Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, PLANETARY SCIENCE ( (c) Oxford University Press USA, 2020. All Rights Reserved. Personal use only; commercial use is strictly prohibited (for details see Privacy Policy and Legal Notice).

date: 31 October 2020

Solar Wind and Terrestrial Planetsfree

  • Edik Dubinin, Edik DubininMax Planck Institute for Solar System Research, Germany
  • Janet G. LuhmannJanet G. LuhmannSpace Sciences Laboratory, University of California, Berkeley
  •  and James A. SlavinJames A. SlavinDepartment of Climate and Space Sciences and Engineering, University of Michigan


Knowledge about the solar wind interactions of Venus, Mars, and Mercury is rapidly expanding. While the Earth is also a terrestrial planet, it has been studied much more extensively and in far greater detail than its companions. As a result we direct the reader to specific references on that subject for obtaining an accurate comparative picture. Due to the strength of the Earth’s intrinsic dipole field, a relatively large volume is carved out in interplanetary space around the planet and its atmosphere. This “magnetosphere” is regarded as a shield from external effects, but in actuality much energy and momentum are channeled into it, especially at high latitudes, where the frequent interconnection between the Earth’s magnetic field and the interplanetary field allows some access by solar wind particles and electric fields to the upper atmosphere and ionosphere. Moreover, reconnection between oppositely directed magnetic fields occurs in Earth’s extended magnetotail—producing a host of other phenomena including injection of a ring current of energized internal plasma from the magnetotail into the inner magnetosphere—creating magnetic storms and enhancements in auroral activity and related ionospheric outflows. There are also permanent, though variable, trapped radiation belts that strengthen and decay with the rest of magnetospheric activity—depositing additional energy into the upper atmosphere over a wider latitude range. Virtually every aspect of the Earth’s solar wind interaction, highly tied to its strong intrinsic dipole field, has its own dedicated textbook chapters and review papers.

Although Mercury, Venus, Earth, and Mars belong to the same class of rocky terrestrial planets, their interaction with solar wind is very different. Earth and Mercury have the intrinsic, mainly dipole magnetic field, which protects them from direct exposure by solar wind. In contrast, Venus and Mars have no such shield and solar wind directly impacts their atmospheres/ionospheres. In the first case, intrinsic magnetospheric cavities with a long tail are found. In the second case, magnetospheres are also formed but are generated by the electric currents induced in the conductive ionospheres. The interaction of solar wind with terrestrial planets also varies due to changes caused by different distances to the Sun and large variations in solar irradiance and solar wind parameters. Other important planetary differences like local strong crustal magnetization on Mars and almost total absence of the ionosphere on Mercury create new essential features to the interaction pattern. Solar wind might be also a feasible driver for planetary atmospheric losses of volatiles, which could historically affect the habitability of the terrestrial planets.

Solar Wind

The terrestrial planets Mercury, Venus, Earth, and Mars are continuously exposed to the flow of mostly hydrogen plasma (solar wind) emitted from the Sun. The solar wind originates in the solar corona and flows outward in all directions, carrying with it the solar magnetic field. Characteristics of solar wind depend on what regions of the solar corona it comes from and on the magnetic activity on the surface of the Sun. The activity of the Sun varies with the solar cycle during which the polarity of the global solar magnetic field reverses. One solar activity cycle lasts about 11 years, although the full magnetic cycle is twice as long. The number of sunspots on the Sun, related to the emergence of magnetic fields from the interior, characterizes solar activity. At solar minimum, when few sunspots are observed, the solar surface field is nearly dipolar, and the solar wind emanating from the high-latitude regions is fast (solar wind velocity Vsw ~ 800 km/s) and tenuous, while at low solar latitudes it is slower (~400 km/s), denser, and much more variable. At these times a clearly defined, near-equatorial heliospheric current sheet separating outward and inward interplanetary fields in the wind from the two hemispheres crosses the ecliptic regularly at a roughly 27-day solar rotation rate. Since the terrestrial planets orbit around the Sun in an almost ecliptic plane which is inclined only 7 degrees from the plane of the Sun’s equator, they are mainly embedded in the slow solar wind. In contrast, near solar maximum the picture becomes much more complex. The size of polar coronal holes decreases, and they may even disappear. Small coronal holes related to sunspots modify the solar wind and current sheet, generating adjacent plasma flows with high and low speeds, which interact. When fast solar wind overtakes plasma from the slow wind, they can produce co-rotating interaction regions (CIRs) characterized by a compression region with a pair of shock waves. Another solar wind phenomenon that strongly affects the planets’ plasma interactions is coronal mass ejections (CMEs) caused by huge eruptions of plasma and magnetic fields from the solar corona. CMEs can travel outward from the Sun at speeds of up to 3,000 km/s, producing strong disturbances in the planetary environments including orders of magnitude solar wind pressure and magnetic field increases. Recent observations made during the last unusually quiet solar minimum of cycle 24 have shown that even at solar minimum, coronal holes outside the polar regions of Sun and multiple streamers can significantly distort slow solar wind.

The solar magnetic field at the base of the solar wind outflows that is effectively “frozen” to the solar wind is called the interplanetary magnetic field (IMF). This field ties the charged particles of plasma, making its behavior similar to a fluid in spite of its collisionless nature. Since the IMF is attached to the Sun at the point where the solar wind originates and the Sun rotates, the magnetic field lines, like the solar wind streams, form a spiral shape known as the Parker spiral. The average direction of the magnetic field at the locations of the different planets is determined by the solar wind speed Vsw, the Sun’s rotation speed, and the distance to Sun. This angle is called the angle of a Parker spiral. Solar wind parameters also vary with distance to Sun. The plasma density (n) and average interplanetary magnetic field (B) radial component vary approximately as 1/r2, while the perpendicular component is reduced by 1/r.

Table 1. The distance to Sun in astronomical units (AU), the solar wind speed (Vsw), the solar wind density (n), the interplanetary magnetic field carried by solar wind (B), the proton temperature (Tp), the electron temperature (Te)


Distance, AU

Vsw, km/s

n, cm−3

B, nT

Tp, 104oK

Te, 104oK





























Table 1 contains a set of “typical” long-term mean parameters measured near Earth, scaled to the orbits of the other terrestrial planets (Slavin & Holzer, 1981). Although this table gives only average values, it is seen that Mercury and Venus are bombarded by much denser solar wind than Mars.

The interaction of the solar wind with terrestrial planets depends not only on solar wind parameters but also on their distance from the Sun and the characteristics of the planets themselves. The solar ultraviolet radiation (EUV), which falls off with heliocentric distance as 1/r2, heats and ionizes neutrals in the planetary upper atmospheres, creating shells of ionized gas (ionospheres). If the planet has a global intrinsic magnetic field like Earth and Mercury, this field can protect it from a direct exposure to the solar wind, forming an effective magnetic cavity (magnetosphere) around the planet. If the planet has no such protection (like Venus and Mars), solar wind interacts directly with its ionosphere/atmosphere. With variations of the solar EUV flux with both heliocentric distance and the solar activity cycle, the ionosphere becomes weaker or stronger, significantly changing the physics of the solar wind/ionosphere interaction.

It has been known since the 1960s that the interaction of solar wind with the magnetized planets is very sensitive to the direction of the IMF. Reconnection between solar wind and planetary magnetic fields can dump energy into the upper atmosphere through the energization and precipitation of electrons and ions and excite aurora. Planets with intrinsic dipole fields are subject to a process known as the Dungey cycle, in which the dipole field lines reconnect with oppositing external interplanetary fields at the dayside, convection to the nightside, and are returned to the dayside following closing of stretched dipole fields on the nightside and sunward convection. Reconnection processes are very important for Earth and also Mercury, which makes their magnetospheres a unique laboratory to study this fundamental phenomenon.

The Sun and solar wind are rather efficient in removing planetary atmospheres. For example, because of low gravity of Mercury, its atmosphere was probably blown away by intense heating close to the sun and scavenging by dense solar wind. Presently, Mercury has only a tenuous atmosphere/exosphere replenished by desorption of surface elements (Na, K). It is also believed that the initial atmosphere of Mars was much denser in past epochs and could have produced conditions that allowed the existence of oceans. A large fraction of this early atmosphere was lost by thermal and non-thermal escape processes. Some of the non-thermal processes are driven by solar wind. Recent observations also suggest that Mars once had an intrinsic magnetic field (see, e.g., Weiss et al., 2008) but lost it several billion years ago, leaving the planetary atmosphere/ionosphere open to direct exposure by solar wind. Although the global magnetic field for Mars is absent, there are large areas with strong remnant magnetization (up to 1,500 nT at 100 km) (Acuña et al., 1999), which significantly influences the global pattern of the solar wind/Mars interaction. Table 2 contains shows the present planetary conditions normalized to Earth values.

Table 2. The planetary radius, the orbital period, the gravity, the escape velocity, the atmospheric pressure on the planetary surface and the main atmospheric species. The values for Mercury, Venus and Mars are normalized to the corresponding values at the Earth



Orbital Period



Escape Velocity

Magnetic Field at the Surface

Atmospheric Pressure at the Surface

Atmospheric Components


6371 km





3000 nT

101 kPa

N2, O2, Ar








<10−14 PE

O2, Na, H, He, K







91 PE

CO2, N, SO2




5.59 EE



Crustal field

0.0063 PE

CO2, O, N2, Ar,

As the supersonic solar wind flows past planets, a bow shock is formed similar to what happens in front of a supersonic jet traveling in the Earth’s atmosphere. The strength of the bow shock is determined by the value of the Mach number, that is, the ratio of the solar wind speed to the “sound” speed in the magnetized solar wind. At the orbits of Venus, Earth, and Mars the solar wind is supersonic and the characteristics of their bow shocks are rather similar. This is not the case for Mercury, where the Mach number becomes rather small, implying very different shock and post-shock plasma conditions. Regardless of the type of obstacle a planet presents to the solar wind, there is a region of shocked solar wind plasma inside the bow shock that is slowed, heated, and deflected around the obstacle. This region, generally referred to as the magnetosheath, is where the close-in solar wind interaction with the planet is defined.

Because of the diverse solar wind and EUV conditions for the orbits of Mercury, Venus, and Mars and their very different planetary characteristics, the interactions of solar wind with these planets have distinctive features. Figure 1 presents a brief timeline of space missions to Mercury, Venus, and Mars on which measurements of solar wind interaction and its effects were studied. It is seen that the missions occurred during very different conditions of the Sun and thus in the solar wind.

Figure 1. Sunspot number, representing the activity of the Sun and dates of the measurements in the space environments of Mercury, Venus, and Mars during the main space missions to these planets.

Solar Wind and Mercury

Mercury is the smallest of the planets, with a mean radius of only 2,440 km (see Table 2). It orbits the Sun in a highly eccentric orbit with perihelion and aphelion of only 0.31 and 0.47 AU, respectively, and an inclination to the ecliptic of ~7 deg. A 2 to 3 resonance exists between Mercury’s orbital and rotational periods, resulting in its “year” and “day” having lengths of 88 and 58.8 days. The tilt of the rotation axis to its orbital axis is quite small, «1 deg. The interior of Mercury is dominated by a ~2,000 km radius, largely iron core, which contributes to it having the highest uncompressed density among the planets, ~ 5.3 versus ~4.0 gm/cm3 for the Earth. The first reconnaissance mission to Mercury was Mariner 10. It carried out three relatively close flybys in 1974 and 1975. Among the many discoveries made by this mission are Mercury’s weak, but highly dipolar, magnetic field and a solar wind interaction that creates a small magnetosphere with many structural similarities to that of Earth. On March 18, 2011, NASA’s MErcury Surface, Space ENvironment, GEochemistry, and Ranging (MESSENGER) spacecraft was inserted into a ~12 h period, 82° inclination, high-eccentricity (~200 × 15,000 km altitude) orbit about Mercury. It continued to orbit Mercury until its fuel was used up and the spacecraft impacted Mercury on April 30, 2015. MESSENGER explored and mapped Mercury’s surface, interior, and magnetosphere.

Figure 2 provides an overview of the structure of the Mercury magnetosphere in relation to the MESSENGER orbit. The mean distance from the center of Mercury to the nose of the magnetopause is only ~1.5 RM, and the downstream diameter of the magnetotail is ~4 RM. While the structure of Mercury’s magnetosphere is very Earth-like, its magnetic dipole is offset to the north by 0.2 RM. Mercury has a very tenuous exosphere with complex composition and poorly understood sources on the planetary surface. Mercury’s magnetospheric plasma is composed of solar wind protons, but heavier ions (i.e., Na+ group and O+ group) created by the photo-ionization of Mercury’s exosphere are present, especially in the cusps and plasma sheet. In fact, the number density of these heavier planetary ions can reach 10% that of the protons.

Figure 2. MESSENGER orbits with periapsis at local noon (dashed green ellipse) and midnight (solid green ellipse) are displayed against a schematic cross section of Mercury’s magnetosphere, with the Sun at left (adapted from Zurbuchen et al., 2011).

The dayside magnetosphere is made up of closed magnetic fields and low beta (i.e., ratio of thermal to magnetic pressure) plasma. At higher latitudes near noon there is a high beta transition region called the magnetospheric cusps. The magnetic field lines in the cusps are in the act of being transferred from the dayside to the nightside magnetosphere. The cusp flux tubes are laden with a mixture of solar wind and magnetospheric plasma. The cusp magnetic flux is carried tailward by the momentum of the solar wind plasma in the magnetosheath and incorporated into the outer layers of the tail lobes. In doing so the solar wind and planetary plasma carried by these flux tubes is added to the high-latitude tail lobes to form the plasma mantle. Just as for Earth, Region 1 field-aligned currents flow along the boundary between the low-latitude closed magnetic field portion of Mercury’s magnetosphere and the high-latitude polar caps encompassing the open magnetic fields in the tail lobes that have one end connected to the solar wind (see Figure 2). These field-aligned currents flow downward on the dawn side of the polar cap and upward on the dusk side have been shown by MESSENGER to close radially through Mercury’s thin crust and across the surface of its highly conducting iron core.

The most important effect of the low Alfven Mach number in the solar wind at Mercury’s orbit is the formation of a thick, very low beta plasma depletion layer behind the bow shock in the inner magnetosheath. Under these conditions the magnetic field intensity varies little or not at all across the magnetopause. When reconnection occurs under these conditions, it is termed “symmetric.” This type of magnetopause reconnection is observed very infrequently for Earth and even less so for the magnetospheres of the outer planets. For symmetric conditions, reconnection can occur whenever the magnetic shear angle across the magnetopause is non-zero. Further, for a given magnetic shear angle across the magnetopause, symmetric reconnection is always faster than when the magnetic field magnitude changes across the magnetopause (i.e., asymmetric reconnection).

MESSENGER also observed Mercury’s response to solar and coronal activity. Two views of Mercury’s dayside magnetosphere during the impact of an interplanetary coronal mass ejection on November 23, 2011 are shown in Figure 3. Figure 3a illustrates some features of the interaction deduced from the analysis of the MESSENGER measurements. During this event reconnection between the interplanetary magnetic field and the dayside magnetic field was determined to occur with a dimensionless reconnection rate of 0.16. In contrast, the asymmetric reconnection rate at Earth is typically several times slower than the rates determined by MESSENGER. Without the electric currents induced on the surface of Mercury’s iron core, the combination of reconnection-driven magnetic flux transfer and the high solar wind dynamic pressure would have eroded and compressed the dayside magnetosphere to the point where solar wind surface impact would have occurred. However, the ICME-driven induction currents in the highly conducting core shield the interior by adding closed magnetic flux (shown in green in Figure 3a) to the flux generated by dynamo action deep in the core (shown in yellow). Hence, the effects of reconnection and induction on November 23, 2011, appear to have been in balance with the nose of the magnetopause ~1.1 RM from the center of the planet.

Figure 3. (a) Increases in solar wind pressure (solar wind dynamic pressure was of 51 nPa) drive induction currents (green loops) on the surface of Mercury’s iron core that generate additional magnetic flux (green field lines). They add to the intrinsic magnetic flux (yellow field lines). Magnetic reconnection at an X-line at magnetopause reduces the effectiveness of the shielding of the surface from the solar wind by transferring flux into the tail (adapted from Slavin et al., 2014). (b) A close-up view of the simulated magnetospheric magnetic fields for November 23, 2011, ICME impact event simulated using a global MHD numerical code. The black lines with arrows show projections of the magnetic field lines onto the Mercury Solar Orbital (MSO) X–Z plane. In the MSO system the X and Y axes lie in Mercury’s equatorial plane, with X directed toward the Sun, Y opposite to planetary motion, and Z completes the right-handed system. Color contours for the electric current density in the y-direction (Jy in nA/m2) and projected onto a 3-D sphere of radius ∼0.8 RM corresponding to the mantle-core boundary. The magenta lines with arrows show the induction currents generated on the core. The green circle of radius 1 RM represents the planetary surface (adapted from Jia et al., 2018).

The results of a global three-dimensional numerical simulation of Mercury’s magnetosphere during the November 23, 2011, ICME impact are displayed in Figure 3b. The simulation also yields a subsolar magnetopause standoff distance of ∼1.1 RM and a dimensionless reconnection rate of ~0.12, both of which are in excellent agreement with the results of the analysis of the MESSENGER data. The cyan and yellow colors at the core boundary in Figure 3b indicating that strong currents are being induced on the surface of Mercury’s iron core in response to compression of the dayside magnetosphere and are remarkably similar to the data analysis–driven illustration in Figure 3a. As expected, the sense of the induced currents in the global MHD simulation is in the direction that shields the conducting core from the compression of Mercury’s dipole magnetic field by the ICME. The results of the global simulation suggest that electric currents induced on the surface of the core will affect not only dayside reconnection and compression of the magnetic field but also internally driven reconfigurations of the nightside magnetospheric magnetic field. Analysis of a larger data set of ICME impacts has determined that the induction currents driven during these events temporarily increase the magnetic moment of Mercury by up to 25%.

Figure 4. MESSENGER magnetic field data and projections of the MESSENGER trajectory are presented in MSM coordinates with red shading marking the location of the spacecraft during the 40 min interval. Three loading—unloading events are indicated by green shading [from Imber & Slavin, 2017]*. The Mercury Solar Magnetospheric (MSM) system defines the X and Y axes to lie in Mercury’s magnetic equatorial plane with X directed toward the Sun, Y opposite to planetary motion and Z completes the right-handed system.

During MESSENGER’s third flyby of Mercury, a series of several-minute-long increases in the intensity of the tail lobe magnetic field followed by decreases of similar duration were observed. During the periods of increasing magnetic field in the tail, the flaring of the magnetic field away from the central axis of the tail was also observed to increase. Conversely, when the tail magnetic field magnitude decreased, the flaring was seen to decrease. These loading—unloading events are key signatures of the Dungey cycle of magnetic flux transfer that powers magnetospheric substorms of the Earth. Once the loading of the Earth’s magnetic tail reaches sufficiently high levels, reconnection commences in the cross-tail current sheet, and the stored magnetic energy is converted into the plasma heating and bulk flows and the magnetic field reconfiguration that constitute magnetospheric substorms.

A total of 438 several-minute-long loading/unloading events were identified in the MESSENGER data over its 4-year mission. Three examples of these events during a 40-min interval of lobe magnetic field measurements are shown in Figure 4. The median duration of the loading/unloading events and the median relative amplitude relative to the background lobe field strength for all 438 events were 195 s and 24%, respectively. The change in the lobe magnetic flux content ranged from 0.2 to 2 MWb, with a median value of 0.59 MWb. Overall, these loading/unloading cycles are 50 to 100 times shorter in duration and two to three times greater in amplitude than is observed during isolated substorm events observed for Earth. On this basis, it was concluded that Mercury exhibits very intense, Earth-like substorm activity, but with a Dungey cycle that is only several minutes long as opposed to its ~1–2 hr for Earth.

The measurements taken by MESSENGER in Mercury’s plasma sheet, like those collected at the forward magnetopause, provide strong evidence for the formation of reconnection X-lines. At these X-lines reconnection takes place between the opposing magnetic flux in the lobes. This newly closed magnetic flux is carried toward Mercury by the fast flow in the sunward exhaust jet from the X-lines, while magnetic flux that is disconnected from the planet are carried tailward to eventually rejoin the solar wind and the IMF. Multiple reconnection X-lines in the cross-tail current sheet naturally lead to the formation of magnetic flux ropes called plasmoids. Just as for Earth, MESSENGER observed plasmoids being carried sunward and anti-sunward by the fast flows in the plasma sheet. Fast sunward flows in the terrestrial plasma sheet are observed to steepen and form dipolarization fronts as they propagate toward the Earth. These fronts are the sites for significant charged particle heating and acceleration. In close association with dipolarization fronts for Mercury, MESSENGER observes not only heating of the thermal ions in the plasma sheet but also the energization of some electrons to energies over 100 keV.

The similarities in the dynamics of the nightside magnetospheres of Mercury and Earth appear to extend even to the observation of analogues to terrestrial auroras for Mercury. MESSENGER observed X-ray emissions from Mercury’s surface in the vicinity of the nightside polar cap boundary. The most likely sources of these nightside emissions have been determined to be Bremsstrahlung emissions generated by the braking of energetic electrons accelerated in the near-tail as they stream down the magnetic field to impact Mercury’s surface. However, it will not be until the dual-orbiter BepiColombo mission arrives in 2025 that simultaneous measurements of the upstream solar wind and the magnetosphere will be available and close comparisons can be made with Earth regarding the response of Mercury’s magnetosphere to conditions in the solar wind.

Solar Wind and Venus

The interaction of Venus with the solar wind was well explored during the ~14-year mission of the Pioneer Venus Orbiter (PVO) and the ~8-year mission of Venus Express. The combination of these missions with comprehensive instrumentation, different orbital sampling, durations of thousands of orbits each, and range of solar activity levels (see Figure 1) provided transformational information on the overall morphology and physics of the Venus space environment. In contrast to Mercury, with its dipole field and tenuous atmosphere, Venus is primarily an ionospheric obstacle to the solar wind flow, as illustrated in Figure 5. In this regard, it is unique among the planets of the solar system. Moreover, because the ionosphere is produced by the solar extreme ultraviolet (EUV) flux and conditions on the Sun and in the solar wind are more variable when the Sun is active, the solar wind interaction and its consequences change over the ~11-year solar cycle.

Flybys of Venus by the early Mariner and Venera spacecraft found that Venus’s magnetic field must be very weak. In fact, the solar wind interaction features provided the clues because they occurred so close to the planet relative to Earth’s. The lowest altitude magnetic field measurements for Venus were made on the PVO orbiter around its ~150 km altitude periapsis. These set an upper limit on a global dipole moment of Venus of ~1/100,000 that of the Earth. The field of this weak dipole is not able to stand off the solar wind by itself, like Earth’s or Mercury’s dipole. Instead, the magnetic fields of currents induced in the Venus’s conducting ionosphere by the passing magnetized solar wind produce an effective “magnetic barrier” to the direct penetration of the solar wind (see Figure 5).

Figure 5 illustrates the major features of the solar wind interaction with Venus’s ionosphere deduced from the solar maximum measurements on PVO. The most basic features are the bow shock that defines the outer boundary of the interaction followed by the large-scale draping of the interplanetary magnetic field over the ionosphere. The related comet-like “induced magnetotail” is made up of the innermost draped fields that have been slowed down where they pass through the atmosphere. These likely originate in the dayside magnetic barrier, whose field piles up until its magnetic pressure is equal to the incident pressure of the solar wind. Inside the magnetic barrier there is an altitude where ionospheric thermal plasma pressure becomes equal to, and then exceeds, the magnetic pressure, defining the ionopause. The dayside ionopause is a sharp boundary when this pressure balance occurs above ~200 km, where collisions between particles are rare. On occasions when the incident solar wind pressure is high compared to the ionosphere pressure, as routinely happened during relatively low solar activity (and hence low solar ionizing flux) or when a strong solar wind stream or disturbance passed by, the magnetic barrier field was observed to penetrate to the lowest point sampled along the orbit. This condition represents a change in the ionospheric current system that makes the magnetic barrier and a broadening of the solar wind/ionosphere interface in the presence of electron collisions with neutrals. Note that the upper (neutral gas) atmosphere of Venus extends out beyond the ionopause into the magnetosheath and solar wind (Figure 5). When Venus atmospheric ions, or ionospheric ions, are produced on these fields by any of the ionization mechanisms—photoionization, electron impact ionization, or charge exchange—they “mass load” the passing solar wind plasma. This mass loading can take several forms, depending on the location within the solar wind interaction, as illustrated in Figure 6. In the ionosphere below the ionopause it is more “fluid-like” (left panel). At higher altitudes, any ions that are produced behave more like isolated particles, gyrating around and moving downstream with the draped magnetic field, e.g., they are “picked up” (right panel). Oxygen and hydrogen ions are particularly important participants in these processes, which enable atmospheric loss to space.

Figure 5. Schematic diagram of the solar wind interaction with Venus (adapted from Russell et al., 2007).

Figure 6. Fluid (a) and kinetic (b) processes play a role in the solar wind energization of the Venus atmosphere. (a) Fluid processes include acceleration by a parallel electric field caused by pressure gradients and the j × B force. (b) Pickup ions on the other hand can be modeled as single test particles in the particle modeling perspective (from Futaana et al., 2017).

Signatures of the Venus obstacle are also detected upstream of the bow shock. Some of the solar wind protons and electrons, and even planetary ions incident on the nose of the bow shock, are accelerated back into the oncoming solar wind along interplanetary magnetic field lines that connect to the shock. This produces a phenomenon called the foreshock, a feature observed as upstream magnetic field and plasma fluctuations and plasma waves.

The ionosphere of Venus both affects and is affected by the direct solar wind interaction. Due to the photochemistry in the mostly CO2 atmosphere of Venus, the ionospheric ion composition is mainly O2+ around its density peak at ~130 km and O+ at altitudes above about 190 km. Around solar maximum, the ionopause is characterized by a well-defined density gradient on the dayside, where pressure balance with the solar wind is established. Figure 7 shows what happens when the solar wind pressure increases and pushes the pressure balance altitude into the collisional region of the upper atmosphere below ~200 km. Altitude profiles of magnetic field magnitude and ionospheric density from the subsolar region like those in Figure 7 provide in situ evidence for magnetic barrier field penetration into the dayside ionosphere and possibly to the surface. There are also consequences for the other properties of the ionosphere, including electron and ion temperatures illustrated here.

Solar wind dynamic pressures for Venus vary according to a distribution dictated by both solar conditions and the distance of Venus from the Sun. Even at solar maximum, the number of times the solar wind pressure exceeds the peak thermal pressure of the ionosphere, a measure of its robustness as an obstacle, is small relative to the situation inferred for Mars. It also appears that under high solar wind pressure conditions, the dayside ionosphere is not simply compressed but rather the upper portion of the ionosphere is missing. It is assumed, and supported by other observations, that the upper ionosphere is removed by solar wind interaction–related erosion processes (e.g., Figures 5 and 6).

Figure 7. Altitude profiles of the electron number density, electron and temperatures and the magnetic field value along the PVO orbits illustrating the response of the ionosphere on increasing solar wind dynamic pressures. Different symbols correspond to different orbits.

(adapted from Luhmann et al., 1987).

Venus rotates so slowly that the same side of the planet and the atmosphere above it is exposed to the Sun for a long time relative to ionization and recombination processes in the ionosphere. As a result, the bulk of the ionospheric flow is antisunward at locations removed from the subsolar point, driven by day-to-night thermal pressure gradients. This both reduces the height of the peak from that of a Chapman layer and provides a robust solar maximum source for a nightside ionosphere. Measurements indicate the maximum measured O+ ion velocities are ~5.5 km/s, whereas ~11 km/s is the escape velocity from Venus. At the top of the ionosphere the other forces related to the solar wind interaction and the draped, penetrating interplanetary field (see Figure 6) can also contribute to antisolar flow, but these forces are less symmetric with solar zenith angle.

The result of the dayside ionosphere production and slow planetary rotation is a nightside ionosphere whose appearance depends on trans-terminator flow. The amount of this flow is controlled by the height of the terminator ionopause, which means the nightside ionosphere density depends on solar wind pressure. In its most robust solar maximum state, when solar wind pressure is normal, it exhibits a well-defined altitude profile except near the antisolar region where one or two deep, vertical bite-outs or “holes” may be seen. These holes contain unusually strong and steady nearly antisunward or sunward magnetic fields that appear to be a part of the draped fields that make up the induced magnetotail. The depth to which holes penetrate is also unknown since they are observed down to the ~150 km periapsis altitude of PVO. It is also not clear whether holes have implications for ionospheric escape.

The other observed state of the nightside ionosphere is related to the high solar wind pressure conditions that lead to low dayside ionopause and magnetized dayside ionosphere. This state, a low-density condition with a relatively strong horizontal nightside magnetic field, has been called a “disappearing” ionosphere. It has been interpreted as the result of the low ionopause shutting off the cross-terminator flow. Disappearing nightside ionosphere conditions are expected to be common at solar minimum when the ionopause should be more generally low.

A faint, patchy, nightside ultraviolet aurora was detected by the PVO ultraviolet spectrometer in a spectral line characteristic of oxygen (130.4 nm). This emission appears to intensify during solar wind disturbances from coronal mass ejections (CMEs). The auroral emissions have been attributed to enhanced electron precipitation into the nightside ionosphere. More recently, ground-based observations of auroral green line (557.7 nm) emissions from Venus’s nightside have been correlated with observations of solar activity, including flares and coronal mass ejections. Mars has diffuse auroral emissions of probably similar nature.

Photochemistry in the primarily O2+ Venus ionosphere leads to production of an extended upper atmosphere or “corona” of hot O atoms that coexists with the hydrogen exosphere. Since the upper atmosphere is always exposed to the solar wind plasma and/or magnetic field, any of the ionization mechanisms that operate there, including photoionization, charge exchange with solar wind protons, or electron impact by solar wind electrons, produce ions that can escape via one of the processes illustrated in Figure 6. Even those ions that instead impact the atmosphere after gaining energy can produce atmosphere losses via the secondary process of sputtering (in which collisional transfer of the ion’s energy to neutral particles ejects a small fraction of whatever is present at the exobase where the atmosphere becomes collisional). The signature large cycloidal trajectories in Figure 6 are most representative of high-altitude pickup of ionized oxygen corona atoms. Smaller cycloids and helices are expected for smaller mass hydrogen pickup and for O+ picked up in the slowed background plasma flows and stronger, draped magnetic fields in the Venus magnetic barrier region. The more fluid-like processes associated with the magnetic tension forces of the draped magnetic fields (j × B) and the magnetic field-aligned polarization electric fields dominate losses from the deeper ionosphere. Figure 8 shows where picked-up O+ ions leaving Venus were seen in the form of downstream cross sections created from many PVO (top) and Venus Express (bottom) orbits. Many questions remain about the O+ and other ion escape rates at Venus, in part because of the limited sampling of the responses of ion escape to the more extreme solar and solar wind conditions.

An ongoing debate concerns the importance of this solar wind-enabled escape for atmosphere evolution at Venus, as for Mars (discussed later). At larger Venus, the gravitational field strength is such that solar wind erosion is the primary way that oxygen can be lost to space. Venus is notoriously dry compared to Earth, and while it is easy to envision the hydrogen from photodissociated water vapor in its upper atmosphere escaping, the oxygen should remain gravitationally bound. Over time, this oxygen would have had to be removed from the atmosphere by some process such as oxidation of the surface. But if Venus originally had an Earth-like inventory of water, it would be difficult for the crust to have accommodated the amount of O by oxidation. In addition, although the average inferred O+ escape rates from observations fall far short of those needed to remove an ocean’s worth of oxygen, periods of disturbed solar wind from coronal eruptions and stream interactions are found to enhance them. Astronomical observations of Sun-like stars indicate the early Sun was generally more active than the present Sun. Thus escape to space is a uniquely viable candidate for oxygen (and, by inference, water) escape from Venus meriting further examination. This historical aspect has motivated broader attention in light of the growing interest in the habitability of terrestrial exoplanets.

Figure 8. (a) Locations of the measured by the PVO spacecraft O+ ions in the YZ-VSE plane sorted by the energy of ions. Ions with energy < 4 keV are shown by open circles; plus symbols—ions with E=4–6 keV; asterisks –ions with E > 6 keV (Luhmann et al., 2006). (b) Detection sites of O+ ions measured by PVO in the XZ-VSO coordinates (Luhmann et al., 2006). (c) Flux of O+ ions in the YZ-VSE plane in the Venus tail measured on VEX. Plasma sheet (PS) and boundary layer (BL) are clearly displayed (Barabash et al., 2007). (d) Fluxes of ions with m/q > 14 in the cylindrical coordinates measured by VEX (Fedorov et al., 2008) White dashed curves depict for comparison the size of Mars and the position of the magnetospheric boundary at Mars, respectively.

Solar Wind and Mars

The exploration of Mars began with the flyby by Mariner 4 and with the Soviet missions Mars 2, 3, and 5, which found a bow shock and magnetotail filled with planetary ions. The next step for study the interaction of solar wind with Mars was taken by the Phobos-2 mission which established that planetary ions are continuously escaping to space. The inferred losses of volatiles of ~1 kg/s indicated a potentially important role of solar wind erosion in the evolution of the atmosphere of Mars. Observations of magnetic fields in the tail region were consistent with Venus-like draping of the IMF, although some observations indicated possible modifications by an intrinsic magnetosphere. The nature of the obstacle to the solar wind was finally resolved by the Mars Global Surveyor (MGS) mission as the previous spacecraft (except the Viking landers, which did not carry an onboard magnetometer) did not approach Mars closer than ∼850 km. The MGS measurements showed that at present Mars does not possess a global intrinsic magnetic field which could be an obstacle for the solar wind as for Earth or Mercury. Instead, MGS has detected localized, rather strong magnetic anomalies of crustal origin. Two space missions, the European Mars Express spacecraft and the American Mars Atmosphere and Volatile EvolutioN (MAVEN) spacecraft, continue to operate in the Mars orbit studying the different aspects of solar wind/Mars interaction.

The absence of a global magnetic field for Mars leads to the nearly direct interaction of the solar wind with its atmosphere and ionosphere and formation of the draped induced magnetotail. The magnetosphere of Mars contains a combination of features associated with comets, Venus, and planets with intrinsic magnetic fields (Figure 9). Similar to the cometary case, the interaction starts at large distances from the planet because of the extended Martian exosphere. Two major neutral populations form the Martian exosphere. The hydrogen corona takes its origin from the surface/subsurface water ice. Hydrogen molecules (H2) diffuse upward into the ionosphere, where they quickly break up into hydrogen atoms. The hydrogen atoms with energies greater than the required energy for their escape (∼0.13 eV) overcome Mars’s gravitational attraction and can be lost to space. The thermal velocity of the heavier atmospheric atoms and molecules is less than the escape velocity, so they are bound by gravity. However, as for Venus, the dissociative recombination of ionospheric O2+ and CO2+ ions (O2+ + e → O* + O* and CO2+ + e → CO* + O*) creates fast neutral oxygen atoms with kinetic energies that can exceed the critical energy (∼2 eV) needed to escape. The hydrogen exosphere and the hot oxygen corona extending out to many Martian radii were observed by remote optical and ultraviolet measurements of sunlight scattered from hydrogen and oxygen atoms. Again as for Venus, ions produced in these extended atmospheres by solar EUV flux, electron impact, and charge exchange are picked up by solar wind and move tailward on cycloidal trajectories (see Figure 6). Another interesting feature of the distant solar wind interaction with Mars is enhanced wave activity upstream of the bow shock. The newborn ions form a beam in solar wind, which generates electromagnetic waves whose frequencies coincide with the proton cyclotron frequency in the solar wind, making them easily identified. An energetic solar wind proton undergoing a charge exchange with exospheric hydrogen atom creates an energetic neutral hydrogen atom (ENA) that can travel large distances unaffected by electromagnetic fields. Such atoms easily cross the magnetosheath and crustal fields and penetrate deep into the Martian atmosphere, where they are partly converted back to energetic positively and negatively charged hydrogen ions. These precipitating energetic hydrogen atoms produce a brightening of the upper atmospheric hydrogen Lyman-α‎ emission (e.g., a proton aurora). The ratio of the average penetrating proton density observed for altitudes of 150 to 250 km to the upstream solar wind density varies from 0.01% to ~1%, implying that solar wind hydrogen deposition early in the solar system’s history when solar wind flux was much stronger might be an important source of atmospheric hydrogen. It is worth noting that a similar chain of processes—charge exchange, ENA production, and back-conversion to energetic positively and negatively charged ions—was observed by the Rosetta spacecraft at the Churyumov-Gerasimenko comet.

Figure 9. Sketch of the Martian magnetosphere containing features of a small comet, Venus and intrinsic magnetosphere.

A fraction of O+ ions originating from the hot oxygen corona can also precipitate back into the Mars upper atmosphere, where neutrals are ejected via a cascade of collisions in Mars’s upper atmosphere with enough energy to exceed the escape velocity at Mars. This coupling of the solar wind with the deep atmosphere by sputtering has been suggested to be a significant atmospheric loss process in Mars’s early history.

The Martian bow shock is located at ~1.5 RM (subsolar distance) and ~2.5 RM (terminator) from the center of Mars. In spite of such features as the presence of pickup ions and the scale size of the interaction region comparable to the ion Larmor radius, the Martian bow shock is rather typical of shocks with high Mach numbers, including its electron and ion foreshocks. Moreover, the dynamics of ions and electrons in the shock transition region indicates that mechanisms responsible for the energy dissipation seem to be similar to those operating at the Earth’s bow shock. However, some new features related to the presence of the exospheres also appear. For example, the impact of the energetic back-streaming electrons on the neutral hydrogen leads to the depletion of the electrons above the energy at which the impact ionization cross section has a maximum.

Closer to Mars the number density of oxygen ions grows. Because of the small size of Mars as compared to the oxygen cyclotron-radius, oxygen ions are deflected and accelerated toward the hemisphere in which the solar wind motional electric field (–Vsw × B) is pointing away from Mars (E+ hemisphere). This ion motion (see Figure 6) provides an excess of transverse momentum in the system, which must be balanced by the motion of the solar wind protons in the opposite direction. Forces responsible for such ion motion are Lorentz-type forces, which appear in multi-ion magneto-hydrodynamic (MHD) equations. These forces are caused by a differential streaming of ion fluids and act on the protons and oxygen ions in the opposite directions. A small lateral deviation of the solar wind velocity (∼5 km/s) in the direction opposite to the direction of the motional electric field indicating a weak mass-loading effect is observed even upstream of the bow shock. Closer to Mars, asymmetric mass-loading effects on features such as solar wind deflection become more pronounced, although there is still a generally symmetric deflection of the plasma flow around the planetary obstacle. At further approach to Mars, the solar wind interacts with the crustal fields and with the ionosphere produced by the photoionization of the neutral atmosphere by the solar EUV irradiance (10–90 nm). As discussed earlier, photochemistry leads to the dominant ion species in the ionosphere of molecular O2+ and atomic O+ oxygen ions. A balance between photoionization and recombination determines the peak electron density at altitudes of about 130 km. At altitudes above ~180 km the ionosphere is no longer in photochemical equilibrium and diffusion and transport processes become much more important. At the nightside the ionospheric density rapidly decreases due to recombination of O2+ ions. Because the thermal pressure in the upper ionosphere of Mars is generally weak compared to incident solar wind pressure, the IMF penetrates deep into the ionosphere and becomes the important factor for the ionospheric dynamics.

Similar to the case of Venus, the draped interplanetary magnetic field lines in the wake form the long magnetic tail on the night side. The well-organized structure of the induced magnetic tail consists of two lobes with oppositely directed magnetic field lines, separated by a plasma sheet. Draping of the interplanetary magnetic field lines around the ionospheric obstacle is asymmetric with respect to the direction of the motional electric field with a shift of the position of the upstream magnetic field pileup toward the –Vsw × B side (E+ hemisphere), and ionosphere asymmetry between E+ and E ionospheres arises. In the E hemisphere, in which the motional electric field is pointed toward the planet, the ionosphere is denser, moves slower, and expands to higher altitudes as compared to the ionosphere in the opposite (E+) hemisphere. This asymmetry propagates downstream to the tail, preferentially supplying by the ionospheric plasma the E hemisphere (Figure 10). Asymmetry of the draped magnetic field also results in asymmetry in the nightside field topology. Similar to the Venus case, the magnetic field lines in the near Mars tail, in the E hemisphere, appear to be wrapped more tightly around the planet than in the opposite hemisphere, thus forming a field reversal region in this portion of the near tail.

Figure 10. (upper panels): Number density and flux of O+ ions in the XZ plane, respectively. Since the orientation of the IMF and the motional electric field varies with time, the relevant coordinate system (MSE) used for the description of the induced draping magnetosphere has the XMSE axis antiparallel to the upstream solar wind flow, the YMSE axis along the cross-flow magnetic field component of the IMF in the solar wind, and the ZMSE axis pointing in the direction of the solar wind motional electric field (−Vsw x BIMF). (bottom panels): Different regions of the tail are clearly displayed while plotting fluxes of oxygen ions in the tail in Bx—ZMSE coordinates. A change of sign of the Bx component of the magnetic field corresponds to the crossing of the central current sheet. These coordinates avoid issues with multiple crossings of the current sheet due to its flapping motion and reduce uncertainties related to a time lag of the current sheet orientation in response to the IMF variations. Higher energy ions accelerated by j × B forces occupy the plasma sheet and the boundary layer, while low-energy ions fill the magnetic lobes (adapted from Dubinin et al., 2017).

As the j × B force of the magnetic shear stresses of the draped field lines is the strongest in the center of the tail (current sheet), the ion energy observed by a spacecraft crossing the tail steadily increases, reaches maximum at the center, and then again decreases. The lobes of the tail are filled by the low-energy ionospheric plasma (similar to Earth’s polar wind), which expands along the field lines driven by the gradient pressure force (–grad Pe)—light electrons trying to run away drag more heavy ions by the arising ambipolar electric field. All these regions in the tail together with a “plume” of pickup ions originating at the dayside are the main channels through which Mars loses ions (Figure 10). Estimates of the total present-day ion escape rate vary from ~2 × 1024 s−1 to 2 × 1025 s−1. Variations in solar wind and solar irradiance significantly affect efficiency of ion losses. The existence of stronger solar wind and higher solar EUV fluxes in the early solar system, inferred from observations of younger Sun-like stars motivates the interest for study of Mars interaction with corotating interaction regions (CIRs) and interplanetary coronal mass ejections (ICMEs) as proxies of the conditions in the past epochs. The observations suggest that escape rates increase by factors of ~2 to 10 times. Diffuse UV auroral emissions also observed in association with these events further suggests related enhancements in energy deposition into the atmosphere at these times. Which precipitating particle species and/or other processes lead to these emissions is still under investigation. In absence of strong solar events at the present unusually weak solar cycles ion losses at extreme conditions might be estimated from the global models. These models suggest an increase of the escape rate up to 1.4 × 1027s−1 during very strong ICMEs.

The existence of strong localized crustal magnetic fields significantly influences the interaction of the solar wind with Mars, adding features typical for planets with a global intrinsic magnetic field. As a result of their presence, features resembling mini-magnetospheres appear. Similar to the Earth, discrete UV auroras caused by precipitation of the energetic electrons can occur above areas with strong crustal magnetization. The crustal magnetic field also strongly reduces mobility of ions and electrons in the ionosphere, making them less exposed to vertical and horizontal motions. As a result, the dayside ionosphere is not axially symmetric but more inflated and dense over regions with strong crustal magnetization. The shape of the ionospheric obstacle to solar wind occurs not smooth but bulges out over regions with strong crustal magnetic field, shifting the local magnetosphere boundary upward, and affecting the position of the bow shock. Crustal magnetic fields also change the magnetic field configuration in the magnetotail. The topology of the field lines becomes very intricate (see Figure 11). Within the draping IMF, new classes of field lines—closed loops of field lines connected to the crustal field and open field lines with one end in solar wind and the another one on the Mars surface formed by reconnection between the IMF and crustal fields—are present. The net result might be the appearance of a “twisted” magnetotail.

Figure 11. Field line topology from two perspectives, with the strong crustal fields on the dayside. The color on the spherical surface is the magnetic magnitude at 150 km. Different types of field lines (except the draped field lines) are shown by colors: purple for closed field lines with both foot points on the dayside, black for closed field lines with both foot points on the nightside, green for one foot point on the dayside (solar zenith angle (SZA) < 90°) and the other on the nightside (SZA > 90°), orange for open field lines attached to the dayside ionosphere, and blue for open field line attached to the nightside (Xu et al., 2017).

The interested reader can find much more information and detail on the solar wind interactions of Mercury, Venus, and Mars (as well as Earth) in specific references. The many scientific results of the MESSENGER mission are summarized in the book Mercury: The View After MESSENGER, including Mercury’s magnetosphere (Slavin et al., 2019), while the specific findings described here are the focus of the listed research articles. Pioneer Venus Orbiter (PVO) mission plasma environment and upper atmosphere results are reviewed in detail in several comprehensive volumes, including the two books, Venus I and Venus II and dedicated issues of Space Science Reviews (e.g., Luhmann, 1986; Luhmann & Cravens, 1991). The Venus Express mission and observations are summarized in a special issue of Planetary and Space Science (2006) and in Space Science Reviews ( e.g., Futaana et al., 2017). The reader can find more details about Martian space in the referenced papers and in several dedicated issues of Space Science Reviews ( e.g., Dubinin et al., 2011) and special issues of Geophysical Research Letters (2015) and the Journal of Geophysical Research (2017) dedicated to the MAVEN observations. The reader is encouraged to explore beyond the short list of references provided here to appreciate the scope of what has been learned about the subject of this brief summary.

Further Reading

Solar Wind
  • Gosling, J. T. (1996). Corotating and transient solar wind flows in three dimensions. Annual Review of Astronomy and Astrophysics, 34, 35–74.
  • Hundhausen, A. J. (1972). Physics and chemistry in space. Vol. 5. Coronal expansion and solar wind. New York: Springer.
  • Luhmann, J. G., Petrie, G., & Riley, P. (2013). Solar origins of solar wind properties during the cycle 23 solar minimum and rising phase of cycle 24. Journal of Advanced Research, 4, 221–228.
  • McComas, D. J., Barraclough, B. L., Funsten, H., Gosling, J. T., Santiago-Munoz, E., Skoug, R. … & Balog, A. (2000). Solar wind observations over Ulysses’ first full polar orbit. Journal of Geophysical Research, 105, 10419–10433.
  • McComas, D. J., Gosling, J. T., & Skoug, R. M. (2000). Ulysses observations of the irregularly structured mid-latitude solar wind during the approach to solar maximum. Geophysical Research Letters, 27, 2437–2440.
  • Parker, E. N. (1958). Dynamics of the interplanetary gas and magnetic fields. Astrophysics Journal, 128, 664–676.
  • Webb, D. F., & Howard, T. A. (2012). Coronal mass ejections: Observations Living Reviews in Solar Physics, 9, 3. 3.
  • Dubinin, E., Fraenz, M., Zhang, T. L., Woch, J., Wei, Y., Fedorov, A., Barabash, S., & Lundin, R. (2013). Plasma in the near Venus tail: Venus Express observations. Journal of Geophysical Research, 118, 7624–7634.
  • Fedorov, A., Barabash, S., Sauvaud, J.-A., Futaana, Y., Zhang, T. L., Lundin, R., & Ferrier C. (2011). Measurements of the ion escape rates from Venus for solar minimum. Journal of Geophysical Research, 116, A07220.
  • Luhmann, J. G., Elphic, R. C., Russell, C. T., Mihalov, J. D., & Wolfe, J. H. (1980). Observations of large scale steady magnetic fields in the dayside Venus ionosphere. Geophysical Research Letters, 7(11), 917–920.
  • Russell, C. T., & Vaisberg, O. (1983). Interaction of the solar wind with Venus, In D. M. Hunton et al. (Eds.), Venus ( pp. 873–940). Tucson: University of Arizona Press.
  • Zhang, T. L., Baumjohann, W., Du, J., Nakamura, R., Jarvinen, R., Kallio, E., Du, M. … & Russell, C. T. (2010). Hemispheric asymmetry of the magnetic field wrapping pattern in the Venusian magnetic tail. Geophysical Research Letters, 37, L14202.
  • Bertaux, J.-L., Leblanc, F., Witasse, O., Quemerais, E. Lilensten, J., Stern, S. A., Sandel, B., & Korablev, O. (2005). Discovery of an aurora on Mars. Nature, 439, 790–794.
  • Brain, D., Barabash, S., Bougher, S. W., Duru, F., Jakosky, B. M., & Modolo R. (2017). Solar wind interaction and atmospheric escape. In R. Haberle et al. (Eds.), The atmosphere and climate of Mars (pp. 464–496). London: Cambridge University Press.
  • Deighan, J., Jain, S. K., Chaffin, M. S., Fang, X., Halekas, J. S., Clarke, J. T., … & Jakosky, B. M. (2018). Discovery of a proton aurora at Mars. Nature Astronomy, 2, 802–807.
  • Dubinin, E., Lundin, R., & Schwingcnschu, K. (1994). Solar wind electrons as tracers of the Martian magnetotail topology. Journal of Geophysical Research, 99, 21233–21240.
  • Dubinin, E., & Fraenz, M. (2015). Magnetotails of Mars and Venus. Geophysical Monograph Series, 207, 43–59.
  • Dubinin, E., & Fraenz, M. (2016). Ultra-low-frequency waves at Venus and Mars. In A. Keiling, D.-H. Lee, & V. Nakariakov (Eds.), Low-frequency waves in space plasmas (pp. 343–364). Wiley.
  • Halekas, J. S., Lillis, R. J., Mitchell, D. L., Cravens, T. E., Mazelle, C., Connerney, J. E. P. … & Ruhunusiri, S. (2015). MAVEN observations of solar wind hydrogen deposition in the atmosphere of Mars. Geophysical Research Letters, 42, 8901–8909.
  • Halekas, J. S Brain, D. A., Luhmann, J. G., DiBraccio, G. A., Ruhunusiri, S., Harada, Y., ... Jakosky, B. M. (2017). Flows, fields, and forces in the Mars-solar wind interaction. Journal of Geophysical Research, 122, 11,320–11,341.
  • Jakosky, B. M., Lin, R. P., Grebowsky, J. M., & 68 more (2015). The Mars Atmosphere and Volatile Evolution (MAVEN) Mission. Space Science Reviews, 195, 3–48.
  • Jakosky, B. M., J. M. Grebowsky, Luhmann, J. G., Connerney, J., Eparvier, F., Ergun, R., … & Yelle, R. (2015). MAVEN observations of the response of Mars to an interplanetary coronal mass ejection. Science, 111, A09213.
  • Johnson, R. E., Schnellenberger, D., & Wong, M. C. (2000). The sputtering of an oxygen thermosphere by energetic O+. Journal of Geophysical Research, 105(E1), 1659–1670.
  • Leblanc, F., Martinez, A., Chaufray, J. Y., Modolo, R., Hara, T., Luhmann, J., … & Jakosky, B. (2018). On Mars’s atmospheric sputtering after MAVEN’s first Martian year of measurements. Geophysical Research Letters, 45, 4685–4691.
  • Luhmann, J. G., Johnson, R. E., & Zhang, M. H. G., et al. (1992). Evolutionary impact of sputtering of the Martian atmosphere by O(+) pickup ions. Geophysical Research Letters, 19, 2151–2154.
  • Luhmann, J. G., Dong, C. F., Ma, Y. J., Curry, S. M., Xu, C., Lee, O., … & Jakosky, B. (2017). Martian magnetic storms. Journal of Geophysical Research, 122, 6185–6209.
  • Lundin, R., Zakharov, A., Pellinen, R, Borg, H., Hultqvist, B., Pissarenko, N., Dubinin, E., & Koskinen, H. (1989). First measurements of the ionospheric plasma escape from Mars. Nature, 341, 609–612.
  • Lundin, R., Winningham, D., Barabash, S., Frahm, R., Brain, D., Nilsson, H., … & Wurz, P. (2006). Plasma acceleration above Martian magnetic anomalies. Science, 311, 980.
  • Mazelle, C. D., Winterhalter, D., Sauer, K., Trotignon, J. G., Acuna, M. H., Baumgartel, K., … & Slavin, J. (2004). Bow shock and upstream phenomena at Mars. Space Science Reviews, 111, 115–181.
  • Mazelle, C., X., Meziane, K., Mitchell, D. L., Garnier, P., Espley, J. R., Hamza, A. M., … & Jakosky, B. (2018). Evidence for neutrals-foreshock electrons impact at Mars. Geophysical Research Letters, 45, 3768–3774.
  • McElroy, M. B., Kong, T. Y., & Yung, Y. (1977). Photochemistry and evolution of Mars’ atmosphere: A Viking perspective. Journal of Geophysical Research, 92(28), 4379–4388.
  • Nagy, A., Winterhalter, D., Sauer, K., Cravens, T. E., Brechts, S., Mazelle, C., … & Trotignon, J. G. (2004). The plasma environment of Mars. Space Science Reviews, 111, 33.
  • Schneider, N., Deighan, J. I., Jain, S. K., Stiepen, A., Stewart, A. I. F., Larson, D., … &Jakosky, B. (2015). Discovery of diffuse aurora on Mars. Science, 350. aad 0313-1-0313-5.