Project Topics
Get Complete Project Material File(s) Now! »
High-speed streams
Another type of solar wind structures a ecting space weather, the solar high-speed streams are created by the fast ow of solar wind plasma along open magnetic eld lines. These regions of open magnetic ux are called coronal holes, and they are almost always present on the Sun, especially at high latitudes [Lowder et al., 2017]. During the declining phase of the solar cycle, coronal holes from the polar regions tend to migrate to lower latitudes, and the resulting high-speed streams of solar wind may ow in the ecliptic plane [Kojima and Kakinuma, 1987]. When the fast solar wind (typically 500{800 km/s) is preceded by a slower stream (200{400 km/s), this creates a compression region characterised by high IMF magnitude, solar wind density and pressure, called corotating interaction region [Richter and Luttrell, 1986]. The terminology came from the observation that such structures often show a 27 day recurrence, corresponding to the solar rotation period near the equator, which implies that a same coronal hole (and hence, the resulting high-speed stream) may persist during several solar rotations. Figure 1.5 shows the IMF magnitude and the solar wind velocity, density and dynamic pressure measured near the Earth during a high-speed stream event from March 2007. The transition from the slow solar wind stream to the high-speed stream contains several local peaks of solar wind pressure and density, in this particular example. The east-west component of the solar wind velocity (not shown here) generally exhibits a sign reversal at the transition from slow to fast solar wind, de ning the stream interface [Kavanagh et al., 2012].
Even though the coronal mass ejections may represent the most violent type of disturbances, it is estimated that the corotating interaction regions and following high-speed streams are more geoe ective, i.e., are more e cient in depositing energy into the ionosphere, even though the energy input into the planetary environments is generally smaller than for CMEs [Turner et al., 2009]. It is suggested that this may arise from solar wind driving di erences between these two types of disturbances, and also possibly from saturation in the magnetospheric response to CMEs. As a result, a CIR may sometimes result in a deposition of energy into the ionosphere comparable to what a CME could produce [Turner et al., 2009]. At Mars, CIRs are more often associated with shocks than at 1 AU [Gosling and Pizzo, 1999]. They are an important cause for atmospheric erosion [Hara et al., 2011].
Magnetospheres
The speed of sound in the solar wind is about 60 km/s near the Earth’s orbit [Hundhausen, 1995], and the Alfven speed is respectively of the order of 50 km/s [Hargreaves, 1995]. This means that the solar wind is supersonic and super-Alfvenic. Therefore, its encounter with obstacles such as planets with an intrinsic magnetic eld or surrounded by an ionised atmosphere results in the formation of a bow shock ahead of the obstacle. At Earth, the nose of the bow shock is located about 14 Earth radii (RE = 6371 km) upstream from the centre of the Earth [Burgess, 1995]
Terrestrial magnetosphere
Since the Earth has an intrinsic magnetic eld, the solar wind cannot penetrate the terrestrial environment and is deviated. The boundary between the geomagnetic eld and the solar wind is de ned as the surface where their magnetic pressures are equal; it is called the magnetopause. The volume inside the magnetopause is the geomagnetic cavity, or the terrestrial magnetosphere. Figure 1.6 gives a schematic illustration of the terrestrial magnetosphere and its main regions. The region located between the magnetopause and the bow shock is called the magnetosheath (in blue in the gure). It consists of solar wind particles forming turbulent plasma, heated by the bow shock crossing. It may play an important role in the coupling of the solar wind and the terrestrial environment by introducing non-linearities in electric eld and Poynting ux transfer between the bow shock and the magnetopause [Pulkkinen et al., 2016].
The existence of the solar wind ow modi es the structure of the geomagnetic eld. From a nearly dipolar eld, it becomes compressed on the dayside and stretched on the nightside. On the dayside, the magnetopause position along the Sun-Earth line is most of the time 10 RE from the centre of the Earth; yet it may be moved to 8 RE when the IMF has a strong southward component [Shue et al., 2001], and even closer to the Earth following the arrival of a CME [e.g., Huttunen et al., 2002]. On the nightside, geomagnetic eld lines are stretched to form the so-called magnetotail, whose extension is estimated to reach beyond several hundreds of Earth radii [Slavin et al., 1983]. On the dayside, two neutral points directly connect the magnetosphere to the Earth; they are called polar cusps. The polar cusps are mapped to the Earth’s surface near magnetic noon at a geomagnetic latitude ranging from 72 to 79 , depending on solar wind conditions, and there is one cusp in each hemisphere [Newell et al., 1989]. They correspond to holes in the magnetic shield of the Earth, through which the solar wind may directly enter the geomagnetic cavity.
Geomagnetic storms and magnetospheric substorms
Ground-based instruments enable the observation of signatures of the geomagnetic eld perturbations and the estimation of the severeness of the disturbance. Magne-tometers (see section 2.1.2) located at low geomagnetic latitudes observe a de ection in the horizontal component of the geomagnetic eld. This is the signature of a geomagnetic storm, which is related to an intensi cation of the ring current (see section 1.3.1) when particles are injected into the inner magnetosphere. Geomagnetic disturbances are often quanti ed using geomagnetic indices, which are generally derived from a network of magnetometer stations. Geomagnetic storm severeness has historically been estimated using the Dst (\disturbance storm time ») index [Sugiura, 1964; Sugiura et al., 1991], which is calculated using magnetometer data from four low-latitude stations and has 1-hour time resolution. More recently, the SYM-H index [Iyemori, 1990] has been preferred, as it provides an equivalent measure of storm severeness but with a higher (1 min) time resolution [Wanliss and Showalter, 2006]. A geomagnetic storm typically follows three phases: short-lived (a few hours) increase in the SYM-H index, followed by a decrease during up to one day, and slow recovery to pre-storm values during several days [Hargreaves, 1995]. The Dst or SYM-H minimum value typically reaches -30 to -50 nT during moderate storms, and may reach under -250 nT during extreme storms [Gonzalez et al., 1999].
In the disturbed magnetosphere, a phenomenon known as substorm takes place, often repeatedly during a geomagnetic storm { even though a substorm may also occur outside of geomagnetic storm periods. It is initiated by magnetic reconnection in the magnetotail, at a distance from the Earth which is still debated. Recent results estimate it to be of the order of 18 RE [Sergeev et al., 2012]. Three main phases are distinguished: the growth, expansion, and recovery phases. In the ionosphere, a substorm has various manifestations that can be observed with ground-based instruments, such as precipitation of particles leading to auroral displays and electron density enhancements in the D, E and F regions, ionospheric currents, and plasma heating. During a typical substorm, the growth phase is associated with quiet auroral arcs in the east-west direction moving equatorwards [Motoba et al., 2015]. Then, during the expansion phase, one of the arcs becomes active and moves polewards; the auroral oval becomes broader. The recovery phase manifests itself as a quietening of the auroral activity, with arcs becoming dimmer and slowly drifting back equatorwards [Akasofu, 1964]. Substorm activity is often measured using the AE (\auroral electrojet ») index, which is calculated using magnetometers located in the auroral oval [Davis and Sugiura, 1966]. It is sensitive to the intensi cation of auroral-zone currents such as the eastward electrojet in the evening sector, the westward electrojet in the early-morning sector, and the substorm electrojet owing westwards in the midnight sector. Typical values for AE during a substorm are of the order of hundreds of nanoteslas but may exceed 1000 nT [Akasofu, 1981].
Some aspects of Mars interactions with the solar wind
Besides the draping of the interplanetary magnetic eld which creates an induced magnetosphere, the solar wind also interacts with Mars in ways that are to some extent analogous to phenomena observed at Earth. Mars Global Surveyor observations suggesting that magnetic reconnection occurs between the Martian crustal elds and the draped IMF were reported by Halekas et al. [2009]. The presence of crustal elds results in the existence of cusps, through which the solar wind electrons may directly access the planetary surface [Dubinin et al., 2008]. These magnetic structures are of small size compared to the terrestrial cusps, hence the regions where direct penetration of solar wind particles is possible are also very localised. Recent observations of ultraviolet auroral emission in the vicinity of these \mini magnetospheres », mostly located in the southern hemisphere, suggest that accelerated solar wind electrons may excite atmospheric species, in a similar way as particle precipitation does at Earth [Brain and Halekas, 2012]. In addition, di use auroral emission was also reported in the northern hemisphere, down to about 60 km altitude [Schneider et al., 2015]. The current understanding is that the discrete type of aurora, on closed eld lines, may be compared to nightside oval aurora at Earth, while the di use aurora, on open eld lines, could be seen as analogous to polar rain aurora at Earth, associated to the precipitation of low-energy (< 1 keV) electrons in the terrestrial polar cap.
In addition, recent results from the Mars Global Surveyor mission indicate that reconnection also takes place in the Martian magnetotail. Eastwood et al. [2012] reported the observation of a chain of three magnetic ux ropes, i.e., twisted magnetic structures containing a owing current, in the magnetotail current sheet. These ux ropes were probably generated by instabilities during a magnetic reconnection event. It is believed that such processes may play an important role in plasma transport in the Martian environment.
Furthermore, Withers et al. [2016] have investigated the e ect of solar wind dynamic pressure uctuations on the Martian environment. They found that higher dynamic pressure results in magnetosphere compression, magnetosheath heating, and topside ionosphere compression.
SGO riometer chain and KAIRA
When particles with energies greater than a few tens of keV precipitate into the ionosphere, they may produce ionisation down to the D region (i.e., below 90 km altitude). This additional ionisation in the D region may be measured using riometers (relative ionospheric opacity metres), as was rst done by Shain [1951].
Riometry is based on the measurement of the power of cosmic radio noise reaching the ground. When cosmic noise originating from a radio source in the sky propagates through the ionosphere, part of its power is absorbed. The absorption in the ionosphere can be derived by isolating the imaginary part of the complex refractive index given by equation (2.1). Under the quasi-longitudinal approximation, the absorption along a path noted l can be calculated in decibels (dB) by Z N dl AdB = 4:6 10 5 2 + (! e !c cos )2 ; (2.4).
with Ne the electron density, the electron collision frequency, ! the angular frequency of the radio wave, !c the electron gyrofrequency, and the angle between the direction of propagation and the magnetic eld vector [Hargreaves, 1969].
In practice, the absorption is negligible in the F and E regions of the ionosphere, since the electron collision frequency is very small. Therefore, cosmic noise absorption is a ected mainly by the electron density in the D region. The power of the radio signal observed by a riometer therefore depends on the distribution of radio sources in the sky, on the electron density pro le in the D region, and on the observed frequency. If a riometer observes along a xed beam, then a same radio source will periodically enter the beam as the Earth rotates. This implies that, under quiet ionospheric conditions with similar solar irradiance daily variations, the cosmic radio noise power measured by a riometer is periodic, the period corresponding to the duration of a sidereal day (Tsid = 86 164 s). However, if the D-region ionisation is enhanced compared to a quiet day, e.g., because of energetic particle precipitation or X-rays originating from a solar are, then a greater amount of cosmic noise power is absorbed. By subtracting the time series of the cosmic radio noise power measured by a riometer from a so-called quiet-day curve (QDC), one may estimate the cosmic noise absorption (CNA) due to enhanced D-region electron density. In publications, CNA generally refers to what should strictly-speaking be called \excess cosmic noise absorption compared to a quiet day ».
Figure 2.4 illustrates the QDC de nition used at SGO, with the example of the QDC de ned for the station located in Ivalo for the month of August 2007. The QDC is estimated manually by superposing the measured power (black dots) during several days over a time period equal to Tsid. This enables one to identify the shape of the signal which would be measured on each sidereal day during that period of the year if there were no disturbances of any kind. This shape gives the quiet-day curve (red line). Then, for a given day, the measured radio noise power is subtracted to the QDC to obtain absorption, which is expressed in decibels. One may notice in Figure 2.4 that the measured power sometimes exceeds the QDC (e.g., between 6 and 15 sidereal local time, SLC). This may be due to solar radio emissions or to radio interferences originating from human-made systems. On the other hand, several downward spikes may also be observed under the QDC in the gure (e.g., near 10 SLC). These are cosmic noise absorption signatures, which are the data of geophysical interest.
It should be noted that QDCs vary according to the period of the year. Indeed, di erent radio noise sources appear in the antenna eld of view throughout a year, as a consequence of the tilt of the Earth rotation axis. In addition, non-geophysical factors, such as the accumulation of snow on the antenna elements in winter, may a ect the riometer beam shape. Therefore, QDCs must be regularly rede ned to process the riometer data.
The CNA data used in Paper II come from four riometers of the SGO chain (Ivalo, Sodankyla, Oulu, and Jyvaskyla), covering auroral and subauroral latitudes (L-shells between 3.8 and 5.7). The instruments are narrow-band wide-beam (60 ) riometers, observing at frequencies near 30 MHz (32.4 MHz for Jyvaskyla). CNA data is available at 1 min time resolution, but in Paper II it is used with 15 min resolution.
Table of contents :
1 Introduction
1.1 Atmospheres and ionospheres
1.1.1 Neutral atmospheres
1.1.2 Ionospheres
1.1.2.1 Photoionisation, chemistry and transport
1.1.2.2 Structures of the terrestrial and Martian ionospheres
1.2 Solar wind
1.2.1 General properties
1.2.2 High-speed streams
1.3 Magnetospheres
1.3.1 Terrestrial magnetosphere
1.3.2 Martian induced magnetosphere
1.4 Planetary environments and solar wind interactions
1.4.1 Reconnection and convection
1.4.2 Geomagnetic storms and magnetospheric substorms
1.4.3 Particle precipitation
1.4.4 Some aspects of Mars interactions with the solar wind
2 Instrumentation
2.1 Ground-based instruments
2.1.1 Sodankyla ionosonde
2.1.2 IMAGE magnetometers
2.1.3 SGO riometer chain and KAIRA
2.1.4 Kilpisjarvi All-Sky Camera
2.2 Satellites
2.2.1 Mars Express
2.2.2 NOAA/POES
2.2.3 ACE
3 Methods
3.1 Superposed epoch analysis
3.1.1 Classical version
3.1.2 Phase-locked version
3.2 Radio-occultation
3.2.1 Principle
3.2.2 Data analysis
4 Results
4.1 Eects of solar wind high-speed streams on the high-latitude ionosphere of the Earth
4.1.1 Solar wind driving
4.1.2 Overview of ionospheric response to high-speed streams
4.1.3 E and F region responses
4.1.4 Energetic particle precipitation
4.2 Modulation of energetic precipitation during pulsating aurora
4.3 Analysis of Mars Express data with the radio-occultation model
4.3.1 Reproduction of the frequency residual prole
4.3.2 Inuence of medium asymmetry
4.3.3 Ion density proles
4.3.4 Improvements introduced in Paper V
5 Conclusion
Bibliography
