enable both observed equatorial and auroral
oval temperatures to be simulated (Majeed et al. 2005, 2008). Overall, the Jupiter ther-
mosphere/ionosphere system is highly variable and is shown to be strongly dependent on
magnetospheric coupling that regulates Joule heating.
Diagnostic heat balance studies utilizing the JTGCM have been employed to quantify in
detail the thermal balance processes required to maintain thermospheric temperatures con-
sistent with multi-spectral observations of Jupiter’s equatorial (Majeed et al. 2005), auroral
oval and polar cap regions (Majeed et al. 2008). It is found that upwelling/divergent winds in
the auroral oval regions provide local cooling (reducing temperatures from values otherwise
expected using 1D models), while downwelling/convergent winds in the equatorial region
provide local heating largely responsible for maintaining the warm temperatures measured.
In general, the significant Jovian auroral plus Joule heating processes appear sufficient to
drive a strong equatorward (meridional) flow that is adequate to overcome Coriolis forces
and to dynamically produce warm equatorial temperatures. This ability of Jovian equator-
ward winds to overcome Coriolis forces is apparently different than for Saturn (Mueller-
Wodarg et al. 2006), where the combined auroral plus Joule heating magnitudes at polar
latitudes are reduced from those at Jupiter (Strobel 2002). Correspondingly, Jovian auroral
and Joule heating at polar latitudes combine to drive local heating and corresponding dy-
namical cooling, together controlling the thermal structure of the ovals and the polar cap
regions. These results are beginning to address the “heat budget problem” of the Jovian
upper atmosphere that has been debated for decades (c.f. Yelle and Miller 2004).
4.2.8 Saturn (STIM)
The first general circulation model of Saturn’s thermosphere and ionosphere was presented
by Mueller-Wodarg et al. (2006) and is referred to as the STIM code. The model forms
part of a collaborative project between Boston University, Imperial College London and
Upper Atmosphere–Ionosphere Modeling 129
University College London. It calculates the response of Saturn’s upper atmosphere to so-
lar heating and ionization, including global dynamics, composition and the thermal struc-
ture. The coupled non-linear equation of momentum, energy and continuity are solved on a
global spherical pressure level grid above the 100 nbar level (800 km above the 1 bar level).
Horizontal and vertical resolutions are flexible, allowing them to be optimized for specific
problems under investigation. A simplified scheme of ion and neutral photochemistry is in-
cluded and recent additions include ion diffusion and full ion-neutral dynamical coupling.
The model has been used to investigate the thermal balance on Saturn in order to investigate
the origin of abnormally large thermospheric temperatures on Saturn and other gas giants.
It was found that the fast rotation of Saturn and strong Coriolis forces prevented energy de-
posited at auroral latitudes in the form of Joule heating from propagating equatorward and
explain the observed low latitude thermosphere temperatures (Mueller-Wodarg et al. 2006;
Smith et al. 2005), as shown in Fig. 11. The importance of global dynamics for understand-
Fig. 11 Diurnally averaged temperatures, horizontal winds and H
2
mixing ratios versus latitude and height
for equinox and solar maximum conditions, as calculated by the STIM GCM. The simulation considers Joule
heating in the auroral regions. U
South
and U
West
are meridional and zonal winds, respectively, defined as
positive southward and westward
130 S.W. Bougher et al.
ing the thermal balance on gas giants such as Saturn makes the use of general circulation
models particularly relevant there. Other studies with the ionospheric module of STIM in-
vestigated the global structure of Saturn’s highly variable ionosphere, considering shadow-
ing by Saturn’s rings (Moore et al. 2004) and effects of water precipitating into Saturn’s
ionosphere from the rings (Moore et al. 2006). These calculations found the presence of wa-
ter to be important to reproduce the dawn dusk asymmetries in electron densities observed
by the Cassini Radio Science experiment (Nagy et al. 2006). A recent study by Moore and
Mendillo (2007) proposed variable water influx rates to be responsible for the high variabil-
ity of Saturn’s ionospheric densities.
4.2.9 Titan (TTGCM)
To understand the global structure and dynamics of Titan’s thermosphere, Mueller-Wodarg
et al. (2000) presented the first general circulation model for Titan’s atmosphere above 600
km altitude. While this model, a collaborative project between University College London,
the University of Arizona and Boston University, originally was based on the terrestrial ther-
mosphere model of Fuller-Rowell and Rees (1980), it turned out that approximations fre-
quently made for Earth are no longer valid on Titan, and ultimately a new model was devel-
oped from scratch. The Titan GCM predicted solar driven day-night temperature differences
on Titan of up to 20 K which drive vigorous thermospheric dynamics. The extended nature
of the thermosphere leads to effects such as continuous solar illumination also on the night-
side at sufficiently polar latitudes. Subsequent studies showed that such winds would effec-
tively redistribute constituents in Titan’s thermosphere, particularly CH
4
, causing large local
time and hemispheric asymmetries in the CH
4
densities (Mueller-Wodarg and Yelle 2002;
Mueller-Wodarg et al. 2003), as shown in Fig. 12. These calculations showed that dynamics
and the distribution of CH
4
in Titan’s thermosphere are intimately coupled. The first in-situ
observations by the Cassini Ion Neutral Mass Spectrometer (INMS) showed that Titan’s
real thermosphere was more complex that suggested by the simple solar-driven calculations.
Current developments of the model include adding the effects of more realistic dynamics at
its lower boundary, which may importantly affect thermospheric dynamics. The model in its
latest version is being constrained by densities observed by the INMS instrument.
Fig. 12 CH
4
mole fractions near
the 1 ×10
3
nbar pressure level
for solstice at solar maximum, as
calculated by the Titan TGCM.
The average height of the
pressure level is 1354 km
Upper Atmosphere–Ionosphere Modeling 131
5 Modeling Frontiers and Problems
5.1 Lower to Upper Atmosphere Coupling
Properly addressing the coupling of the lower and upper atmospheres of planetary environ-
ments is a difficult modeling task. “Whole atmosphere” models are ultimately required to
capture the physical processes (e.g., thermal, chemical, dynamical) throughout the entire at-
mosphere from the ground to the exobase. However, diffusion processes are much different
above and below the homopause, requiring a method to be employed to bridge the transition
between the homosphere and heterosphere regions. In addition, timescales for chemical and
radiative processes vary greatly throughout the atmosphere, typically requiring small time-
steps within finite-difference codes. Numerical stability (while utilizing longer time-steps)
can be achieved in a number of ways; e.g., by employing implicit solvers and various nu-
merical filters. Finally, exercising of multi-dimensional codes on multi-processor computers
can also reduce the wall clock time for global simulations.
5.1.1 Separate but Coupled Model Frameworks vs. Whole Atmosphere Model Frameworks
Two approaches have been employed to date to capture the physics of the entire atmosphere
(ground to exobase): (a) coupling of separate lower and upper atmosphere codes; and (b) sin-
gle framework “whole atmosphere” model codes. Each approach has advantages and dis-
advantages. The coupling of separate codes permits the unique physical processes (and
timescales) of the lower and upper atmospheres to be addressed separately within codes
which can be optimized for this purpose. Molecular diffusion is one example for the upper
atmosphere, for which an implicit (vertical) formulation permits a longer model time-step to
be used. However, linking two separate models across an interface is not “seamless”. By this
we refer to the lack of an exact match of thermal and dynamical processes (e.g., solar heat-
ing, IR cooling, diffusion, numerical filtering) across this interface. Furthermore, both up-
ward and downward coupling (i.e., constituent fluxes) is not easily activated across separate
models. Whole atmosphere models obviate the need for an “artificial” boundary between
2-separate codes, while at the same time providing a continuous application of processes
throughout the ground to exobase model domain. Small time-steps may be needed to ac-
commodate disparate processes and their timescales throughout the model domain. Finally,
whole atmosphere model simulations can be visualized from “top to bottom” with a single
post-processor.
Examples of both modeling approaches are presented: (Sect. 5.1.2) the whole atmosphere
model approach for Earth (NCAR WACCM), (Sect. 5.1.3) the coupled separate model ap-
proach for Mars (NASA MGCM and NCAR MTGCM), and (Sect. 5.1.4) the upward ex-
tended LMD-MGCM. The coupled model approach for Mars is a precursor to new Mars
whole atmosphere models that are presently being developed and validated (see Table 1;see
Sects. 4.2.3 and 5.1.4).
5.1.2 NCAR WACCM (Earth)
The WACCM model version 3 (WACCM3) is a state-of-the-art climate model developed at
NCAR that extends from the Earth’s surface to the lower thermosphere. This model is an out-
growth of three independent models developed separately across three divisions at NCAR.
It combines the major features of these three independently developed models of the at-
mosphere, the Middle Atmosphere Community Climate Model (MACCM) (Boville 1995),
132 S.W. Bougher et al.
the chemical model MOZART (Brasseur et al. 1998) and the TIME-GCM (Roble 2000).
This model is one of the few high-top general circulation models that include the Hamburg
Model of the Neutral and Ionized Atmosphere (HAMMONIA) (Schmidt et al. 2006)and
the extended Canadian Middle Atmosphere Model (CMAM) (Fomichev et al. 2002). These
models have been used to study problems such as the solar influence on Earth’s climate, con-
stituent transport and trends in the middle atmosphere, the influence of the stratosphere on
the tropospheric climate and the connection between climate change and polar mesospheric
clouds. WACCM3 extends between the surface and the lower thermosphere near 140 km.
But work is now progressing to move the upper boundary to 500–700 km by incorporating
the aeronomy of the thermosphere and ionosphere from the TIME-GCM into an upward
extended WACCM.
A number of studies are underway with this new model but one, Sassi et al. (2004),
showed a coupling between El-Nino/Lanina ocean influences on the stratosphere/meso-
sphere region and another (Richter et al. 2008) showed the importance of gravity wave
forcing on the basic structure of the upper atmosphere. Details of the model can be found on
the web site http://www.cgd.ucar.edu/research/models/waccm.html.
5.1.3 NCAR Coupled MGCM-MTGCM (Mars)
The coupled NASA Ames MGCM and the NCAR MTGCM models constitute a numerical
framework of 2-independent multi-dimensional codes linked across an interface at 1.32-
microbars (60–80 km) in the Mars atmosphere (see Sect. 4.2.5). This coupled configura-
tion permits both thermal and large scale dynamical processes to be linked across the lower
and upper atmospheres of Mars (e.g., Bougher et al. 2004, 2006b). The 2-model treatment
is designed to be a testbed for addressing coupling processes in advance of the develop-
ment and validation of a comprehensive Mars “whole atmosphere” model framework (e.g.
Sect. 4.2.3).
The coupled MGCM-MTGCM system itself has been used successfully to address/
interpret an assortment of spacecraft observations, including Mars Global Surveyor, 2001
Mars Odyssey, and Mars Reconnaissance Orbiter thermosphere and/or ionosphere data sets
(Bougher et al. 1999b, 2000, 2004, 2006b). For example, the recently discovered winter
polar warming features of the Mars lower thermosphere (100–130 km) are found to vary
greatly over the Mars seasons (e.g., Keating et al. 2003; Bougher et al. 2006b). Figure 13 il-
lustrates coupled MGCM-MTGCM simulations for Ls =90 (aphelion) and 270 (perihelion)
conditions, demonstrating that the basic features of the Martian thermospheric winter polar
warming are controlled by seasonal changes in the solar plus tidal forcing, the correspond-
ing variations in the strength of the inter-hemispheric Hadley circulation, and the resulting
changes in the magnitude of the adiabatic heating near the winter poles. Calculations of polar
warming show that perihelion adiabatic heating can be highly variable from one Mars year to
the next, and more than twice as strong as that for aphelion conditions (Bougher et al. 2006b;
Bell et al. 2007). Finally, without the deep inter-hemispheric Hadley circulation made pos-
sible using these coupled lower and upper atmosphere simulations, winter polar warming
features in the Mars thermosphere would not be reproduced at all (Bell et al. 2007).
Several studies are underway utilizing this coupled MGCM-MTGCM framework. For
example, the role of interannual variations in horizontal and vertical dust distributions in af-
fecting the thermospheric temperature and wind distributions is being investigated. Factors
influencing the seasonal variation in the Mars mesopause heights and minimum tempera-
tures are also being determined (McDunn et al. 2007, 2008).
Upper Atmosphere–Ionosphere Modeling 133
Fig. 13 MGCM-MTGCM zonal averaged temperature slices as a function of height and latitude: (a) Ls =90
and (b) Ls = 270. Contour intervals are 10 K. Color shading is coordinated between these plots. From
Bougher et al. (2006b)
5.1.4 LMD-MGCM (Mars)
The Mars LMD-GCM is based on the Terrestrial climate GCM of the Laboratoire de Meteo-
rologie Dynamique (Sadourny and Laval 1984). It has been adapted by Hourdin et al. (1993)
and Forget et al. (1999), who developed the first model covering the Martian atmosphere up
to 80 km. The transport equations for the dynamics are directly taken from the LMD ter-
restrial GCM. They are based on a finite-difference formulation of the classical “primitive
equations” of meteorology which are a simplified version of the general equations of fluids
based on three main approximations: (1) the atmosphere is assumed to be a perfect gas; (2) it
is supposed to remain vertically in hydrostatic equilibrium; and (3) the vertical dimension of
the atmosphere is supposed to be much smaller than the radius of the planet (thin-layer ap-
proximation). The vertical discretization is based on σ -coordinates, where σ = p/p
s
is the
pressure p at a given grid point normalized by its local value p
s
at the surface of Mars (nor-
malized pressure coordinates). The grid is chosen to have good coverage of the atmospheric
boundary layer.
The energetics solved in the model includes the effects of suspended dust and CO
2
in the
infrared in the Mars atmosphere. For the carbon dioxide, the thermal infrared and the near
infrared absorption are solved using a NLTE approximation based on a parameterization of
the heating and the cooling. Concerning the effect of dust, the radiative transfer of the solar
radiation, including absorption and scattering, is modelled through a multi-stream approach.
Then, the scattering of the thermal infrared outside the CO
2
15 µm band is modelled using a
two-stream radiative transfer model. The CO
2
condensation-sublimation cycle is related to
the thermal balance and is likely to a be source or sink of energy, mainly through the latent
heat release associated with the change of state. This process is realistically included in the
model and the different phases of carbon dioxide are managed consistently (in particular
energy and mass conservation).
As mentioned above, the necessary discretization of time and space in the numerical
model implies that mechanisms cannot be consistently modelled, because they occur at
scales much below the lower limits of the model. They have to be added in an ad hoc way.
The relevant processes that have been included near the surface of the planet at sub-grid
scales are turbulent diffusion in the planetary boundary layer, convection, orography, and
low-level drag.
134 S.W. Bougher et al.
Fig. 14 Difference of temperatures between an LMD-GCM simulation without photochemistry and a simu-
lation with photochemistry
The LMD-GCM extension to the thermosphere and the exosphere is a collaborative work
with the University of Oxford and the Instituto de Astrofísicade Andalucía. The model has
been successively extended from the ground up to a height of approximately 120 km (Ange-
latsiColletal.2003), 240 km (González-Galindo et al. 2004) and finally up to the exosphere
(González-Galindo et al. 2007). The extension to the upper altitudes has been done in such
a way that the processes that are important either in the mesosphere or in the thermosphere
are taken into account.
In rarefied regions (thermospheres), each component of the neutral gas has an individ-
ual behaviour, which can be modelled using the Enskog approach, because this region is
still collision dominated. We must distinguish horizontal from vertical dynamics, since the
temporal and spatial scales are different. Temperature can be assumed to be the same for
all the species, so a single equation for the temperature, which accounts for the thermal
conductivity of the mean gas and the UV heating can be used. We do the same for the hor-
izontal velocity, using a mean molecular viscosity. However, vertical diffusion is important
in the upper atmosphere, and a multi-species Chapman-Enskog approach is used for multi-
component diffusion.
At upper altitudes, it is demonstrated that photochemistry becomes a very important fea-
ture of the atmospheric dynamics. A 1-D photochemical model using a complex photo-
chemistry scheme, which includes the 12 major constituents of the C, O and H families, and
accounts for 27 reactions between them has been developed and included in the LMD-GCM
model. Figure 14 shows the impact that the photochemistry may have on the energy balance
in the thermosphere. Temperature differences up to 35 K can be obtained above 200 km if
some chemical processes are neglected.
The infrared processes related to CO
2
, already mentioned above, are relevant processes
for the mesosphere, and strongly NLTE in that region. Some approximation and parameteri-
zation has been done to include them, without increasing dramatically the computation time.
The main heating source of the Martian thermosphere is the UV heating. In order to be able
Upper Atmosphere–Ionosphere Modeling 135
to reproduce the thermal structure of the Martian upper atmosphere, which is critical for
the hydrostatic equilibrium, as well as for dynamics, a parameterization has been included
based on a full 1-D UV heating model. This full model includes the absorption by CO
2
,O
2
,
atomic oxygen, H
2
,H
2
O, H
2
O
2
and O
2
in the UV–visible range.
The GCM model has been used to build a database for Martian atmosphere (Lewis et al.
1999), which has become the ESA reference model for the atmosphere of Mars.
5.2 Thermosphere/Ionosphere to Exosphere Coupling
5.2.1 Previous Exosphere Modeling Approaches
As discussed in detail by Johnson et al. (2008), the hot coronae of atomic H, O and C have
either been observed or postulated to exist at both Venus and Mars. The major reasons for
interest in these hot populations are their importance in the long-term evolution of each
atmosphere as well as the role they play in the general solar-wind interaction of each planet.
The history of Venus and Mars exospheric modeling approaches is contained in several
early studies (e.g. Cravens et al. 1980;Nagyetal.1981, 1990, 2001; Nagy and Cravens
1988; Kim et al. 1998; Hodges and Tinsley 1981, 1986; Hodges 2000). Both 1D models
(based upon 2-stream calculations and Liouville’s equation) plus multi-dimensional Monte
Carlo particle trajectory models have been employed in these exospheric simulations. Recent
1D spherical Monte Carlo models have been constructed by Cipriani et al. (2007), extend-
ing the original work of Hodges (2000) to include hot O, C, CO
2
and CO. They also have
examined more carefully the relative sources of dissociative recombination and atmospheric
sputtering (Luhmann and Kozyra 1991)byO
+
pickup ions. Very recently Chaufray et al.
(2007) published a study running a 3D Monte Carlo model for exospheric species, where
they use the 1D ion profiles from the model of Krasnopolsky (2002) and as well as night-
side ionosphere profiles obtained from Viking 1 measurements obtained from Zhang et al.
(1990). They extrapolated these profiles around Mars for various SZAs and explored the 3D
aspects of the related solar wind interaction processes.
5.2.2 Michigan Mars Exosphere Codes
In this section, the new approach of Valeille et al. (2007a, 2007b, 2008) is described which
probes the effects of the 3D structure of the thermosphere on the exosphere. They use the re-
sults of a general circulation model, but perform 2D axisymmetric Direct Simulation Monte
Carlo calculations for the SZA variation of the hot oxygen exosphere. The next major ad-
vance will be for the use of a 3D thermosphere-ionosphere model and a 3D exosphere model.
In order to merge the many local processes into a global picture, a model is needed,
which can include these important processes on small spatial scales and which can diversify
the different energy scales. The Michigan multi-species kinetic model is based on a tech-
nique called the Direct Simulation Monte Carlo (DSMC) method (Bird 1963, 1994). DSMC
was developed to simulate the transition regime, where the mean free path of particles is
too large for continuum hydrodynamics to be applicable. Individual particles are simulated
as they move around within a grid, colliding with other particles and with solid objects
(if any). Macroscopic properties, such as density, velocity and temperature are computed
by appropriately averaging particle masses, locations, velocities, and internal energies. Mo-
mentum and energy exchanges with surfaces allow for chemical reactions and sputtering
effects. Mass production can also be introduced as a boundary condition. DSMC is based
on the “rarefied-gas” assumption that over a short time ‘step’ the molecular motion and
136 S.W. Bougher et al.
Fig. 15 Escape and upward and downward exobase fluxes of hot oxygen atoms for different solar condi-
tions. Plotted on the left is the escape flux as a function of solar zenith angle for our new 2D axisymmetric
Direct Simulation Monte Carlo model of the exosphere of Mars. In addition to the escape fluxes we are also
calculating the return flux to the thermosphere and compare with what is normally assumed in thermosphere
models. The separate upward and downward fluxes at the exobase are given in the plot on the right. Left
panel curves: EHP (Equinox, High Solar Activity, Polar cut)—top curve; EHE (Equinox, High Solar Activ-
ity, Equatorial)—bottom curve
the intermolecular collisions are uncoupled and therefore can be calculated independently.
Molecules are moved over the distances appropriate for this time step, followed by the cal-
culation of a representative set of collisions. The time step is small compared to the mean
collision time, and the results are independent of its actual value.
The Michigan DSMC code was developed with the dusty-gas comet coma as the first
science application (Tenishev et al. 2008) improving in a number of fundamental and tech-
nical ways over the original comet DSMC of Combi (1996). It was also developed as a
general-purpose gas kinetic solver and tested against a number of standard aerodynamical
gas kinetic problems (Tenishev and Combi 2003a, 2003b) as well as comets. The code has
been applied both to Mars’ exosphere (Valeille et al. 2007a, 2007b, 2008) and plumes of
Enceladus (Tenishev et al. 2007, 2008).
Valeille et al. (2007a, 2007b, 2008) have applied DSMC to the hot atomic oxygen corona
of Mars, using the 3D MTGCM of Bougher et al. (2006b) (see Sect. 4.2.5) to understand
the overall day-night structure of the exosphere as well as to explore day-night, local time,
meridian, latitudinal, seasonal and solar activity variations. The goals are to understand the
range and limits on the escape rate, exosphere distribution, as well as the return fluxes of
exospheric O to the thermosphere.
A set of 2D axisymmetric DSMC models have been run with coordinates of radius and
SZA for a variety of upper thermosphere conditions taken from the 3D MTGCM: for quiet
and active sun, for equinox conditions, for solstice conditions (both at aphelion and per-
ihelion), and for thermosphere distributions around the equator and along the polar noon-
midnight meridian. The escape rates and exobase upward and downward fluxes as a function
of SZA for an equinox geometry at solar high (active) and solar low (quiet) conditions are
shown in Fig. 15. In addition, Fig. 16 shows the O density distribution in the corona for
the solar maximum and solar minimum models for the equinox geometry. Finally, Fig. 17
shows the O atom escape rates for solar maximum and minimum conditions for the new
results by Valeille et al. (2007a, 2007b, 2008) compared with published values from the
literature.
Upper Atmosphere–Ionosphere Modeling 137
Fig. 16 Hot oxygen density for different solar conditions. Shown is a false color contour plot of the hot
oxygen atom exospheric density for (a) solar active and (b) solar quiet conditions of the MTGCM model.
Using the same color-table highlights the differences of the exosphere for active and quiet solar conditions
Fig. 17 Comparison of O atom escape rates from various models (from Valeille et al. 2008). Shown are
the solar minimum and solar maximum escape rates from several models with the thermosphere conditions
used, given in parentheses. Kim et al. (1998) is corrected by a factor of 6.5 according to Nagy et al. (2001).
The Valeille et al. (2008) escape rate, averaged over the solar cycle and Mars seasons, is 1.0 ×10
26
s
1
.
Detailed Mars orbiter photochemical escape measurements are needed to validate these models
The general structure of the exosphere and escape fluxes is similar from time to time,
but the actual values of the escape rates in the extreme model cases can range over nearly a
factor of 50 when comparing conditions from solar minimum to maximum, and from Mars
perihelion and aphelion solstices with the equinoxes that are between the two extremes.
These results clearly imply that the next step in modeling Mars’ exosphere needs to be done
with a 3D exosphere model using a realistic 3D thermosphere for its base assumption.
138 S.W. Bougher et al.
References
N. Achilleos, S. Miller, J. Tennyson, A. Aylward, I. Mueller-Wodarg, D. Rees, J. Geophys. Res. 103, 20089–
20112 (1998). doi:10.1029/98JE00947
N. Achilleos, S. Miller, R. Prangé, G. Millward, M.K. Dougherty, N.J. Phys. 3 (2001). doi:10.1088/
1367-2630/3/1/303
M. Angelats i Coll, F. Forget, M.A. López-Valverde, P.L. Read, S.R. Lewis, J. Geophys. Res. 109, 1011
(2003). doi:10.1029/2003JE002163
J.M. Bell, The dynamics of the upper atmospheres of Mars and Titan. Ph.D Thesis, University of Michigan
(2008)
J.M. Bell, S.W. Bougher, V. De LaHaye, J.H. Waite Jr., A. Ridley, Eos Trans. Am. Geophys. Union 87(Jt.
Assem. Suppl.) (2006), abstract U52A-06
J.M. Bell, S.W. Bougher, J.R. Murphy, J. Geophys. Res. 112, E12002 (2007). doi:10/1029/2006JE002856
J.-L. Bertaux et al., Science 307, 566–569 (2005)
G.A. Bird, Phys. Fluids 6, 1518–1519 (1963)
G.A. Bird, Molecular Gas Dynamics and the Direct Simulations of Gas Flows (Clarendon, Oxford, 1994)
V. Boqueho, P.-L. Blelly, J. Geophys. Res. 110, A01313 (2005). doi:10.1029/2004JA010414
S.W. Bougher et al., Icarus 68, 284–312 (1986)
S.W. Bougher, R.E. Dickinson, E.C. Ridley, R.G. Roble, Icarus 73, 545–573 (1988)
S.W. Bougher, J.C. Gerard, A.I.F. Stewart, C.G. Fesen, J. Geophys. Res. 95, 6271–6284 (1990)
S.W. Bougher, W.J. Borucki, J. Geophys. Res. 99, 3759–3776 (1994)
S.W. Bougher, D.M. Hunten, R.G. Roble, J. Geophys. Res. 99, 14609–14622 (1994)
S.W. Bougher, M.J. Alexander, H.G. Mayr, in Venus II, ed. by S.W. Bougher, D.M. Hunten, R.J. Philips
(University of Arizona Press, Tucson, 1997), pp. 259–292
S.W. Bougher, S. Engel, R.G. Roble, B. Foster, J. Geophys. Res. 104, 16591–16611 (1999a)
S.W. Bougher et al., Adv. Space Res. 23(11), 1887–1897 (1999b)
S.W. Bougher, S. Engel, R.G. Roble, B. Foster, J. Geophys. Res. 105, 17,669–17,689 (2000)
S.W. Bougher, R.G. Roble, T.J. Fuller-Rowell, in Atmospheres in the Solar System: Comparative Aeronomy,
ed. by M. Mendillo, A. Nagy, H. Waite. Geophysical Monograph, vol. 130 (American Geophysical
Union, Washington, 2002), pp. 261–288
S.W. Bougher, S. Engel, D.P. Hinson, J.R. Murphy, J. Geophys. Res. 109, E03010 (2004). doi:10.1029/
2003JE002154
S.W. Bougher, J.H. Waite, T. Majeed, G.R. Gladstone, J. Geophys. Res. 110, E04008 (2005). doi:10.1029/
2003JE002230
S.W. Bougher, S. Rafkin, P. Drossart, Planet. Space Sci. 54, 1371–1380 (2006a)
S.W. Bougher, J.M. Bell, J.R. Murphy, M.A. López-Valverde, P.G. Withers, Geophys. Res. Lett. 33, L02203
(2006b). doi:10.1029/2005GL024059
S.W. Bougher, A. Brecht, C. Parkinson, S. Rafkin, B. Foster, 37th COSPAR Scientific Assembly 2008, C32-
0003-08 Abstract (2008)
B.A. Boville, J. Geophys. Res. 100, 9017–9039 (1995)
G.P. Brasseur et al., J. Geophys. Res. 103, 28265–28290 (1998)
A. Brecht, S.W. Bougher, S. Rafkin, B. Foster, Eos Trans. Am. Geophys. Union 88(52), (Fall Meeting Suppl.)
(2007), abstract P33B-1299
J.Y. Chaufray et al., J. Geophys. Res. 112
, E09009 (2007). doi:10.1029/2007JE002915
R.H. Chen, A.F. Nagy, J. Geophys. Res. 83, 1133–1140 (1978)
R.H. Chen, T.E. Cravens, A.F. Nagy, J. Geophys. Res. 83, 3871–3876 (1978)
F. Cipriani, F. Leblanc, J.J. Berthelier, J. Geophys. Res. 112, D07001 (2007). doi:10.1029/2006JE002818
M.R. Combi, Icarus 123, 207–226 (1996)
T.E. Cravens, T.I. Gombosi, A.F. Nagy, Nature 283, 178–180 (1980)
T.E. Cravens, A.J. Kliore, J.U. Kozyra, A.F. Nagy, J. Geophys. Res. 86, 11,323–11,329 (1981)
V. DeLa Haye, Coronal formation and heating efficiencies in Titan’s upper atmosphere: Construction of a
coupled ion, neutral and thermal structure model to interpret the first INMS Cassini data. Ph.D. Thesis,
University of Michigan, Ann Arbor (2005)
Y. Deng, A.D. Richmond, A.J. Ridley, H. Liu, Geophys. Res. Lett. 35, L01104 (2008). doi:10.1029/
2007GL032182
A.L. Dobbin, A.D. Aylward, M.J. Harris, J. Geophys. Res. 111, A07314 (2006). doi:10.1029/2005JA011543
A. Eviatar, A.D. Barbosa, J. Geophys. Res. 89, 7393–7398 (1984)
P.R.Fieldetal.,J.Atmos.Sol.Terr.Phys.60, 523–543 (1998)
M.O. Fillingim et al., Geophys. Res. Lett. 34, L12101 (2007). doi:10.1029/2007GL029986
V.I. Fomichev et al., J. Geophys. Res. 107 (2002). doi:1029/2001JD000479
Upper Atmosphere–Ionosphere Modeling 139
J.M. Forbes, F.G. Lemoine, S.L. Bruinsma, M.D. Smith, X. Zhang, Geophys. Res. Lett. 35, L01201 (2008).
doi:10.1029/2007GL031904
F. Forget et al., J. Geophys. Res. 104, 24155–24176 (1999). doi:10.1029/1999JE001025
J.L. Fox, Icarus 51, 248–260 (1982)
J.L. Fox, Planet. Space Sci. 36, 37–46 (1988)
J.L. Fox, in Venus and Mars: Atmospheres, Ionospheres and Solar Wind Interaction, ed. by J.G. Luhmann,
M. Tatrallyay, R. Pepin. Geophysical Monograph, vol. 66 (AGU Press, Washington, 1992), pp. 191–222
J.L. Fox, J. Geophys. Res. 109, A11310 (2004). doi:10.1029/2004JA010380
J.L. Fox, A. Dalgarno, J. Geophys. Res. 84, 7315–7333 (1979)
J.L. Fox, A. Dalgarno, J. Geophys. Res. 86, 629–639 (1981)
J.L. Fox, K.Y. Sung, J. Geophys. Res. 106(A10), 21305–21336 (2001)
J.L. Fox, K.E. Yeager, J. Geophys. Res. 111 (2006). doi:10.1029/2006JA011697
J.L. Fox, P. Zhou, S.W. Bougher, Adv. Space Res. 17, (11)203–(11)218 (1995)
T.J. Fuller-Rowell, J. Geophys. Res. 103, 3951–3956 (1998)
T.J. Fuller-Rowell, D. Rees, J. Atmos. Sci. 37, 2545–2567 (1980)
T.J. Fuller-Rowell, D. Rees, Planet. Space Sci. 10, 1209–1222 (1983)
T.J. Fuller-Rowell, D. Rees, S. Quegan, R.J. Moffett, G.J. Bailey, J. Geophys. Res. 92, 7744–7748 (1987)
T.J. Fuller-Rowell et al., in STEP Handbook on Ionospheric Models, ed. by R.W. Schunk (Utah State Univer-
sity, Logan, 1996)
T.J. Fuller-Rowell, G.H. Millward, A.D. Richmond, M.V. Codrescu, J. Atmos, Sol.-Terr. Phys. 64, 1383–1391
(2002)
F. González-Galindo, M.A. López-Valverde, M. Angelats i Coll, F. Forget, J. Geophys. Res. 110, 9005 (2004).
doi:10.1029/2004JE002312
F. González-Galindo, F. Forget, M.A. López-Valverde, M. Angelats i Coll, S.W. Bougher, LPI Contributions
1353, 3099 (2007)
R.M. Haberle et al., J. Geophys. Res. 104(E4), 8957–8974 (1999)
W.B. Hanson, S. Sanatani, D.R. Zuccaro, J. Geophys. Res. 82, 4351–4367 (1977)
M.J. Harris, A new coupled middle atmosphere and thermosphere general circulation model: Studies of dy-
namic, energetic and photochemical coupling in the middle and upper atmosphere. Ph.D Thesis, Uni-
versity College London (2001)
R.R. Hodges, J. Geophys. Res. 105, 6971–6981 (2000)
R.R. Hodges, Geophys. Res. Lett. 29(3), 1038 (2002). doi:10.1029/2001GL013852
R.R. Hodges, B.A. Tinsley, J. Geophys. Res. 86, 7649–7656 (1981)
R.R. Hodges, B.A. Tinsley, J. Geophys. Res. 91, 13649–13659 (1986)
F. Hourdin, P. Le Van, F. Forget, O. Talagrand, Meteorological variability and the annual pressure cycle on
Mars.J.Atmos.Sci.50, 3625–3640 (1993)
W.B. Hubbard et al., Astron. Astrophys. 269, 541–563 (1993)
R.E. Johnson et al., Space Sci. Rev. (2008, this issue)
W.T. Kasprzak et al., in Venus II, ed. by S.W. Bougher, D.M. Hunten, R.J. Philips (University of Arizona
Press, Tucson, 1997), pp. 225–257
G.M. Keating, N.C. Hsu, Geophys. Res. Letts. 20, 2751–2754 (1993)
G.M. Keating, J.Y. Nicholson III, L.R. Lake, J. Geophys. Res. 85, 7941–7956 (1980)
G.M. Keating et al., Science 279, 1672–1676 (1998)
G.M. Keating et al., Brief review on the results obtained with the MGS and Mars Odyssey 2001 Accelerom-
eter Experiments. International Workshop: Mars Atmosphere Modeling and Observations, Inst. de As-
trofis. de Andalucia, Granada, Spain, paper (2003)
D. Kella, L. Vejby-Christenson, P.J. Johnson, H.B. Pedersen, L.H. Andersen, Science 276, 1530–1533 (1997)
J.T. Khiel et al., J. Clim. 11, 1131–1149 (1998)
J. Kim, A.F. Nagy, T.E. Cravens, A.J. Kliore, J. Geophys. Res. 94, 11,997–12,002 (1989)
J. Kim, A.F. Nagy, J.L. Fox, T.E. Cravens, J. Geophys. Res. 103(12), 29,339–29,342 (1998)
Y.H.Kim,S.Son,Y.Yi,J.Kim,J.Kor.Ast.Soc.34, 25–29 (2001)
A.J. Kliore, G.S. Levy, D.L. Cain, G. Fjeldbo, S.I. Rasool, Science 205, 99–102 (1967)
V.A. Krasnopolsky, Photochemistry of the Atmospheres of Mars and Venus (Springer, Berlin, 1982)
V.A. Krasnopolsky, J. Geophys. Res. 107(E12), 5128 (2002). doi:10.1029/2001JE001809
M.A. Krestyanikova, V.I. Shematovitch, Sol. Syst. Res. 39, 22–32 (2005)
S. Kumar, D.M. Hunten, J. Geophys. Res. 79, 2529–2532 (1974)
S.R. Lewis et al., J. Geophys. Res. 104, 24,177–24,194 (1999)
H.-L. Liu, R.G. Roble, J. Geophys. Res. 107(D23), 4695 (2002). doi:10.1029/2001JD001533
H.-L. Liu, R.G. Roble, Geophys. Res. Lett. 32, L13804 (2005). doi:10.1029/2005GL022939
K. Lodders, B. Fegley Jr., The Planetary Scientists Companion (Oxford University Press, New York, 1998)
140 S.W. Bougher et al.
M.A. López-Valverde, D.P. Edwards, M. López-Puertas, C. Roldán, J. Geophys. Res. 103(E7), 16,799–16,812
(1998)
J.G. Luhmann, J.U. Kozyra, J. Geophys. Res. 96, 5457–5467 (1991)
Y. Ma, A.F. Nagy, I.V. Sokolov, K.C. Hansen, J. Geophys. Res. 109, A07211 (2004). doi:10.1029/
2003JA010367
T. Majeed, J.H. Waite Jr., S.W. Bougher, G.R. Gladstone, J. Geophys. Res. 110, E12007 (2005).
doi:10.1029/2004JE002351
T. Majeed, J.H. Waite Jr., S.W. Bougher, G.R. Gladstone, J. Geophys. Res. (2008, submitted)
C.R. Martinis, J.K. Wilson, M.J. Mendillo, J. Geophys. Res. 108(A10), 1383 (2003). doi:10.1029/
2003/JA009973
E. Mazarico, M.T. Zuber, F.G. Lemoine, D.E. Smith, Eos Trans. Am. Geophys. Union 88(Fall Meeting
Suppl.) (2007), abstract P32A-04
T.L. McDunn et al., The 7th International Conference on Mars, Pasadena, CA (2007), abstract XX
T.L. McDunn et al., AGU Chapman conference on the solar wind interaction with Mars. San Diego (2008),
abstract A-07
M.B. McElroy, Astrophys. J. 150, 1125–1138 (1967)
M.B. McElroy, J. Geophys. Res. 73, 1513–1521 (1968a)
M.B. McElroy, J. Atmos. Sci. 25, 574–577 (1968b)
M.B. McElroy, J. Geophys. Res. 74, 29–42 (1969)
M.B. McElroy et al., Science 194, 1295–1298 (1976)
M.B. McElroy, T.Y. Kong, Y.L. Yung, A.O. Nier, Science 194, 1295–1298 (1976)
G.H. Millward, R.J. Moffett, S. Quegan, T.J. Fuller-Rowell, in STEP Handbook on Ionospheric Models,ed.
by R.W. Schunk (Utah State University, Logan, 1996a)
G.H. Millward, H. Rishbeth, R.J. Moffett, S. Quegan, T.J. Fuller-Rowell, J. Geophys. Res. 101, 5149–5156
(1996b)
G.H. Millward et al., J. Geophys. Res. 106(A11), 24,733–24,744 (2001)
G.H. Millward, S. Miller, A.D. Aylward, I.C.F. Mueller-Wodarg, N. Achilleos, in Comparative Atmospheres
in the Solar System, ed. by M. Mendillo, A. Nagy, J.H. Waite (American Geophysical Union, Washing-
ton, 2002a), pp. 289–298
G.H. Millward, S. Miller, T. Stallard, A.D. Aylward, Icarus 160, 95–107 (2002b)
G.H. Millward, S. Miller, T. Stallard, N. Achilleos, A.D. Aylward, Icarus 173, 200–211 (2005)
L. Moore, M. Mendillo, Geophys. Res. Lett. 34, L12202 (2007). doi:10.1029/2007GL029381
L. Moore et al., Geophys. Res. Lett. 33, L22202 (2006). doi:10.1029/2006GL027375
L.E. Moore, M. Mendillo, I.C.F. Mueller-Wodarg, D.L. Murr, Icarus 172, 503–52 (2004)
I.C.F. Mueller-Wodarg, R.V. Yelle, Geophys. Res. Lett. 29 (2002). doi:10.1029/2002GL016100
I.C.F. Mueller-Wodarg, R.V. Yelle, M. Mendillo, L.A. Young, A.D. Aylward, J. Geophys. Res. 105, 20833–
20856 (2000)
I.C.F. Mueller-Wodarg, R.V. Yelle, M. Mendillo, A.D. Aylward, J. Geophys. Res. 108(A12), 1453 (2003).
doi:10.1029/2003JA010054
I.C.F. Mueller-Wodarg, M. Mendillo, R.V. Yelle, A.D. Aylward, Icarus 180, 147–160 (2006)
I.C.F. Mueller-Wodarg et al., Space Sci. Rev. (2008, this issue)
A.F. Nagy, T.E. Cravens, Geophys. Res. Lett.
15, 433–435 (1988)
A.F. Nagy, J. Kim, T.E. Cravens, Ann. Geophys. 8, 251–256 (1990)
A.F. Nagy, T.E. Cravens, S.G. Smith, H.A. Taylor, H.C. Brinton, J. Geophys. Res. 85, 7795–7801 (1980)
A.F. Nagy, T.E. Cravens, J.H. Yee, A.I.F. Stewart, Geophys. Res. Lett. 8, 629–632 (1981)
A.F. Nagy, M.W. Liemohn, J.L. Fox, J. Kim, J. Geophys. Res. 106(10), 21,565–21,568 (2001)
A.F.Nagyetal.,J.Geophys.Res.111, A06310 (2006). doi:10.1029/2005JA011519
A.O. Nier, M.B. McElroy, Science 194, 1298–1300 (1976)
L. Qian, S.C. Solomon, R.G. Roble, T.J. Kane, Geophys. Res. Lett. 33, L23705 (2006). doi:10.1029/
2006GL027185
L. Qian, S.C. Solomon, R.G. Roble, T.J. Kane, Geophys. Res. Lett. 35, L07811 (2008). doi:10.1029/
2007GL033156
S. Quegan et al., J. Atmos. Terr. Phys. 44, 619–640 (1982)
S. Rafkin, A. Stern, S. Bougher, A. Brecht, Venus Express Science team meeting. Abstract, Thuile, Italy,
March 18–24 (2007)
A.D. Richmond, E.C. Ridley, R.G. Roble, Geophys. Res. Lett. 19, 601–604 (1992)
J.H. Richter, F. Sassi, R.R. Garcia, K. Matthes, C.A. Fischer, J. Geophys. Res. (2008, in press)
A.J. Ridley, Y. Deng, G. Toth, J. Atmos. Sol. Terr. Phys. 68, 839 (2006)
H. Rishbeth, I.C.F. Mueller-Wodarg, Ann. Geophys. 17, 794–805 (1999)
H. Rishbeth et al., Ann. Geophysicae 18, 945–956 (2000a)
H. Rishbeth, R.V. Yelle, M. Mendillo, Planet. Space. Sci. 48, 51–58 (2000b)
Upper Atmosphere–Ionosphere Modeling 141
R.G. Roble, in The Upper Mesosphere Lower Thermosphere: A Review of Experiment and Theory, ed. by
R.M. Johnson, T.L. Killeen. Geophysical Monograph, vol. 87 (American Geophysical Union, 1995),
p. 1
R.G. Roble, in Atmospheric Science Across the Stratopause. Geophysical Monograph, vol. 123 (American
Geophysical Union, 2000), pp. 53–67
R.G. Roble, R.E. Dickinson, Geophys. Res. Lett. 16, 1441–1444 (1989)
R.G. Roble, E.C. Ridley, Geophys. Res. Lett. 21, 417–420 (1994)
R.G. Roble, E.C. Ridely, R.E. Dickinson, J. Geophys. Res. 92, 8745–8758 (1987)
C. Roldan, M.A. Lopez-Valverde, M. Lopez-Puertas, D.P. Edwards, Icarus 147, 11–25 (2000)
R. Sadourny, K. Laval, in New Perspectives in Climate Modeling, ed. by A. Berger, C. Nicolis (Elsevier,
Amsterdam, 1984), pp. 173–197
F. Sassi, D. Kinneson, B.A. Boville, R.R. Garcia, R.G. Roble, J. Geophys. Res. 109, D17108 (2004).
doi:10.1029/2003JD004434
H.G. Schmidt et al., J. Climate 19 (2006). doi:10.1175/JCLI3829.1
G. Schubert et al., in Exploring Venus as a Terrestrial Planet. Geophysical Monograph, vol. 176 (American
Geophysical Union, 2007), pp. 101–120
R.W. Schunk, A.F. Nagy, Ionospheres: Physics, Plasma Physics and Chemistry. Cambridge Atmospheric and
Space Science Series (Cambridge University Press, Cambridge, 2000)
R. Shimazaki, M. Shimizu, Rep. Ionos. Space Res. Jpn. 24, 80–98 (1970)
H. Shinagawa, T.E. Cravens, J. Geophys. Res. 93, 11,263–11,277 (1988)
H. Shinagawa, T.E. Cravens, J. Geophys. Res. 94, 6506–6516 (1989)
C.G.A. Smith, S. Miller, A.D. Aylward, Ann. Geophys. 23, 1943–1947 (2005)
R.W. Stewart, J. Atmos. Sci. 25, 578–582 (1968)
R.W. Stewart, J. Atmos. Sci. 28, 1069–1073 (1971)
D.F. Strobel, in Atmospheres in the Solar System: Comparative Aeronomy, ed. by M. Mendillo, A. Nagy,
H. Waite. Geophysical Monograph, vol. 130 (American Geophysical Union, Washington, 2002), pp.
7–22
V.T. Tenishev, M.R. Combi, AIAA Paper 2003-3776 (2003a)
V.T. Tenishev, M.R. Combi, RGD, 23rd Int. Symp. 663, 696 (2003b)
V.T. Tenishev, M. Combi, H. Waite, Eos Trans. Am. Geophys. Union 88(23) (Jt. Assem. Suppl.) (2007),
abstract P43A-02
V.T. Tenishev, M.R. Combi, B. Davidsson, Astrophys. J. (2008, in press)
R.H. Tolson et al., J. Spacecr. Rockets 44(6), 1172–1179 (2007)
A. Valeille, V. Tenishev, S.W. Bougher, M.R. Combi, A.F. Nagy, Eos Trans. Am. Geophys. Union 88(23) (Jt.
Assem. Suppl.) (2007a), abstract SA31B-02
A. Valeille, V. Tenishev, S.W. Bougher, M.R. Combi, A.F. Nagy, B. A. A. S. 39, #24.01 (2007b)
A. Valeille, M.R. Combi, V.T. Tenishev, S.W. Bougher, A.F. Nagy, Icarus (2008, submitted)
U. von Zahn et al., J. Geophys. Res. 85, 7829–7840 (1980)
J.H. Waite et al., Science 308, 982–986 (2005)
O. Witasse et al., Space Sci. Rev. (2008, this issue)
P.G. Withers, Geophys. Res. Letts. 33, L02201 (2006). doi:10.1029/2005GL024447
R.V. Yelle, S. Miller, in Jupiter: The Planet, Satellites, and Magnetosphere, ed. by F. Bagenal, T.E. Dowling,
W.B. McKinnon (Cambridge Univ. Press, New York, 2004), pp. 185–218
M.H.G. Zhang, J.G. Luhmann, A.J. Kliore, J. Kim, J. Geophys. Res. 95, 14829–14839 (1990)
S. Zhang, S.W. Bougher, M.J. Alexander, J. Geophys. Res. 101, 23195–23205 (1996)
... For the Earth-like case, the 1 bar atmosphere of TOI-700 d is akin to that of the modern Earth, i.e., primarily N 2 and O 2 . In this study, we utilize and rescale the 1D averaged upper atmospheric profiles from the thermosphere general circulation models of Venus and Earth (Bougher et al. 2008). In the case of Venus/Venus-like planets, the dominant neutral species transitions from molecular CO 2 to atomic O with increasing altitude in the thermosphere. ...
... By utilizing a five-species multifluid MHD model, we solve e − , H + , O + , + O 2 , and + CO 2 for the Venus-like case, and the dominant three fluid species (e − , H + , and O + ) for the Earthlike case, with these species specified by the dominant ionospheric components (Bougher et al. 2008); we deliberately adopt a Venus-like atmospheric composition given its importance in exoplanetary science (Kane et al. 2019). At the model inner boundary, the species densities satisfy the photochemical equilibrium conditions (Schunk & Nagy 2009). ...
... The angular (i.e., horizontal) resolution is chosen to be 3°×3°. However, as the upper atmosphere of a Venus-like planet is cooler than that of a Earth-like planet due to the efficient CO 2 cooling (Bougher et al. 2008), the scale height associated with the former's atmosphere is correspondingly smaller than the latter; therefore, we adopt the smallest radial resolution of 5 km and 10 km for Venus and Earth analogs, respectively, to span the whole upper atmospheric region. The code is run in the planet-star-orbital coordinate system, where the x-axis is directed from the planet toward the star, the z-axis is perpendicular to the planet's orbital plane, and the y-axis completes the right-hand system. ...
Article
Full-text available
The recent discovery of an Earth-sized planet (TOI-700 d) in the habitable zone of an early-type M-dwarf by the Transiting Exoplanet Survey Satellite constitutes an important advance. In this Letter, we assess the feasibility of this planet to retain an atmosphere—one of the chief ingredients for surface habitability—over long timescales by employing state-of-the-art magnetohydrodynamic models to simulate the stellar wind and the associated rates of atmospheric escape. We take two major factors into consideration, namely, the planetary atmospheric composition and magnetic field. In all cases, we determine that the atmospheric ion escape rates are potentially a few orders of magnitude higher than the inner solar system planets, but TOI-700 d is nevertheless capable of retaining a 1 bar atmosphere over gigayear timescales for certain regions of the parameter space. The simulations show that the unmagnetized TOI-700 d with a 1 bar Earth-like atmosphere could be stripped away rather quickly (<1 gigayear), while the unmagnetized TOI-700 d with a 1 bar CO2-dominated atmosphere could persist for many billions of years; we find that the magnetized Earth-like case falls in between these two scenarios. We also discuss the prospects for detecting radio emission of the planet (thereby constraining its magnetic field) and discerning the presence of an atmosphere.