Journal of Geophysical Research: Planets
RESEARCH ARTICLE
10.1002/2013JE004491
Key Points:
• Comparison of Mars and Earth
thermospheres
• Mars exosphere density variability
from 1999 to 2010
• Daily mean densities derived from
orbit perturbation analysis
Supporting Information:
• Readme
Correspondence to:
S. Bruinsma,
[email protected]
Citation:
Bruinsma, S., J. M. Forbes, J.-C. Marty,
X. Zhang, and M. D. Smith (2014),
Long-term variability of Mars’ exosphere based on precise orbital analysis
of Mars Global Surveyor and Mars
Odyssey J. Geophys. Res. Planets, 119,
doi:10.1002/2013JE004491.
Received 24 JUL 2013
Accepted 28 DEC 2013
Accepted article online 4 JAN 2014
Long-term variability of Mars’ exosphere based on precise
orbital analysis of Mars Global Surveyor and Mars Odyssey
Sean Bruinsma1 , Jeffrey M. Forbes2 , Jean-Charles Marty1 , Xiaoli Zhang2 , and Michael D. Smith3
1
Department of Terrestrial and Planetary Geodesy, Centre National d’Etudes Spatiales, Toulouse, France, 2 Department of
Aerospace Engineering Sciences, University of Colorado Boulder, Boulder, Colorado, USA, 3 NASA Goddard Space Flight
Center, Greenbelt, Maryland, USA
Abstract A long-term perspective on Mars’ exosphere variability at 405 km is provided by merging
together density data derived from precise orbit determination of the Mars Global Surveyor and Mars
Odyssey (MO) satellites extending from 2001 to 2010. These data are heavily weighted toward afternoon
local times at high latitudes in the Southern Hemisphere. Clear long-term solar and annual variations are
well captured by empirical formulas. Residuals from the empirical fit show evidence for relative depletions
in exosphere density around Mars’ closest approach to Earth, which would be consistent with a scavenging
mechanism that is dependent on solar wind dynamic pressure. Superimposed on this variation with
Mars-Sun distance are positive density residuals during Mars year (MY)25, MY27, and MY29 that are
apparently due to elevated dust levels in Mars’ middle atmosphere. However, during MY24, MY26, and MY28
there are dust level increases without any corresponding increase in exosphere density. We suspect that this
inconsistency is related to a variable ability to sense the response to dust-related effects, imposed by the
high-latitude limitations of our measurements combined with interference between the mechanisms that
translate middle atmosphere heating to an exosphere response. Evidence also supports the hypothesis that
winter helium bulge effects contributed to the inferred interannual density variability during the 2007–2009
solar minimum period, when the O-He transition height likely resided near the ∼405 km orbit of MO.
1. Introduction
Thermosphere and exosphere temperatures and densities are fundamental parameters of planetary
atmospheres that reflect the effects of variable solar energy inputs over timescales ranging from hours to
decades. Quantifying and understanding contemporaneous changes in the thermospheres and exospheres
of Earth and Mars to changes in solar flux serve to constrain comparative planetary thermosphere simulations and may help resolve existing uncertainties in thermal balance processes [Bougher et al., 2009], as
well as factors controlling chemical composition and escape [Krasnopolsky, 2010]. Such comparisons have
been performed using densities derived from precise orbit determination of the Mars Global Surveyor (MGS)
spacecraft at Mars and the CHAMP satellite at Earth, with respect to 27 day solar rotation variations during
2002 and longer-term solar cycle variations between 1999 and 2005 [Forbes et al., 2006, 2008]. However,
solar activity reached a deep minimum during 2007–2009, and it remains unknown how Mars’ thermosphere and exosphere responded to this unusual depression in solar flux, and whether any differences in
response between Mars and Earth might have occurred.
Exosphere densities at Mars could also be affected by solar forcing of another type—that of pressure pulses
in the solar wind. Recent analyses [e.g., Lundin, 2008a, 2008b; Dubinin et al., 2009; Edberg, et al., 2010] of
ion fluxes by the ion mass analyzer instrument on Mars Express (MEx) reveal that dynamic pressure pulses,
mostly in the form of corotating interactions regions (CIRs), significantly increase the escape rates of heavy
planetary ions (O+ , O+2 , and CO+2 ) from Mars. It appears that both EUV solar flux and solar wind dynamic
pressure play important roles [Lundin, 2008a]. The solar flux increases ionization levels, inflates the thermosphere, ionosphere, and exosphere, thus presenting a larger “target” or “obstacle size” for subsequent stripping by the CIR. The CIRs on the other hand, do not appear to significantly affect the obstacle size, but the
momentum transfer between the solar wind and Mars’ exosphere, and hence the escape rate, increases as
the dynamic pressure increases. Scavenging has been observed at least down to the 270–300 km altitudes
accessed by MEx [Dubinin et al., 2009]. The question remains, might neutral densities in the exosphere be
affected by such events as well?
BRUINSMA ET AL.
©2014. American Geophysical Union. All Rights Reserved.
1
Journal of Geophysical Research: Planets
10.1002/2013JE004491
Figure 1. Mars Odyssey dayside local solar time (h) at the equator.
In this paper we combine previous exosphere densities derived from precise orbit determination of MGS
[Forbes et al., 2008] during 1999–2004, with newly derived similar data from Mars Odyssey (MO) during
2004–2010, to create a unified data series at 405 km extending from 1999 to 2010. This represents the
longest record in existence concerning the state of Mars’ exosphere. These data enable us to examine the
variation in Mars’ exosphere density during the transition from the solar maximum of 2001 through the
deep solar minimum of 2007–2009 and to compare with the contemporaneous solar response at Earth
as determined by the CHAMP satellite from 2001 to 2010. In addition, our data set offers the opportunity
to examine possible relationships with dust storms and connections with solar wind dynamic pressure, as
described above. These data should constitute an invaluable long-term reference for anticipated neutral
composition measurements to be made between 120 and 350 km by the Neutral Gas Ion Mass Spectrometer instrument on the Mars Atmosphere Volatile Evolution (MAVEN) mission (http:lasp.colorado.edu/maven).
MAVEN measurements are scheduled to commence in late 2014.
As detailed by Krasnopolsky [2010], disagreement exists within the literature as to the solar cycle dependence of exosphere temperatures that are internally consistent with the long-term variation of exosphere
densities derived from orbital analysis of MGS [Forbes et al., 2008] between 1999 and 2005. While firstprinciples numerical models [Bougher et al., 2009; Gonzalez-Galindo et al., 2009] are in basic agreement
with temperatures derived from the Drag Temperature Model (DTM)-Mars empirical model [Bruinsma and
Lemoine, 2002], Krasnopolsky [2010] maintains that taking proper account of eddy diffusion, chemistry, and
escape leads to a much smaller variation of exosphere temperature for the same range of exosphere densities. For this reason, the current study is confined to total mass densities derived from MO, and no attempt
is made to derive exosphere temperatures. Krasnopolsky [2010] does suggest that some of the above discrepancy could be resolved if the MGS densities at lower than average solar activity represent overestimates
of the true densities, perhaps due to the difficulties of estimating densities under these difficult conditions.
The above controversy points to the importance to Mars thermosphere and exosphere science of providing
MO densities through the deep solar minimum of 2007–2009 and to the need to provide reliable uncertainty estimates as part of our analysis. The following section describes our procedure for data processing
and uncertainty estimation.
2. Odyssey Density Derivation
The Mars Odyssey (MO) spacecraft has been in a near-polar sun-synchronous orbit at a mean altitude
near 410 km since February 2002, with periapsis near 85◦ S. From the beginning of the mission through
September 2008, MO was in a 5:00 A.M./5:00 P.M. orbit. After a series of maneuvers, the local solar time of
the orbit was changed to 3:45 A.M./3:45 P.M. as of June 2009. The dayside local time at equator crossings is
shown in Figure 1.
We have derived mean neutral densities of Mars’ exosphere using estimated density scale factors derived
from MO precise orbit computations. The accuracy of the estimated drag scale factors depends on the
BRUINSMA ET AL.
©2014. American Geophysical Union. All Rights Reserved.
2
Journal of Geophysical Research: Planets
10.1002/2013JE004491
Figure 2. Mars density (gcm−3 ) at 405 km altitude versus year, 2002–2010. Black dots represent MO densities, and the black line represents 81 day running means of these values. The blue line corresponds to Mars-GRAM 2005 model densities. The solid and dotted red
lines denote the mean F10.7 flux at Mars and Earth-Sun-Mars angles, respectively.
quality of the trajectory determination and the tracking system. The former requires accurate gravity field
and radiation pressure models, whereas the latter requires accurate and evenly distributed tracking data.
Leakage of nondensity related errors into the estimated drag scale factors takes place if these conditions
are not met. A second concern is the drag coefficient, which models the momentum exchange between
the satellite surface and the colliding particles and so affects the deduced density directly. Our modeling
includes a 110 × 110 spherical harmonic model of the Mars gravity field (MRO110B) [Konopliv et al., 2011]
and a nine-plate macromodel (spacecraft bus, solar array, and high-gain antenna) to model the surface
forces, including the effect of self-shadowing. MO attitude information in the form of telemetered quaternions were used to orient the macromodel in inertial space and to orient both the articulating solar array
and the high-gain antenna with respect to the spacecraft body. Tracking data from the Deep Space Network
(DSN) were processed in daily arcs (i.e., 24 h). The adjusted parameters in each arc include the spacecraft
state, and a drag scale factor (nominally 1.0) and solar radiation pressure reflectivity coefficient (nominally
1.0) per arc. Angular momentum desaturation (AMD) maneuvers are modeled explicitly. The drag scale
factor is used to scale a model density, which results in a mean density with temporal resolution of 1 day and
a spatial resolution limited to the MO orbital plane and altitude. In this study, Mars-Global Reference Atmospheric Model (GRAM) version 2005 (see see.msfc.nasa.gov/tte/model_Marsgram.html), an engineering-level
atmosphere model, was used. It is based on first-principles model results: on the NASA AMES Mars General
Circulation Model from 0 to 80 km, and on the Mars Thermospheric General Circulation Model at higher
altitudes. We did not explicitly model the attenuating effect of dust on the radiation pressure as in Konopliv
et al. [2011], but it is very limited in time and at least partly corrected for through the estimation of a radiation pressure scaling factor. Second, tests verified that the drag scale factor is unaffected by the radiation
pressure scaling factor.
Daily densities from the orbital decay analysis of MO are presented in Figure 2, along with an 81 day running
mean of the daily densities (black line). The quasi 2 year variations in density and F10.7 in Figure 2 are Mars
orbital effects. The MO densities shown are selected based on the Root Mean Square (RMS) of fit and the
number of DSN Doppler residuals of the arc (< 0.004 Hz and > 500, respectively), as well as the value of the
estimated scale factors and their formal errors. The Earth-Sun-Mars angle 𝛼ESM is also shown in this figure,
since the DSN observations become very noisy near conjunction (𝛼ESM =0), and more data tend to be rejected
and reflect greater variability around these periods. For a study focusing on long-term variability, opposition
effects do not impose a significant shortcoming. More scatter is also seen extending from 2008–2010. This
is due to the simultaneous combination of low densities occurring during this minimum in solar activity,
BRUINSMA ET AL.
©2014. American Geophysical Union. All Rights Reserved.
3
Journal of Geophysical Research: Planets
10.1002/2013JE004491
Figure 3. Variation of Mars density at 405 km from 1999 to 2010 and Earth density at 360 km from 2002 to 2010. The solid black line
represents 81 day running mean densities combining both MGS and MO data into a single time series. The black dotted line corresponds
to a least squares fit to these data, and the vertical bars are uncertainty estimates as explained in the text. The CHAMP density data are
represented by the blue line, and the F10.7 solar flux adjusted to Mars is given by the red line; both are 81 day running means.
the major orbit maneuver mentioned earlier, Earth-Sun opposition, and fewer tracking measurements
over 2009. Computing longer arcs in order to increase the (cumulative) effect of drag in this case does not
improve the estimation of the drag scale factors because of the AMD maneuvers every (or every other) day.
Also shown in Figure 2 are the corresponding 81 day mean Mars-GRAM densities, which are seen to severely
underestimate the MO densities during 2005 and 2009. During 2005, the spread in MO densities about the
running mean is rather small, and the model-data discrepancy is without question. Although the scatter
in densities during 2009 is much greater, the mean of the density points is well above that of Mars-GRAM.
The local time change from 2007 to 2009, i.e., 2 h earlier in the afternoon, causes a density increase of only
5–15% according to DTM-Mars [Bruinsma and Lemoine, 2002] or Mars-GRAM 2005, leaving more than 50%
of the observed amplitude unexplained. These model-data discrepancies underscore the value and need for
the type of data presented here.
3. Merged MGS and MO Density Data Set and Analysis
In Figure 3, we present a merged data set combining MGS densities presented previously [Forbes et al., 2008]
with the 81 day mean densities shown in Figure 2. The MGS densities were first converted from 390 km to
405 km using Mars-GRAM according to the following formula:
𝜌405 = 𝜌390
𝜌model 405
𝜌model 390
(1)
Since the MGS and MO data were least noisy during the overlapping year of 2004, this year was used
to determine a calibration scale factor of 1.8 for the MO densities that was carried through to following years. This scaling is justified because of the different orbit determination programs, i.e., different
force models and satellite macromodels, used to derive MGS and MO densities. The MO data were scaled
to MGS because it is an already published data set. So, the data before 2004 in the merged curve are
MGS densities, and those after 2004 originate from the MO data analysis after scaling, whereas 2004 is
the average of MGS and MO. Periapsis of MGS and MO are confined to high southern latitudes between
1400 and 1800 local time. However, one is reminded that the densities retrieved from the orbits of
MGS and MO are not point densities, but represent some average drag distributed along the orbit. The
displayed densities are heavily weighted toward afternoon local times at high southern latitudes, but
do contain significant influence from other latitudes and even the nightside of the planet. This is shown
BRUINSMA ET AL.
©2014. American Geophysical Union. All Rights Reserved.
4
Journal of Geophysical Research: Planets
10.1002/2013JE004491
Figure 4. Density ratios along the MO orbit of three latitude bands to the 75◦ S band and the average ratio (green), and Ls , in degrees
(orange; right axis).
in Figure 4, which displays the ratios of Mars-GRAM densities at 75◦ S to those in three 30◦ latitude bands
along the MO orbit centered at 30◦ S, 30◦ N, and 75◦ N. When this ratio approaches unity, densities in both latitude bands are nearly equal. The higher densities in the 75◦ N latitude band during Northern Hemisphere
winter relative to Southern Hemisphere summer (red curves in mid-2007 and mid-2009 dipping to near 1.0
around Ls = 270) are due to the so-called winter helium bulge [Keating and Prior, 1968], which is modeled
in Mars-GRAM. Note also that in Southern Hemisphere winter (around Ls = 90) that the red curve reaches
almost 6.0 during mid-2008, indicating significant He abundance in the Southern Hemisphere. These
extrema in the 75◦ S/75◦ N density ratio from mid-2007 to mid-2009 are likely due to a lowering of the He-O
transition height to near-MO orbital altitudes during this solar minim period, similar to the situation with the
GRACE satellite at Earth [Bruinsma and Forbes, 2010; Liu et al., 2013; Thayer et al., 2012].
Figure 3 includes t interval uncertainty estimates for each independent 81 day mean density. These were
calculated based on the central limit theorem and Student’s t distribution/t test [Venables and Ripley, 2002].
Basically, with a certain, e.g., 95% confidence applied here, the real mean will be within the t interval which
is the mean ±t∗ √𝜎N , where t∗ is determined by the degree of freedom N − 1 and the confidence level, and 𝜎 is
the standard deviation of the N data points (N = 81). Also included in Figure 3, for reference, are the 81 day
mean solar flux adjusted to Mars and the contemporaneous 81 day mean orbit-averaged densities observed
at Earth by the accelerometer on the CHAMP satellite. The dotted line represents a nonlinear least squares
fit to the merged densities (correlation coefficient = 0.933), as given by the following formula:
𝜌405 km = [0.81 + 0.1613F10.7] × [1.0 + 0.411 cos(Ls − 247◦ )]
(2)
The form of this equation follows that of Krasnopolsky [2010], who found it to be superior to the one originally employed in the Forbes et al. [2008] study. We also examined a quadratic dependence on the solar flux,
but it did not appreciably improve the quality of fit. Following Krasnopolsky’s [2010] example, we display
in Figure 5 the annual term and solar flux term separately, and provide the solar flux dependence of the
CHAMP densities for comparison. We note that for a factor of 10 density increase at Earth (from F10.7 = 65 to
215) that the Mars density increases by a factor of 3.6 for the corresponding solar fluxes at Mars (from
F10.7 = 35 to 110).
There is evidence that the 10.7 cm solar flux does not capture the degree to which the EUV flux minimized
during the deep solar minimum of 2007–2009 [Chen et al., 2011; Solomon et al., 2011; Thayer et al., 2012].
As a point of reference, Thayer et al. [2012] found that a downward adjustment of F10.7 by about 14 solar
flux unit (sfu) (from 67 sfu to 53 sfu) was needed to reconcile Mass Spectrometer Incoherent Scatter (MSIS)
model densities with CHAMP and GRACE measurements during 2008. This adjustment is consistent with
studies by Solomon et al. [2011] and Chen et al., [2011] using different data and models. For Mars, this would
BRUINSMA ET AL.
©2014. American Geophysical Union. All Rights Reserved.
5
Journal of Geophysical Research: Planets
10.1002/2013JE004491
Figure 5. (top) Density variability with respect to solar flux received at Mars (symbols and linear fit in black) and Earth (symbols and
linear fit in blue). (bottom) Annual density variation at Mars.
imply assigning to densities during 2008 an F10.7 value that is lower by 6 sfu, obviously not enough to
change the fit displayed in Figures 3 and 5. As we show below, there are more important effects to consider
when interpreting the density levels and variability during the solar minimum period of 2007–2009.
Figure 6 illustrates the differences between the fit and the 81 day mean densities shown in Figure 3. These
density differences represent that part of the density variability that is not captured by solar flux and annual
variations. Also indicated in this figure is the Mars-Sun distance (RSM , in units of 1 AU) and dust optical
depth measured by the MGS/Tropospheric Emission Spectrometer (TES) (1999–2003) [Smith, 2004] and
MO/Thermal Emission Imaging System (THEMIS) (2003–2010) [Smith, 2009] instruments, averaged over all
longitudes and between ±30◦ latitude. The latter serves as a proxy for dust content in the lower and middle
atmosphere of Mars.
Examination of all three curves in Figure 6 leads to some interesting as well as perplexing results. First,
the density differences appear to follow RSM , but with distinct interruptions in this behavior between
Ls = 270 ± 90 (south hemisphere summer) during Mars year (MY)25, MY27, and MY29. Also note that all
BRUINSMA ET AL.
©2014. American Geophysical Union. All Rights Reserved.
6
Journal of Geophysical Research: Planets
10.1002/2013JE004491
Figure 6. The black line in this figure illustrates differences between the fit and the 81 day mean densities shown in Figure 2. The red
line represents Mars-Sun distance, and the dotted line is dust optical depth measured by the MGS/TES (1999–2005) and MO/THEMIS
(2005–2010) instruments, averaged over all longitudes and between ±30◦ latitude.
the associated maxima in the RSM curve occur when the dust opacity is at a minimum. A decrease in exosphere density during closest approach to the Sun would be consistent with enhanced solar wind dynamic
pressure and exosphere escape if a mechanism exists that connects neutral atmosphere escape with the
ionospheric scavenging discussed briefly in section 1. We furthermore note that the “interruptions” to this
apparent behavior take the form of positive density differences in MY25, MY27, and MY29. Given the coincidence with enhancements in dust opacity during these periods, one might conclude some connection with
enhanced middle atmosphere solar radiation absorption/heating, and some type of “atmospheric inflation”
that extends to thermosphere and exosphere altitudes. The perplexing aspect of these results is the absence
of any such response during MY28, the period of most intense dust activity at Mars since 2002 (MY25). We
also note here that the helium bulge effects discussed previously act in the direction to enhance the above
interannual residual during 2007–2009. The above issues are discussed further in the following section.
4. Summary, Discussion, and Conclusions
In this paper we present a long-term perspective on exosphere density behavior at Mars, corresponding
to 405 km altitude at high southern latitudes during 1999–2010, and heavily weighted toward afternoon
(∼1400–1800 LT) conditions. These densities were obtained from precise orbit determination of the MGS
(1999–2004) and MO (2004–2010) satellites. Much of the observed variability in the 81 day mean densities
is captured in a least squares sense by the product of a linear dependence on the 81 day mean 10.7 cm
solar radio flux, F10.7, and an annual term that is also linearly dependent on F10.7. Contemporaneous
density measurements at 360 km at Earth are also presented and compared with those at Mars. We note
that for a factor of 10 density increase at Earth (from F10.7 = 65 to 215) that the Mars density increases by a
factor of 3.6 for the corresponding solar fluxes at Mars (from F10.7 = 35 to 110). This comparative planetary
result is of potential importance to establishing the veracity of first-principles models that seek to produce
consistent results.
Examination of those aspects of the density variability not accounted for by single mathematical relationships that depend on F10.7 reveals two interesting points. First, there is a tendency for Mars exosphere
density to decrease near perihelion (Ls ≈ 251), that is, when Mars is closest to the Sun. We speculate that
this might be a neutral atmosphere effect that accompanies scavenging of the ionosphere by solar wind
dynamic pressure as described, e.g., by Lundin [2008a, 2008b]. Given the tilt of Mars’ rotation axis, around
perihelion, Mars’ daytime high-latitude exosphere is tilted in the direction of the solar wind, which suggests
at least heuristically maximum exposure to atmospheric scavenging of solar wind origin. Based on Mars
Express, dayside electron density measurements near 400 km [Duru et al., 2008], and under the assumption
BRUINSMA ET AL.
©2014. American Geophysical Union. All Rights Reserved.
7
Journal of Geophysical Research: Planets
10.1002/2013JE004491
of charge neutrality with O+ , only about 1% of the satellite drag sensed by MO is due to ionospheric ions;
therefore, we are referring to a potential neutral atmosphere effect here.
Second, around perihelion, other events occur that tend to obscure a more definitive interpretation in terms
of scavenging, as proposed above. Southern Hemisphere summer (Ls ≈ 210–330) is dust storm season at
Mars. It is well-known that elevated dust levels are accompanied by significantly elevated solar radiation
absorption in Mars’ middle atmosphere. This leads to thermal expansion of Mars’ atmosphere [Bougher et al.,
1997] and to enhanced excitation of the solar semidiurnal tide [Bougher et al., 1997; Forbes and Miyahara,
2006]; both of these effects extend into the thermosphere. It is also likely that gravity wave fluxes are
enhanced during dust storms, and it is now known that gravity waves exert a significant influence on the
mean structure and circulation of the thermosphere [Medvedev and Yigit, 2012]. Observational evidence furthermore exists that quantify the thermosphere density increases that occur below 200 km in connection
with dust storms at Mars [England and Lillis, 2012], and reveals the density response to be maximum at low
latitudes and relatively small at high latitudes. According to Mars thermosphere GCM’s [Bougher et al., 1997,
2006], thermosphere density response to elevated dust is more complicated than just thermal expansion
or direct response to the solar semidiurnal tide. Along with differential heating, momentum deposition by
the semidiurnal tide and possibly gravity waves drive a circulation system in the thermosphere that produce
important temperature and density variations that result from adiabatic heating and cooling due to the vertical winds that accompany this modified circulation [Bougher et al., 1997, 2006]. In fact, while Bougher et al.
[1997] find factors of 5–10 increases in density near 110 km, near the exobase (≈ 220 km) a 10–20 K cooling
is found in the dayside Southern Hemisphere, a ∼50 K warming at northern polar latitudes, and a ∼20–50 K
warming on the nightside at all latitudes.
It is not clear how the exosphere density response at 405 km at high latitudes under afternoon summer
conditions fits into the current observational and modeling picture. Based on what we do know, the sampling of Mars’ exosphere obtainable by precise orbit determination of MGS and MO is certainly not optimum
for measuring the response to elevated dust levels in Mars’ middle atmosphere. The absence of an observable response in MY28, but detection of an apparent response during the less intense dust storm periods
in MY27 and MY29 is perplexing. Note also that there is no apparent response to enhanced dust levels during MY24 and MY26, years of dust elevation similar to MY27 and MY29. It may be that the combined effects
of atmospheric inflation, direct perturbations by the solar semidiurnal tide, and adiabatic heating and cooling effects due to momentum deposition by a variety of dissipating waves, combined with the high-latitude
limitations of our measurements, makes our ability to sense the response to dust-related effects sensitive to
all of these factors and therefore high variable. It is also possible that the exosphere response may depend
on how high into the middle atmosphere that the dust is distributed. A sensitivity analysis by one or more
whole atmosphere-exosphere models could shed some light here.
According to our analysis, and despite known shortcomings, the use of F10.7 as a proxy for EUV flux during the solar minimum of 2007–2009 has not significantly affected our interpretation of density levels or
variability during this period. There are other factors that overshadow this effect. First, beginning in late
2008 an orbit maneuver occurred that shifted the local time by about two hours earlier; we believe that this
contributed at about the 20% level to the increase that is measured during 2009. In addition, during this
solar minimum period, it appears that contraction of the thermosphere caused the O-He transition height
to descend to the vicinity of the ∼405 km orbital altitude of MO, leading the inferred densities to include
a contribution from the so-called “winter helium bulge” that is more well studied at Earth, particularly for
this most recent solar minimum [Bruinsma and Forbes, 2010; Liu et al., 2013; Thayer et al., 2012]. Evidence
indicates that during mid-2007 and mid-2009, during Southern Hemisphere summer, that enhanced winter
He densities contributed to the drag on MO about equally to that at periapsis near 75◦ S. We have therefore
overestimated the densities attributable to 75◦ S during mid-2007 and mid-2009, the latter adding to the
local time enhancement during 2009 as noted above. In addition, during mid-2008 (Southern Hemisphere
winter) inferred densities at 75◦ S are enhanced relative to those that would have existed in absence of the
He bulge.
The above effects have implications in terms of using the data presented here to validate Mars thermosphere models. A correct data-model comparison would involve comparing densities inferred from orbital
analysis with densities sampled and integrated along the MGS/MO orbits within the model, taking into
account local time changes and effects due to the winter helium bulge. Our results certainly raise questions
BRUINSMA ET AL.
©2014. American Geophysical Union. All Rights Reserved.
8
Journal of Geophysical Research: Planets
10.1002/2013JE004491
as to what type of exosphere environment might be viewed by MAVEN, and may assist in development of
observational strategies for that mission.
The MGS and MO densities presented in this paper are available in ASCII format as a supplemental file to this
paper.
Acknowledgments
This work was supported in part from
NASA grant NNX12AQ20G under
the Mars Data Analysis Program
to the University of Colorado, and
CNES/TOSCA.
BRUINSMA ET AL.
References
Bougher, S. W., J. Murphy, and R. M. Haberle (1997), Dust storm impacts on the Mars upper atmosphere, Adv. Space Res., 19, 1255–1260.
Bougher, S. W., J. M. Bell, J. R. Murphy, M. A. Lopez-Valverde, and P. G. Withers (2006), Polar warming in the Mars thermosphere: Seasonal
variations owing to changing insolation and dust distributions, Geophys. Res. Lett., 33, L02203, doi:10.1029/2005GL024059.
Bougher, S. W., T. M. McDunn, K. A. Zoldak, and J. M. Forbes (2009), Solar cycle variability of Mars dayside exospheric temperatures:
Model evaluation of underlying thermal balance, Geophys. Res. Lett., 36, L05201, doi:10.1029/2008GL036376.
Bruinsma, S. L., and J. M. Forbes (2010), Anomalous behavior of the thermosphere during solar minimum observed by CHAMP and
GRACE, J. Geophys. Res., 115, A11323, doi:10.1029/2010JA015605.
Bruinsma, S., and F. G. Lemoine (2002), A preliminary semiempirical thermosphere model of Mars: DTM-Mars, J. Geophys. Res., 107(E10),
5085, doi:10.1029/2001JE001508.
Chen, Y., L. Liu, and W. Wan (2011), Does the F10.7 index correctly describe solar EUV flux during the deep solar minimum of 2007–2009?
J. Geophys. Res., 116, A04304, doi:10.1029/2010JA016301.
Dubinin, E., M. Fraenz, J. Woch, F. Duru, D. Gurnett, R. Modolo, S. Barabash, and R. Lundin (2009), Ionospheric storms on Mars: Impact of
the corotating interaction region, Geophys. Res. Lett., 36, L01105, doi:10.1029/2008GL036559.
Duru, F., D. A. Gurnett, D. D. Morgan, R. Modolo, A. F. Nagy, and D. Najib (2008), Electron densities in the upper ionosphere of Mars from
the excitation of electron plasma oscillations, J. Geophys. Res., 113, A07302, doi:10.1029/2008JA013073.
Edberg, N. J. T., H. Nilsson, A. O. Williams, M. Lester, S. E. Milan, S. W. H. Cowley, M. Fränz, S. Barabash, and Y. Futaana (2010), Pumping out
the atmosphere of Mars through solar wind pressure pulses, Geophys. Res. Lett., 37, L03107, doi:10.1029/2009GL041814.
England, S. L., and R. J. Lillis (2012), On the nature of the variability of the Martian thermospheric mass density: Results from electron
reflectometry with Mars Global Surveyor, J. Geophys. Res., 117, E02008, doi:10.1029/2011JE003998.
Forbes, J. M., and S. Miyahara (2006), Solar semidiurnal tide in the dusty atmosphere of Mars, J. Atmos. Sci., 63, 1798–1817.
Forbes, J. M., S. Bruinsma, and F.G. Lemoine (2006), Solar rotation effects in the thermospheres of Mars and Earth, Science, 312,
1366–1368.
Forbes, J.M., F. G. Lemoine, S. L. Bruinsma, M. D. Smith, and X. Zhang (2008), Solar flux variability of Mars’ exosphere densities and
temperatures, Geophys. Res. Lett., 35, L01201, doi:10.1029/2007GL031904..
Gonzalez-Galindo, F., F. Forget, M. A. Lopez-Valverde, M. Angelats i Coll, and E. Millour (2009), A ground-to exosphere Martian general circulation model: 1. Seasonal, diurnal, and solar cycle variation of thermospheric temperatures, J. Geophys. Res., 114, E04001,
doi:10.1029/2008JE003246.
Keating, G. M., and E. J. Prior (1968), The winter helium bulge, Space Res., 8, 982–992.
Konopliv, A. S., S. W. Asmar, W. M. Folkner, O. Karatekin, D. C. Nunes, S. E. Smrekar, C. F. Yoder, and M. T. Zuber (2011), Mars
high resolution gravity fields from MRO, Mars seasonal gravity, and other dynamical parameters, Icarus, 211, 401–428,
doi:10.1016/j.icarus.2010.10.004.
Krasnopolsky, V. A. (2010), Solar activity variations of thermosphere temperatures on Mars and a problem of CO in the lower atmosphere,
Icarus, 207, 638–647.
Liu, X., J. P. Thayer, A. Burns, W. Wang, and E. Sutton (2013), Altitude variations in the thermosphere mass density response to
geomagnetic activity during the recent solar minimum, J. Geophys. Res. Space Physics, in press.
Lundin, R., S. Barabash, A. Fedorov, M. Holmstroöm, H. Nilsson, J.-A. Sauvaud, and M. Yamauchi (2008a), Solar forcing and planetary ion
escape from Mars, Geophys. Res. Lett., 35, L09203, doi:10.1029/2007GL032884.
Lundin, R., S. Barabash, M. Holmstroöm, H. Nilsson, M. Yamauchi, M. Fraenz, and E. M. Dubinin (2008b), A comet-like escape of
ionospheric plasma from Mars, Geophys. Res. Lett., 35, L18203, doi:10.1029/2008GL034811.
Medvedev, A. S., and E. Yiğit (2012), Thermal effects of internal gravity waves in the Martian upper atmosphere, Geophys. Res. Lett., 39,
L05201, doi:10.1029/2012GL050852.
Smith, M. D. (2004), Interannual variability in TES atmospheric observations of Mars during 1999–2003, Icarus, 167, 148–165.
Smith, M. D. (2009), THEMIS observations of Mars aerosol optical depth from 2002–2008, Icarus, 202, 444–452.
Solomon, S. C., L. Qian, L. V. Didkovsky, R. A. Viereck, and T. N. Woods (2011), Causes of low thermospheric density during the 2007–2009
solar minimum, J. Geophys. Res., 116, A00H07, doi:10.1029/2011JA016508.
Thayer, J. P., X. Liu, J. Lei, M. Pilinski, and A. G. Burns (2012), The impact of helium on thermosphere mass density response to
geomagnetic activity during the recent solar minimum, J. Geophys. Res., 117, A07315, doi:10.1029/2012JA017832.
Venables, W. N., and B. D. Ripley (2002), Modern Applied Statistics with S. 4th ed., Springer, New York.
©2014. American Geophysical Union. All Rights Reserved.
9